Improving Health Sector Performance: Institutions, Motivations and Incentives - The Cambodia Dialogue 9789814311854

Citation preview

Improving Health Sector Performance

The GMS (Greater Mekong Subregion) Development Series is an initiative of the Cambodia Development Resource Institute (CDRI), Cambodia’s leading independent development policy research institute. Established in 1990, CDRI’s research focuses on five areas: economy, trade and regional cooperation; poverty, agriculture and rural development; democratic governance and public sector reform; natural resources and the environment; and social development.

The Institute of Southeast Asian Studies (ISEAS) was established as an autonomous organization in 1968. It is a regional centre dedicated to the study of socio-political, security and economic trends and developments in Southeast Asia and its wider geostrategic and economic environment. The Institute’s research programmes are the Regional Economic Studies (RES, including ASEAN and APEC), Regional Strategic and Political Studies (RSPS), and Regional Social and Cultural Studies (RSCS). ISEAS Publishing, an established academic press, has issued more than 2,000 books and journals. It is the largest scholarly publisher of research about Southeast Asia from within the region. ISEAS Publications works with many other academic and trade publishers and distributors to disseminate important research and analyses from and about Southeast Asia to the rest of the world.

ii

00 Improving Health Prelims

2

8/24/11, 4:20 PM

GMS Development Series 1 A CDRI Publication

Improving Health Sector Performance Institutions, Motivations and Incentives The Cambodia Dialogue Edited by

Hossein Jalilian and Vicheth Sen

Institute of Southeast Asian Studies Singapore and

First published in Singapore in 2011 by ISEAS Publishing Institute of Southeast Asian Studies 30 Heng Mui Keng Terrace Pasir Panjang Singapore 119614 E-mail: [email protected] Website: http://bookshop.iseas.edu.sg All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the Institute of Southeast Asian Studies. © 2011 Institute of Southeast Asian Studies, Singapore The responsibility for facts and opinions in this publication rests exclusively with the authors and their interpretations do not necessarily reflect the views or the policy of the publishers or their supporters. USAID Disclaimer The research for this publication was made possible by the support of the American people through the United States Agency for International Development (USAID). The findings of this study are the sole responsibility of Cambodia Development Resource Institute, and do not necessarily reflect the views of USAID or the United States Government. ISEAS Library Cataloguing-in-Publication Data Improving health sector performance : institutions, motivations and incentives— The Cambodia dialogue / edited by Hossein Jalilian and Vicheth Sen. “… based on a collection of papers presented at a two-day international conference on ‘Improving Health Sector Performance: Institutions, Motivations and Incentives’ in Phnom Penh from 26–27 April 2010”—Acknowledgements. 1. Public health—Asia—Congresses. 2. Public health—Pacific Area—Congresses. 3. Public health—Cambodia—Case studies—Congresses. 4. Community health services—Asia—Congresses. 5. Community health services—Pacific Area—Congresses. 6. Community health services—Cambodia—Case studies—Congresses. I. Jalilian, Hossein. II. Sen, Vicheth. III. International Conference on Improving Health Sector Performance : Institutions, Motivations and Incentives (2010 : Phnom Penh, Cambodia) RA525 I34 2011 ISBN 978-981-4311-84-7 (soft cover) ISBN 978-981-4345-52-1 (hard cover) ISBN 978-981-4311-85-4 (e-book PDF) Cover photo taken on a visit to a health care unit in a village in Cambodia. Source: CDRI. Typeset by Superskill Graphics Pte Ltd Printed in Singapore by Mainland Press Pte Ltd

iv

00 Improving Health Prelims

4

8/24/11, 4:21 PM

CONTENTS Acknowledgements

ix

Foreword

xi

List of Contributors

xiii

Introduction

xvii

Part I Overview 1. What Incentives Are Effective in Improving Deployment of Health Workers in Primary Health Care in Asia and the Pacific? 3 Vivian Lin, Lee Ridoutt, Emily Brink, and Bruce Hollingsworth 2.

Reforming Provider Behaviour through Incentives: Challenges and Reflections from the U.K. Experience Alan Maynard

19

Part II Organizational Arrangements: Purchasing Health Services 3. The Transition to Semi-Autonomous Management of District Health Services in Cambodia: Assessing Purchasing Arrangements, Transaction Costs and Operational Efficiencies of Special Operating Agencies 45 Khim Keovathanak and Peter Leslie Annear 4.

Vouchers as Demand-side Financing Instruments for Health Care: A Review of the Bangladesh Maternal Voucher Scheme and Implications for Incentives for Human Resource Management Jean-Olivier Schmidt and Atia Hossain

76

v

00 Improving Health Prelims

5

8/24/11, 4:21 PM

Contents

vi 5.

6.

Social Health Insurance in Cambodia: An Analysis of the Health Care Delivery Mechanism Sopheap Ly Purchasing Health Services in New Zealand Toni Ashton and Maria C.G. Bautista

136

Part III Optimal Health Workers Contracts 7. A Civil Service That Performs: Primary Health Care in Curitiba, Brazil Geoffrey Shepherd 8.

9.

10.

Increasing Uptake of Reproductive Health Services Using Innovative Financing Models: Experiences of Marie Stopes International Che Katz and Thoai D. Ngo

13.

151

184

Understanding Rural Health Service in Cambodia: Results of a Discrete Choice Experiment Bundeth Seng, Neath Net, Nonglak Pagaiya, and Sethea Sok

202

Contracting Health Workers to Underserved Areas: Indonesian Approaches to a Distributional Challenge Hjalte Sederlof and Rooswanti Soeharno

248

Part IV Managing Doctors and Nurses 11. How Managers Manage in Cambodia’s Public Health Sector Neath Net, Roger Hay, and Klaus J. Broesamle 12.

101

The Impact of Management Training and Education on the Performance of Health Care Providers: What Do We Know? Orvill Adams and Leanne Idzerda

265

304

Incentive Systems in Public Health Care Organizations in Italy 328 Alberto Asquer

vi

00 Improving Health Prelims

6

8/24/11, 4:21 PM

Contents

vii

Part V Health Service Consumer Behaviour 14. Factors Influencing Health-Seeking Behaviour in Siem Reap: A Qualitative Analysis Sarah C. Smith 15.

353

Villagers’ Evaluation of a Community-based Health Insurance Scheme in Thmar Pouk, Cambodia 365 Sachiko Ozawa and Damian Walker

APPENDIX Health Equity Funds Implemented by URC and Supported by USAID Tapley Jordanwood, Maurits van Pelt, and Christophe Grundmann Index

387

421

vii

00 Improving Health Prelims

7

8/24/11, 4:21 PM

viii

00 Improving Health Prelims

8

8/24/11, 4:21 PM

ACKNOWLEDGEMENTS This volume comprises chapters based on a collection of papers presented at a two-day international conference on “Improving Health Sector Performance: Institutions, Motivations and Incentives” in Phnom Penh from 26–27 April 2010. The event was co-organized by the Cambodia Development Resource Institute (CDRI) and the Oxford Policy Institute (OPI), with financial support from the Department for International Development (DFID), United Kingdom. The conference would not have been possible without the efforts of many people within the CDRI and OPI. Significant contributions to the organization of the conference were provided by the Social Development Programme, with support from Administration and Support Services, External Relations, and Publications Units, CDRI. Special thanks also are due to the late Roger Hay, former director of OPI, for his active support during the initial stage of the project. Overall guidance was provided by the senior management of CDRI. Most of the chapters in this volume have benefited substantially from peer review by a number of experts in the health sector. Thanks are due to the following reviewers for their invaluable contributions to this process: Peter Annear, Kimberley McClean, Vivian Lin, Maryam Bigdeli, Ann Robins, and Geoffrey Shepherd. Thanks are also due to Allen Myers for his editorial assistance. We would also like to thank H.E. Dr Tia Phalla, vice-chair of the National AIDS Authority, Cambodia, for contributing the Foreword to this volume.

ix

00 Improving Health Prelims

9

8/24/11, 4:21 PM

Acknowledgements

x

Finally, we would also like to thank University Research Co., LLC, especially Dr Christophe Grundmann, for financial support of this publication. Hossein Jalilian (PhD) Director of Research Cambodia Development Resource Institute Vicheth Sen Programme Coordinator Social Development Programme Cambodia Development Resource Institute

x

00 Improving Health Prelims

10

8/24/11, 4:21 PM

FOREWORD The Cambodian Government considers health as the heart of the Millennium Development Goals (MDGs) because healthy people constitute the basis for human resource development and sustainable socio-economic progress. The chapters in this volume come from papers presented at an international conference organized by the Cambodia Development Resource Institute in April 2010. With participation from local and international experts, the conference aimed at collecting major experiences and innovative solutions from inside and outside the country to improve health sector performance, with particular focus on institutions, motivations and incentives. The growing gap between the supply of health care professionals and the demand for their services is recognized as a key issue for health and development worldwide. The Cambodian National Strategic Plan for Health notes that in 2008, for every 10,000 members of the population there were only 0.10 physicians, 0.55 secondary midwives, and 1.21 secondary nurses. This shortage of health human resources is becoming a high priority issue on the political agenda. The introduction of the Priority Operating Costs (POC) system in July 2010 will allow for the motivation of health care workers to deliver better public services. Within the framework of Cambodian ownership and leadership, this public administration reform aims at facilitating the harmonization and alignment of practices by development partners in this area. The general principles of POC aim at “management which is result-based with payments that are made against achievements”. It is further characterized as “a scheme which provides means to appointees to do their work efficiently in a spirit of ownership and leadership”.

xi

00 Improving Health Prelims

11

8/24/11, 4:21 PM

Foreword

xii

The chapters in this volume deal with both an overview of the issues and with the four broad themes of the conference: the purchasing of health services; health worker contracts; managing doctors and nurses; and health service consumer behaviour. Through the chapters and discussions, the conference offered a unique forum for policy makers, health officials, civil society organizations, United Nations agencies, and development partners to discuss with local and international researchers major issues concerning the performance of the health sector. The lessons drawn will enable Cambodia to plan and manage human resources in the health sector better and to address major issues of organization, motivations and incentives. From the conference, four critical issues of human resources management in health — numbers, quality, attitudes, location — were identified as key factors for the implementation of national health strategies and for attaining the MDGs that are health related. Dr Tia Phalla Vice-Chair National AIDS Authority

xii

00 Improving Health Prelims

12

8/24/11, 4:21 PM

LIST OF CONTRIBUTORS Dr Orvill Adams is director of Orvill Adams & Associates Incorporated, Health Systems Policy and Workforce Planning, Ottawa, Ontario, Canada. Dr Peter Leslie Annear, PhD, is senior research fellow, Nossal Institute for Global Health, University of Melbourne, Australia. Professor Toni Ashton, PhD, is associate professor and director, Centre for Health Services Research and Policy, School of Population Health, University of Auckland, New Zealand. Dr Alberto Asquer is assistant professor, Department of Business Management, Faculty of Economics, University of Cagliari, Italy. Dr Maria C.G. Bautista is senior tutor, Centre for Health Services Research and Policy, School of Population Health, University of Auckland, New Zealand. Emily Brink is a PhD student, School of Public Health, La Trobe University, Melbourne, Australia. Klaus J. Broesamle is consultant for the Oxford Policy Institute, the United Kingdom. Dr Christophe Grundmann (PhD) is chief of party, University Research Co. (URC), Cambodia. Roger Hay (deceased) was the former executive director of the Oxford Policy Institute, the United Kingdom.

xiii

00 Improving Health Prelims

13

8/24/11, 4:21 PM

List of Contributors

xiv

Professor Bruce Hollingsworth is professor and director of the Centre for Health Economics, Monash University, Melbourne, Australia. Atia Hossain is senior advisor in health economics, Deutsche Gesellschaft für Technische Zusammenarbeit (GTZ) GmbH, Bangladesh. (P.O. Box 6091, Gulshan I, Dhaka, Bangladesh). Leanne Idzerda is research assistant at the Centre for Global Health, Institute for Population Health, University of Ottawa, Ontario, Canada. Hossein Jalilian, PhD, is a reader in Economic Development at the University of Bradford. Over the period from April 2007–Sept. 2010, on leave of absence from the University of Bradford, he was research director at CDRI. Tapley Jordanwood (M.Sc.) is health financing programme manager, University Research Co. (URC), Cambodia. Che Katz is programme director, Marie Stopes International Cambodia. Keovathanak Khim is a PhD student, Nossal Institute for Global Health, University of Melbourne, Australia. Professor Vivian Lin, PhD, is professor of Public Health, School of Public Health, La Trobe University, Melbourne, Australia. Ly Sopheap is currently working as a team leader of the Public Sector Strengthening Component for the Cambodia MSME Project, USAID Cambodia. She received her LLD in 2009 from the Graduate School at Nagoya University, Japan, majoring in Public Administration. Professor Alan Maynard, PhD, is professor of health economics at the Department of Health Sciences and the Hull-York Medical School at the University of York, England. He was chair of the York Hospitals NHS Trust from 1997 to 2010. Net Neath is research fellow of the Social Development Programme, Cambodia Development Resource Institute (CDRI), Cambodia.

xiv

00 Improving Health Prelims

14

8/24/11, 4:21 PM

List of Contributors

xv

Ngo Thoai D. is Asia regional research manager, Research & Metrics, Health System Department, Marie Stopes International. Sachiko Ozawa, PhD, is an assistant scientist, Department of International Health, Health Systems, Johns Hopkins Bloomberg School of Public Health, Johns Hopkins University, the United States. Dr Pagaiya, Nonglak, PhD, is a researcher for the International Health Policy Programme (IHPP), Ministry of Public Health, Thailand. Lee Ridoutt is director of Human Capital Alliance, Sydney, Australia. Jean-Olivier Schmidt is programme director, Deutsche Gesellschaft für Technische Zusammenarbeit (GTZ) GmbH, Bangladesh. (#74/1, 2nd Main Road, 8th Block, Jayanagar, Bangalore 560082, India). Hjalte Sederlof is advisor to the Ministry of Health, Indonesia, on health systems development. Sen, Vicheth is the programme coordinator of the Social Development Programme, Cambodia Development Resource Institute (CDRI), Cambodia. Dr Seng, Bundeth is research associate of the Social Development Programme, Cambodia Development Resource Institute (CDRI), Cambodia. Geoffrey Shepherd is an economist with experience at the World Bank, the United States and the University of Sussex and a consultant in the areas of political economy and public sector reform. Sarah C. Smith is a PhD candidate in medical anthropology at the University Queensland, School of Social Sciences, Australia. Dr Rooswanti Soeharno is advisor to the Ministry of Health, Indonesia, on health systems development. Sok, Sethea is research associate of the Social Development Programme, Cambodia Development Resource Institute (CDRI).

xv

00 Improving Health Prelims

15

8/24/11, 4:21 PM

List of Contributors

xvi

Dr Phalla Tia is vice-chair of the National AIDS Authority (NAA), Cambodia. Maurits van Pelt (M.Sc., L.L.M.) is a consultant for the University Research Co. (URC), Cambodia. Professor Damian Walker is associate professor, Johns Hopkins Bloomberg School of Public Health, Johns Hopkins University, the United States.

xvi

00 Improving Health Prelims

16

8/24/11, 4:21 PM

INTRODUCTION There is growing international evidence that the effectiveness of health services stems primarily from the extent to which the incentives facing providers and consumers are aligned with “better health” objectives. Efficiency in health service provision requires that providers and consumers have incentives to use health care resources in ways that generate the maximum health gains. Equity, in at least one sense, requires that consumers requiring the same care are treated equally, irrespective of their ability to pay. Efficiency in the use of health services requires that consumers are knowledgeable about the services on offer and which are most appropriate for their needs. Although these principles are enshrined in the design of every health system in the world, they have proven extremely difficult to apply in practice. Health care providers have financial obligations to their families as well as professional obligations to their patients. Health service consumers generally lack information about both their health and health services so that they underconsume or overconsume health care. Matters are unlikely to improve until there is more empirical evidence of the way health service providers and consumers in different parts of the world are actually motivated and how these incentives are best modified to improve health service efficiency and equity. Health provider incentives are generated by the institutions (rules) that govern their professional behaviour. These can be classified broadly into institutions that generate extrinsic incentives (organizational formats, managerial arrangements, social norms, and remuneration), and intrinsic incentives arising from values such as altruism, beliefs, reputation, and work satisfaction. The incentives that consumers have to use health services efficiently stem from intrinsic incentives arising from concern for their health, knowledge about the available health services, and social institutions, such as the trust and respect that mediate their transactions with providers. xvii

00 Improving Health Prelims

17

8/24/11, 4:21 PM

Introduction

xviii

This volume is based on a collection of papers from both Cambodian and international scholars, researchers, and practitioners in the health sector presented at a two-day international conference co-organized in Phnom Penh on 26–27 April 2010 by the Cambodia Development Resource Institute (CDRI) and the Oxford Policy Institute (OPI) with support from the Department for International Development (DFID), the United Kingdom. Under the theme, “Improving Health Sector Performance: Institutions, Motivations and Incentives”, the conference brought about sixty Cambodian researchers and health sector managers together with researchers from the region and elsewhere. Structured according to the programme of the conference, this collection focuses on four themes related to the incentives that motivate health service providers and consumers: Theme 1: Alternative organizational arrangements for delivering health services, including commissioning and purchasing There is an international trend towards active commissioning and contracting with both government and non-government providers for the supply of health services. Recent reforms in Cambodia are introducing such arrangements across the health sector. At the same time, provider units in Cambodia are being transformed into semi-autonomous organizations (Special Operating Agencies). This is intended to increase activity, improve service quality, focus service provision on priorities, and so improve allocative efficiency by providing better specified and stronger incentives for health service managers and providers. However, contract management also carries additional transaction costs. What is the evidence for better technical efficiency, taking into account increased activity and service quality? Theme 2: Incentives associated with alternative health worker employment contracts The way clinicians are employed and paid creates powerful incentives for resource management, productivity (or shirking), and service quality. Commonly, government health services are staffed by government employed clinicians, often with civil service status. The principal-agent theory suggests that perverse incentives arise and are common when provider and patient objectives do not coincide. How, and to what extent, can health worker contracts, including obligations to allow observation of their work, mitigate such agency problems? What is the relationship between salaried employment, capitation, fee-for-service, and “payment xviii

00 Improving Health Prelims

18

8/24/11, 4:21 PM

Introduction

xix

by results” remuneration and health service volumes and quality? In a region where government salaries are typically low, what is the relationship between levels of remuneration, labour productivity, and service quality? Theme 3: Health worker management regimes Clinicians have long prized and defended their independence. However, this independence is being increasingly eroded by clinical protocols, service delivery targets, and managed working practices. Some would argue that this has undermined professional ethics and the intrinsic incentives of altruism and reputation. Do “managed” clinicians perform better? Are doctors and nurses best at managing doctors and nurses? How does clinician management in government and non-government sectors differ and with what results? What is the role of organizational leadership in clinician motivation, in innovation, and in the improvement of organizational performance? Theme 4: Incentives associated with the effective use of health care by service consumers Most empirical demand-side health service research has focused on price elasticities and the demand for the different types of health care. But the demand for health services is also influenced by consumer knowledge, access to information (health care markets are characterized by great informational asymmetries between patients and providers), peer pressure, and trust in health care providers. In particular, this collection of papers will consider and examine three sets of issues: 1. 2.

3.

What is the effect of information on the incentives people have to improve their lifestyles and to use health services more efficiently? What are the effects of subsidies and taxes on health care-seeking behaviour? Should people be paid to adopt healthier behaviours or seek preventative care, such as Conditional Income Transfers (CIT)? What are the incentives for people, well and sick, to enrol in voluntary health insurance schemes?

We hope that the chapters in this volume will contribute to our understanding of what is known about the institutions and incentives moderating the behaviour of health service providers and consumers in xix

00 Improving Health Prelims

19

8/24/11, 4:21 PM

Introduction

xx

Asia and beyond, and how these can be incorporated into national policies for the improvement of health services in Cambodia and other countries. It is also hoped that the book will provide lessons learnt from Cambodia and other countries to increase our knowledge of incentives and of their relationships with the behaviour of health service providers and consumers that could prove useful for formulating policy recommendations for Cambodia and other nations. Moreover, we hope this volume provides different research methodologies that are reliable and could be recommended for future research in the health sector in Cambodia and elsewhere. Finally, we expect that the findings and conclusions from different studies in this collection will promote further study of issues in the health sector, especially the institutions and incentives that relate to health service strategy.

xx

00 Improving Health Prelims

20

8/24/11, 4:21 PM

Incentives for Deployment of Primary Health Care Workers

I Overview

01 Improving Health

1

8/24/11, 4:21 PM

1

V. Lin, L. Ridoutt, E. Brink, and B. Hollingsworth

2

01 Improving Health

2

8/24/11, 4:21 PM

Incentives for Deployment of Primary Health Care Workers

3

1 WHAT INCENTIVES ARE EFFECTIVE IN IMPROVING DEPLOYMENT OF HEALTH WORKERS IN PRIMARY HEALTH CARE IN ASIA AND THE PACIFIC?1 V. Lin, L. Ridoutt, E. Brink, and B. Hollingsworth

KEY MESSAGES Improving access to primary health care is important for prevention, early detection, and management of health conditions and also contributes to more efficient utilization of resources in a health system. However, workforce shortages and high staff turnover rates can be a major barrier to the access to quality primary health care. Health systems in Asian and Pacific countries are dominated by urbanbased tertiary care facilities that are considered by consumers to offer superior quality services, and by health practitioners to represent higher status employment. Experiences in both developed and developing countries suggest that personal factors (such as place of origin and location of training) are

01 Improving Health

3

8/24/11, 4:21 PM

V. Lin, L. Ridoutt, E. Brink, and B. Hollingsworth

4

important for health workforce recruitment, but workforce retention strategies need to pay attention to professional factors. Individual financial and non-financial incentives can be effective in improving the deployment of health workers in primary health care in rural settings,2 and could be promising for urban primary health care, but these incentives need to be used in combination within the context of the local environment so that incentives specifically target individual and location-specific requirements. Financial incentives for individuals tend to be more effective for recruitment than retention. Organizational incentives to improve the work environment and give better professional support to individual health workers are more important for workforce retention. A strategy that seeks to combine the skill bases of highly trained health professionals, lower skilled health workers, and the informal sector (including regulating traditional healers) can improve access to primary health care and, in doing so, may assist with workforce retention as well. Creating a workforce with adequate interest in and commitment to primary health care requires strategies to realign the curriculum and training locations of educational institutions.3 Ultimately, the status of primary health care within a health system will be important in attracting and retaining high-quality health workers. Adopting a system where primary health care is the first point of contact and referral, including for financing purposes (that is, a “primary health care as gatekeeper” system) is likely to be effective in this regard. Overall, individual, organizational, and system incentives should be applied. But workforce issues should be addressed concurrently with other building blocks of the health system. In the short term, policies directed towards improving individual material and professional incentives are most likely to yield a quicker return. However, policies that focus on systemic changes, such as new service models and the place of primary care within the health system, are best placed to deliver sustainable workforce results.

PURPOSE OF PAPER AND METHODOLOGY A major global health policy goal is the achievement of universal access to primary health care as a cost effective strategy integrating prevention, detection, and continuing care (Atun 2004; Starfield 2009). Policy measures

01 Improving Health

4

8/24/11, 4:21 PM

Incentives for Deployment of Primary Health Care Workers

5

required to achieve universal access will entail minimizing financial barriers to access, ensuring appropriately skilled human resources for health care, promoting appropriate care-seeking behaviour, and maintaining a good standard of health services. This paper summarizes a policy brief that focuses on one of the health workforce policy challenges for achieving universal access to primary health care (World Health Organization 2007), specifically: What are the most effective incentives to improve deployment and retention of health workers in primary health care in Asia and the Pacific region? The original longer policy brief synthesizes relevant literature concerning workforce incentives in primary health care in the Asia-Pacific region, and also draws from broader international literature where lessons may be relevant. Key databases were utilized, as well as grey literature from agencies such as the World Health Organization and World Bank. Evidence was derived from systematic reviews and high quality narrative reviews. However because of important contextual factors in the AsiaPacific region, lower quality narrative reviews and single studies from this region were also considered important. Primary health care (PHC) is seen here as first-contact care for prevention, detection, and treatment in a community setting, and the primary health care workforce is seen as those engaged in this in the formal health sector. The search revealed an abundance of literature globally, focusing on rural and remote settings, which are often PHC settings. However, there is very little literature on urban, underserved PHC settings for low- and middle-income countries, let alone the Asia-Pacific region in particular. While further research into the recruitment and retention in urban areas is clearly needed, this paper has drawn upon available literature about rural areas to the extent that they may be applicable, or at least hold lessons, for both urban and rural PHC.

CURRENT CHALLENGES TO WORKFORCE RECRUITMENT AND RETENTION Health systems in Asia and the Pacific are diverse — some are predominantly public and others are largely private. Primary health care is variously provided by solo practitioners and large organizations, by well regulated general practitioners and mid-level practitioners, unregulated private providers, and traditional healers.

01 Improving Health

5

8/24/11, 4:21 PM

V. Lin, L. Ridoutt, E. Brink, and B. Hollingsworth

6

Despite diversity in health care financing and governance arrangements, most health systems in Asia and the Pacific are dominated by the tertiary care sector. Primary health care services in general tend to be seen as low quality and unattractive to consumers. The preference of consumers to self-treat or seek attention from private and non-allopathic providers, or turn to tertiary hospitals (Boupha et al. 2005; Soeters and Griffiths 2003; Nguyen et al. 2005) reflects users’ perception of the low status of primary health care services, and, by implication, the workers in those services. From a labour market perspective, the primary health care services segment is, in many countries, a less attractive employment option (The World Bank 2009). While Asia has about 50 per cent of the world’s population, it only has 30 per cent of the global stock of doctors, nurses, and midwives (Joint Learning Initiative 2004), though the actual composition and challenges relating to the health workforce differ significantly from country to country across the region. Country data (World Health Organization 2006; World Health Organization 2009) suggest that the region has a lower worker density of doctors, nurses, and other health workers than any other region in the world, excluding sub-Saharan Africa. In many disadvantaged communities in both rural and urban areas, primary health care workers are scarce (Lao Ministry of Health 2005; Nguyen et al. 2005), and the utilization of their services may be low as well. Despite policy measures to improve the availability and distribution of the health workforce to primary care facilities, health workers still prefer not to go to or stay in primary health care (especially in underserved areas) for a range of complex reasons. Henderson and Tulloch (2008) have suggested reasons for health workers in the Asia-Pacific (Fiji, Tonga, Samoa, Vietnam, Papua New Guinea, Cambodia, Thailand, Pakistan, India, Sri Lanka, Vanuatu) not wanting to stay with their employers (low retention) as including: • Income: low salaries and the lack of adequate allowances; • Working conditions: inadequate facilities, shortages of drugs/ equipment, heavy workload, and difficulties with transportation; • Professional support: inadequate clinical supervision, weak support and management, few opportunities for professional development, limited scope to upgrade qualifications, and mismatches in skills and tasks;

01 Improving Health

6

8/24/11, 4:21 PM

Incentives for Deployment of Primary Health Care Workers

7

• Career prospects: shortage of job prospects and lack of promotion prospects/career structure; • Social amenities: inadequate living conditions, risk of violence/lack of safety, political instability, family members far away, and poor education prospects for children.

USE OF INCENTIVES — A POLICY FRAMEWORK Countries have used a variety of incentives to try to alter workforce behaviour and address the reasons for workforce shortages in underserved areas (Dolea et al. 2010), usually in rural regions and not necessarily in primary health care settings. These incentives have focused on financial and material benefits, as well as non-financial assistance, such as safety concerns, professional gains, and psychosocial benefits. Table 1.1 summarizes the most common incentives reported in the literature. Studies globally have generally shown that individual benefits (such as pay and extras) are important for recruitment, but professional benefits (including relationships and opportunities) are more important for retention (Dambisya 2007; Henderson and Tulloch 2008; Lehmann et al. 2008; WillisShattuck et al. 2008; World Health Organization and Asia Pacific Action Alliance on Human Resources for Health 2009). However, what incentives work will vary considerably among various health service delivery contexts and models of care, within health systems, and between individual workers. As seen in Table 1.1, incentives have been used to entice as well as coerce individual health workers to work in primary health care settings. Organizational incentives have also been applied to alter the work environment, and system incentives to alter the position, status, and nature of primary health care. Incentives are also used to increase performance and thus service quality, which may in turn have an indirect effect on both access to primary health care and workforce retention. Depending on the nature of the workforce shortage, different combinations of incentives will be required. Given the complexity of factors which influence health worker behaviour, Ridoutt and Santos (2006) suggest that Maslow’s hierarchy of needs may be a useful framework — as seen in Table 1.2 — for pointing to the range of factors that influence health workers’ decisions and suggest policy issues that require attention in order to attract and keep health workers.

01 Improving Health

7

8/24/11, 4:21 PM

V. Lin, L. Ridoutt, E. Brink, and B. Hollingsworth

8

TABLE 1.1 Common Incentives Incentives

Training/Recruitment

Financial — individual

Scholarship Mandatory service Higher salaries & bonuses

Financial — organizational

Non-financial — individual

Retention Higher salaries Dual practice

Public-private partnerships Pay for performance Working conditions Social amenities/ services

Non-financial — organizational

Continuing professional development Performance management Social amenities/services Working conditions Teamwork Clinical supervision Outreach services

Non-financial — health system

Workforce substitution Localized training

Regional health services Primary care gatekeeping and specialization

Source: Summarized from sources used in this re view.

EVIDENCE FOR EFFECTIVENESS OF INCENTIVES The most commonly used financial incentives in both developed and developing countries are incentives directed at individual workers. These can be financial and non-financial and include: (1) targeted scholarships (Bärnighausen and Bloom 2009; Sempowski 2004), (2) mandatory training and service requirements (which may be an indirect incentive insofar as exposing health workers to new experiences and environments) (Health and Human Resources Development Center 2007; Henderson and Tulloch 2008), and (3) higher salaries and other benefits (Health and Human Resources Development Center 2007; Christianson et al. 2009). These have all proved to be important in recruitment, but not sufficient for retention (Drager et al. 2006). For example, non-financial incentives —

01 Improving Health

8

8/24/11, 4:21 PM

Incentives for Deployment of Primary Health Care Workers

9

TABLE 1.2 Maslow’s Hierarchy Applied to Workforce Policy Development MASLOW’s NEEDS

WORKFORCE ISSUE

Physiological

Recruitment, absenteeism

Safety

Recruitment, absenteeism

Social

UNDERLYING CONCERN

ISSUE FOR ATTENTION

Job security, safety

Salary, working conditions, safety

Recruitment, turnover

Social environment, professional climate

Working relations, social networks, team work

Self-esteem

Retention, productivity

Professional status and recognition, future prospects

Feedback and rewards, career progress, attractive salary

Self-actualization

Retention, adaptiveness

Professional, social, and organizational climate

Professional challenge, self-development, autonomy, responsibility

Source: Adapted from Ridoutt and Santos 2006, p . 31.

such as improved living conditions, social networking, and opportunities for continuing education and professional development — are just as, or even more, important than financial incentives, in achieving retention (Henderson and Tulloch 2008; Lehmann et al. 2008). Some non-financial incentives can be directed at individuals (such as professional development support), but are also often applied through organizational and system reforms. Table 1.3 provides an overview of the most common incentives adopted in Asia-Pacific Region countries that are directed at the individual. It is at the organizational level that service quality and performance are linked to workforce motivation and possible retention. Promising incentives or strategies include: (1) creating a positive work environment through improved teamwork and clinical support, better leadership and management style, performance management systems, continuing education, career development, and the provision of social amenities and services (Dieleman et al. 2009; Henderson and Tulloch 2008; Human Capital

01 Improving Health

9

8/24/11, 4:21 PM

V. Lin, L. Ridoutt, E. Brink, and B. Hollingsworth

10

TABLE 1.3 Individual Incentives Used for Health Workforce Commonly Adopted in Asia-Pacific Countries Financial — individual

Non-financial — individual

Terms and conditions of emplo yment: • Income increases (Philippines) • Job security/tenure (Indonesia) • Pension (Indonesia) • Insurance (e.g. health) • Allowances (e.g. housing, clothing, child care , transportation, parking) (Indonesia) • Paid leave Performance payments for: • Achievement of performance targets (Cambodia, China) • Length of service (Indonesia) • Location or types of w ork (e.g. remote locations) (Thailand, New Zealand, Indonesia, Philippines) Other financial support: • Fellowships • Loans: approval, discounts • Scholarships (Australia, New Zealand, Nepal) • Payment of postgraduate training (Mongolia)

Support for career and professional development: • Effective supervision • Coaching and mentoring structures • Access to/support for training and education (Pacific Islands, China, Indonesia, Sri Lanka, India) • Sabbatical and study leave (Fiji) Access to services and amenities: • Health • Child care and schools • Family/Spouse occupational opportunities • Recreational facilities • Housing • Attractive social and community networks (Thailand, China)

Sources: Based on references used in the re view.

Alliance 2005; Lehmann et al. 2008); (2) ensuring the availability of necessary equipment and supplies (Chopra et al. 2008; Willis-Shattuck et al. 2008); and (3) innovative and collaborative new service models that increase access for underserved populations (Wakerman et al. 2008; World Health Organization 2008) and, at the same time, motivate the workforce. At the system level, it is possible to adopt structural and financing reforms. Structural reforms that are likely to be effective include: (1) reorienting primary health care as a specialization and as the “gatekeeper”,4 thus improving its status (Pagaiya and Noree 2009; Asian Development Bank 2008); (2) new system service models, such as regionalized networks that provide for career pathways; (3) developing a system of mid-level practitioners who are up-skilled to carry out some of the key tasks of high-level professionals who are difficult to recruit and

01 Improving Health

10

8/24/11, 4:21 PM

Incentives for Deployment of Primary Health Care Workers

11

retain (Pashen et al. 2007); (4) addressing labour market issues, including the location, recruitment, and training/socialization strategies of health care training institutions (Brooks et al. 2002; Henry et al. 2009; Laven and Wilkinson 2003; Rabinowitz et al. 2008), and using strategies that make the most of workers who are already there or are more willing to work in underserved areas (Chopra et al. 2008). For example, upskilling existing health workers (Araoyinbo and Bateganya 2008; Beney et al. 2000; Laurant et al. 2005), or training lay health workers (Haines et al. 2007; Hongoro and McPake 2004; Lewin et al. 2010); and (5) integrating informal, private, and traditional providers into the formal health system (Sharma 2001; World Health Organization 2004). Reforms of financing and payment mechanisms (such as health insurance, contracting and other public-private partnerships, and payment for performance) are less likely to have direct impacts unless they result in primary health care being the gatekeeper and user fees are removed. It is likely that a package of policy interventions is needed, rather than the adoption of singular approaches. The key interventions that may be of value are: developing the primary care gatekeeping role, providing competitive remuneration, developing positive work environments that offer autonomy but also strengthen teamwork opportunities, supporting small teams or solo practitioners with professional development, improving workplace infrastructure, and supporting a commitment to quality of care.

ASSESSING AND CHOOSING POLICY OPTIONS Following a review of the evidence, the authors derive conclusions about the best ways to assess and choose policy options: The choice of policy interventions for recruiting and retaining health workers in primary health care facilities will depend on consideration of available evidence as well as whether they can be adapted to local contexts. The key criteria for choosing policy options should include: • How likely they (the options) are to make a difference to workforce deployment; • Whether they will improve access (and equity of access) for appropriate disadvantaged groups (including those determined by age, gender, or ethnicity) to primary health care;

01 Improving Health

11

8/24/11, 4:21 PM

V. Lin, L. Ridoutt, E. Brink, and B. Hollingsworth

12

• Whether the cost is reasonable for the expected return; • Whether they are implementable. Policies that aim to provide the best incentives for the health worker (supply side) should be linked with policies that can increase access (demand side) to the appropriate quantity and quality of primary health care because incentives for the health workforce are only one element of a comprehensive strategy to build a high-performing primary health care system. Incentives for health workers complement such strategies as: (1) removing or minimizing financial barriers to access to primary health care (e.g. reducing out-of pocket cost, especially for those who do not have the ability to pay); (2) improving service delivery and providing innovative service models that are responsive to community needs; (3) having legislative frameworks to regulate provider behaviour so as to engender trust in health care institutions and health workers; and (4) enhancing health literacy and the community’s ability to use health resources well. Incentives for the health work force at whatever level can be initiated at any point, from local service management to countrywide programmes. However, workforce incentives need to be addressed alongside other policy measures to strengthen the primary health care system. The mix of incentives depends on the health system and organizational contexts. Furthermore, the degree to which the quality of health workers and health care services is deemed important by the community depends on the level of economic development. Despite these complexities, some indicative policy interventions for both recruitment and retention could be adopted according to health system contexts and practice settings, as shown in Table 1.4. Overall, a carefully focused package of financial and non-financial incentives needs to be devised on the basis of local context, organizational structure, institutional capacity, culture, and wider social values and expectations. Such a package must take into account the ease of implementation and monitoring, the cost and time frame for the package to take effect, and the sustainability of the package. Over time the policy measures adopted will need to evolve as the character of the health system changes. For example, Table 1.5 displays how policy measures evolve with increasing country development, changing population distribution, and the resultant changing characteristics of the health service.

01 Improving Health

12

8/24/11, 4:21 PM

Incentives for Deployment of Primary Health Care Workers

13

TABLE 1.4 Policy Interventions in Context Labour Market Context

Practice Setting

Workforce shortage

No workforce shortage

Solo practice

Better pay and other amenities Clinical support Continuous professional development Workforce substitution Outreach services/visiting specialists Telemedicine

Clinical support and upskilling Continuous professional development

Health centre

Outreach services Visiting specialists Higher salaries and better conditions Workforce substitution

Higher salary and better conditions Primary care as gatekeeper New service models (e.g. regional service networks) Clear role definition Pay for performance (linked to specific preventive services and outcomes) Performance management systems Performance monitoring

Source: Authors’ conclusions based on liter ature review.

For most low- and middle-income countries, a likely pathway of phased interventions would be: • In the short term, a focus on individual material and professional benefits, including targeted scholarships, mandatory training and service, better pay and working conditions and better clinical support, would appear to be the most beneficial. • In the longer term, a focus on changing the status of primary care and professional rewards for working in primary care is appropriate. Possible measures include: (1) reorienting primary health care as the “gatekeeper”; (2) multidisciplinary teams and new service models; (3) addressing the location; recruitment, and training strategies of undergraduate training institutions; and (4) developing a system of workforce substitution.

01 Improving Health

13

8/24/11, 4:21 PM

V. Lin, L. Ridoutt, E. Brink, and B. Hollingsworth

14

TABLE 1.5 Evolving Policy Measures According to Country Context Country context

Characteristics of health service

Possible policy measures

Low income, high proportion of population in rural areas

Limited facilities, low utilization of primary health care services, strong belief in traditional health care, brain drain of more skilled health workers to other countries

System/labour market: primary care workers’ substitution for doctors and other skilled workers, local training and placement in rural areas Individual: mandatory service in public health facilities and in underser ved areas; rural allowances

Medium income, high proportion of population in rural areas

Need to expand rural primary health care, growing consumer desire for quality services, some high standard tertiary care facilities in urban areas

System/labour market: increased production of doctors, nurses, primary care workers Individual: specialist training for career progression, but linked to rural service, social network to support doctors Organizational: improve infrastructure and equipment in pr imary health care services

Medium income, moderate proportion of population in rural areas

Well developed urbanbased tertiary care facilities, emerging private sector services, competition between public and private sectors for labour, growing consumer preference for quality Western medicine

Organizational: improve infrastructure and equipment in public sector Individual: career advancement measures for rural doctors, financial incentives for public sector practice and rural hardship System/labour market: local recruitment/training/placement, increased production of pr imary care workers and nurses

High income, low proportion of population in rural areas

Universal coverage provided, but increased concern about quality

Individual: special allowances for doctors, dentists, pharmacists, and nurses System/labour market: local recruitment/training/placement System: primary care as gatekeeper

Source: Adapted from Pagaiya & Norree (2009).

01 Improving Health

14

8/24/11, 4:21 PM

Incentives for Deployment of Primary Health Care Workers

15

Notes 1. This draft policy brief was commissioned by WHO and its Western Pacific Region and has been supported technically and financially by the Asian Development Bank and the World Bank. These three agencies are working together to establish an Asia Pacific Observatory on Health Systems and Policies, which will publish a full version of this brief once the review is finalized. 2. There is good evidence for rural areas, but a scarcity of evidence for underserved urban primary health care settings. More research is needed into recruitment and retention in underserved urban primary health care settings. 3. There is strong evidence for institutions and training in rural areas. Their location in poorer urban areas has not been adequately researched. 4. A gatekeeper role means that primary health care is known as the first point of contact, and may be financed to do so, and therefore manages referrals to higher levels as appropriate.

References Araoyinbo, I. & M. Bateganya. “Substitution of nurses for doctors in managing HIV/AIDS antiretroviral therapy” (Protocol). Cochrane Database of Systematic Reviews (Online), 2008. Available at (accessed 28 October 2010). Asian Development Bank. Programme Performance Evaluation Report in Mongolia. Mongolia: Health Sector Development Program; Mandaluyong City: ADB Operations Evaluation Department, 2008. Atun, R. “What Are the Advantages and Disadvantages of Restructuring a Health Care System to Be More Focused on Primary Care Services?” Copenhagen: World Health Organization Regional Office for Europe, 2004. Bärnighausen, T. & D. E. Bloom. “Financial Incentives for Return of Service in Underserved Areas: A Systematic Review”. BMC Health Services Research (NLMMEDLINE), 9 (2009): 86. Beney, J., L. Bero, & C. Bond. “Expanding the Roles of Outpatient Pharmacists: Effects on Health Services Utilisation, Costs, and Patient Outcomes”. Cochrane Database of Systematic Reviews (Online), 2000. Available at (accessed 28 October 2010). Boupha, B., S. Phomtavong, K. Akkhavong, & S. Xaisida. Strengthening the Quality of Human Resources for Health Oriented toward the District and Village Levels in Lao People’s Democratic Republic. Bangkok, Thailand: Lao Democratic Republic Ministry of Health National Institute of Public Health, 2005.

01 Improving Health

15

8/24/11, 4:21 PM

V. Lin, L. Ridoutt, E. Brink, and B. Hollingsworth

16

Brooks, R. G., M. Walsh, R. E. Mardon, M. Lewis, & A. Clawson. “The Roles of Nature and Nurture in the Recruitment and Retention of Primary Care Physicians in Rural Areas: A Review of the Literature”. Academic Medicine, no. 77 (2002): 790–98. Chopra, M., S. Munro, J. Lavis, G. Vist, & S. Bennett. “Effects of Policy Options for Human Resources for Health: An Analysis of Systematic Reviews”. The Lancet, 371, 2008. Christianson, J., K. Sutherland, & S. Leatherman. Financial Incentives, Healthcare Providers and Quality Improvements: A Review of the Evidence. London: The Health Foundation, Quest for Quality and Improved Performance (QQUIP). Dambisya, Y. “A Review of Non-financial Incentives for Health Worker Retention in East and Southern Africa”. EQUINET discussion paper no. 44, with ESC A-HC. Harare, Zimbabwe: Regional Network for Equity in Helath in East and Southern Africa (EQUINET) and the East, Central and Southern African Health Community, 2007. Dieleman, M., B. Gerretsen, & G. J. Van Der Wilt. “Human Resource Management Interventions to Improve Health Workers’ Performance in Low and Middle Income Countries: A Realist Review”. Health Research Policy and Systems, 2009, p. 7. Dolea, C., L. Stormont, & J.-M. Braichet. “Evaluated Strategies to Increase Attraction and Retention of Health Workers in Remote and Rural Area”. Bulletin of the World Health Organization, no. 88 (2010): 379–85. Drager, S., M. Dal Poz, & D. Evans. “Health Workers Wages: An Overview from Selected Countries”. Background paper for the World Health Report 2006 — “Working Together for Health”. Geneva: Evidence and Information for Policy, World Health Organization, 2006. Haines, A., D. Sanders, U. Lehmann, A. K. Rowe, J. Lawn, S. Jan, D. Walker, & Z. Bhutta. “Achieving Child Survival Goals: Potential Contribution of Community Health Workers”. The Lancet, 369, no. 2121, 2007. Health and Human Resources Development Center. Annual Review of HRH situation in Asia-Pacific Region 2006–2007. Beijing: Ministry of Health, People’s Republic of China, 2007. Henderson, L. & J. Tulloch. “Incentives for Retaining and Motivating Health Workers in Pacific and Asian Countries”. Human Resources for Health, 2008, p. 6. Henry, J., B. Edwards, & B. Crotty. “Why do Medical Graduates Choose Rural Careers?” Rural and Remote Health, 2009, p. 9. Hongoro, C. & B. Mcpake. “How to Bridge the Gap in Human Resources for Health”. The Lancet, 364, no. 1451, 2004. Human Capital Alliance. “Recruitment and Retention of Allied Health Professionals in Victoria — A Literature Review”. Report to Department of Human Services. Melbourne: Human Capital Alliance, 2005.

01 Improving Health

16

8/24/11, 4:21 PM

Incentives for Deployment of Primary Health Care Workers

17

Joint Learning Initiative. Human Resources for Health: Overcoming the Crisis. Cambridge: Harvard University, Global Equity Initiative, 2004. Lao Ministry of Health, N. I. O. P. H. Strengthening the Quality of Human Resources for Health Oriented toward the District and Village Levels in Lao People’s Democratic Republic. Lao Ministry of Health, National Institute of Public Health, 2005. Laurant, M., D. Reeves, R. Hermens, J. Braspenning, R. Grol, & B. Sibbald. “Substitution of Doctors by Nurses in Primary Care”. Cochrane Database of Systematic Reviews (Online), 2005. Available at (accessed 28 October 2010). Laven, G. & D. Wilkinson. “Rural Doctors and Rural Backgrounds: How Strong is the Evidence? A Systematic Review”. Australian Journal of Rural Health, no. 11 (2003): 277–84. Lehmann, U., M. Dieleman, & T. Martineau. “Staffing Remote Rural Areas in Middle- and Low-income Countries: A Literature Review of Attraction and Retention”. BMC Health Services Research, no. 8. Lewin, S., S. Munabi-Babigumira, C. Glenton, K. Daniels, X. Bosch-Capblanch, B. Van Wyk, J. Odgaard-Jensen, M. Johansen, G. Aja, M. Zwarenstein, & I. Scheel. “Lay Health Workers in Primary and Community Health Care for Maternal and Child Health and the Management of Infectious Diseases”. Cochrane Database of Systematic Reviews (Online), 2010. Available at (accessed 29 October 2010). Nguyen, B., B. Nguyen, & N. Lh. “Human Resources for Health in Vietnam and the Mobilization of Medical Doctors to Commune Health Centres”. Asia Pacific Alliance on Human Resources for Health country reviews (Online), 2005. Available at (accessed 16 June 2010). Pagaiya, N. & T. Noree. Thailand’s Health Workforce: A Review of Challenges and the Experiences (version 2). Bangkok: International Health Policy Program, Ministry of Health, 2009. Pashen, D., B. Chater, R. Murray, V. Sheedy, C. White, L. Eriksson, S. De La Rue, & M. Du Rietz. The Expanding Role of the Rural Generalist in Australia — A Systematic Review. Brisbane: Australian College of Rural and Remote Medicine, 2007. Rabinowitz, H., J. Diamond, F. Markham, & J. Wortman. “Medical School Programs to Increase the Rural Physician Supply: A Systematic Review and Projected Impact of Widespread Replication”. Academic Medicine, no. 83, 2008. Ridoutt, L. & T. Santos. Workplace Health Promotion and Organisational Change in Recruitment & Retention of the Health & Community Services Workforce. Melbourne: Department of Human Services, 2006. Sempowski, I. “Effectiveness of Financial Incentives in Exchange for Rural and Underserviced Area Return-of-service Commitments: Systematic Review of the Literature”. Canadian Journal of Rural Medicine, no. 9 (2004): 82–88.

01 Improving Health

17

8/24/11, 4:21 PM

V. Lin, L. Ridoutt, E. Brink, and B. Hollingsworth

18

Sharma, D. C. “India to Promote Integration of Traditional and Modern Medicine”. The Lancet, 358, no. 1524, 2001. Soeters, R. & F. Griffiths. “Improving Government Health Services through Contract Management: A Case from Cambodia”. Health Policy and Planning, 18, no. 1 (March 2003): 74–83. Starfield, B. “Toward International Primary Care Reform”. Canadian Medical Association Journal, no. 180 (2009): 1091–92. The World Bank. “What Health Workers Want: Challenges and Choices in Human Resource Management in Lao PDR”. First draft for discussion on 13 September 2009. Vientiane: World Bank, 2009. Wakerman, J., J. Humphreys, R. Wells, P. Kuipers, P. Entwistle, & J. Jones. “Primary Health Care Delivery Models in Rural and Remote Australia — A Systematic Review”. BMH Health Services Research, no. 8, 2008. Willis-Shattuck, M., P. Bidwell, S. Thomas, L. Wyness, D. Blaauw, & P. Ditlopo. “Motivation and Retention of Health Workers in Developing Countries: A Systematic Review”. BMC Health Services Research, no. 8 (2008): 247. World Health Organization. Review of Traditional Medicine in the South-East Asia Region. New Delhi: WHO South East Asia Regional Office, 2004. World Health Organization. The World Health Report 2006. Working Together for Health. Geneva: World Health Organization, 2006. World Health Organization. “Everybody’s Business. Strengthening Health Systems to Improve Health Outcomes”. WHO’s framework for action. Geneva: WHO, 2007. World Health Organization. World Health Report: Primary Health Care — Now More Than Ever. Geneva: WHO, 2008. World Health Organization. “WHO Global Atlas of the Health Workforce” (Online). Geneva: WHO, 2009. Available at (accessed 15 September 2009). World Health Organization & Asia Pacific Action Alliance on Human Resources for Health. Conference proceedings in the joint WHO/AAAH conference on Getting Committed Health Workers to Underserved Areas: A Challenge for Health Systems”. Hanoi, Vietnam, 2009.

01 Improving Health

18

8/24/11, 4:21 PM

Reforming Provider Behaviour through Incentives

19

2 REFORMING PROVIDER BEHAVIOUR THROUGH INCENTIVES Challenges and Reflections from the U.K. Experience Alan Maynard

INTRODUCTION Before exploring theories of motivation, trust, incentives and regulation, it is important to clarify a number of fundamental terms in health policy. These terms are often used by policymakers and academics in an inappropriate manner which confuses policy debate. Following this, the similarities of public and private health care systems throughout the world are outlined, and common deficiencies in health care markets are reviewed. Next the motivation of health care professionals and the roles of trust and incentives in delivering health care are explored. The regulation of health care markets, and interventions such as pay for performance (P4P) incentive schemes, reflect an implicit judgment by policymakers that they do not trust the professionals who determine the nature and timing of

02 Improving Health

19

8/24/11, 4:22 PM

Alan Maynard

20

health care. The dangers this creates are addressed before financial incentives are reviewed. This discussion poses questions such as whether it is more efficient to incentivize individual practitioners, teams, or organizations such as the hospitals in which they work. Are bonuses more or less efficient than penalties in inducing changed behaviour? Throughout this paper, the need for careful experimentation and systematic evaluation in the development of health policy and reform is emphasized. Without this, powerful incentives may produce perverse outcomes that fail to give policymakers and societies expenditure control, value for money, efficiency, and equity or access to care for the population.

1. WORDS HAVE MEANINGS: SOME DEFINITIONS 1.1 Structure, Process, and Outcome The U.S. health services researcher Donabedian (1966) emphasized the need to distinguish between structure, process, and outcome in health care. The complex policy issue is whether the reform of the organizational structures through which health care is delivered affects patient care, or outputs, and if so, whether the changes improve patient outcomes. There is a propensity among policymakers to assume that if they alter organizational structures, this will improve care and benefit patients. But this is generally an evidence-free assumption. If reform takes place, these linkages have to be evaluated and proved rather than assumed.

1.2 Policy Goals When societies invest in health care, what goals are they pursuing? How are these goals ranked? And what trade-offs have to be made among competing goals?

1.2.1 Equity In Europe during the late nineteenth century (Bismarck) and early twentieth century (Lloyd George and Beveridge), the social goal of these innovative policymakers was to make the population healthier by improving access to health care. More recently the Chinese and even (!) the United States have reformed their legislation in order to improve access to health care for the poor and, like the Europeans before them,

02 Improving Health

20

8/24/11, 4:22 PM

Reforming Provider Behaviour through Incentives

21

made such reforms both for altruistic and also self-interest reasons, for example, to maintain social cohesion or “harmonious development” and also to have a healthy workforce. A difficult issue for reformers pursuing this goal is the extent to which health, the desired outcome, is affected by health care inputs and processes. The major determinants of population health appear to be genetic endowment, individual behaviour, income, and education (Evans et al. 1994). The health care industry clearly has an important role in prevention policies such as immunization, vaccination, and basic maternal care, and “repair interventions”, following trauma and for common surgical conditions. Beyond that, what is the relative cost effectiveness of investing in health care or these other determinants of health?

1.2.2 Efficiency Efficiency or “value for money” is a relationship between the value of what is given up when a patient is treated (what economists call opportunity cost) and the value of what is gained, the health outcome benefit to patients (Maynard 1997). This concept differs from a common definition of clinical value: clinical effectiveness or whether the intervention affects the length and quality of the patient’s life. It is useful to remember that what is clinically effective may not be efficient, but what is efficient is always clinically effective (Maynard 1997).

1.2.3 Expenditure Control Macroeconomic expenditure control is always a concern for policymakers because fiscal ineptitude can, on occasion, create local political turmoil as well as the possibility of World Bank-IMF intervention in fiscal and monetary policy. There is a simple accounting identity between household incomes and health care expenditure. Households can contribute to health care funding via general taxation, social insurance contributions (a form of disguised taxation), private insurance premiums, and user charges. The choice of funding will be determined by local ideology, tradition, and public acceptability. Restricting one funding “pipe” may affect flow volume in another pipe, thereby possibly frustrating financial control.

02 Improving Health

21

8/24/11, 4:22 PM

Alan Maynard

22

The funds from these sources determine incomes and employment of providers such as health care professionals, hospitals, and pharmaceutical companies. Thus what households give up is what providers receive as rewards: the income-expenditure identity. Typically, single payer or taxfinanced systems of health care funding offer better expenditure control, with private insurance systems usually experiencing premium inflation rates two to three times the rate of general inflation (Wennberg 2002). Private insurers, like public purchasers, tend to be poor at controlling inflation and ensuring the efficient use of scarce resources.

1.2.4 Trade-offs What is efficient may not be equitable. For example, keeping very low birth weight babies alive or carrying out cardiac surgery on smokers may not be efficient, but society may choose to care for such people, thereby giving up health gain (efficiency) to achieve its view of equitable access. Choices such as these tend to be implicit (that is, the opportunity costs are often unclear), with clinicians exercising local discretion.

1.3 Productivity Throughout the world, but particularly in countries significantly affected by the world recession, there is a new categorical imperative: “to improve productivity” in public services. But what does this mean? Is productivity a relationship between inputs and outputs (activities or processes) or, in health care, is it the relationship between inputs and patient outcomes? This crucial distinction is often ignored. Real productivity gain involves either the use of less inputs such as labour, drugs, and hospital time to produce the same patient outcomes or health gains for patients, or increased gains in patient health with the same inputs. Policymakers who focus on the input-output relationship risk producing more activity or processes of care, but not improving patient health. Greater focus on the input-patient outcome relationship is necessary, but difficult given the evidence base and poor measures of whether spending on health care (a process) actually improves patient health (the desired outcome). Focusing on what is measurable may mean the production of more activity in the health care system but little or no production of improved health outcomes for the population.

02 Improving Health

22

8/24/11, 4:22 PM

Reforming Provider Behaviour through Incentives

23

1.4 Overview The reform of health care systems tends to focus on structural reorganization and the delivery of health care processes with too little regard for whether structural and process changes improve patient health. Reformers may pursue improvements in equity (particularly in access to care), efficiency, and macroeconomic expenditure control. Trade-offs between these goals are usually implicit, but a ranking in terms of relative importance is useful as a means of informing the evaluation of the effects of reform.

2. COMMON PROBLEMS IN HEALTH CARE SYSTEMS There are five common problems in all public and private health care systems: 1. 2. 3. 4. 5.

a poor evidence base about “what works” or what is clinically effective, let alone cost effective, in health care; large variations in what clinicians deliver to similar patients with similar needs and characteristics; avoidable medical errors that endanger patient safety; failure to exploit the potential of skill mix changes; a paucity of outcome measures, that is, demonstration that the use of inputs and processes improves patient outcomes and shows improvements in their length and quality of life.

2.1 What “Works” in Health Care? Which health care interventions improve patient health? Is this information used to prioritize investments in care? The evaluation of medical care used in developed countries indicates that there are many interventions that are widely used, but have never been evaluated in well designed and executed randomized clinical trials. An example of this is from the British Medical Journal Evidence Centre (2009), which concludes that: 1.

02 Improving Health

11 per cent of interventions have been evidenced as beneficial to patients;

23

8/24/11, 4:22 PM

Alan Maynard

24 2. 3. 4. 5. 6.

23 per cent of interventions are likely to be beneficial; 7 per cent of interventions carry a trade-off between benefits and harm; 5 per cent of interventions are unlikely to be beneficial; 3 per cent of interventions are likely to be ineffective or harmful; 51 per cent of interventions are of unknown effectiveness.

These startling “guesstimates” of what works in medicine have two important implications. First, those interventions in the 51 per cent category may improve patient health, but this cannot be easily tested in randomized controlled trials because the use of such a technique would be unethical in many health care markets. Consequently, practitioners may use such techniques and support their choices of treatment with data drawn from “experience” or non-trial evaluative methods. The second implication is that there is a need to focus new investment on the delivery of interventions that are known to benefit patients, that is, the 11 and 23 per cent categories. Sadly, this priority has proved difficult to implement, with the result that proven, efficacious interventions such as vaccination, immunization, and the delivery of cheap generic drug interventions to patients with chronic diseases are not provided in most health care systems, resulting in avoidable mortality and morbidity.

2.2 Variations in Clinical Practice For more than forty years, researchers in North America and Europe have been demonstrating large variations in clinical practice. The problem is that, while variation is normal, patients with similar personal characteristics and health needs are being given very different medical interventions. Hence, even when there is an evidence base of clinical effectiveness and efficiency (the 11 and 23 per cent categories in the preceding section), health care systems and the practitioners working in them fail to deliver those interventions that patients need and which would benefit them. For instance, a study by the Rand Corporation concluded that the majority of people in the United States were not receiving the health care they needed for the conditions they had (McGlynn et al. 2003). The evidence about these wasteful clinical practice variations is considerable. Both in the United States (Wennberg & Gittelsohn 1973; Wennberg et al. 1987; Fisher et al. 2003) and Britain (Department of Health

02 Improving Health

24

8/24/11, 4:22 PM

Reforming Provider Behaviour through Incentives

25

and Social Security 1976; Maynard & Bloor 2003; NHS Institute for Innovation and Improvement 2009), the issue has been chronicled for decades. However, the policy response has been muted and mostly characterized by advocating change rather than incentivizing decision makers, in particular, health care professionals, so that they alter their behaviour. The magnitude of the savings that could be captured if clinical practices were more uniform and evidence-based is much debated. The group at the Dartmouth Medical School in the United States have analysed Medicare data for several decades and are convinced that perhaps 30 per cent of spending could be saved if the overproviders changed their practices to those of safe, conservative physicians (Fisher et al. 2003). These researchers also pointed out in 2009 that such savings could fund a considerable part of the Obama administration’s health care reforms (Fisher et al. 2009). In England the NHS Institute for Innovation and Improvement has argued that if practitioners changed their practice, savings of up to £9 billion would be achievable (NHS Institute for Innovation and Improvement 2009). The problem of clinical practice variations has been discussed for decades, and the waste they produce is visible in all health care systems. It is curious that policies to reduce clinical practice variations and free scarce resources have been weak and ineffectual worldwide.

2.3 Medical Errors and Patient Safety Errors in manufacturing and service industry production processes can waste money and kill citizens. In health care systems, errors are also costly and deadly. There is a consensus that in Western health care systems, medical error rates are about 10 per cent of all hospital admissions, that is, one in ten patients is damaged in some way. In large developed countries, this is equivalent to avoidable mortality rates of several jumbo jet disasters each day. There are many types of errors whose rates can be reduced by measurement, benchmarking and improved management by clinical teams, for example: 1.

02 Improving Health

wrong site surgery — e.g. recently an Irish surgeon removed the healthy kidney rather than a diseased kidney from a nine-year-old boy;

25

8/24/11, 4:22 PM

Alan Maynard

26 2. 3.

wrong drug (pharmaceutical) or the wrong dose of the right drug; hand hygiene: poor practices lead to avoidable infections.

International attention was drawn to these issues by a US report, entitled “To Err is Human” (Committee on Quality Health Care in America 1999) and changes in practice internationally have been energized by the Institute for Healthcare Improvement (IHI) (www.ihi.org). They have demonstrated that some simple changes in practices can generate significant patient and financial benefits. However, the economic evidence base for many of the multitude of policy initiatives advocated by the IHI has yet to be proven.

2.4 Skill Mix Issues The quantity and quality of training received by physicians and nurses varies from country to country, and this complicates analysis of skill mix options. In Germany physicians are omnipotent, and their power and the system of their remuneration (fee for service) make the use of skilled nurses as complements, let alone substitutes, very difficult. In Britain 30,000 nurses have the authority to prescribe the full pharmaceutical formulae and carry out many tasks regarded as physician domains. The policy issues in skill mix centre on the notions of general and specific training (Becker 1962). What medical skills might we expect of physicians, nurses, and other skill groups? — the ability to act humanely and considerately through good communication skills, the ability to measure and interpret body temperature and blood pressure, the ability to administer pharmaceuticals, the ability to vaccinate, immunize, and set up drips, and domestic skills in feeding, toileting, and cleaning patients. These general skills are complemented by specific skills of particular professions. For instance, physicians need skills in assessing patient data, signs, and symptoms of illness, and specific training to determine efficient treatment and prognosis. Many medical tasks are routine and mundane, requiring general skills that can be acquired at low cost. Where specific skills are claimed to exist, there is a tendency for unwarranted claims about complexity, sometimes used as a means of maintaining craft monopolies and income — for example, many common surgical procedures, routine anaesthesia, radiology, and pharmaceuticals. The policy issue in

02 Improving Health

26

8/24/11, 4:22 PM

Reforming Provider Behaviour through Incentives

27

these areas is not who should carry out these roles, but ensuring that patient outcomes are produced efficiently, regardless of the skill used to produce care. As countries become more affluent, the division of labour between the health professions is reinforced by separate “trade” organizations and separation in regulation. This “separateness” is reinforced by training and education systems, and the result is that the professions become separated into silos where self-interest rather than public interest may dominate. Furthermore, this separation and the notions of “clinical autonomy” and the lack of transparency and accountability inhibit the efficient use of teamwork, which in chronic conditions is usually an essential part of patient care. The early evidence on physician-nurse substitution in primary care is nearly forty years old (Spitzer et al. 1974). The authors of the systematic review of randomized trials by the Cochrane Collaboration concluded that nurses could carry out most primary care tasks as well as physicians. Nurses appeared to be slower than physicians, but better liked by patients (Laurant et al. 2005). The cost implications of substituting nurses for physicians are poorly researched. In addition to this literature demonstrating substitution possibilities in primary care, there is an emerging knowledge base on the use of nurses in secondary care, for example, in conducting endoscopy (Williams et al. 2009). Despite this emerging knowledge base, the use of nurses to replace physicians and the use of nursing assistants to replace nurses is slow, being inhibited by powerful professions that seek to protect their incomes and employment. Many routine and common tasks in health care do not need physicians, but often local politics results in debates about “physician shortages” when none exists if substitution possibilities are exploited and inefficient skill and qualification inflation is avoided.

2.5 Outcome Measurement It is much easier to benchmark performance using process-output data and avoid consideration of whether increased activity and health care outputs actually improve patient health. This issue was well discussed by British policy reformers in the 1840s. They designed the Lunacy Act (Lumley 1845) which required all practising physicians to report annually whether their patients were:

02 Improving Health

27

8/24/11, 4:22 PM

Alan Maynard

28 1. 2. 3. 4.

dead recovered relieved unrelieved

This simple system of outcome measurement was reinforced by what is now called “pay for performance” (P4P). Physicians failing to record these data could be fined £2 (nearly £10,000 in 2010 prices). There was much opposition to this system of outcome measurement although it was adopted in acute as well as psychiatric hospitals in subsequent decades. An important issue was whether and how to casemix adjust the reported data. Was it acceptable to report crude data? Or was it necessary to adjust the reported results for the risk characteristics of patients, for example, differences in age, gender, co-morbidities, income, and education? This debate continues today in the form of how best to report standardized mortality rates (SMRs) of hospitals and physicians. President Reagan’s administration was the first to publish mortality rates for U.S. hospital used by Medicare patients (the elderly). Many hospital managers rejected these data as inaccurate, but the government pointed out that they were derived from hospital returns. So if they were inaccurate, it was caused by the managers filing poor data! Currently the Institute for Health care Improvement is focused on reducing inpatient mortality rates. Again there is the complex issue of how to adjust crude mortality rates. Gradually a consensus is emerging in Britain as competing groups seek to agree on an acceptable common Hospital SMR (HSMR). No risk adjustment will achieve “the right result”, but agreement provides a benchmark by which performance can be interrogated. A consensus measure will facilitate greater clinical and management focus on outliers and explain their behaviour. Furthermore, such a consensus HSMR provides a measurement mechanism by which experimental interventions can be evaluated. For instance, a recent study using evidence-based bundles of care for conditions causing death in hospitals was shown to reduce HSMR, at least in the short run, and at an unknown cost (Robb et al. 2010). Mortality data are an incomplete outcome measure. The purpose of investing in health care is to avoid premature mortality and to improve the quality of life of patients. Many health care interventions improve the

02 Improving Health

28

8/24/11, 4:22 PM

Reforming Provider Behaviour through Incentives

29

physical and psychological functioning of patients. Clinical trialists over the last four decades have designed and fine-tuned patient-reported outcome measures (PROMs). Generic measures such as EQ5D (www.euroqol.org) and short form 36 (www.sf36.org), which can be used across different clinical therapies, have been used in thousands of clinical trials and translated into dozens of languages. In addition, measures specific to particular conditions have been used extensively. In an effort to emulate the architects of the 1845 Lunacy Act, the English Government has funded the collection of PROMs data for four simple surgical conditions since April 2009. The types of surgery covered are listed in Table 2.1, as are the quality of life measures used. As can be seen, the government after some piloting work has chosen (Browne et al. 2007) to use EQ5D as its preferred generic measure and the Oxford Hip Score as a specific measure for hip and knee replacement procedures. The results from this initial PROMs investment will appear cumulatively from 1 September 2010. For each of the procedures in Table 2.1, PROMs data are collected before, and at three to six months after the intervention. Pilot work has shown that quality of life gains for hip and knee replacements are generally reported by 80 per cent of patients. Reported success rates in the “pilot” for hernia repair were lower, due to surgery for asymptomatic patients. Data from the full rollout of the policy will be used to inform patient choice by revealing the comparative success rates of clinicians and hospitals. Also the data will inform clinical audit and revalidation of physicians by the General Medical Council.

TABLE 2.1 Measuring Patient Outcomes in the English NHS Procedure

Condition-specific

Generic

Primary unilateral hip replacement Primary unilateral knee replacement Groin hernia repair Varicose vein procedures

Oxford Hip Score Oxford Hip Score None Aberdeen Varicose Vein Questionnaire

EQ5D EQ5D EQ5D EQ5D

Plus a standard set of patient-specific questions in all cases Source: DH Operating Framework, Guidance on the routine collection of patient-repor ted outcome measures, Department of Health 2009.

02 Improving Health

29

8/24/11, 4:22 PM

Alan Maynard

30

2.6 Overview The provision of health care is inefficient worldwide. The industry consumes large amounts of each country’s GDP, but the evidence base of many interventions is absent or incomplete. What is recognized as clinically effective and efficient is often not delivered to patients. There are large variations in clinical practice, and this waste deprives potential patients of care from which they could benefit. Medical errors are ubiquitous, poorly managed, and dangerous to patients’ health. Management focus on the efficient skill mix to deliver care is absent, and professional groups defend their interests rather than focus on the interests of patients and taxpayers. Outcome measurement is poor as decision makers focus on outputs rather than demonstrably improving patient welfare. All these issues have been researched and debated for at least four decades, but policymakers have failed to reform professional practices in ways that use resources efficiently. The medical professions and the powerful provider organizations in which they work continue to demand, and often receive, more resources even though the population health gain is not demonstrable. Politicians and policymakers erroneously assume that more inputs will produce more outputs or processes of care, which will automatically improve patient outcomes. Relationships between structure, process, and outcome have to be proved and prioritized in relation to their relative efficiency, rather than assumed by vote-maximizing politicians, and profit-maximizing professions and public and private bureaucracies. This “faith”-based policymaking is being challenged by addressing the issue of incentives. All health care systems have incentives built into their structures, processes, and professions. It is these incentives that create the perverse behaviours and inefficiency detailed above. Could the reform of incentives lead to increased efficiency in the use of society’s health care resources? An essential element in answering this question is the use of robust evaluation techniques (Maynard & Chalmers 1997; Cook & Campbell 1979; Campbell & Russo 1999). Without such rigour, there is a risk that health care reform will be little more than “shifting the deckchairs on the Titanic”, that is, altering health care structures and processes with no knowledge of, or learning about, whether such changes economize the use of inputs and/or improve patient outcomes.

02 Improving Health

30

8/24/11, 4:22 PM

Reforming Provider Behaviour through Incentives

31

3. INCENTIVES ARE IMPORTANT Markets are networks of buyers and sellers. These buyers and sellers may be public government agencies or private organizations such as insurers, pharmaceutical companies, and professional associations. The behaviour of these supply-side and demand-side organizations is determined by incentives, but incentives can be explicit and implicit, and financial and non-financial. Incentives are important because they strongly influence how individuals, professions, and institutions perform. The effects of incentives have to be evaluated in relation to society’s objectives, in particular, equity in access to care, the efficient use of resources, and expenditure control. Changed incentives should be evaluated on how they improve performance in meeting objectives. There are many types of incentives. The primary distinction is between non-financial and financial ones.

3.1 Non-financial Incentives The creation of professional associations is usually advocated on the basis of quality control and patient safety. However worthy this purpose, professional associations also tend to become the means by which professionals improve their employment and remuneration. Thus there is a potential clash between the public interest of patient safety and the enhanced quality of care and the self-interest of the members of the profession. The Irish playwright and social activist, George Bernard Shaw, described professions as “a conspiracy against the laity” (Shaw 1908). The libertarian economist, Milton Friedman, was also very critical of professions and their capacity to use their power to reward themselves rather than benefit the consumer (Friedman 1962). Market regulation is a statement that the buyers and sellers in a market are no longer trusted to maximize the interests of the consumer. Trust is the primary means of transaction in most markets. Confucius argued that if a ruler has weapons, food, and trust, he should give up weapons first, food second, and trust last. He emphasized that “without trust, we cannot stand”.

02 Improving Health

31

8/24/11, 4:22 PM

Alan Maynard

32

All contracts are incomplete: that is, they cannot cover all aspects of transactions in the marketplace. The more detailed the contract, the higher the cost of policing compliance with its terms and conditions. This is the same with market regulation by professions and government: the more complex the attempted control of prices, quality, and quantities, the higher the costs of monitoring behaviour. Furthermore, the greater the complexity of regulation and contracts, the greater the incentive for the funders and providers of health care to “game” or cheat. With contracts incomplete and regulation of public and private market activity so costly, the potential benefits of trust are obvious. If professions regulate themselves, this may be more efficient than regulation by an outside body, which would have to invest significantly in information acquisition, analysis, and policing. If organizations such as clinics and hospitals have rigorous systems of internal performance management and can be trusted to deliver high quality and safe care for patients, this lessens the need for government regulation. The principle of trust is thus important and may be an efficient mechanism for the governance of transactions. For most people, we might hypothesize that each day they go to work to do a good job (Akerlof & Kranton 2010). If employers, professions, and other regulators scrutinize their activity oppressively, this action, which demonstrates lack of trust between employer and employee, may undermine efficiency. The problematic and pragmatic issue faced by public and private regulators is how far trust can be relied on to govern transactions in any market. The “failures” of the health care market are evidence of the inability of trust relationships to deliver efficiency, equity, and expenditure control. The self-interest of professionals, workers, clinics, hospitals, and the pharmaceutical industry has eroded trust and created regulatory interventions. The issue is whether these interventions improve market efficiency or reinforce the problems they are designed to mitigate. Their effect is determined by their design and implementation and by learning from careful evaluation of such “social experiments” (Campbell 1969). Two obvious issues in non-financial incentives are transparency and accountability. Transparency requires the collection of performance data and benchmarking performance. This can be simple or complex. For instance, creating accurate data about work attendance and performance of primary duties (for example, being present and processing particular numbers of patients at each clinic session) provides time series information

02 Improving Health

32

8/24/11, 4:22 PM

Reforming Provider Behaviour through Incentives

33

and affords cross-sectional comparison. Data such as this enable professional groups and managers to identify “outliers” and the possibility of learning from high performers and improving the delivery rate of low performers. Simple activity rate data can be augmented by performance criteria that can be monitored. For instance, the care package for a patient who has a myocardial infarction (heart attack), heart failure, pneumonia, or who needs a hip replacement, should be explicit and evidence based. Adherence to such protocols can be monitored. Given that patients differ, professional peers and managers might expect a 70–80 per cent adherence to protocol, with a more detailed review of exceptions and adverse patient outcomes. The pursuit of accountability is complicated by populations living longer. Longer life for citizens is a major achievement of economic and social progress. However, this triumph is characterized by the elderly having complex morbidities that may be managed over many years. The current state of medical science is characterized by treatment protocols designed for the “best” treatment of individual morbidities. There are few protocols for the treatment of patients with multiple morbidities. Consequently, there is a risk that following one protocol may adversely affect the patient’s welfare in relation to another condition; for example, if a patient has renal failure and cancer, the optimum treatment of the cancer may reduce renal function. In such circumstances, clinical choices of treatment involve very careful trade-offs in order to do what is best for the patient. With professionally agreed, evidence-based protocols, transparency is increased. Provided this transparency is agreed by the professions and evidenced, accountability can be enhanced. The age of “clinical autonomy”, when it meant that clinicians could do what they liked for patients, is dead, and, as expressed by a U.K. surgeon, “no-one need regret its passing” (Hampton 1983). Physicians and other health care professionals are responsible for rationing health care. Rationing care involves depriving patients of care from which they could benefit and which they wish to consume. In a world of scarce economic resources, inefficiency in rationing care on grounds of cost effectiveness is unethical because such waste deprives potential patients of care from which they could benefit. The collection of performance data in relation to process protocols and outcomes, when presented simply to practitioners and managers, can be used to encourage reduction in poorly performing outliers and the improvement of average performance. The use of simple production

02 Improving Health

33

8/24/11, 4:22 PM

Alan Maynard

34

engineering techniques introduced to the Japanese car industry by Deming (www.deming.org) and later systematized by Motorola as “six sigma”, can transform quality and safety in manufacturing and in health care. Six sigma involves creating statistical distributions of processes and outcomes and then focusing on their “tails”, that is, on where there are “good” performers (more than three standard deviations above the mean performance score) to learn how to improve performance, and focusing on “poor” performers (more than three standard deviations below the mean) so as to work with them to improve performance. Approaches such as this produce an emphasis on “lean” production techniques that involve the analysis of existing processes and subjecting them to debate and change: for example, how much time does a nurse spend with patients as opposed to doing other tasks? In the United Kingdom, nurses appear to spend only 25 per cent of their time in direct patient care. By careful redesign of the work environment and task processes, this can be increased to 40 per cent and more. In summary, non-financial incentive mechanisms dominate management efforts to improve equity, efficiency, and expenditure control in public and private health care systems. Despite the gradual spread of techniques such as “lean” and “six sigma”, public and private health care systems continue to waste resources. As a consequence, there is increased international interest in augmenting non-financial incentives with financial incentives.

3.2 Financial Incentives Given the limitations of the evidence base, any approach to introducing financial incentives should be cautious or marginal; it may be best to “nudge” decision makers to change (Thaler et al. 2010). Whether or not such an approach is used, there are some fundamental issues that have to be considered.

3.2.1 What Kind of Financial Incentive Should Be Used? For instance, do bonuses have a greater effect on behaviour than penalties? Economic theory predicts that a small threat to income may have a more substantial effect on behaviour than the promise of a slightly increased income (Kahneman & Tversky 1979).

02 Improving Health

34

8/24/11, 4:22 PM

Reforming Provider Behaviour through Incentives

35

The decline in reputation that financial penalties incur affects market position and professional ranking. An extreme form of such a penalty would be refusal to pay when the treatment of a patient fails. If there is avoidable mortality in a hospital, should payment not be made? Some would argue that avoidable mortality is an indicator to be used to direct managerial enquiry and that incentivizing it might not be sensible (Nolte & McKee 2003). However, adjusting hospital tariffs to induce behaviour changes may be productive. In many countries, tariffs are based on diagnosis-related groups (DRGs), which, in turn, are usually estimated on the basis of the average costs of providing particular types of care across hospitals, for example, a stroke, an appendectomy, or the treatment of a myocardial infarction. But why use the average cost to determine the tariff? In England from April 2010, those hospitals that achieve high levels of day case surgery for gall bladder removal are offered an addition to their normal DRG tariff. If these hospitals follow a particular stroke care protocol, they also receive a bonus. There are four “best practice” tariffs that are being used to drive clinical practice towards what is seen as greater efficiency. In an effort to shift care from the hospital to the community, another incentive is being used: the two-part tariff. This rewards hospitals at full tariff for emergency care up to a volume equal to 2008–09 volumes. Above that level of activity, hospitals receive only 30 per cent of the tariff. This frees resources for commissioners to fund community and primary care, and, it is hoped, reduces the level of hospital emergency admissions. The issue is whether penalties for failing to provide care in this way (that is, payment below the tariff) would have equal, better or worse success in altering behaviour.

3.2.2 Who Should Be Incentivized? Should the incentives be targeted at institutions or at physicians? If incentives are used to alter the behaviour of institutions such as clinics and hospitals, what are the mechanisms by which the behaviour of micro decision makers can be altered and made consistent with the organization’s goals? In England there is an incentive system targeted at groups of physicians delivering primary care and an incentive system targeted at institutions in secondary care.

02 Improving Health

35

8/24/11, 4:22 PM

Alan Maynard

36

In primary care, payments are made to groups of physicians if they deliver specified services to patients, for example, control of blood pressure or control of diabetes. The incentives are focused on the group to ensure that practitioners police one another’s behaviour, as failure of one results in decreased income for all. This “quality outcomes framework” (QOF) created a debate about the volume and efficiency of incentives in producing improvements in health outcomes. As ever with fee for service remuneration systems, it is essential to incentivize activities that produce significant health gains at least cost. Fleetcroft and Cookson (2006), for example, have argued that some aspects of the QOF were poor value for money. The QOF was costly, inflating primary care remuneration by nearly £1 billion, but it did have a significant effect in inducing physician compliance with targets. In the first year a target of 95 per cent was achieved, and there is evidence that this generous scheme reduced practice variation, with poor performers improving their activity levels (Doran et al. 2008). While considerable effort has been put into designing and funding the QOF for the U.K. NHS, no such system has been introduced to incentivize the behaviour of hospital physicians. In the United Kingdom, these practitioners are eligible for “clinical excellence awards”, which are an inefficient system of performance-related pay, allocated on the basis of incomplete self-reporting of NHS activity, teaching, and training.

3.2.3 What Behaviours Are Being Targeted by Financial Incentives? Are they process incentives such as adherence to treatment protocols, or measures of patient outcomes, or are they “never” events — events which should never happen in a “well-managed” clinic or hospital? In U.S. Medicare since 2008 and in the English NHS from 2010, there is a list of “never” events for which tariff (DRG) reimbursement is refused. Table 2.2 lists the “never” events for England. This list is quite modest, and the government has indicated that it will be extended. Possibilities for extensions include the use of the wrong drug or the wrong dose of a pharmaceutical, catheter associated urinary tract infections, and central line infections in intensive care units. The management of “never” events requires good reporting systems to detect practice variations. There is an obvious risk that if such issues are

02 Improving Health

36

8/24/11, 4:22 PM

Reforming Provider Behaviour through Incentives

37

TABLE 2.2 Never Events — Non-Payment of Tariff, England 2010–11 Wrong site surgery Retained instrument post-operation Wrong route of administration of chemotherapy Misplaced naso or orgastr ic tube not detected prior to use Inpatient suicide by use of non-collapsible rails In hospital maternal death from post partum haemorrhage after elective caesarean section Intravenous administration of mis-selected concentr ated potassium chloride Source: Department of Health 2009, p . 34.

incentivized, practitioners and institutions will “game” the system by under-reporting adverse events. This in turn may create the need for control systems that consume resources. An alternative approach to incentivizing change and reducing clinical practice variations is to use P4P in relation to preferred and, it is hoped evidence-based protocols and activities. In U.S. Medicare, the Premier programme has incentivized five clinical areas: heart failure, acute myocardial infarction, pneumonia, coronary bypass grafts, and hip and knee replacements. Hospitals are required to report their performance data related to quality indicators associated with each of these areas and are then given a composite score. Their ranking in the performance league table is linked to payments. Those hospitals in the top 10 per cent of performers get a bonus of 2 per cent on their DRG tariffs. Those in the second decile get a bonus of 1 per cent. Initially, the scheme threatened penalties of 2 per cent loss for the bottom decile and 1 per cent for the second least well performing decile. The effect of the scheme was such that the penalties have not been levied. However, even though performance on these indicators has improved, a review of the few robust quantitative analyses of Premier has shown that greater adherence to process protocols has not led to reduced costs or reduced mortality. The opportunity costs of the Premier programme, like for the UK-GPQOF, are unclear. If targets are incentivized and performance is improved in relation to these targets, what, if anything, is given up? The absence of data on this issue is cause for concern. Despite such caveats about the P4P Premier model, it is being copied extensively in the English NHS as part of a programme called CQUIN or

02 Improving Health

37

8/24/11, 4:22 PM

Alan Maynard

38

commissioning for quality and innovation. In 2009–10 hospitals were obliged to collect baseline data for performance items such as those in Table 2.3. Failure to collect these data could reduce tariff income by 0.5 per cent. In 2010–11 these CQUIN activities are benchmarked and improvements in performance are required. Failure to achieve the targets results in hospitals losing 1.5 per cent of tariff income. The 2010–11 NHS Operating Framework for the English NHS proposes that in future as much as 10 per cent of tariff income will be at risk if hospitals fail to achieve their CQUIN targets.

TABLE 2.3 Commissioning for Quality & Innovation 2010–11 (regional example) 1. 2. 3. 4. 5. 6. 7. 8.

Venous-thrombeoembolism (VTE) risk reduction Improve responsiveness to personal needs from adult inpatient sur vey Maternity: reduce rates of intrapartum still birth, reduce the number of babies admitted to the neonatal unit, reduce caesarean sections (to 18 per cent) Improve end of life care Adhere to hip and knee best pr actice care bundle (protocol) Adhere to acute myocardial infarction best practice care b undle (protocol) Nutritional screening Improve pressure sore care in line with protocols

Source: NHS Yorkshire and the Humber, Regional CQUIN indicators, 2010–11.

The CQUIN programme is radical, but in light of the unimpressive early results from the US Premier programme, its effects may be modest. Without doubt it is important to evaluate such initiatives carefully to avoid international emulation based on desperation, fashion, and faith, rather than empirical evidence of efficiency (Maynard & Bloor 2010).

3.3 Conclusions Intuitively, economists favour the use of financial incentives as mechanisms for “nudging” decision makers to alter their behaviour. Worldwide there is considerable interest in these policies among policymakers because of observable long-term failures in public and private health care markets,

02 Improving Health

38

8/24/11, 4:22 PM

Reforming Provider Behaviour through Incentives

39

and the effects of recession on health care budgets. However as Akerlof and Kranton (2010) have emphasized, P4P has some limitations: • It may be difficult to determine whether employees and institutions make the right decisions because, for instance, the results of decisions may not be evident for many years; that is, the short- and long-term effects of P4P may be different. • P4P systems may attract risk takers rather than those who prefer steady employment and will perform steadily over time. • Employees can “game” the system; for example, in the English NHS, penalties for poor performance in infection control for Methicillin-resistant Staphylococcus aureus and C. difficile were “negotiated” away locally even though some hospitals failed to meet their targets. • P4P crowds out intrinsic rewards, that is, the natural inclination for employees to do a good job. The lessons from international experience are: • Incentives change decision makers’ behaviours. Like dynamite, financial incentives have to be handled with care! • Both financial and non-financial incentives, such as trust, are important. The evidence base for both approaches is very limited. Care in designing, implementing, and evaluating change in incentive mechanisms is essential. • The changes in behaviour produced by these incentives may not always improve equity in access to care or efficiency in the use of resources and expenditure control. Evaluation of such social experiments requires not only careful management and measurement, but also clear articulation and ranking of objectives in order to evaluate success.

4. CONCLUDING COMMENTS The purchasing of health care by private insurers and governments is inefficient. The provision of health care by private and government providers is inefficient. This inefficiency is the product of poorly designed and implemented incentives, both financial and non-financial. These

02 Improving Health

39

8/24/11, 4:22 PM

Alan Maynard

40

incentives, besides protecting inefficiency, produce inequity in access to care and expenditure inflation. Government and private institutions recognize these problems that have been well documented by researchers over decades (Maynard 2005). However, they remain unwilling to challenge those who benefit from inefficiency. Instead of relying merely on institutional reform and increased spending, a focus on transparency and accountability in delivering care on the basis of evidence of efficiency is still needed and remains elusive in societies rich and poor.

References Akerlof, G.A., and R.E. Kranton. Identity Economics: How Our Identities Shape Our Work, Wages, and Well-Being. Princeton: Princeton University Press, 2010. Becker, G.S. “Investment in Human Capital: A Theoretical Analysis”. Journal of Political Economy LXX, no. 9 (1962): 49. Browne, J., L. Jamieson, J. Lewsey et al. Patient Reported Outcome Measures (PROMs) in Elective Surgery: Report to the Department of Health. London: London School of Hygiene and Tropical Medicine, 2007. British Medical Journal Evidence. Clinical Evidence Handbook. London: British Medical Journal Publishing, 2009. Campbell, D. & M.J. Russo. Social Experimentation. Thousand Oaks, California: Sage, 1999. Campbell, D.T. “Reforms as Experiments”. American Psychologist 24, no. 4 (1969): 409–29. Committee on Quality Health Care in America. “To Err Is Human: Building a Safer Health System”. Washington, DC: Institute of Medicine, 1999. Cook, T.D. & D.T. Campbell. Quasi-experimentation: Design and Analysis Issues for Field Settings. Boston MA: Houghton Mifflin, 1979. Department of Health and Social Security. “Priorities in Health and Personal Social Services in England: A Consultative Document”. London: HMSO, 1976. Department of Health. “Operating Framework 2010–11”. London: Department of Health, 2009. ———. “Operating Framework 2008–09”. London: Department of Health, 2007. Donabedian, A. “Evaluating the Quality of Medical Care”. Milbank Memorial Fund Quarterly 44, no. 3 (part 2), 1966. Doran, T., C. Fullwood, E. Kontopantelis, and D. Reeves. “Effect of Financial Incentives on Inequalities in the Delivery of Primary Clinical Care in England: Analysis of Clinical Activity Indicators for the Quality and Outcomes Framework”. The Lancet, 372, no. 9640 (2008): 728–36.

02 Improving Health

40

8/24/11, 4:22 PM

Reforming Provider Behaviour through Incentives

41

Evans, R.G., M.L. Barer, and T. Marmor. Why Are Some People Healthy and Others Not? The Determinants of Health of Populations. New York: Aldine de Gruyter, 1994. Fisher, E.S., J.P. Bynum, and J.S. Skinner. “Slowing the Growth of Health Care Costs — Lessons from Regional Variation”. New England Journal of Medicine 360, no. 9 (2009): 849–52. Fisher, E.S., D.E. Wennberg, T.A. Stukel, D.J. Gottlieb, F.L. Lucas, and E.L. Pinder. “The Implications of Regional Variations in Medicare Spending Part 2: Health Outcomes and Satisfaction with Care”. Annals of Internal Medicine 138, no. 4 (2003): 288–98. Fleetcroft, R. & R. Cookson. “Do the Incentive Payments in the New NHS Contract for Primary Care Reflect Likely Population Health Gains?” Journal of Health Services Resources Policy 11, no. 1 (2006): 27–31. Friedman, M. Capitalism and Freedom. Chicago: University of Chicago Press, 1962. Hampton, J. “The End of Clinical Freedom”. British Medical Journal 287 (1983): 1237–38. Kahneman, D. & A. Tversky. “Prospect Theory: An Analysis of Decision under Risk”. Econometrica 47, no. 2 (1979): 263–92. Laurant, M.G.H., D. Reeves, R. Hermens, J. Braspenning, R. Grol, and B. Sibbald. “Substitution of Doctors by Nurses in Primary Care” (review). The Cochrane Library, no. 2. United Kingdom: John Wiley and Sons Ltd, 2005. Lumley, W.G. Lunacy: The New Lunacy Acts, with an introduction, abstracts, notes and index; and an appendix containing the statutes relating to insane criminals and prisoners, and a statement of the different forms of insanity. London: Shaw & Sons, Fetter Lane, 1845. Maynard, A. & K. Bloor. “Trust and Performance Management in the Medical Market Place”. Journal of the Royal Society of Medicine 96 (2003): 532–39. ———. “Will Financial Incentives and Penalties Improve Hospital Care?” British Medical Journal 340 (2010): c88. Maynard, A. & I.C. Chalmers. Non-random Reflections on Health Services Research: On the 25th Anniversary of Archie Cochrane’s ‘Effectiveness and Efficiency’. London: British Medical Journal Books, 1997. Maynard, A. “Evidence-based Medicine: An Incomplete Method for Informing Treatment Choices”. The Lancet 349 (1997): 126–28. ———. “The Public-Private Mix for Health: Plus ca Change, plus c’est la Meme Chose”. London: Radcliffe Medical Press, 2005. McGlynn, E.A., S.M. Asch and J. Adams. “The Quality of Health Care Delivered to Adults in the United States”. New England Journal of Medicine 348 (2003): 2635– 45. NHS Institute for Innovation and Improvement. Focus on Productivity and Efficiency. London: NHS III, 2009.

02 Improving Health

41

8/24/11, 4:22 PM

Alan Maynard

42

NHS Yorkshire and the Humber. Regional CQUIN Indicators, 2010–11. See . Nolte, E. & M. McKee. “Measuring the Health of Nations: Analysis of Mortality Amenable to Health Care”. British Medical Journal 327, no. 7424 (2003): 1129. Robb, E., B. Jarman, G. Suntharalingam, C. Higgens, R. Tennant, and K. Elcock. “Using Care Bundles to Reduce In-hospital Mortality: Quantitative Survey”. British Medical Journal 340 (2010): c1234. Shaw, G.B. The Doctor’s Dilemma. Harmondsworth: Penguin, 1946. Spitzer, W.O., D.L. Sackett, and J.C. Sibley. “The Burlington Randomized Trial of the Nurse Practitioners”. New England Journal of Medicine 290 (1974): 251–56. Thaler, R.H. & C.R. Sunstein. Nudge: Improving Decisions about Health, Wealth, and Happiness. Newhaven,Connecticut: Yale University Press, 2010. Wennberg, J.E., J.L. Freeman, and W.J. Culp. “Are Hospital Services Rationed in New Haven or Over-utilised in Boston?” The Lancet (23 May 1987): 1185–88. Wennberg, J.E., and A. Gittelsohn. “Variations in Medical Care among Small Areas”. Science 246, no. 4 (1973); 100–11. Wennberg, J.E. “Unwarranted Variations in Healthcare Delivery: Implications for Academic Medical Centres”. British Medical Journal 325, no. 7370 (2002): 961– 64. Williams, J., I. Russell, D. Durai, W.Y. Cheung, A. Farrin, K. Bloor, S. Coulton, and G. Richardson. “Effectiveness of Nurse-delivered Endoscopy: Findings From Randomised Multi-Institution Nurse Endoscopy Trial (MINuET)”. British Medical Journal (2009): 338:b231.

02 Improving Health

42

8/24/11, 4:22 PM

Semi-Autonomous Management of Cambodia’s District Health Ser vices

43

II Organizational Arrangements: Purchasing Health Services

03 Improving Health

43

8/24/11, 4:22 PM

Khim Keovathanak and Peter Leslie Annear

44

03 Improving Health

44

8/24/11, 4:22 PM

Semi-Autonomous Management of Cambodia’s District Health Ser vices

45

3 THE TRANSITION TO SEMIAUTONOMOUS MANAGEMENT OF DISTRICT HEALTH SERVICES IN CAMBODIA Assessing Purchasing Arrangements, Transaction Costs, and Operational Efficiencies of Special Operating Agencies Khim Keovathanak and Peter Leslie Annear

AIMS OF THIS PAPER Under the theme, “health sector institutions”, this chapter focuses on new arrangements for the management of health districts in Cambodia, called Special Operating Agencies (SOAs). SOAs are a form of internal contracting recently initiated as part of the government’s wider public sector reform. The health outcomes of Cambodians are poor. The move to SOAs deserves special attention as an attempt to tackle inefficiency in health

03 Improving Health

45

8/24/11, 4:22 PM

Khim Keovathanak and Peter Leslie Annear

46

service delivery, improve the quality of health services, increase access to and utilization of services, and bring about better health outcomes. It is therefore important to devise an approach to assess whether the new model actually delivers the results that were intended. The original paper on which this chapter is based constitutes a first attempt to make a theoretical and practical description of the move to SOAs. It analyses the concept of contracting, describes the arrangements for SOAs, and proposes an approach to evaluate them. The paper draws on previous experience with the contracting in and contracting out (external contracting) of health services in Cambodia. We develop a framework for the evaluation of SOAs based on an assessment of their additional costs and benefits compared with the standard form of health service delivery administration through the Ministry of Health (MOH).

PRINCIPLES OF CONTRACTING Since the 1980s, a New Public Management (NPM) approach has been adopted by governments across the world. New Public Management is based on the introduction of market mechanisms into the public sector with the aim of improving the efficiency of public services provision (Hood 1991). Competitive tendering, contracting, and performance-based incentives are among the core elements of this approach, which has been widely applied in the health sector. In principle, the three main elements of health care financing are revenue collection, pooling, and purchasing (WHO 2000). Contracting is a form of purchasing. Purchasing deals with the different means and methods of paying health care providers for the interventions they deliver. Purchasing should be strategic to ensure cost-effective service delivery and the achievement of health goals. Contracting is based on the principal-agent theory (Figueras et al. 2005). The contract usually involves two parties: the principal, who wishes to secure the provision of goods or services, and who contracts the agent, who will deliver them (Perrot 2004). Contracting is a process of fulfilling an agreement between the parties, and involves the implementation of activities by one side (the agent) and the provision of reward or compensation by the other (the principal). The contract binds the parties in the transaction to perform their roles as agreed in the contract.

03 Improving Health

46

8/24/11, 4:22 PM

Semi-Autonomous Management of Cambodia’s District Health Ser vices

47

Contracting employs the free market principle, in which competition is believed to favour the better and more efficient performers. Given the scarcity of resources for health and the often high expectations about outcomes, efficiency is one of the main incentives for the adoption of contracting (Harding and Preker 2003). Contracting aims to produce better results in terms of both health services management and the quality of service delivery, and to contain costs by improving efficiency. Other arguments put forward in support of contracting include promoting better planning and policy development by improving the flow of information about issues such as volume of goods, services, costs, quality, and needs. It is also argued that contracting increases the predictability of the price of goods and services because the price is negotiated and agreed on before the contract is made. Moreover, the principal may stipulate that services be delivered to those most in need, which improves the equity of service delivery and provides improved access to the targeted populations (Loevinsohn & Harding 2005). The main tool in the process, the contract, can take different forms. The classical form of the contract tends to be rigid and complete, prescribing in detail, among other items, the quantity, quality, and timing of the product or service to be delivered. In a relational form, on the other hand, the contract tends to be less complete, less rigid, and more flexible. In this form, it is less enforceable, leaving room for renegotiation, and is dependent on trust and cooperation between the parties (WHO 2008). There are various forms of contracting: contracting in, contracting out, internal contracting, institutional and individual performance contracts, and other forms (Perrot 2004). All of these forms have been carried out at varying levels and scope within the health sector, both for medical service provision (ranging from field health education and outreach activities to complicated surgeries) (Hu et al. 2008) and non-medical services (such as cleaning or food provision for patients) (Mills 1998). Despite a trend towards greater use, the experience with contracting has also raised several concerns. Mills and Broomberg (1998) and Harding and Preker (2003) identify a number of issues: (i) transaction costs can be high, eclipsing the gain from contracting; (ii) a low capacity to manage contracts may mean that the optimal benefits cannot be achieved; (iii) if there are only low-capacity bidders, contracting may end up with suboptimal providers, resulting in low quality services; (iv) setting the price requires access to adequate and good quality information to avoid

03 Improving Health

47

8/24/11, 4:22 PM

Khim Keovathanak and Peter Leslie Annear

48

overestimation of costs, overpayment of providers, and waste of resources; and (v) contracting requires rigorous monitoring and evaluation that need both resources and capacity. Contracting often involves an increased level of autonomy in implementation for the agent. Extending greater autonomy to health institutions, including responsibility for staff training, provision of staff performance incentives, and monitoring, are among the approaches that have been used for enhancing service delivery in the past (Oliveira-Cruz et al. 2003). In Cambodia, different models of contracting for delivery of district health services have been piloted in recent years. The first was a comparative trial of contracting in and contracting out, followed by a hybrid form of contracting based mainly on the contracting in of health management services. These will be described in the following section. More recently, a new relationship between the central MOH and its units has been initiated. Called Special Operating Agencies (SOA), the new form has been implemented as a pilot in a few districts that formerly used external contracting. The motivation for moving to contracting was to improve the delivery of services to the public, but the earlier examples also involved higher input costs than standard delivery through MOH.

CONTRACTING AND SPECIAL OPERATING AGENCIES IN CAMBODIA The Cambodian Health System In the last two decades, Cambodia has piloted several policy initiatives and interventions designed to bring about improvements in health services delivery. In the early 1980s, following the Khmer Rouge period, Cambodia barely had a health system. Health facilities were nothing more than ruined, empty buildings, and the health care staff in the whole country numbered fewer than a hundred. In the following years, under the then socialist government, the MOH led the reconstruction of the health system with extremely few resources while confronting an international blockade, isolation, and domestic armed conflict (Annear 1998). Cambodia has made relatively huge progress, following the U.N.sponsored national elections and national unification in 1993. Since then health service improvement has received a great deal of attention from development partners and the government alike, and investment in the

03 Improving Health

48

8/24/11, 4:22 PM

Semi-Autonomous Management of Cambodia’s District Health Ser vices

49

health sector has increased markedly. The government’s recurrent budget and development partners’ respective funding increased from US$2.10 and US$3.00 per capita in 2000 to US$4.00 and US$8.30 per capita in 2005 (Lane 2007). The government budget has increased significantly since then, with government and donor funding now almost on parity. A number of major reforms have taken place since the early 1990s. The Health Coverage Plan was implemented after 1995 to divert investment in health care services to lower levels, namely the district, to respond better to the needs of the population. Under the plan, service coverage is characterized by the establishment of health Operational Districts (OD), which include a referral hospital (RH), health centres (HC) and health posts. Currently, there are 77 ODs, 75 RHs, 967 HCs, and 108 health posts, and the numbers are gradually growing (MOH 2008b). Consecutive health sector strategic plans have been developed and implemented by MOH since 1996. Along with these, strategic plans according to disease or intervention areas, such as HIV/AIDS, malaria, and reproductive health, were developed and implemented with considerable support from bilateral and multilateral partners, including the U.K. Department for International Development (DfID), the Japanese International Cooperation Agency (JICA), the U.S. Agency for International Development (USAID), the World Bank, the Global Alliance for Vaccines and Immunization (GAVI), and the Global Fund to Fight AIDS, TB and Malaria (GF). On the financing side, facility-based user fees were approved under MOH regulations from 1996, and locally based health equity funds that reimburse facilities for the cost of fee exemptions for the poor were established from 2000 with support from MOH. The successful implementation of these interventions has provided lessons for policymakers in terms of sustainable financing and equity in service delivery. At the same time, foreseeing the need to accommodate the demand for health services within the informal sector and the growing urban middle class, a number of different forms of private and community health insurance have been piloted in various ODs since 1998, in accordance with MOH national guidelines and regulations (MOH 2006a, 2008a).

Previous Experience with Contracting In an endeavour to improve the efficiency of health service delivery and improve health outcomes, while harnessing the principle of market

03 Improving Health

49

8/24/11, 4:22 PM

Khim Keovathanak and Peter Leslie Annear

50

economy, a number of contracting models have been piloted and implemented in Cambodia during the last decade. Phase One of contracting was piloted in five ODs from 1999 to 2002 under the Asian Development Bank/DfID-funded Health Sector Support Project (HSSP). The project tested both the contracting in and contacting out of management services at the district level in a total of five ODs (Loevinsohn & Harding 2005). Under contracting in, the management of government health services was outsourced to a local or international NGO while all facilities, staffing, drug supply, etc., remained under the government system. Under contracting out, the local or international NGO was given the responsibility for managing the entire operational district health system, including staffing and the procurement and distribution of supplies, while fixed assets (such as buildings and facilities) remained under government ownership. Evaluation of the contracting pilot indicated improved outcomes at contracted ODs compared with MOH-run ODs in terms of coverage rates, increased service delivery, and reduced out-ofpocket expenditure (Schwartz et al. 2005). Phase Two of contracting was implemented under the second phase of the HSSP from 2004 to 2008, with funding from the World Bank and DfID. A hybrid form of contracting was implemented in eleven ODs. Under this model, NGOs were contracted to manage government health services, employing government staff, and facilities. While there has been no independent and comprehensive assessment of this form of contracting so far, a number of improvements were identified in different reviews (AFP 2007; MOH 2006b), including: (i) clearer lines of authority established between actors, with sanctions enforced; (ii) increased ability to boost staff performance through a performance-based incentive system and innovative performance management procedures; (iii) improved accountability and transparency in decision making; (iv) improved efficiency resulting from an innovative approach to streamlining operational procedures; and (v) better human resources management. Under the Phase Two contracting model, an internal contracting model was piloted in five more ODs as part of a health systems strengthening project supported by the Belgium Technical Cooperation (BTC). Internal contracting differed from the hybrid model in that through internal contracting the ODs were managed by MOH staff with support from the Provincial Health Department (PHD) and BTC technical advisers. Performance-based incentives and monitoring were the main features of this arrangement, along with proactive forms of management.

03 Improving Health

50

8/24/11, 4:22 PM

Semi-Autonomous Management of Cambodia’s District Health Ser vices

51

An assessment of the internal contracting model showed improved worker motivation, higher utilization of health services, better maternal and child health outcomes, as well as higher efficiency and transparency in management (Keller et al. 2008). The more recent move to SOAs for some ODs was based on these initial contracting experiences. This new form is in line with the MOH view that contracting with external institutions would not continue indefinitely and that service delivery at public facilities should remain the responsibility of the ministry (ADB 2004). The aim is to institutionalize internal contracting within ODs and provincial hospitals, beginning with the eleven former HSSP contracting ODs and five BTC ODs. All the parties to the contractual arrangement under the SOAs are units of MOH, with different roles, responsibilities, and mandates within the health system.

Transition to Special Operating Agencies In making its decisions about the form of contracting appropriate to Cambodian conditions, the MOH has followed a number of basic principles that characterize its approach to public health service delivery. According to its Health Strategic Plan 2008–2015, the MOH’s vision is to enhance sustainable development of the health sector and contribute to poverty alleviation and socio-economic development. Its mission includes stewardship of the entire health sector, together with a commitment to improved public health service delivery. Key values include decentralized service delivery and management functions, social health protection and access for the poor, a client-focused approach to health service delivery, an integrated approach to public health interventions, effective human resources management, and good governance (MOH 2008c). These principles are reflected in the MOH approach to contracting in three important ways: • Resource constraints: The evaluation of Phase One of the contracting pilot pointed to the success of contracting out in increasing health coverage, compared with contracting in and normal MOH service delivery, but at almost twice the cost. It was the view of the MOH that this was not financially sustainable given government budget constraints and its dependence on donor contributions for the development of the social sector. By 2005, donors’ contributions to the health sector were twice that of the government’s (US$8.30 vs. US$4.00 per capita per annum, respectively).

03 Improving Health

51

8/24/11, 4:22 PM

Khim Keovathanak and Peter Leslie Annear

52

• Capacity development: Evidence gathered through discussions with the MOH, PHD, and OD staff reveals the perceived downsides of contracting. Interviewees pointed to both the lack of involvement in the implementation of externally managed contracts and the lack of information sharing with OD staff by the contractor. This led to resentment among MOH staff and a failure in knowledge and skills transfer. • National ownership: There is a view within MOH, partly reflecting a perceived expectation among the population, that the ministry has the responsibility both for stewardship and for service delivery within the health system (ADB 2004). In practice, the SOA is a form of internal contracting that uses different actors or agents, but with similar arrangements to earlier contracting procedures. Moving to the SOA form represents a return to management by government staff at the OD level, while under previous contracting models (contracting in, contracting out, and hybrid) they were not directly responsible for management and service delivery. Table 3.1 highlights key differences in the management arrangements between the three main forms of organization: the standard MOH OD, Phase Two of contracting, and the SOA. The columns show the key features of the models. All three models were (in the case of contracting ODs), or are, mandated to deliver the Minimum Package of Activities (MPA) and Complementary Package of Activities (CPA). MPA includes both curative and preventive services delivered at health centres and health posts, while CPA is delivered at referral hospitals. There are three levels of district referral hospitals: CPA1, CPA2, and CPA3. They differ in the number of beds, number of technical staff, range of services, and ability to cope with complicated procedures. With regard to management approach, financial resources, human resources (HR), and staff performance management, the former contracting ODs and the new SOAs differ from the standard government ODs while the contracting ODs and SOAs are identical. To begin with, all ODs follow government guidelines on management, whereas, the contracting ODs and SOAs are given some autonomy over the management of HR, staff incentives, and supplies. As a result, some contracting ODs banned private practice by government staff and made allocation of resources and redistribution of income (incentives) transparent. Both the contracting ODs and the SOAs hire additional staff

03 Improving Health

52

8/24/11, 4:22 PM

03 Improving Health

53

8/24/11, 4:22 PM

MPA CPA

Special Operating Agencies in 16 ODs (2009–present) MOH, PHD, contracted OD office (SOA) (RH/HC)

MOH, PHD, contracted NGO OD office (RH/HC)

MOH, PHD, OD office (RH/HC)

Key stakeholders Official budget and funding HR and staff performance management

Standard management guidelines PLUS some autonomy over HR, supply and incentive management, monitoring of performance-based incentives approach

Standard management guidelines PLUS autonomy over HR, supply and incentive management, monitoring of performancebased incentives approach

Government budget, facilities and staff PLUS SDG from donors’ pooled fund for performancebased incentives, management, TA, drugs and/or contracted staff

Government budget, facilities and staff PLUS HSSP (donors’ fund) for performance-based incentives, management, technical assistance (TA), drugs and contracted staff

Government budget, facilities, staff PLUS SDG from donors’ pooled fund for performance-based incentives, contracted staff

Government budget, facilities, staff PLUS SDG from donors’ pooled fund for performancebased incentives, with contracted staff

Standard OD management Government salaries with Civil servant codes of as in government guidelines line item budget from MOH conduct (rarely enforced)

Management approach

Abbreviations: CPA — complementary package of activities; HC — Health centre; HSSP — Health Sector Support Project; MPA — minimum package of activities; NGO — non-government organization; PHD — provincial health department; RH — referral hospital; SDG — ser vice delivery grant.

MPA CPA

MPA CPA

Standard government OD

Contracting period 2004–08

Service package

Type of OD

TABLE 3.1 Key Differences between Models of Health District Management Semi-Autonomous Management of Cambodia’s District Health Ser vices

53

Khim Keovathanak and Peter Leslie Annear

54

to work in the public facilities and have the authority to reshuffle government staff members who do not meet performance criteria. Staff performance management is the landmark intervention used in contracting ODs and SOAs, but is absent in the standard ODs. The contracting and SOA approaches involve transparency in the allocation of the workforce, clear job descriptions, performance monitoring and evaluation, and performance-based incentives. As staff conditions are set in the contract, with clear criteria for determining levels of performance in contracting ODs and SOAs, including penalties for underperformance, staff performance is monitored more rigorously. With regard to resources, standard ODs have only two major income streams — government budget and user fees. The contracting ODs receive these two revenue streams, plus additional funds from the World Bank/ MOH HSSP, used mainly for performance monitoring, incentives, and capacity development. In the case of SOAs, this additional funding is called a Service Delivery Grant (SDG) and is used in the same way; the extra funding accounts for approximately 40 per cent of the total budget managed by the ODs. According to the mid-term review of the second phase of the HSSP, the sustainability of the contracting model was an area of concern, and contractor exit strategies (handover to MOH administration) were still unclear despite the imminent end of the project in 2008. Wide variability in incentive payment systems and amounts was noted. It appeared that contractors had not functioned within the overall MOH provincial annual operational planning, and contractors instead tended to substitute their own targets and priorities for the ministry’s national targets. No quality of care indicators had been included in the contracts, and support for broader systems development at the OD level and below was absent. A number of issues related to staff capacity building and remuneration were also raised (MOH 2006b).

Key Features of Special Operating Agencies The SOA concept was formalized by Royal Decree 346 in 2008, which approved the application of the procedure across all government services (Royal Government of Cambodia 2008). The SOA is a Cambodian form of the broader concept of internal contracting. The concept involves a new set of instruments that are

03 Improving Health

54

8/24/11, 4:22 PM

Semi-Autonomous Management of Cambodia’s District Health Ser vices

55

derived from government policy on public services delivery and are designed to improve the quality of public health service delivery. Under the royal decree, local government administrations and lower levels of the MOH administration are given a degree of autonomy in making decisions about the best use of their human, physical, and financial resources to deliver the highest possible quality of services in the most effective way and to enhance performance and accountability through streamlining administration to be more transparent and more responsive to people’s needs. To facilitate implementation in the health sector, the Council of Ministers issued Sub-Decree 69 in April 2009, under which certain MOH institutions are designated SOAs (Royal Government of Cambodia 2009a). While the sub-decree also identifies provincial hospitals, regional training schools, and national programmes for SOA status, in most cases their actual conversion is pending; currently, only the sixteen former contracting ODs and eight provincial hospitals have been given SOA status. Despite being a form of internal contracting, the allocation of SOA status involves no competitive bidding. According to the manual for SOA implementation in the health sector (MOH 2009), the objectives of the SOA are to: • Improve the quality and delivery of government health services in response to health needs; • Change the behaviour of health staff gradually towards the principles of motivation, loyalty, service delivery, and professionalism; • Promote prudent, effective, and transparent performance-based management; and • Develop sustainable service delivery capacity with the available resources. In assessing the emergence and role of the SOA, the analysis in this paper adopts generally accepted criteria used in assessing contractual arrangements — that is, the actors involved, their roles and responsibilities, the form of the contract, and the nature of funding.

Actors SOAs are operational service delivery units with a clear policy mandate and willingness to improve performance, while the SOA management and

03 Improving Health

55

8/24/11, 4:22 PM

Khim Keovathanak and Peter Leslie Annear

56

staff remain in the public service. In the health sector, SOA denotes the management team of the OD, or the OD Office. This management team is given jurisdiction over its subsidiary facilities (RH and HC) and responsibility for service delivery to the OD catchment population. The implementation of the SOA involves a number of actors (as principal or agent), including the minister of health or representative, the provincial health director, the OD Office (which becomes the SOA directorate), and the staff of the health facilities operating under the SOA (referral hospital, health centres, and health posts). These are all parties to contracts that bind the SOA and its subordinate units. Within the current Cambodian health system, communities and patients are represented by village health support groups who attend health centre management committee meetings, of which they provide comment and feedback to the health service providers. These user groups are not directly involved in SOA contracting arrangements.

Form of Agreement At the national level, SOAs are awarded authority from the Office of the Council of Ministers (the executive arm of the Cambodian government), from the Ministry of Economy and Finance, and from their parent ministry (MOH), in return for a commitment to enhanced performance and accountability. First, the central MOH (principal) and the PHD (agent) sign a performance agreement that outlines the commitments and the responsibilities of the parties. Then, the PHD (principal) signs a service delivery management contract with the SOA (OD Office, as agent). The SOA head is given some discretion over management arrangements; the SOA head may then choose to make an agreement with subsidiary health services units (RH and HC) for service delivery. In all these contracts, each institution is represented by its head or director. Information obtained during field visits to two SOAs indicated that such arrangements generally take two forms: a three-tier contract arrangement (MOH-PHD-SOA) and a four-tier contract arrangement (MOH-PHD-SOA-RH/HC). These two forms were used previously during phase two of contracting, and each form has its own advantages and disadvantages (Soeters & Griffiths 2003). Under the three-tier arrangement, the SOA signs a performance agreement with each staff member at the health facility (third tier).

03 Improving Health

56

8/24/11, 4:22 PM

Semi-Autonomous Management of Cambodia’s District Health Ser vices

57

This arrangement follows the model used by Save the Children Australia, an international NGO, during the second phase of contracting, and emphasizes the central role of the SOA head in negotiating performance contracts with individual staff members. Under the four-tier arrangement, the SOA director signs an agreement with the head of each health facility (RH and HCs). The head of each facility then signs an agreement with each staff member working in the facility. The four-tier contract follows the model used by HealthNet International, another international NGO, during the second phase of contracting and emphasizes the devolution of power to heads of facilities in contract negotiation and management. All contracts within the system are of the relational type. Table 3.2 summarizes the three-tier and four-tier arrangements.

Roles and Responsibilities The manual for SOA implementation in the health sector provides guidance on the role and responsibility of each actor in the contract (MOH 2009). Under the performance agreement between MOH and PHD, MOH must ensure transparent and timely allocation of funds from all sources, TABLE 3.2 Principal and Agent to Tiered Contracts under the SOA Three Tier

Tier of contracting Principal Tier 1

Minister of Health

Tier 2

Tier 3

Agent

Principal

Agent

Commissioner/ PHD

Minister of Health

Commissioner/ PHD

Commissioner/ PHD

OD/SOA Director Commissioner/ PHD OD/SOA Director

OD/SOA Director All RH, HC, and HP directors and staff

Tier 4

Four Tier

Director of RH Chiefs of HCs

OD/SOA Director

Director of RH and chiefs of HCs

Staff of RH Staff of HCs and health posts

Abbreviations: HC — health centre; HP — health post; OD — Oper ational district; PHD — provincial health department; RH — Referral hospital; SOA — special operating agency.

03 Improving Health

57

8/24/11, 4:22 PM

Khim Keovathanak and Peter Leslie Annear

58

including incentive and contract payments to the OD/SOA, and maintain drug supplies. It must provide assistance to the PHD for a variety of tasks, including: workforce planning and resourcing workforce requirements; reviewing and endorsing national performance targets and three-year planning objectives; and providing the PHD with timely information about policies, guidelines, and enforcement of health legislation, professional ethics, and codes of conduct. In practice, at the central MOH, the project management unit (PMU) of the HSSP is responsible for coordinating these activities and for managing the budget for SOA services. SOAs also receive funds in the form of an SDG from HSSP funds pooled by development partners, and managed and monitored by the PMU. The PMU also provides a team for monitoring and evaluating SOA implementation. The PHD is charged with the role of commissioner (principal), acting on behalf of the MOH to oversee and support the SOA. Under this agreement, the PHD is responsible for resolving issues related to implementation within its own jurisdiction (the province). The PHD (principal) then signs a service delivery management contract with the SOA (OD Office, as agent). Under the service delivery management contract, the responsibilities of the PHD include timely and transparent provision of financial resources and funds destined for the SOA (regardless of source, without charging any form of fees or deductions), timely provision of drugs, assistance in human resources management, and performance management. The PHD also monitors the SOA, usually on joint visits with the PMU monitoring team. The SOA (as agent) is responsible for ensuring that the management of resources (financial, pharmaceutical, and human) is undertaken with transparency and accountability at all facilities (RH and HC). The SOA operates semi-autonomously, with the power to manage its own staff, including hiring and firing contract staff (those who are not civil servants but work on contract with the MOH). It is responsible for demonstrating fairness and transparency in organization, financial management, and the distribution of staff incentives. The SOA must adhere to financial procedures, ensure that all personnel behave in an equal manner to all patients (paying and non-paying), without discrimination because of religion, ethnicity, social status, HIV/AIDS status, gender, age, or political affiliation, and guarantee the effective delivery of services by ensuring a good service environment and staffing, as well as achieving performance

03 Improving Health

58

8/24/11, 4:22 PM

Semi-Autonomous Management of Cambodia’s District Health Ser vices

59

and quality targets, and operating under good governance principles. According to stated rules, staff members are expected not to take underthe-table payments, poach patients or run private practices in public facilities, embezzle medicine or materials from the facilities, or pay facilitation fees or commission in return for contractual payments. What makes SOA management differ from the management of standard ODs is the increased authority given to the SOA/OD and the new strategies employed to enhance performance. This semi-autonomous management means that the SOA has the authority to decide how revenues and some funds are used, to determine gaps in human resources, to fill those gaps, and to issue and monitor contracts with providers and heads of facilities on their performance and incentives.

The SOA Performance Contract Generally speaking, the contract should assure the minister that the SOA management has a clear sense of direction, is dealing with the right issues, and is managing its affairs accordingly (CAR 2008). Under the SOA guidelines, the service delivery management contract between the PHD and SOA defines measurable results using performance indicators. A typical PHD-SOA contract includes: a brief description of major issues confronting the SOA; its proposed strategy and the specific authorities granted; a situation analysis, including internal strengths and weaknesses; objectives and performance targets; key assumptions (economy, market, resources, demand) and priorities (with a proposed rate and fee structure); operational priorities and main activities; performance indicators, including service quality and operational efficiency; the nature and level of user fees; human resource requirements; discipline procedures, rewards and sanctions; and reporting modalities. The contract defines performance targets, standards (nature and quality) for the services provided by the SOA, service fees and delivery time for users, main activities to be implemented by the SOA, the resources that the SOA needs, indicators for expected results, procedures for monitoring and evaluation, internal rules, reporting regime to the parent ministry, and rewards and sanctions. The PHD-SOA agreement requires rolling contracts to be submitted annually as part of the ministry planning process. The contract requires the preparation of a business plan for the SOA for the coming year according to the terms of reference submitted along with the contract.

03 Improving Health

59

8/24/11, 4:22 PM

Khim Keovathanak and Peter Leslie Annear

60

A typical personal service contract between the SOA and its health facility staff includes the terms and conditions of appointment; a commitment to values, ethics, and discipline; a commitment to achieve results based on the operational plan; levels of remuneration and benefits; position descriptions and competency profiles; a brief statement of objectives for the position; a list of major tasks and related indicators; and required competencies. In all these contracts, the monitoring and evaluation of health service delivery performance is an essential element, and indicators, benchmarks, types of data needed, frequency of data collection, and times of review are clearly spelled out. The same is true of staff performance and incentives. Monitoring is of several layers. At the SOA level, following previous monitoring practices during the contracting period, a monitoring group ranging in number from four to six persons is established to monitor service delivery performance at facility level, conducting quarterly assessments that track progress and determine the incentive levels for staff. As commissioner, the PHD has a role in monitoring. A PHD monitoring team has been established and charged with this task. However, little information about their monitoring functions is available. At the central level, five service-delivery monitoring groups (SDMG) have been set up to perform monitoring at all ODs, including SOA districts. The SDMGs are made up of officials working in various departments of the MOH. Given the pre-occupation of officials with their duties at the MOH, it remains to be seen how well these groups carry out the extra monitoring functions. In addition to these several monitoring groups, the SDG manual states that regular auditing of the technical and financial aspects of contract performance will be carried out by an independent external agency. Monitoring can be viewed as enforcement of the contract. While it appears that every level has a role in monitoring, little information is available as to how these various groups coordinate among themselves and carry out their monitoring roles: how frequently they perform this role, what methodology is used at each level and with whom the monitoring report is shared. How effectively these groups perform their monitoring roles, cross-check and use monitoring information to improve contract performance remains to be seen.

03 Improving Health

60

8/24/11, 4:22 PM

Semi-Autonomous Management of Cambodia’s District Health Ser vices

61

All these contracts (PHD-SOA and SOA-facilities/staff) are relational contracts. Within the inherently strong hierarchy of Cambodia’s health system, compliance in terms of rules and protocols can be anticipated. It remains to be seen, however, how effectively these contracts will be enforced should services delivery performance targets not be complied with.

Funding According to the SOA decree, the SOA budget is one part of the annual budget of the parent ministry. The SOA budget is based on planned operations. Each SOA formulates plans based on the standard planning guidelines — annual operational plan and three-year rolling plan. These plans are submitted to MOH for approval. Funding is provided quarterly and is conditional on submission of a quarterly plan and review reports. In the health sector, the SOA (including its facilities) has various sources of revenue, including the national budget allocation (via the central MOH), SDGs under the HSSP, a proportion of user fees collected at facilities, receipts from health equity funds and other community insurance mechanisms, contributions from development partners, and donations from NGOs and other charities. The national budget and SDGs provide the bulk of SOA/OD funding, usually accounting for over 80 per cent of the total budget. At least initially, funds from the health budget will be provided by line item, under MOH guidelines. However, SOAs have been designated as early adopters of new practices and processes advocated through national public financial management reform. Under this reform, the SOA management should implement the allocated budget using a programmatic approach (linking activities, results, and resources). While all ODs receive a portion of national funds through the budget office of MOH, the SDG is a purpose-specific additional source of funds for SOAs and is released through a separate office (the HSSP PMU) that coordinates donors’ pooled funds. Other funds, such as those from NGOs, charities, or user fees, are used at the discretion of the OD by applying the standard procedures for managing revenues. Table 3.3 summarizes the funding sources for a typical SOA, the Ponhea Kraek-Dambae SOA, for 2009.

03 Improving Health

61

8/24/11, 4:22 PM

Khim Keovathanak and Peter Leslie Annear

62

TABLE 3.3 Funding Sources for Ponhea Kraek-Dambae SOA, 2009 Source

Purpose

Proportion (%)

Programme activities and monitor ing, staff salaries, social support/allowances

69.99

User fees

Incentives for staff, facility maintenance, operations

5.92

Health equity fund

Cover costs of health ser vices used by the poor

1.85

Incentives and staff training and technical assistance

22.24

Government budget

a

Service delivery grant (through HSSP) Total

100%

Note: a. About a quar ter of this (programme activities and social suppor t) is managed by the SOA.

A FRAMEWORK FOR EVALUATION OF SOAs Introduction The stated aim of SOAs is to promote cost-effective and client-centred services, improve management of resources, and encourage a change in the attitude and behaviour of health staff and health service providers (CAR 2008). The experience with health sector SOAs is still elementary. Delays in preparing and signing contracts for the first sixteen ODs designated as SOAs meant that implementation began only in the third and fourth quarters of 2009. In practice, there is so far no established mechanism for judging the cost and consequences (hereafter referred to as benefits) of the SOA approach. To evaluate SOA performance, however, there is a need to address questions related to effectiveness and efficiency (as well as equity implications) in this form of public health service administration. In this section, we therefore propose a conceptual framework for evaluating the costs and benefits of SOA implementation. Rather than an economic evaluation, the framework provides a method for approaching policymaking with regard to SOAs as a form of internal contracting. We raise a number of issues that contribute to the development of such a

03 Improving Health

62

8/24/11, 4:22 PM

Semi-Autonomous Management of Cambodia’s District Health Ser vices

63

framework in the spirit of supporting the work of public administration and strengthening the roles of MOH and its subsidiary institutions in public health service management. Our aim is to compare SOAs with the standard OD management model, not to explore all possibilities for OD management. The framework first presents in totality the whole range of costs incurred in preparing and implementing SOAs and the potential benefits; we later look at the application of the framework using specific, measurable indicators.

Identifying Costs and Benefits The costs associated with SOAs are both direct (the budgeted costs of service delivery and management) and indirect (or “transaction” costs, which are often hidden in non-specified ancillary items). It is critical to include the hidden transaction costs in the calculation in order to make a full and accurate evaluation. The benefits associated with SOAs derive from greater efficiencies in administration, but also relate to gains in population access and health status resulting from more effective and efficient service provision by the SOA, and the flow-on benefits in public confidence and national policy development. The proposed framework employs a marginal approach, looking not only at the total cost of SOA implementation and the total benefits derived, but also considering the additional costs (or savings) in implementing SOAs compared with existing systems, and the additional benefits gained as a result. This will be further explained in the detailed analysis below. The main categories of cost and benefit are listed in Table 3.4. On the cost side, there are implementation and transaction costs. On the benefit side, there are benefits to the population, the health care system, and the government and policymakers.

Costs SOAs involve both service delivery costs and considerable transaction costs. Service delivery costs are the direct costs of the actual delivery of services. They may include staff, drug supplies, facilities, and miscellaneous expenses, for example, for staff lodging and transportation.

03 Improving Health

63

8/24/11, 4:22 PM

Khim Keovathanak and Peter Leslie Annear

64

TABLE 3.4 Cost and Benefit Variables for Assessment of SOAs Cost variables

Benefit variables

Services delivery costs Supplies of drugs and medical equipment Facilities and equipment and their maintenance Utilities: water, electricity, gas Travel and accommodation for staff Service delivery monitoring and evaluation Subsidies/exemptions for poor patients Salaries and incentives of both government and hired staff Transaction costs Contract negotiation: staff time (oppor tunity cost) and funds to prepare contr act, cover daily subsistence allowance, and hold meetings with other par ties Contract monitoring: personnel and funds to conduct monitoring and super vision activities to ensure performance and quality of services delivered Measurement: meetings and staff time to analyse data and compile inf ormation Staffing: costs of recr uitment, e.g. meetings, advertisement, testing Capacity development: capacity building activities for performance, HR, financial management Hiring technical assistance for the first years (1–2 years)

To the population: Better health outcomes Increased utilization of health ser vices Increased access to quality care Lower out-of-pocket spending by households Guaranteed access for the poor To the service delivery system: More efficient and lower cost of ser vice delivery Enhanced management performance Enhanced, institutionalized system for HR performance and incentives Better staff performance and lower staff turnover Skill and capacity gains Reduced costs of unnecessar y treatments Savings from preventing mismanagement of funds and supplies To policymakers: Restoring public confidence in public health services Experience and lessons for policymaking and directions National ownership

Transaction costs derive from all those activities that are required to draft and negotiate the contract, monitor compliance, measure outputs against contract requirements, and build the necessary administrative and personnel capacity. These are commonly hidden and absorbed within direct administrative and other costs of the contracting parties and not itemized in the contract (Forder et al. 2005). These costs are unavoidable because without them no contract can be formed. The parties would otherwise not be aware of their contractual obligation, or their roles and

03 Improving Health

64

8/24/11, 4:22 PM

Semi-Autonomous Management of Cambodia’s District Health Ser vices

65

responsibilities, and there would be no information regarding the services required or how these services should be delivered or the appropriate level of quality required. Also, contract implementation must be monitored to ascertain that the required goods or services are delivered in the required quantity and quality. Transaction costs are a hallmark of the contracting approach and an unavoidable part of SOAs. These costs are not listed in SOA administration or line-item budgeting, or included in the contract price agreed between the SOA principal and the agent, but may be found as overheads in salary or administration budgets elsewhere within the MOH.

Benefits We propose here that the benefits associated with implementing SOAs are based primarily on consideration of the types of likely beneficiaries, grouped into three categories: population, health system, and government and policymakers. The population is the primary beneficiary of SOAs. The benefits may include better health outcomes and/or savings in out-of-pocket household spending as a result of better service delivery. Better health outcomes may be measured, in the longer term, by lowered mortality or morbidity rates and increased life expectancy, or through disability-adjusted life years. In the short term, benefits may accrue with improved access to services. In a country where the costs of health care are a major cause of impoverishment, more efficient service delivery through SOAs could mean decreased household expenditure on health, contributing significantly to poverty reduction. Benefits may also derive from a more efficient service delivery system. By service delivery system we refer to the structure, arrangements, and people working within the system. The benefits could include higher outputs from services, institutionalizing performance-based pay, staff and management gaining skills and capacity, and even savings on health care that could be used for other priorities. SOA implementation may provide lessons for future management improvement by providing experience on such policy questions as: What are the best contractual arrangements? What is the best way to provide incentives and monitor staff performance? What is the best way to arrange and contain transaction costs? Has technical assistance been adequately

03 Improving Health

65

8/24/11, 4:22 PM

Khim Keovathanak and Peter Leslie Annear

66

provided and used? What is the best way to improve effectiveness and efficiency of health care delivery? What can be done to improve health outcomes, including attaining MDG targets? The political gains from SOA implementation may be significant. The government and policymakers may see a benefit in regaining public trust in government health services. There may be a greater sense of national pride as Cambodian health officials restore sovereignty as the key decision makers and administrators within the health system, reduce aid dependency, and build national capacity for health system improvement.

The Evaluation Framework To date, no official framework has been established for evaluation of SOAs. As a result, mechanisms to collect baseline data, measure progress, or manage contract compliance have not yet been established. The MOH Health Information System and population-based data available in some SOAs from the evaluation of previous contracting interventions are the only sources of existing information. A purposeful evaluation framework is needed. The primary goal and major benefit of SOA implementation is increased efficiency in health service delivery. Efficiency refers to the ability to deliver optimal outputs with given resources. If the SOA model is efficient, the health system should deliver an increased or improved level of services at a lower average cost. Here we propose an approach to measuring the efficiency of SOAs that is based on intermediate service delivery outputs rather than higher level health outcomes. While measurement of health outcomes can be achieved only in the longer term, information about the efficiency of SOAs is required immediately. The evaluation framework proposed here is illustrated in Figure 3.1. It is a retrospective pre- and post-intervention time series comparison with controls.

The Framework Follows Six Steps Step 1. Identify a counterfactual or control against which SOAs can be compared (Eldridge & Palmer 2009). Past studies indicate the need for a strong design for evaluation of schemes that employ the arrangements of pay for performance models (such as SOA); the design should include an impact evaluation that makes a

03 Improving Health

66

8/24/11, 4:22 PM

Semi-Autonomous Management of Cambodia’s District Health Ser vices

67

FIGURE 3.1 Framework for SOA Evaluation

SOA

Costs (CSOA )

SOA Benefit (BSOA )

COUNTERFACTUAL

Standard OD (comparator)

Costs (CSOD )

Marginal cost = CSOA-CSOD

Standard OD Benefit (BSOD)

Marginal benefit = BSOA-BSOD

Abbreviations: C — costs; B — benefits; SOA — special operating agency; SOD — standard operational district.

comparison between intervention and non-intervention groups (Eichler 2006). The control can be one or several standard ODs. However, they must, as far as possible, have the same characteristics as the SOA for comparisons to produce valid results through matched control (Rossi et al. 1999). Of particular interest are the external investments in the OD or SOA by NGOs or charity organizations, which sometimes go unnoticed and unrecorded. These investments are used for a range of community interventions, including promoting better health-seeking behaviours. These investments and interventions should be well documented and incorporated into the assessment framework. Step 2. Determine which benefits are to be measured and what indicators of benefits are to be used. Service delivery outputs are the immediate benefits from the change in management through SOAs. Identify the key indicators of services delivery output and management improvement for use in comparisons. As illustrated in Table 3.1, both standard ODs and SOAs are tasked to deliver the same packages of activities (CPA and MPA). All or some of the items in the package of activities may be selected as indicators.

03 Improving Health

67

8/24/11, 4:22 PM

Khim Keovathanak and Peter Leslie Annear

68

Step 3. Identify cost indicators in both the standard OD and the SOA and estimate the marginal cost (MC) of moving to the SOA arrangement (Drummond et al. 1997). These costs must include both implementation and transaction costs. The marginal cost is the difference between the cost of the standard OD (SOD) and the SOA, and can be written as: MC = CSOA – CSOD where CSOD is the cost of standard OD and CSOA is the cost of SOA A positive MC means that the cost of SOA is greater than that of SOD. A negative MC means that the cost of SOA is less than that of SOD. Step 4. Identify the benefits of both the standard Operational District and the SOA and estimate the marginal benefit of the SOA compared with that of the OD (Eldridge & Palmer 2009; Drummond et al. 1997). The analysis is based on the key output indicators selected in Step 2. Here, we do not propose to use a composite index as in cost-effectiveness analysis or to value the benefits in monetary terms, but rather to measure the change in the selected indicators as a simple percentage. While this departs from strict cost-benefit or cost-effectiveness analysis, we believe it is necessary because the use of the framework in ongoing monitoring and evaluation will require ease of implementation and analysis using available routine data. The marginal benefit is the difference between the benefit of the standard Operational District and the SOA, and can be written as: MB = BSOA – BSOD where BSOD is the benefit of the SOD and BSOA the benefit of the SOA Step 5. A positive MB means that the benefit from the SOA is greater than from the standard Operational District. A zero MB means that there is no difference in the level of benefits. A negative MB means that the benefit from the SOA is less than that from the standard Operational District. Step 6. Assess the marginal cost of and the marginal benefit derived from SOA implementation as a means to determine whether the investment in the SOA is worthwhile (WHO 2003). In this framework, the assessment requires a value judgment on the part

03 Improving Health

68

8/24/11, 4:22 PM

Semi-Autonomous Management of Cambodia’s District Health Ser vices

69

of health administrators and policymakers who must decide if they believe the additional costs are justified by the marginal change in benefits. In this assessment, the following factors should be included: • The environment in which the ODs operate (geography, infrastructure, population) and the level of inputs into the ODs (financing from different schemes) • Trade-offs between compared variables: how to make a sensible judgement about success when changes in a number of indicators are positive while others may be negative or remain constant • Weighting given to indicators: indicators that are closely associated with mortality or morbidity should be given the highest weight • Sustaining the level of benefits: whether the benefits received will continue in the next years if everything remains the same; this also includes the benefits for the health system — for example, continuing the system of performance management and/or staff expertise acquired in the implementation • Long-term versus short-term effects: knowing which model is more efficient and likely to produce a higher impact on mortality and morbidity in the long term.

Evaluation Indicators Table 3.5 summarizes the main indicators recommended for the proposed evaluation framework. As each of the service delivery indicators makes a different contribution to changes in health outcomes, it is necessary to assign weights. The benefit indicators are restricted to services delivery and management outputs. While there are broader benefits (as shown in Table 3.3), it is very difficult to measure them. Comparison of indicators may employ statistical techniques that confirm whether or not the change in indicator level is statistically significant, taking into consideration possible sampling and data errors and extraneous factors in the surveys (Eldridge & Palmer 2009; Rossi et al. 1999). One particular concern is the availability and quality of data for these indicators. While most, if not all, of these indicators are included in the Health Information System (HIS), the quality of data may not be adequate.

03 Improving Health

69

8/24/11, 4:22 PM

Khim Keovathanak and Peter Leslie Annear

70

TABLE 3.5 Selected Evaluation Indicators CPA Outpatient consultations for adults Outpatient consultations for children under five Bed occupancy rate Average length of stay Delivery by surgery Anti-retroviral treatment coverage Malaria treatment by public health facility TB cure rate Diabetes treatment at public health facility Hypertension treatment at public health facility

MPA

Management

Outpatient consultations for adults Outpatient consultations for children under five Immunization coverage for children under one Antenatal care coverage Birth spacing utilization Delivery by trained health professionals Children under five receiving Vitamin A

Quarterly planning and review conducted Capacity to manage finances Availability of drugs at facilities Labour power for MPA and CPA delivery Contract management and contract monitoring Implementation of integrated supervision Quality improvement Client satisfaction with MPA and CPA services

The accuracy of HIS reports from facilities has not been high and may be prone to manipulation because they are the basis for estimating input needs and reviewing performance. In non-SOA districts, monitoring, supervision, and spot checking may not be carried out regularly. Another source of data for verification is warranted and may come from health facility and population surveys.

Opportunities and Challenges of the Framework The rationale for SOAs is that they will deliver the same or improved government health services at a lower economic cost. Caution is needed. If the cost of service delivery is reduced only for the government health budget (direct costs), but is greater in total (across the sector or the economy, including transaction costs), then apparent efficiency gains are not real and the overall economic results may be negative. In such a case, the SOA form of management could not be regarded as cost-effective. This evaluation framework aims to provide both a comprehensive picture of potential costs and benefits, as well as a practical approach to

03 Improving Health

70

8/24/11, 4:22 PM

Semi-Autonomous Management of Cambodia’s District Health Ser vices

71

assessing the efficiency and effectiveness of SOAs. The approach is based on a number of important considerations: First,

the framework uses service delivery outputs as intermediate indicators of SOA benefit; studies have shown the strong correlations between health services delivery and health outcomes. While identifying changes in health outcomes as a result of SOA intervention will commonly take longer than three years, it is important to know in the short term whether the delivery mechanism really does its job.

Second, this evaluation approach is appropriate for the administrative capacity currently available and does not require a high level of technical expertise. Cambodia faces constraints in carrying out economic evaluations such as cost-effectiveness or cost-benefit analysis using disability-adjusted life years. There are also broader debates about using such measures (Anand & Hanson 1997). Third,

there are limitations in the types of data suitable and available for cost analysis. The current structure of financial management, record keeping, and budgetary arrangements would present barriers to complex analysis.

However, this approach to evaluating contracting also has a number of limitations. First,

the use of several indicators to compare efficiency between models can be subject to bias if the weights assigned to each indicator of outcome or performance are not accurate or consistent.

Second, some results are likely to reflect a trade-off between the variables measured: for example, a 15 per cent increase in antenatal care coverage along with an 8 per cent drop in immunization coverage would make the results appear inconclusive. Third,

03 Improving Health

the judgement as to whether improvement in outputs is worth the extra investment associated with SOAs remains subjective: whether an increase of 12 per cent in deliveries at public health facilities, a 15 per cent increase in outpatient consultations, and

71

8/24/11, 4:22 PM

Khim Keovathanak and Peter Leslie Annear

72

a 10 per cent increase in immunization coverage is worth an extra, say, US$30,000 depends on the opinion of policymakers. An objective approach would require the use of a composite index of benefits to compare with the cost (Drummond et al. 1997), but any such exercise is beyond the current administrative capacity. One unknown factor currently is the impact of a recent decision by the government under Sub-Decree 206 of December 2009 to withdraw government contributions to incentive schemes, such as the Merit Based Performance Incentive (MBPI) and Priority Mission Group (PMG), some of which are used by SOAs as incentives to frontline health workers (Royal Government of Cambodia 2009a, 2010). Since incentives account for more than one third of the total SOA budget, this change, if applied to SOAs, will impact on their development. The effect of this change remains to be seen.

CONCLUSIONS The SOA is an organizational form based on policy considerations linked to the previous experiences of contracting. It is both an attempt to improve the efficiency of government service delivery and to re-establish control over resources through the MOH. The SOA arrangement does not take the form of a classical contract between buyer and seller. Rather, it is a new form for the administration of government health service delivery based on line-item budgeting, with supplementary sources of revenue and incentives and sanctions related to performance and outputs. The purpose is fundamentally to reform the current centralized and administrative approach to management of health service delivery through the MOH. The SOA establishes a relationship between different levels of the MOH that may otherwise be termed a performance agreement or relational contract. The contract is defined by the SOA terms of reference, funding streams, monitoring procedures, rewards and sanctions, performance standards, and a programmatic approach linked to MOH annual operational plans. SOAs incur significant transaction costs for contract negotiation, capacity development, and monitoring and evaluation. The cost of these activities is sometimes hidden between the line items of SOA budgets. Such transaction costs must be made explicit.

03 Improving Health

72

8/24/11, 4:22 PM

Semi-Autonomous Management of Cambodia’s District Health Ser vices

73

If it cannot be shown clearly that the SOA form provides economic efficiencies and net benefits in comparison to existing forms of service delivery, then its continuation cannot be guaranteed. Firm evidence is needed to show that the SOA produces additional benefits even though costs may be increased, measures hidden transaction costs appropriately, and provides an overall economic, as well as health, benefit. Equity and access are critical issues in the delivery of health services. Access to health services for the very poor is restricted by the cost of services and the associated transport, food, and opportunity costs. Health equity funds have provided a solution to this dilemma by funding the cost of exemption from user fees for the poor. The costs and benefits of SOA implementation must therefore include the necessary subsidies for those who otherwise could not afford to access health care. By defining the necessary cost categories and potential benefits, the proposed framework can be used effectively for guiding financial and programmatic planning, despite its limitations. Even so, measuring the efficiency of SOAs remains a challenge that requires further careful investigation. The use of the framework remains dependent on the questions one seeks to answer, the availability of quality data, and agreement on the measurement of outcomes, and it is often the case that a lack of consensus among parties in defining cost and benefit indicators is the major stumbling block to evaluation. This framework, nonetheless, provides a starting point for further discussion, data collection, and analysis to support an evidence-based approach to policymaking.

References ADB. Project Completion Report on the Basic Health Services Project. Phnom Penh: Asian Development Bank, 2004. AFP. Cambodia Health Services Contracting Review: First Draft Final Report. Phnom Penh: Agence Française de Développement, 2007. Anand, S. & K. Hanson. “Disability-adjusted Life Years: A Critical Review”. Journal of Health Economics 16 (1997): 685–702. Annear, P. “Health and Development in Cambodia”. Asian Studies Review 22, no. 2 (1998): 193–221. CAR. Special Operating Agencies Implementation Guide: Performance and Accountability, General Secretariat. Phnom Penh: Council for Administrative Reform, 2008. Drummond, M., B. O’Brien, G. Stoddart, & G. Torrance. Methods for the Economic Evaluation of Health Care Programmes. Oxford: Oxford University Press, 1997. Eichler, R. Can “Pay for Performance” increase utilisation by the poor and improve

03 Improving Health

73

8/24/11, 4:22 PM

Khim Keovathanak and Peter Leslie Annear

74

quality of health services? Discussion paper. Center for Global Development. Washington, DC, 2006. Eldridge, C. & N. Palmer. “Performance-based Payment: Some Reflections on the Discourse, Evidence and Unanswered Questions”, Health Policy and Planning 24, no. 3 (2009): 160–166. Figueras, J., M. Elias, & R. B. Saltman (eds.). Purchasing to Improve Health Systems Performance. European Observatory on Health Systems and Policies Series. Berkshire: Open University Press, 2005. Forder, J., R. Robinson, & B. Hardy. “Theories of Purchasing”. In Purchasing to Improve Health Systems Performance, edited by J. Figueras, R. Robinson, & E. Jakubowski, pp. 83–101. Berkshire: Open University Press, 2005. Harding, A. & A. Preker (eds.). Private Participation in Health Services. Washington DC: World Bank, 2003. Hood, C. “A Public Management for all Seasons”. Public Administration 69 (1991): 3–19. Hu, S., S. Tang, Y. Liu, Y. Zhao, M-L. Escobar, & D. de Ferranti. “Reform of How Health Care is Paid for in China: Challenges and Opportunities”. The Lancet 372, no. 9652 (2008): 1846–53. Keller, S., J. de Jong, & J. Thomé. Assessment of Performance Contracting in Kampong Cham Province, Cambodia. Phnom Penh: Belgium Technical Cooperation, 2008. Lane, C. Scaling Up for Better Health in Cambodia — A Country Case Study for the World Health Organization in the follow-up High Level Forum on Health Millennium Development Goals. Phnom Penh: WHO and Ministry of Health, 2007. Loevinsohn, B. & A. Harding. “Contracting for the Delivery of Primary Health Care in Cambodia: Design and Initial Experience of Large Pilot-Test”. The Lancet 366, no. 9486 (2005): 676–681. Mills, A. “To contract or Not to Contract? Issues for Low and Middle Income Countries”. Health Policy and Planning 13, no. 1 (1998): 32–40. Mills A. & J. Broomberg. Experiences of Contracting: An Overview of the Literature. Geneva: World Health Organization, 1998. MOH. Guidelines for the Implementation of Community Based Health Insurance Schemes. Phnom Penh: Ministry of Health, World Health Organization, German Technical Cooperation, 2006a. ———. Mid-Term Review Report 2003–June 2006: Health Sector Support Project. Phnom Penh: Ministry of Health, 2006b. ———. Guideline for Implementation of Health Equity Funds and Government Subsidy Schemes. Phnom Penh: Ministry of Health, 2008a. ———. Annual Health Financing Report. Phnom Penh: Ministry of Health, 2008b. ———. Health Strategic Plan, 2008–2015: Accountability, Efficiency, Quality, Equity. Phnom Penh: Ministry of Health, 2008c.

03 Improving Health

74

8/24/11, 4:22 PM

Semi-Autonomous Management of Cambodia’s District Health Ser vices

75

———. Special Operating Agency Manual. Phnom Penh: Ministry of Health, 2009. Oliveira-Cruz, V., K. Hansen, & A. Mills. “Approaches to Overcoming Constraints to Effective Health Service Delivery: A Review of the Evidence”. Journal of International Development 15, no. 1 (2003): 41–65. Perrot, J. The Role of Contracting in Improving Health Systems Performance. Geneva: World Health Organization, 2004. Rossi, P., H. Freeman, & M. Lipsey. Evaluation: A ASystematic Approach. Thousand Oaks, London, New Delhi: Sage Publications, 1999. Royal Government of Cambodia. Royal Decree on the Common Principles in the Establishment and Operation of Special Operating Agencies. Phnom Penh: RGC, 2008. ———. Sub-decree 69 on the Designation of Health Institutions as SOA. Phnom Penh: RGC, 2009a. ———. The Cancellation of Priority Mission Group and Merit-based Performance Incentive, in Sub-decree 206. Phnom Penh: RGC, 2009b. ———. Royal Government’s Decision to Review Incentive Arrangements and Pay Reform to Accelerate Public Administration Reform. Phnom Penh: RGC, 2010. Schwartz, J. B., I. Bhushan, E. Bloom, and B. Loevinsohn. “Contracting Health Care Services for the Rural Poor — the Case of Cambodia”. In Development Outreach. Washington: World Bank, May 2005. Soeters, R. & F. Griffiths. “Improving Government Health Services through Contract Management: A Case from Cambodia”. Health Policy and Planning 18, no. 1 (2003): 74–83. WHO. World Health Report 2000. Geneva: World Health Organization, 2000. ———. Making Choices in Health: WHO Guide to Cost-Effectiveness Analysis. Geneva: World Health Organization, 2003. ———. Consultation on Strategic Contracting in Health Systems. Geneva: World Health Organization, 2008.

03 Improving Health

75

8/24/11, 4:22 PM

Jean-Olivier Schmidt and Atia Hossain

76

4 VOUCHERS AS DEMAND-SIDE FINANCING INSTRUMENTS FOR HEALTH CARE A Review of the Bangladesh Maternal Voucher Scheme and Implications for Incentives for Human Resource Management1 Jean-Olivier Schmidt and Atia Hossain

INTRODUCTION The challenge of reaching the health related Millennium Development Goals has led to the search for new mechanisms of channelling resources to demonstrate the efficacy of pilot interventions for future scaling up and to benefit needy groups. Demand-side financing (DSF) mechanisms are an alternative way of targeting services to the poor instead of channelling funds to the supply side of health care. DSF mechanisms have received growing attention from governments and development partners (Gwatkin et al. 2005). Mixed evidence of the impact of government budget

04 Improving Health

76

8/24/11, 4:23 PM

Vouchers as Demand-side Financing Instr uments for Health Care

77

mechanisms in achieving goals (Chabot 2009) gives added impetus to the use of such instruments as DFS in Bangladesh. These discussions in the literature are particularly pertinent given that the international debate on how to improve aid effectiveness increasingly turns attention to providing direct budget support. DSF instruments attempt to transfer purchasing power to specified groups for defined goods and services rather than allocating all resources to the supply side, on the assumption that those in need are able to access services (Standing 2002; Pearson 2001;2 Ensor 2004). The interest in using DSF mechanisms rests on the observation that decades of investing in the supply side has not made a sufficient contribution to improved health care services and related improvements in health outcomes. Vouchers,3 a key demand-side instrument, can potentially address a number of public policy issues, including the underutilization of essential services (changing consumer behaviour), improving targeting of public subsidies to the needy (linking demand to supply), empowering consumers with a choice of providers (freedom of choice), and promoting provider competition and responsiveness (changing provider behaviour and serving as a human resource incentive to promote desired behaviour). Therefore, the use of vouchers can lead to the improved quality of health care services and ultimately the desired improvement in health outcomes. While most past and recent initiatives to channel public funding to the demand side through vouchers has taken place in low-income countries (Nicaragua, Kenya, Uganda, Tanzania, Cambodia and Bangladesh), this instrument has also been applied and tested in highincome countries to increase the efficiency of public spending on social services (Meng et al. 2006; Van den Berg & Hassink 2008). While international evidence so far has shown a moderate success in the use of DSF mechanisms to increase the utilization of key services among priority groups, there is far less positive evidence concerning the effect of DSF on service quality as a consequence of greater competition (Ensor 2004). The overall health system in each of these countries differs markedly in resource generation, risk pooling, and purchasing. Therefore, vouchers need to be adapted to fit different national financing contexts. In the context of these questions, this chapter, which discusses the Bangladeshi model for stimulating demand for maternal and neonatal health services, provides another very interesting experience that contributes to the global discussion on how vouchers can act as incentives for increasing service utilization and changing provider behaviour. In addition, this chapter

04 Improving Health

77

8/24/11, 4:23 PM

Jean-Olivier Schmidt and Atia Hossain

78

discusses the extent to which competition can stimulate improvements in health service quality. It presents the findings of a rapid assessment of the Maternal Health Voucher Scheme (DSF pilot) in 2008 in Bangladesh, which is currently being implemented by the government in thirty-three sub districts (upazila). This assessment was conducted by the Health Economics Unit of the Ministry of Health and Family Welfare with support from GTZ. It assessed the potential and limitations of the voucher scheme and provided lessons for other developing countries facing similar challenges. In addition, this chapter draws on findings from an economic evaluation of the same scheme, as well as findings from a study of non-financial incentives for human resources in the Bangladeshi health system. Section 1 provides the context in which the scheme is implemented while Section 2 describes the voucher scheme and Section 3 describes the study methodology. Results of the rapid evaluation are presented in Section 4 and discussed in Section 5. The scheme is placed in a broader context, including its repercussions on human resource incentives, in the conclusion in Section 6. This review sought to provide early evidence concerning the voucher system in Bangladesh in order to offer insights into how it was being implemented, offer recommendations for improvement, as well as provide input for the international debate on vouchers as a demand-side mechanism to finance health services. Further details of the voucher scheme can also be found in Koehlmoos et al. (2008).

CONTEXT In the past decade, substantial progress in improving health outcomes has been made in Bangladesh. For example, infant mortality rate has decreased steadily from 87 per 1,000 live births (1989–93) to 52 (2002/06), the fertility rate has dropped, and the nutritional status of children has improved (NIPORT 2009). However, considerable challenges remain, particularly in maternal and neonatal health. Estimates of the maternal mortality rate vary. The Maternal Mortality Survey of 2001 indicated a rate of 570 deaths per 100,000 (380–760, 95 per cent CI [Confidence Interval]), whereas estimates based on large-scale data collected by fieldworkers estimated a lower rate of 320. Even if the lower figure is reliable, a large reduction from the current rate is still required to achieve the national Millennium Development Goal of 143 deaths per 100,000 by 2015 (ICDDRB 2006;

04 Improving Health

78

8/24/11, 4:23 PM

Vouchers as Demand-side Financing Instr uments for Health Care

79

NIPORT et al. 2003). However, reductions in maternal mortality have occurred despite persistently high levels of home births (80 per cent), most of these without a trained birth attendant (79 per cent) (NIPORT 2009). In addition to gains in women’s education and socio-economic improvements, studies in Matlab Upazila in Chandpur District, Bangladesh, suggest that the recent reduction in maternal mortality may also have been due to reductions in the incidence of short interpregnancy intervals, fewer abortion related deaths, and better access to essential obstetric care (EOC) (Chowdhury et al. 2007; Rahman et al. 2009). On the other hand, another report from Matlab Upzila suggests that maternal mortality increased due to increases in the proportion of first births (Hale et al. 2009). Further improvements within the health care sector may require increased distribution of skilled care and affordable health care for the poor. There exist s a substantial equity divide among those who receive services. Pregnant women in the top income quintile are almost four times more likely to have received ante-natal care (ANC) from a skilled attendant and are ten times more likely to be attended at childbirth by a medically trained person than those in the bottom quintile (Gwatkin et al. 2007). Access to maternal health services is influenced by a range of demandand supply-side factors. Demand side factors include lack of information about when to seek treatment and the options available, high direct and indirect costs, transport costs, intrahousehold preference, and sociocultural norms (Ensor & Cooper 2004). On the supply side, reducing bottlenecks in resource availability and the correct use of resources to deliver services to the needy is required. A recent study in Bangladesh suggests that, even in areas where the training of skilled birth attendants has received priority, the socio-economic status of households continues to be strongly associated with care seeking behaviour (Anwar et al. 2008). In Bangladesh, the main strategy for improving access to health services, particularly for the poor, has been the implementation of supply-side interventions, funded from tax revenue and donors. Yet it is evident that supply-side financing has not been able to address even simple supplyside problems, such as inadequate supply of drugs and medical supplies (FMRP 2007), lack of functional equipment (Simed International 2008), vacant positions, and high absenteeism among health professionals in public facilities (Chaudhury & Hammer 2004). Furthermore, patients have to make substantial out-of-pocket payments for medicine and informal fees (Data International 2003). Public sector supply-side subsidies have often failed to reach the poor as better off groups receive a greater share of

04 Improving Health

79

8/24/11, 4:23 PM

Jean-Olivier Schmidt and Atia Hossain

80

these subsidies. Much of the funding allocated for improving access to health care is either not spent at all or spent only late in the year and not always on the most effective services (OPM 2005).

DESCRIPTION OF VOUCHER SCHEME The apparent failure of supply-side subsidies to increase essential service utilization sufficiently, particularly among the poor, has led to a growing interest in demand-side financing (DSF), as demonstrated by the inclusion of proposals for DSF in the implementation plans for the current national health programme (Health Nutrition Population Sector Programme, HNPSP) that began in 2005 (MOHFW 2005). The Ministry of Health and Family Welfare (MOHFW), with support from the WHO, developed a maternal health voucher scheme in order to increase utilization of quality maternal care among poorer members of the population from designated providers in selected subdistricts. The pilot was initially started in twenty-one subdistricts in 2006, though implementation only began in 2007, first in two subdistricts, and later extended to the remaining subdistricts. A further twelve subdistricts were added in June 2007. The pilot scheme area now covers a population of 10.3 million in thirty-three subdistricts, around 7 per cent of the country’s population. Initially, the intention of the voucher scheme was to means-test recipients prior to voucher distribution. Following a suggestion from an independent review team of the HNPSP, it was decided that in the poorest subdistricts, vouchers should be made available to all pregnant women (Martinez et al. 2007). The voucher provides free service access for three ANC check-ups; safe delivery, including caesarean section and complication management; post-natal care; and cash benefits, including a transport allowance, and inkind benefits (Table 4.1). Vouchers are distributed by skilled birth attendants and other primary level health workers during ANC checks. In means-tested areas, a committee at the ward (union) level decides on eligibility, based on criteria, including landholdings and family income. To avoid introducing an incentive to conceive, vouchers were to be distributed only to women for first and second births who utilize family planning to ensure two years’ spacing between births. These conditions have been criticized as being difficult to monitor and penalizing women who have the most vulnerable pregnancies.

04 Improving Health

80

8/24/11, 4:23 PM

Vouchers as Demand-side Financing Instr uments for Health Care

81

TABLE 4.1 Details of DSF Scheme (Initial Design and 2007 Modification) Entitlement of Beneficiaries Initial version

Modified version (end of 2007)

3 ANC check-ups

3 ANC check-ups

Safe delivery (institutional/skilled birth attendant [SBA] at home)

Safe delivery (institutional/SBA at home) BDT4 2000 for buying nutritious food Gift box worth BDT500 (baby soap, big towel, 2 sets of babywear, 1 large malt drink [Horlicks])

1 post-natal check-up

1 post-natal check-up

Transportation allowance

BDT300 @ BDT100 per ANC visit BDT100 for facility-based delivery BDT100 for one post-natal visit

Policymakers and development partners wanted to operate the voucher scheme through government channels as far as possible. Administratively, a national DSF committee oversees programme implementation (Figure 4.1). Subdistrict and ward committees report through this hierarchy. Donor funding for the scheme is released initially by the World Bank (which manages the development partners’ pooled fund) to the Ministry of Health and Family Welfare and then through a government-owned commercial bank (Sonali). Subdistrict health offices operate four accounts for four distinct types of benefits: vouchers for women, public facilities, transportation, and cash incentives for health care workers in public facilities (as outlined in Table 4.2). Vouchers received by public facilities are submitted to the bank branch for reimbursement of incentives and the funds are credited to the subdistrict account after first being endorsed by the resident medical officer. Details on voucher transactions and funding flows are made public through monthly reports from the subdistrict to the national DSF committee. Initially, some seed funding is provided to facilities to improve services. Government staff are permitted to benefit directly from cash incentives for some services, including ANC care, safe delivery and caesarean sections — payments they receive in addition to their regular salary. A similar system is used to allocate funding to designated non-government providers, although no seed funding is available for them. Furthermore, public and private facilities were checked with a standardized normative checklist on the availability of services.

04 Improving Health

81

8/24/11, 4:23 PM

04 Improving Health

82

8/24/11, 4:23 PM Payments for delivery services

Designated Private and NGO providers

Reimbursing

“Maternal Health Voucher Scheme” Disbursement account

Sonali Bank Upazila Branch (for provider payment)

Facility (Upazila Health Complex)

Referral Transport

Private provider

Facility staff

Up to Tk 500

Reimbursing public providers

“Seed Fund” account

One time allocation

Line Director, Essential Service Delivery

Line Director, Policy Reform

Pooled Fund MOHFW

Reimbursing 50 per cent Drugs & supplies

Family Welfare Assistant/ Health Assistant (HA)

Tk 10 per registration

Voucher holders

Gift box – Tk 500

1. Transportation cost 2. Cash Incentives

Sonali Bank Upazila Branch (for beneficiary payment)

FIGURE 4.1 Organizational Structure of the DSF System5

82 Jean-Olivier Schmidt and Atia Hossain

Vouchers as Demand-side Financing Instr uments for Health Care

83

TABLE 4.2 DSF Scheme Payments to Providers Provider

Purpose

Amount (Taka)

Incentives for Public Providers

Voucher distributors (e.g. SBA/Family Welfare Visitor (FWV)/HA)

Registering voucher holder

10

Blood testing service Public providers get 50 per cent, i.e. BDT17.50; nongovernment 100 per cent

Two pre-delivery blood tests

BDT70 (BDT35 per test)

SBA using equipment and chemicals from public facilities gets BDT17.50, but if using own equipment and reagents gets BDT35

Urine testing service Public providers get 50 per cent i.e. BDT17.50; nongovernment 100 per cent

Two pre-delivery urine tests

BDT70 (BDT35 per test)

SBA using equipment and chemicals from public facilities gets BDT17.50, but if using own equipment and reagents gets BDT35

Antenatal service (e.g. SBA/FWV/ UHC/other selected private or NGO clinic)

3 visits

150 (BDT50 per visit)

SBA/FWV gets 50 per cent, i.e. BDT25 per visit

Delivery (e.g. SBA/ FWV/UHC/other selected private or NGO clinic)

Normal delivery

300

BDT150 is provided to public providers: SBA/ FWV gets BDT150 divided in public facilities: BDT60 for doctors, BDT 40 for nurse, BDT 25 for aya/ ward boy, 25 for cleaner

Post-natal service (e.g. SBA/ FWV/ UHC/other selected private or NGO clinic)

Post-natal care

50

SBA/ FWV gets BDT25

continued on next page

04 Improving Health

83

8/24/11, 4:23 PM

Jean-Olivier Schmidt and Atia Hossain

84

TABLE 4.2 — cont’d Provider

Purpose

Amount (Taka)

Incentives for Public Providers

Medicine provider (e.g. SBA/FWV/ UHC/other selected private or NGO clinic)

Medicine

100

Selected government facilities/private or NGO clinic

Caesarean section

6,000

BDT3,000 is given to public provider as follows: 1,100 for surgeon, 600 for anaesthetist, 500 for surgical assistant, 500 for senior staff nurses, 100 for aya, 100 for ward boy, 100 for cleaner

Selected government facilities/private or NGO clinic

Forceps/vacuum etc. delivery

1,000

BDT500 is given to public providers: 300 for doctor, 100 for nurse, 50 for aya/ ward boy, 50 for cleaner

Private or NGO clinic

Treatment of eclampsia

1,000

Public provider does not receive any incentive for this service

In principle, voucher distribution should encourage consumer choice by offering a range of designated facilities from which to choose. Payments are made to facilities depending on the service offered (Table 4.2). For nongovernment facilities, the full payments go directly to the service providers and can be used flexibly to motivate staff and improve services. In government facilities, direct payments were not possible since government rules preclude facilities from collecting and holding additional revenue — all should be returned to the treasury. Instead, subdistricts set up funds into which money is paid; 50 per cent of this can thereafter be used for the general development of the facility while the other half is either given to staff (with a differentiation based on qualification) or to the family welfare visitor directly. In the case of a Caesarean section, the additional payment for qualified personnel is quite substantial: a surgeon obtains BDT1,100 for a Caesarean.

04 Improving Health

84

8/24/11, 4:23 PM

Vouchers as Demand-side Financing Instr uments for Health Care

85

STUDY OBJECTIVES AND METHODS The novelty of the voucher scheme, keen interest among the government and development partners, and expectations that the scheme might lead to similar mechanisms in other health services, prompted a rapid assessment to be carried out under the purview of the Health Economics Unit, Ministry of Health and Family Welfare, and supported by GTZ in March 2008. This was substantiated by an in-depth economic evaluation in mid-2009. The rapid assessment was aimed at assessing the early implementation, including problems in distributing vouchers, releasing money, and obtaining services, and at adjusting implementation features. The study also sought to provide recommendations on aspects of the potential outcome and impact on which a more extensive evaluation should focus in the future. Trends in utilization were examined in voucher and non-voucher areas in order to provide early evidence of associations between vouchers and service access, specifically by the poor. Hence, this study provides insight into the operational details of existing voucher schemes for utilization and provision of maternal health services, and provides the first assessment of whether this mechanism is adequately designed to meet the objectives of the scheme. However, this study is not designed to assess the economic efficiency of the voucher scheme nor its impact on maternal mortality reduction. The economic evaluation sought to explore cost-effectiveness. A mix of quantitative and qualitative methods was used. The assessment sought to explore the operation of the voucher scheme, including bottlenecks in allocating resources and distributing vouchers, and the effect that this might have on service utilization, especially by the poor. In addition to the survey data, this chapter makes use of data from the country-wide management information system to explore trends in overall health service utilization. Valid quantitative data could be obtained from fifteen out of twentyone DSF pilot subdistricts for the seven months from October 2007 to April 2008. For qualitative data, a survey of targeted pregnant women or those who had recently given birth was conducted. The qualitative methods included in-depth interviews and focus group discussions with beneficiaries, service providers, and government officials. Six subdistricts were selected to represent the entire fifteen subdistricts by clustering areas into above average and below average performers, based on their record in voucher distribution, ANC coverage, safe deliveries, and

04 Improving Health

85

8/24/11, 4:23 PM

Jean-Olivier Schmidt and Atia Hossain

86

Caesarean section rates. In each of the six subdistricts, the three unions with the highest number of pregnant women registered and vouchers distributed between late 2007 and February 2008 were included in the rapid assessment.

RESULTS This section presents the findings regarding the key indicators for measuring the success of the voucher scheme in terms of its impact on (1) coverage of, and utilization by women, (2) perceived and technical quality of maternal child health services, and (3) perverse incentives for conducting Caesarean sections. During the initial months of the scheme, uptake was relatively low.6 Coverage of voucher distribution increased substantially in early 2008 following a government order to increase benefits and permit incentives for public providers, and an improvement in funds disbursement (Figure 4.2). However, coverage of voucher distribution fell markedly off in areas where distribution was universal, while coverage remained high in areas in which poor women were specifically targeted. While the sample is too small; for definite correlation, interviews with key informants suggest that users in universal voucher distribution areas were not well informed of the scheme’s benefits or how to claim them. Another factor that helps explain why voucher uptake was less than expected is that the voucher is only redeemable in the woman’s area of residence, whereas it is expected that a woman will stay with her parents, who likely live in a different area, for delivery. In addition, parity restrictions previously not observed began to be enforced. The choice of providers is extremely limited in practice. In some areas, only the government hospital is registered to accept vouchers. There are various reasons for this: there might be no private or NGO provider able to meet an adequate standard of care, particularly in poorer areas at which the scheme is targeted; or private providers are reluctant to take part at the prevailing rates of reimbursement. The assessment attempted to look at the possible perverse incentive of remunerating practitioners personally for conducting Caesarean sections. The national rate of Caesarean sections is low by international standards, but is rising, particularly in urban areas (NIPORT 2009). Practitioners in public facilities receive BDT3,000 (US$42) for each Caesarean section,

04 Improving Health

86

8/24/11, 4:23 PM

04 Improving Health

87

0

20

40

60

80

100

120

140

160

180

Oct 07

19

39

Nov 07

11

37

Universal (N=5)

10

26

Dec 07

Means-tested (N=10)

24 Jan 08

33

Voucher distribution as % of monthly

Feb 08

67

147

Mar 08

124

135

Apr 08

91

168

FIGURE 4.2 Monthly Voucher Distribution in Universal and Means-Tested Areas Voucher Distribution as % of Monthly Targets

Vouchers as Demand-side Financing Instr uments for Health Care

8/24/11, 4:23 PM

87

Jean-Olivier Schmidt and Atia Hossain

88

quite a substantial sum considering a doctor’s salary is approximately BDT20,000 per month. The effect of voucher distribution on Caesarean section rates is ambiguous. Within the six subdistricts studied in the assessment, the institutional rate of Caesarean sections was found to be 10 per cent, substantially lower than the rate in facilities across the country (23 per cent, according to the Monitoring and Information System of the government). Official reporting of Caesarean section rates in EOC facilities in DSF distribution areas suggests that the rate may have risen. Caesarean sections increased threefold between January 2007 and the end of 2008, while in other subdistricts no clear trend is apparent (Figure 4.3). Yet these data provide a partial picture. Substantial voucher distribution did not occur until January 2008, suggesting that the rise in rates is associated with other factors. Furthermore, Caesarean sections as a proportion of deliveries show no clear trend in DSF areas and are below the rate in non-DSF areas. Finally, the overall rate of Caesarean sections in the population (number divided by expected births) remained at the same relatively low level (3.2 per cent) in 2008 and the first half of 2009 despite voucher distribution. What may be happening is that more pregnancies are being channelled into facilities where a proportion of them require surgical intervention, leading to the increased volume. This suggests that surgical delivery in the DSF areas, which were deliberately chosen as relatively deprived, have caught up with other areas as a consequence of factors that are, at most, partly associated with the voucher scheme. The potential for perverse incentive remains, however, and it will be important for the impact evaluation to focus on whether any general change in demand for surgical delivery is clinically indicated. The assessment suggested a general satisfaction with the quality of the services received by beneficiaries. At the same time, women suggested the need for improved ward facilities, such as more toilets and beds. At a technical level, a lack of specialized staff and equipment was cited as a severe impediment to the uptake of services. Outcome indicators are lacking, and where they exist their interpretation is problematic. Maternal deaths are extremely rare, and the data, patchy. Information is available on stillbirths and neonatal deaths. This data suggest that rates, while fluctuating, are falling slightly, with no clear difference between DSF and non-DSF areas (Figure 4.4). More deliveries are being

04 Improving Health

88

8/24/11, 4:23 PM

89

8/24/11, 4:23 PM

80

180

280

380

480

580

Jun 2007

CS (DSF)

Oct 2007

CS (Non-DSF)

Jan 2008

CS% (DSF)

Apr 2008

Jul 2008

CS% (Non-DSF)

0.0%

2.0%

4.0%

6.0%

8.0%

10.0%

12.0%

14.0%

16.0%

FIGURE 4.3 Monthly Caesarean Sections in Voucher and Non-voucher Areas (Index Jan 2007=100, and as % of deliveries)

Monthly Caesar section index (Jan 2007=100)

Jan 2007 Feb 2007 Mar 2007 Apr 2007 May 2007

Jul 2007 Aug 2007 Sep 2007

Nov 2007 Dec 2007

Feb 2008 Mar 2008

May 2008 Jun 2008

Aug 2008 Sep 2008 Oct 2008 Nov 2008

04 Improving Health

Caesarean sections as percentate of total facility deliveries

89

Vouchers as Demand-side Financing Instr uments for Health Care

90

0%

1%

2%

3%

4%

5%

6%

7%

Mar 2008

Apr 2008

Feb 2008 Jan 2008

Dec 2007

Nov 2007

Jul 2007

May 2007 Jun 2007

Apr 2007

% still births & neonatal deaths (non-DSF)

May 2008

Complication % (Non-DSF)

Jul 2008

% still births & neonatal deaths (DSF)

Jun 2008

Complication % (DSF)

Sep 2008

Oct 2008

Aug 2008

Oct 2007

Sep 2007

Aug 2007

Mar 2007

Feb 2007

Jan 2007

04 Improving Health Nov 2008

20%

40%

60%

80%

100%

120%

140%

FIGURE 4.4 Stillbirth and Neonatal Death Rates and Complication Rates in DSF and non-DSF Areas

90 Jean-Olivier Schmidt and Atia Hossain

8/24/11, 4:23 PM

Vouchers as Demand-side Financing Instr uments for Health Care

91

admitted, and it may be that the reduction is simply the result of an increased use of facilities for uncomplicated delivery. Yet rates of complications show no clear trend, suggesting that the case mix remains approximately the same and that the outcome of care, as reflected in neonatal deaths, is not deteriorating. However, caution is required when interpreting these figures. The data set is limited, and the DSF voucher scheme was implemented at the same time as other interventions that have led to a general investment in maternal health (including skilled birth attendant training and upgrading of EOC services). The assessment showed a number of discrepancies in the operation of the scheme, including vouchers distributed to ineligible women, designated providers not maintaining agreed service standards, and unofficial fees collected from voucher holders. These occurrences are not unexpected, and some level of such behaviour is to be expected in any system. It can be reduced where there are clear and well functioning monitoring and management arrangements. Scheme administration would need further strengthening and capacity development to pick up the perverse behaviours reported. Much of the burden of the administration falls to the subdistrict health and family planning officer, who oversees the payment of incentives to women. No additional allowance or staffing is provided for this administration. In a number of cases, the local DSF committees were not meeting regularly. Providers experienced delays in reimbursement because payments had to be co-signed by a member of the subdistrict committee and a representative in the designated bank. In addition, a lack of consistency in reporting and authorization rules was highlighted. Furthermore, data are collected by the DSF organizers but are not incorporated within the existing management information systems of the public sector.

DISCUSSION There are three principal reasons for introducing vouchers to increase utilization of essential maternal care: (1) to provide more accurate targeting of the poor; (2) to expand choice and quality of providers through competition; and (3) to channel funding to facilities on the basis of performance (activities). Despite the weaknesses revealed by the rapid assessment, the routine management information system suggests that total facility-based deliveries

04 Improving Health

91

8/24/11, 4:23 PM

Jean-Olivier Schmidt and Atia Hossain

92

have risen since the implementation of the voucher scheme was started (Figure 4.5). In DSF areas, facility-based deliveries have risen by about two and half times, and much faster than in other areas. Furthermore, a rapid increase appears to have occurred after January 2008, corresponding to the increase in voucher distribution. Therefore, an association between voucher scheme coverage and facility-based delivery could exist, though not in a strict causal sense since other factors were not controlled. In addition, it is unclear how far the increase in facility-based delivery has occurred at the expense of home-based deliveries assisted by skilled attendants. A concern is raised that a rapid increase in the use of subdistrict and district facilities resulting from the voucher scheme may place an unsustainable burden on public facilities, given the low numbers of hospital beds and emergency obstetric care teams (gynaecologist and anaesthetist) in these facilities.7 The additional funding to health facilities from the voucher scheme does not allow health facilities to react adequately to the increasing demand by augmenting, for example, health care personnel, drugs, and equipment. Improving funding to the demand side of health care systems (consumers or their agents) can be effective, however, only if the supply side has the capacity to react to the growing demand and deliver better quality services. Part of the inducement may be automatic in that, if more funding is delivered to services through DSF, then improved quality becomes affordable. Yet the nature of health care means that market mechanisms may not be automatic. Several issues are pertinent here. One is to ensure that providers maintain acceptable standards through an accreditation mechanism. This rapid assessment suggests that this does not always happen and that continued enforcement is required to ensure standards are maintained. A related issue concerns competition. A lack of genuine alternatives, such as in the private sector, may be one reason providers fail to maintain standards. In other countries it has been shown that with relatively simple services, such as the bed net scheme in Tanzania or treatment for STIs in Nicaragua, ensuring competition can be feasible even in poor rural areas since the cost of market entry is low (Mushi et al. 2003). Providing a complex service, such as comprehensive obstetric care, by contrast, requires substantial investment, and competition is often genuinely feasible only in richer, urban areas. It is clear that in most of the areas covered by the evaluation, there is little effective competition between providers since the

04 Improving Health

92

8/24/11, 4:23 PM

Number of deliveries 0

50

100

150

200

250

300

350

400

Total deliveries (Non-DSF) Total deliveries (DSF-U)

Total deliveries (DSF)

May 2007 Jun 2007

Apr 2007

450

Jul 2007

93 Aug 2007

500

Sep 2008

Oct 2008

Aug 2008 Jun 2008

Jul 2008

May 2008

Mar 2008

Apr 2008

Feb 2008 Jan 2008

Dec 2007

Nov 2007

Oct 2007

Sep 2007

Mar 2007

Feb 2007

Jan 2007

04 Improving Health Nov 2008

FIGURE 4.5 Deliveries in Health Facilities in Universal DSF (DSF-U), Selective DSF and non-DSF Areas (January 2007=100)

Vouchers as Demand-side Financing Instr uments for Health Care

8/24/11, 4:23 PM

93

Jean-Olivier Schmidt and Atia Hossain

94

number of designated providers is small. The reason for this is partly that existing non-government providers are unwilling to deliver the services at the low tariffs set by the DSF committee. Since no proper costing of maternal, newborn, and child health services has been done, it is unclear what level of incentive would be required for facilities to feel it is worth their competing. More fundamentally, the subdistricts in which the pilot DFS voucher schemes have been introduced are generally areas where there is little incentive to develop substantive, independent, private sector capacity because of the level of the service quality and the availability of required services (for example, emergency obstetric care) in the private sector. Much of the private sector provision is from government doctors delivering private, out-of-hours services. This is convenient for doctors since it allows them to charge substantial fees for those able to pay, but at the same time, obtain a public salary. Encouraging private sector providers who are wholly independent from the public sector and capable of reaching the designated standards is likely to require substantially larger payments than the current voucher scheme provides. This implies a cost of ensuring competition that is probably not affordable on a large scale, though this would need further investigation. Under the DFS voucher scheme, Bangladesh initially chose priority areas and then targeted the services at poor households based on a means test. Later it was suggested that, since most households in some areas were probably poor and certainly had inferior health status, targeting should be universal. Universal targeting was suggested because the cost and inconvenience of identifying poor families was thought likely to outweigh the administrative burden of universal targeting. The voucher scheme necessitated establishing a substantial parallel administration to manage the funds. A management structure from national to local level directs policy on vouchers, including payment rates and means-test criteria, oversees payment of providers and individuals, assesses facility designation, and decides on those that should receive benefits. The multiple layers of administration, particularly the necessity to obtain medical officer authorization for all payments to the subdistrict fund and the bank accounts of non-government providers, slows the release of funds and adds a further disincentive to join the scheme. It is not yet clear how much the voucher funding mechanism costs, although it is likely that this is reflected as much in an additional burden on existing staff as in direct financial costs. The economic evaluation

04 Improving Health

94

8/24/11, 4:23 PM

Vouchers as Demand-side Financing Instr uments for Health Care

95

estimated the direct average costs for providing maternal health services at about US$41 per voucher. In an attempt at a cost-effectiveness analysis, the economic evaluation compared control and DSF areas and found that each additional delivery with a qualified provider in a DSF area that can be attributed to the DSF programme, cost roughly US$70. Given the lack of competition, the extra administrative burden is incurred largely to direct funds to a very small number of existing facilities, and to provide incentives to women. This raises the important question of whether the additional cost is justified by the benefits. The main advantage of the scheme is in channelling funds to facilities on the basis of outputs rather than inputs — a first step towards performance-based financing. This could indeed have substantial impact and provide a learning experience. However, since it is a government scheme and, more often than not, it is government facilities that are benefiting, one could also investigate whether output-based funding could be used more efficiently to allocate the funds through the existing public allocation system. There is also a danger of drawing the scarce human resources of the system into only one specific type of intervention. The study (see footnote 1) on how to use incentives to improve retention and performance of public sector doctors and nurses showed that the overall weak technical environment and lack of social recognition decreased satisfaction and motivation to work. Improving this situation could potentially do more for work motivation and health system performance than paying for targeted types of behaviour. The history of permitting user charge retention provides some of the answer. For many years, various projects have experimented with user charges to benefit health facility care. Yet when consideration was given to extending the pilots and retaining funding in health facilities, permission was refused (Dave Sen et al. 2000). The voucher scheme offers a way to circumvent the system of public resource allocation and direct resources to facilities on the basis of activity (and potentially performance). It provides flexible funding to subdistrict seed funds that is not tied to specific line items. In this sense, vouchers represent a pragmatic response to an unyielding financing system. Yet the question remains, as with the user fee pilots, as to what will happen in the future. The permission to provide incentives directly to providers was given only for the pilot, and there is no guarantee that the Ministry of Finance will permit this to continue. In other countries (for example, Uganda, Kenya), independent voucher management agencies have been established to steer the implementation

04 Improving Health

95

8/24/11, 4:23 PM

Jean-Olivier Schmidt and Atia Hossain

96

of the voucher programme for maternal, newborn and child health, as well as family planning services. While this relieves much of the administrative burden from the public health system, these programmes struggle with integration into the public health care system to ensure sustainability. In this perspective, the Bangladesh DSF experience may provide interesting lessons to countries considering institutional solutions for voucher schemes.

CONCLUSION This rapid assessment suggests that the DSF voucher scheme, once initial funding bottlenecks were reduced, led to a rapid distribution and utilization of vouchers. This is associated with an increase in facility-based normal delivery. Continued concern that surgical delivery would be increased unnecessarily does not appear to have matured, although the pattern requires further exploration and continued monitoring. This review shows that the expectation that the pilot scheme would lead to greater choice and competition in delivery services and thus promote quality of services has not occurred and indeed is not really feasible given the lack of service providers, particularly those providing basic essential obstetric care services. Management difficulties are to be expected in a new scheme, and the assessment revealed a number of issues that require resolving. Of greater concern is whether the elaborate administrative structure developed for the operationalization of the scheme can be made functional without considerable investment in parallel structures. Making the scheme universal, at least in some areas, has reduced some of the bureaucracy, but considerable complexity remains. Complexity is indeed a feature of many voucher schemes and raises a wider debate about whether these mechanisms can be cost-effective and how, in the long term, they can be integrated into the existing health care system. The voucher scheme in Bangladesh follows the recent trend in low-income countries to empower the demand side of health financing. Yet the experience of the past and present implementation suggests that rigidities in the public expenditure system are also an important reason for implementation of DSF in order to show demonstrated impact on reducing rigidities. Overall impact evaluation will address the effect of the scheme on health outcomes compared with the existing system. It will also be important, however, to consider whether changes in the regular

04 Improving Health

96

8/24/11, 4:23 PM

Vouchers as Demand-side Financing Instr uments for Health Care

97

budget and public finance and allocation criteria might be equally or more cost-effective.

ACKNOWLEDGMENTS We would like to thank the following people for providing valuable information and comments on the scheme: Professor A.K.M. Azad, director, MIS, DGHS; Ashraful Islam, MIS, DGHS; Dr Nazrul Islam, deputy programme manager, line director, ESD; Tahmina Begum, consultant for GTZ for explaining some of the operational parts of the scheme; the Health Economics Unit, Ministry of Health and Family Welfare, for facilitating the data collection. This chapter is based on findings from a rapid assessment conducted by the Health Economics Unit with the support of GTZ (Deutsche Gesellschaft für Technische Zusammenarbeit) in April 2008. Data collection was done by Data International under the guidance of Dr Najmul Hossain and by the International Centre for Diarrhoeal Disease Research, Bangladesh, under the guidance of Dr Tracey P. Koehlmoos. Dr Frank Paulin (WHO) gave valuable comments on the terms of reference for this study and the report on the rapid assessment.

Notes 1. This chapter is a modified version of an original paper submitted to Health Policy by Jean-Olivier Schmidt, Tim Ensor, Atia Hossain, and Salam Khan: “Vouchers as Demand Side Financing Instruments for Healthcare: A Review of the Bangladesh Maternal Voucher Scheme”. It incorporates additional findings from two other studies in Bangladesh commissioned by GTZ and executed by ABT Associates: (1) Economic Evaluation of Demand-Side Financing (DSF) for Maternal Health in Bangladesh (2010) by Laurel Hatt, Ha Nguyen, Nancy Sloan, Sara Miner, Obiko Magvanjav, AshaSharma, Jamil Chowdhury, Rezwana Chowdhury, Dipika Paul, Mursaleena Islam, and Hong Wang and (2) Incentives to Improve Retention and Performance of Public Sector Doctors and Nurses in Bangladesh (2010) by Marc Luoma, Jobayda Fathema, Jamil H. Chowdhury, and Hong Wang. 2. In the original reference list, both of these were numbered 3. 3. DSF instruments are unconditional cash transfer through cash subsidy and tax redistribution, health cards, social health insurance, conditional cash transfers. 4. BDT or Bangladesh taka is the monetary unit of the Bangladeshi Taka; US$1 = 70 BDT or Tk. 5. Variations in this mechanism have been introduced in some areas.

04 Improving Health

97

8/24/11, 4:23 PM

Jean-Olivier Schmidt and Atia Hossain

98

6. Calculated as a proportion of women expected to be pregnant and, in the case of means-tested areas, poor. This is a rough calculation based on the district birth rate and proportion expected to be in poverty, and levels exceeding 100 per cent can arise because local rates differ. 7. A recent study found that for areas of similar size Sri Lanka has 3.08 emergency obstetric care teams per 1,000 population while Bangladesh has 0.24 (Somanathan et al. 2006).

References Anwar, I., M. Sami, N. Akhtar, M. Chowdhury, U. Salma, M. Rahman, & M. Koblinsky. “Inequity in Maternal Healthcare Services: Evidence from Homebased Skilled-Birth-Attendant Programmes in Bangladesh”. Bulletin of the World Health Organisation 86, no. 4 (2008): 241–320. Chabot, J. “Bangladesh Health, Nutrition and Population Sector Programme (HNPSP), Annual Programme Review”. Dhaka: Prepared for Ministry of Health and Family Welfare and Donor Consortium, 2009. Chaudhury, N. & J.S. Hammer. “Ghost Doctors: Absenteeism in Rural Bangladeshi Health Facilities”. World Bank Economic Review 18, no. 3 (2004): 423–44. Chowdhury, M.E., R. Botlero, M. Koblinsky, S.K. Saha, G. Dieltiens, & C. Ronsmans. “Determinants of Reduction in Maternal Mortality in Matlab, Bangladesh: A 30-year Cohort Study”. The Lancet 370, no. 9595 (19 October 2007): 1320–28. Data International. Bangladesh National Health Accounts 1999-2001. Dhaka: Health Economics Unit, Ministry of Health and Family Welfare, 2003. Dave Sen, P., E. Karim, J. Martin, & N. Abedin. “Proposal to Ministry of Finance for Local Utilisation of User Fee Revenue on a Pilot Basis within HPSP”. Research Note 22. Dhaka: Health Economics Unit, Ministry of Health and Family Welfare, 2000. Ensor, T. “Consumer-led Demand Side Financing in Health and Education and its Relevance for Low and Middle Income Countries”. International Journal of Health and Planning Management 19 (2004): 267–85. Ensor, T. & S. Cooper. “Overcoming Barriers to Health Service Access: Influencing the Demand Side”. Health Policy and Planning 19, no. 2 (1 March 2004): 69–79. FMRP (Financial Management Reform Programme). “Governance, Management and Performance in Health and Education Facilities in Bangladesh: Findings from the Social Sector Performance Qualitative Study”. Draft report. Dhaka: Ministry of Finance, 2007. Gwatkin, D.R., S. Rutstein, K. Johnson, E. Suliman, A. Wagstaff, & A. Amouzou. Socio-Economic Differences in Health, Nutrition, and Population. Washington, DC: World Bank, 2007. Gwatkin, D., A. Wagstaff, & A.S. Yazbeck. Reaching the Poor Health, Nutrition, and

04 Improving Health

98

8/24/11, 4:23 PM

Vouchers as Demand-side Financing Instr uments for Health Care

99

Population Services: What Works, What Doesn’t, and Why. Washington: World Bank, 2005. Hale, Lauren, Julie DaVanzo, Abdur Razzaque, & Mizanur Rahman. “Which Factors Explain the Decline in Infant and Child Mortality in Matlab, Bangladesh?”. Journal of Population Research 26, no. 1 (March 2009): 3–20. ICDDRB (International Centre for Diarrhoeal Disease Research, Bangladesh). “Selected Maternal Health Indicators Obtained by Geographical Reconnaissance”. Health and Science Bulletin 4, no. 2 (2006): 12–17. Koehlmoos, T.L.P., A. Ashraf, H. Kabir, Z. Islam, R. Gazi, N.C. Saha, & J. Khyang. “Rapid Assessment of Demand-side Financing Experiences in Bangladesh”. Working Paper 170. Dhaka: International Centre for Diarrhoeal Disease Research, Bangladesh, 2008. Martinez, J., T. Chowdury, N. Faiz, I. Pathmanathan, T. Ensor, M. Ali, A. Wahab, C. Minett, & J. Martin. Annual Programme Review: Main Consolidated Report. Dhaka: World Bank, 2007. Meng, H., B. Friedman, A.W. Dick, B.R. Wamsley, G.M. Eggert, & D. Mukamel. “Effect of a Voucher Benefit on the Demand for Paid Personal Assistance”. Gerontologist 46 (2006): 183–92. National Institute of Population Research and Training. “Utilisation of Essential Service Delivery”. Dhaka: NIPORT, 2009. MOHFW. “Health, Nutrition and Population Sector Programme: Revised Programme Implementation Plan July 2003 to June 2010”. Dhaka: Planning Wing, Ministry of Health and Family Welfare, 2005. Mushi, A.K., J.R.A. Schellenberg, H. Imponda, & C. Lengeler. “Targted Subsidy for Malaria Control with Treated Nets Using a Discount Voucher System in Tanzani”. Health Policy and Planning 18, no. 2 (2003): 163–71. National Institute of Population Research and Training, ORC Macro, Johns Hopkins University & ICDDR,B. Bangladesh Maternal Health Services and Maternal Mortality Survey 2001. Dhaka and Calverton, MD: NIPORT, ORC Macro, Johns Hopkins University and ICDDR,B, 2003. OPM. Social Sector Performance Surveys: Primary Health and Family Planning in Bangladesh — Assessing Service Delivery. Oxford: FMRP, 2005. Pearson, M. Demand Side Financing for Healthcare. London: DFID Health Systems Resource Centre, 2001. Rahman, M., J. DaVanzo, A. Razzaque, K. Ahmed, & L. Hale. Demographic, Programmatic, and Socioeconomic Correlates of Maternal Mortality in Matlab, Bangladesh. Pathfinder International, 2009. Simed International. Bangladesh Medical Equipment Survey. Dhaka: Simed International, 2008. Somanathan, A., R. Rannan-Eliya, N. Hossain, B.R. Pande, B.P. Sharma, & L. Sikurajapathy. “Optimization of the Configuration of Public Hospital

04 Improving Health

99

8/24/11, 4:23 PM

Jean-Olivier Schmidt and Atia Hossain

100

Infrastructure at the District Level in South Asian Countries”. Paper submitted to South Asia Network of Economic Research Institutes. Colombo: Health Policy Associates (Pvt) Ltd, 2006. Standing, H. (2002), “An Overview of Changing Agendas in Health Sector Reform”. Reproductive Health Matters 11, no. 10(20) (2002): 19–28. Van den Berg, B. & W.H.J. Hassink. “Cash Benefits in Long-Term Home Care”. Health Policy 88, no. 2-3 (December 2008): 209–21.

04 Improving Health

100

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

101

5 SOCIAL HEALTH INSURANCE IN CAMBODIA An Analysis of the Health Care Delivery Mechanism Sopheap Ly

1. INTRODUCTION This chapter is based on a research paper that aimed to analyse various mechanisms of health care delivery, both in theory and in practice, in order to suggest appropriate mechanisms for a social health insurance (hereafter SHI) scheme. The main question addressed by the chapter therefore concerns government policy for health care delivery when the SHI scheme is implemented in 2015. Should the government engage private providers in delivering services to the insured by allowing the SHI scheme to contract with private providers, or should it instead invest in building more health facilities and recruiting more staff for public facilities? Although providing services through public facilities would reduce transaction costs, building more public facilities to accommodate rising demand adequately would not be feasible and would consume a lot of time and resources.

05 Improving Health

101

8/24/11, 4:23 PM

Sopheap Ly

102

When we take into account (i) the intention of the government, (ii) the behaviour of health service users and (iii) the capacity of public facilities, the proposition that the private sector should be engaged in the delivery of health services to the insured under the SHI scheme is convincing. The 2005 Master Plan for Social Health Insurance in Cambodia articulated an intention to encourage private providers to contract with insurance schemes as soon as an accreditation system is established (MoH 2005). The master plan, however, stated clearly that the package will initially cover only services delivered by public providers within the province, meaning that neither national hospitals nor private facilities will be covered under those insurance schemes (MoH 2005) currently. The engagement of private sector service providers, therefore, depends on how soon an accreditation system is put in place. As for the behaviour of health service users, research in 2006 found that the utilization rate of public facilities to seek first treatment was very low, accounting for only 22 per cent, while that of the private sector accounted for 69 per cent1 (48 per cent used private hospitals, clinics, pharmacies, and private consultations and 21 per cent used non-medical shops and outlets) (National Institute of Public Health 2006).2 Such a low utilization rate reflects low trust in and satisfaction with public facilities. In addition, there were many complaints from public facility users about inadequate medicines, expensive services, long distances, poor staff attitude, low service quality, and inappropriate opening hours (National Institute of Statistics 2004; Annear 2006, pp. 15–18). Another big challenge facing the government is scaling up the number of public health facilities. The Ministry of Health has voiced great concern over the current shortage of health staff and lack of capital investment which makes it difficult to cover operational costs and invest in health infrastructure over the period 2008–15 (MoH 2008a). According to the 2008 Health Strategic Plan, the number of commune health centres is expected to increase by about 77 per cent from 957 in 2007 to 1,697 in 2015, and that of district referral hospitals is expected to increase by about 20 per cent during the same period (MoH 2008). For health centres alone, at least 6,640 new staff will need to be employed over the same period. In this regard, the Health Strategic Plan suggests that the existing criteria should be reviewed and focused on among other factors, (i) the need in major urban areas vs. remote/rural areas; (ii) the potential of private providers; and (iii) the alignment of Health Coverage Plan with the Decentralization and Deconcentration requirements (MoH 2008a).

05 Improving Health

102

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

103

By considering the above constraints and factors, it seems that the government is considering the potential of the private sector. It is unclear, however, whether the subdecree on social health insurance will engage the private sector or not. According to some officials in the Ministry of Health’s department of planning, it is most likely that the private sector will not be engaged as health care providers under the SHI scheme. The chapter, therefore, aims to suggest appropriate health care delivery mechanisms for the social health insurance system by considering the possibility of including private providers in the SHI scheme. Two main subquestions are addressed. The first asks whether health care delivery should be organized according to hierarchy or market mechanisms. The second relates to the incorporation of competition in each mechanism. Given the limitations of this chapter, the question of how to ensure the coherent application of the SHI scheme with other insurance schemes, such as community-based health insurance, social assistance, and private health insurance schemes, is not addressed. Nor does the chapter touch on other aspects (other than structuring health care delivery) of the SHI scheme, including issues of the calculation and collection of contributions, benefit packages, payment methods, and purchasing of health services. In this chapter, the hierarchy mechanism refers to public monopoly, meaning that health facilities are public facilities with the possibility of employing private physicians; the relationship between the public purchasers and private providers is one between employer and employee and their contract is a “contract of service” (Batt 1967).3 The market mechanism refers to contracting with private providers, meaning that the government and/or insurers can contract with both public and private providers. In this case, the relationship between public purchasers and private providers is the relationship between purchaser and contractor, and their contract is a “contract for service” (Batt 1967). Competition can be incorporated or excluded in both mechanisms, depending on how incentives are designed.

2. UNDERSTANDING CAMBODIAN SOCIAL HEALTH INSURANCE With the belief that a social health insurance system could lead to universal health care coverage, Cambodia introduced a Master Plan for Social Health Insurance in 2005. The plan incorporates compulsory, voluntary and social assistance schemes. Implementation of the voluntary, and social assistance

05 Improving Health

103

8/24/11, 4:23 PM

Sopheap Ly

104

schemes is progressing gradually while the compulsory scheme has not yet started. The community-based health insurance scheme is implementing voluntary health insurance. Up to 2008, there were nine CBHI schemes enrolling 45,882 people (MoH 2008b). As for the social assistance scheme, the 2005 master plan stated that equity funds and government funds be used to purchase health insurance cards for the poor. This practice is now being piloted, for example, by GTZ in Kampot and Kompong Thom (Annear 2008). A health equity fund has been implemented since 1997 (Bautista 2003) and is recognized as an effective and widespread form of social health protection for the poor (Annear 2008). It is considered to be one of the social assistance schemes; about 152,213 poor people received assistance from the equity fund scheme in 2007(MoH 2008c). The compulsory scheme is due to start operating in 2015 (MoH 2009b) and is currently awaiting a subdecree on social health insurance. The scheme will target public officers and private employees in the formal sector and their dependents. This compulsory scheme is called “social health insurance”. The term, “social health insurance”, caused some confusion when it was used as the title of the 2005 Master Plan for Social Health Insurance, which encompasses three main schemes — compulsory, voluntary, and social assistance. Therefore, it is necessary to define the term “social health insurance” as used in this chapter clearly.

Definition of Social Health Insurance Peter Annear (2008) points out that the term, “social health insurance”, as used in the master plan is not appropriate and may cause confusion because it covers both the insurance scheme and the social assistance scheme. He thus suggests changing the term “social health insurance” to “social health protection” in the new master plan. In the final draft of the new master plan in May 2009, “social health insurance” was replaced by “social health protection”, that is, the “Master Plan for Social Health Protection.”4 The term, “social health insurance” (SHI) has been clearly defined according to its characteristics by Gottret and Schieber (2006) in that the SHI has four main features. First, if the government does not implement a universal coverage plan, the SHI scheme is only compulsory for certain designated professions and employments. Second, there is a direct link between the contributions and benefits and thus only the contributors can

05 Improving Health

104

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

105

have access to the benefits. Third, the institution running the SHI scheme can be a single entity or multiple entities, which are usually independent from the government though the government regulates its role and organization, benefit packages, and the calculation and collection of contributions. Fourth, the SHI scheme requires strong solidarity among members, which makes it possible to pool funds and cross-subsidize. Based on the discussion of the definition of social health insurance above, this chapter only looks at the compulsory scheme or social health insurance to be implemented for civil servants and private workers in 2015. Although the SHI scheme is unavailable to them for the time being, those civil servants and private workers are currently protected by some forms of social security system.

The Development of Social Health Insurance in Cambodia To date, there are a number of laws and regulations providing some kind of social security system for both civil servants5 and private employees. The Royal Decree on the Social Security Benefits of the Civil Servants in 2008 stipulates six benefits for civil servants, including retirement, disability, maternity, work injury, death, and dependents benefits (Art 2). Despite the fact that the Royal Decree is silent about the SHI, according to the final draft of the master plan for social health protection, the SHI for civil servants will be under the responsibility of the Ministry of Social Affairs (MoH 2009b). The safety and health benefits of private employees are protected by the Labour Law, adopted in 1997. The government has recently introduced a social security system for private employees who are the subjects of the Labour Law through the Law on Social Security adopted in 2002. This Social Security Law stipulates a pension system and occupational risk benefits, but does not clearly stipulate what SHI exists for workers. Although there is no explicit provision regarding SHI, Article 1 of the Social Security Law leaves this matter open-ended by stating that “other subject matters shall be subsequently determined by Sub-Decrees on the basis of actual condition of the national economy”. Therefore, it could be understood that SHI would be one of the “subject matters” and thus would be covered by a subdecree. In 2007 a subdecree on the Establishment of a National Social Security Fund (NSSF) was issued to implement the Social Security Law.6 However, in the first phase starting from 2008, the NSSF covers only occupational risk. It is planned that the NSSF will cover

05 Improving Health

105

8/24/11, 4:23 PM

Sopheap Ly

106

SHI for private sector workers in the second phase and workers’ pensions in the third phase. The SHI scheme for private sector workers will come under the responsibility of the Ministry of Labour. In addition, some private employees of large enterprises, international organizations, and non-government organizations are insured for their health care and/or work injuries by a number of for-profit commercial private health insurance companies. So far, there is no study on how many people are covered by private health care insurance schemes or which private insurance provider offers the most attractive health benefit plans. In summary, there are several regulations on the social security system for public and private employees. Those regulations have clear provisions regarding pension schemes, work-related accidents, and sickness, but seem to be silent about SHI. The subdecree on SHI to be issued in 2015 will introduce social health protection for civil servants and private employees. How to make this scheme comprehensive and effective depends on many factors; one of them is to structure health care delivery appropriately, which is the main purpose of this chapter. In order to address the above questions, we will review and analyse a number of organizational theories, including transaction costs theory, and the experiences of several developed and developing countries.

3. MECHANISMS FOR THE DELIVERY OF HEALTH SERVICE 3.1 Hierarchy Mechanism or Market Mechanism?

3.1.1 Theoretical Analysis The choice of hierarchy mechanism or market mechanism refers to whether a transaction should be made in-house or on the market. It is a “make or buy” decision. Here it is necessary to analyse four main theories: modern bureaucracy, bounded rationality, transaction costs, and new public management. The theories of bureaucracy and bounded rationality support the hierarchy mechanism, while the new public management theory supports the market mechanism. The transaction costs theory seems to be in the middle, suggesting that the organization should balance transaction and production costs before deciding whether to expand production or resort to contracting; however, its tendency seems to favour in-house production.

05 Improving Health

106

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

107

Both Weber’s theory of bureaucracy and Simon’s theory of bounded rationality claim that efficiency can be best achieved through “organization”, which means in-house transaction because the organization gives a short cut for effective and efficient interaction and communication among its members (Weber 1947; Simon 1997). The principle of hierarchical office is one of six fundamental principles of modern bureaucratic administration and refers to “a clearly established system of super- and subordination in which there is a supervision of the lower offices by the higher ones” (Weber 1947).7 Weber suggested that bureaucracy is indispensable in order to gain maximum efficiency in a modern organization, regardless of type, and is very important in a large and complex organization (Weber 1947; Mommsen 1989). In one remarkable statement, Weber (1947) wrote that the purely bureaucratic administrative organization is “capable of attaining the highest degree of efficiency and is formally the most rational known means of carrying out imperative control over human beings”. Simon’s theory of bounded rationality suggested that in order to attain the highest efficiency, an individual must make a rational choice (Simon 1997). The question is how individuals can make a rational decision while their rationality is bounded. In answer, Simon warned that perfect rationality is “impossible”, but approaching it is “possible” through the establishment of organization. He contended that “it is only because individual human beings are limited in knowledge, foresight, skill, and time that organisations are useful instruments for the achievement of human purpose” (Simon 1957). Coase and Williamson also agreed that efficiency can best be achieved through organization or in-house transactions because the market is likely to be risky and more costly where there is uncertainty (especially for the transaction of services), bounded rationality of the purchaser in foreseeing the consequences of the contract, and information asymmetry between purchasers and providers, who are mostly opportunists (Coase 1937; Williamson 1971; Williamson 1975). According to Coase, there are two principal reasons that the transaction is profitably internalized in the organization. First, there is a cost in using price mechanisms or transaction costs, including costs of discovering the relevant prices or costs of negotiating and concluding a separate contract for each exchange transaction, and costs of making a long-term contract (Coase 1937). Shifting from the market into the organization means that a series of separate contracts will be replaced by one incomplete contract,

05 Improving Health

107

8/24/11, 4:23 PM

Sopheap Ly

108

that is, the employment contract, which is more adaptable and flexible (Coase 1937; Williamson 1975). The second reason is uncertainty; Coase (1937) contended: “It seems improbable that a firm would emerge without the existence of uncertainty”. It appears that there is no need for market transactions according to the above argument of Coase. Unfortunately, everything has its own limits as Coase (1937) pointed out: “A firm will tend to expand until the costs of organizing an extra transaction within the firm become equal to the costs of carrying out the same transaction by means of an exchange on the open market or the costs of organizing in another firm.” Transaction costs theory was also advocated by Williamson, who went further than Coase in pointing out the various critical problems of marketbased exchange. He identified three main problems when considering the question of internalizing production or moving to market transaction: rationality, uncertainty, and opportunism (Williamson 1975). It is difficult to conclude a long-term contract because the parties are bounded by their inability to foresee future risks and consequences. Also, the problem depends on the degree of uncertainty of the subject of the transaction. As a result, if bounded rationality is problematic and the uncertainty or complexity of the transaction is high, vertical integration is better than market transaction: “When…..transactions are conducted under conditions of uncertainty/complexity, in which event, it is very costly, perhaps impossible, to describe the complete decision tree, the bounded rationality constraint is binding and an assessment of alternative organisational modes, in efficient respects, becomes necessary.” (Williamson 1975) Another assumption of Williamson is that when there is little competition, or in a case of “asset specificity” or monopoly in the market, individuals are likely to engage in opportunistic behaviour, which he referred to as “a lack of candour or honesty in transactions” (Williamson 1975). Therefore, Williamson favoured internal organization when uncertainty, bounded rationality, and opportunism are critical. He argued that internal organization “promotes convergent expectations” and that it “often has attractive properties in that it permits the parties to deal with uncertainty/ complexity in an adaptive, sequential fashion without incurring the same types of opportunism hazards that market contracting would pose. Such adaptive, sequential decision processes economise greatly on bounded rationality” (Williamson 1975, emphasis in original). Vertical integration reduces opportunism problems by attenuating information asymmetry.

05 Improving Health

108

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

109

In addition, Williamson (1971) claimed that the firm or vertical integration has three major advantages in comparison with the market: incentives, controls, and inherent structure. First, internal organization will reduce the chance of protracted bargaining between independent parties. Second, a firm has “the wider variety and greater sensitivity of control instruments for intrafirm”; has “the constitutional authority and low-cost access to the requisite data which permit it to perform more precise own-performance evaluations”; and has its own “reward and penalty instruments” (Williamson 1971). Third, it has possession of “efficient conflict resolution machinery” — ”fiat” — which is a more efficient way to settle conflicts, especially minor ones (Williamson 1971). Therefore, Coase and Williamson suggested that vertical integration will economize greatly on prices discovery, negotiating and concluding separate contracts, monitoring outcomes, and resolving disputes. At the same time, they also warned of unlimited integration because at some point in the organization’s expansion, internal production costs may outweigh market transaction costs. Weber and Simon favour the internalization of transaction in order to attain the highest efficiency. The transaction costs theory seems to be neutral, but at the same time suggests paying attention to uncertainty and other market failure issues before deciding on transacting production in the market. Now let us examine the new public management (NPM) theory. According to Hood (1991), the NPM theory originated from “a marriage of two different streams of ideas”: (i) “new institutional economics”, which built on the public choice theory, principal-agent theory, and transaction costs theory; and (ii) “managerialism”, which concerned the influence of business-type management on the public sector. The development of the NPM theory was considered “a rejection of the Weberian theory that government is best served by bureaucratic organisation” (Lane 2000) and a result of the critique of bureaucracy and hierarchical rules-based systems (Yamamoto 2003). In a nutshell, the NPM theory is about “managerialism” and “contractualism”, which mainly focus on contract making and enforcement (Lane 2000). However, contractualism does not replace all the elements of public policy and public administration (Lane 2000). The NPM theory concerns the modern governance of the public sector. Its concept is not only about contracting with the private sector or privatization, but mainly

05 Improving Health

109

8/24/11, 4:23 PM

Sopheap Ly

110

about changing the way government achieves efficiency either by importing private management theory into the public sector, or by contracting with the private sector to deliver goods and services.

3.1.2 Implications for Health Care Now the question is how we should apply these theories to health care services. How can we know whether it is efficient to structure health care delivery via a hierarchy or a market mechanism? In order to answer this question, it is necessary to look at factors such as (i) the nature of health services; (ii) the degree of information asymmetry, uncertainty, bounded rationality, and opportunistic behaviour associated with health services; and (iii) the ability of the government to tackle these problems. It is widely acknowledged that information asymmetry, uncertainty, and opportunism are very acute in the health sector. Information asymmetry occurs between the physician and patients usually because the physicians possess much greater medical knowledge. Asymmetric information leads to uncertainty over the quality of service provision (Arrow 1963). Risk and uncertainty have a close link, and are the “significant elements in medical care hardly needs argument” (Arrow 1963). In addition, uncertainty and opportunism often result from the complexities of health care, from the existence of specific assets, and from the fact that future health status is usually unknown (Bartlett & Le Grand 1993). Concerning health risk, Arrow (1963) argued that “the most distinguishing characteristics of an individual’s demand for medical services is that it is not steady in origin as for food or clothing, but irregular and unpredictable”. He added: “There are two kinds of risks involved in medical care: the risk of becoming ill, and the risk of total or incomplete or delayed recovery. The loss due to illness is only partially the cost of medical care. It also consists of discomfort and loss of productive time during illness, and, in more serious cases, death or prolonged deprivation of normal function” (Arrow 1963). Nevertheless, according to Flood (2000), some scholars argue — drawing on the theory of “contestable markets” — that hierarchical command and control are often inefficient and, thus, the government should always contract out to the private sector. But the theory of contestable markets is criticized by Williamson for overlooking opportunistic behaviour (Flood 2000).

05 Improving Health

110

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

111

Arrow (1963) argued strongly for the objection that the profit motive should not be allowed in the provision of medical care. He emphasized “the relative unimportance of profit-making in hospitals” saying that “the very word, ‘profit’, is a signal that denies the trust relations”. He added that medical care “is based on non-market relations that create guarantees of behaviour which would otherwise be afflicted with excessive uncertainty” (Arrow 1963). Le Grand (2003b) also observed that there is concern about the profit motive because it undermines the morality and equity of health service. Taking into consideration the theories discussed above and the nature of health care, which is characterized by high uncertainty and highly asymmetric information, it is possible to conclude that the market mechanism appears inappropriate to handle health service provision. First, transaction costs are likely to be high because of the high degree of uncertainty of illness. Second, it is hard to design a long-term contract due to lack of information and difficulty in predicting future costs. Third, service providers tend to enter into a contract opportunistically because of high information asymmetry and difficulty in monitoring the quality of service. Fourth, there is a strong objection over the profit motive of market relation which can lead to relations of distrust between health care providers and health care users. However, the question that remains is to what extent the hierarchy mechanism in health care is always effective. The transaction costs theory argues that when the levels of bounded rationality, uncertainty, opportunism, and information asymmetry are low, the transaction can be made in the market. Therefore, at a certain level, health care may be contracted to the market if the government has enough ability to address those market failure problems. A common concern over working with the private sector is the high transaction cost of contractual relationships; however, Loevinsohn (2008) noted after reviewing six case studies of contracting with non-state providers in developing countries that the performance of those non-state providers was better than that of public providers. The cases that Loevinsohn reviewed were mostly related to management contracts or contracting in, not a contract to purchase the whole service. Moreover, the contractors were mostly not-for-profit providers. The outcome may be different if the contract is for purchasing health services and the contractors are for-profit providers.

05 Improving Health

111

8/24/11, 4:23 PM

Sopheap Ly

112

3.2 Competition or No Competition? The command-and-control or hierarchy mechanism is often criticized as inefficient because of the absence of competition. In the hierarchy mechanism, the public providers will have a monopoly and thus have no incentive to improve services. Le Grand (2007) claimed that users’ choice can influence the performance of service providers and further argued that “the government can use the ‘invisible hand’ of choice and competition to achieve its aims in public services”. Among the four means to ensure the provision of services — trust, targets, voice, and choice and competition — Le Grand (2007) argued that having choice and competition appears to have “the best prospect of delivering a good service” because this can ensure not only responsiveness, efficiency, and high quality of services, but also equity and social justice. The other three means are likely to be incomplete on their own or can merely address short-term problems (Le Grand 2007). One of the criticisms of choice and competition in public service delivery is that choice should not be encouraged in the public domain. Marquand and Titmuss are among the scholars who criticize choice and competition in the public domain. Marquand (2004) argued: “People are consumers only in the market domain; in the public domain they are citizens. Attempts to force these relationships into a market would undermine the service ethic, which is the true guarantor of quality in the public domain.” Titmuss (1997) contended: “It is the responsibility of the state, acting sometimes through the processes we call ‘social policy’, to reduce or eliminate or control the forces of market coercions which place men in situations in which they have less freedom or little freedom to make moral choices, and to behave altruistically if they so will.” So for Marquand and Titmuss, the private and public sectors are two distinct spheres with different values. Users in the private sphere are considered to be consumers while providers are profit-bound and selfinterested, whereas in the public sphere, users are considered as citizens served by public officers who are linked by service ethic and moral choice. Le Grand (2003) refers to consumers and providers in the private sphere as queen and knave, and regards citizens and public officers as pawn and knight.8 Regarding “exit” and “voice” as two instruments to influence public service delivery, Hirschman (1970) suggested using them carefully.

05 Improving Health

112

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

113

He highlighted “exit” as a means to ensure the accountability of decision makers through a competitive market, “voice” as a means to change a firm or organization from within, and “loyalty” as a means to build trust and encourage people to use voice for improvement. Hirschman argued that even though exit was a cheap means for an individual because it did not require direct communication between the dissatisfied customer and the firm, its success depended on the assumption that consumers had the information needed to make efficient choices. In this respect, he noted that if exit is too easy an option, then a crucial number of customers may leave the firm before any correction takes place. Therefore, there will be a welfare loss to both consumers and firms, which, in this case, requires firms to build loyalty. In contrast, exit cannot work in monopoly markets and may work less well in oligopolistic markets due to price collusion. On the other hand, voice is a messy and costly means, and its effectiveness depends on “the influence and bargaining power that customers and members can bring to bear within the firm from which they buy or the organizations to which they belong” (Hirschman 1970). Mills et al. (2001) suggested a middle approach between exit and voice, which depends on the type of service. Where there are no economies of scale and consumers can make reasonable judgments about the quality of a service, exit may help improve government service provision. Where there are reasonable economies of scale, it may not be possible to create consumer choice, and giving consumers voice within the system may be effective. Mills also argued that in order to encourage competition, a system for rewarding providers and paying them for providing care is required (Mills 1995; see also Broomberg 1994). However, there are three major concerns over the capacity of developing countries to promote competition: market structure, contract management, and effects on the whole health system. The concern over market structure lies with shortages of health facilities and difficulties gaining physical access to them, especially in rural areas (Mills 1995). In such a case, Mills suggested that competition can be promoted in urban areas for primary care services because the private sector is more capable of investing in primary care than in hospital services. However, since most developing countries are not yet able to ensure and control the quality of the private sector, she argued that the private, not-for-profit sector (NGOs) may be potential partners (Mills 1995).

05 Improving Health

113

8/24/11, 4:23 PM

Sopheap Ly

114

The second concern is that the inability of the government to negotiate and manage contracts and monitor the quality and performance of contractors may result in high transaction costs (Mills 1995). The third concern involves the effects of competition on the health care system as a whole, such as the movement of trained staff from public sector to private; the inflexibility of contracts, which can lock the government into an unfavourable situation; and the distortion of the behaviour of public providers to focus more on competition than on the welfare of patients (Mills 1995). Mills proposed that the conditions of developing countries are not favourable enough to a nationwide competition mechanism. However, selective contracting with the private sector may be feasible, such as contracts in urban areas and for non-clinical services, for management service, and for primary care services. Mills (1995) considered contracting for non-clinical and primary care as “likely to be the easiest to implement, at least in those countries with a reasonably-sized private sector”. Although Mills supports some contracting with the private sector, she opposes the introduction of competition in developing countries because of poor information systems and the low ability to create a competitive environment. She suggests instead that long-term relationships with providers be developed. Broomberg (1994) endorsed the same proposal, arguing: “… most of the conditions required for successful implementation of these reforms are absent in all but a few, richer developing countries, and … the costs of [managed market reforms], particularly in equity terms, are likely to pose substantial problems”.

3.3 Four Mechanisms The existence of competition does not define either the hierarchy mechanism or the market mechanism. Both can incorporate competition, as Le Grand (2007) made clear: “Introducing competition and quasi-markets into public services should not be confused with the so-called privatization of those services. For it is perfectly possible to have competition between publicly owned or non-profit entities without any participation from the private sector. It is the presence of competition that matters, not the ownership structure of providers.” It is quite convincing that competition can be an efficient instrument to encourage the public sector to perform efficiently and responsively.

05 Improving Health

114

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

115

However, through a review of literature mentioned above, there is less support for incorporating the element of competition in the health care sector for three reasons. The first reason concerns the issues of ethics, morality, and altruism raised by Frederickson (1999), Titmuss (1997), and Marquand (2004). The second reason is that there are many studies showing that most of those who work for the public health care sector are motivated by altruism (Le Grand 2003a). The third reason is that, if competition means profit making, it will build distrustful relations, because, as Arrow (1963) said, “The association of profit-making with the supply of medical services arouses suspicion and antagonism.” Nevertheless, Le Grand (2003a) suggested that where demand is high, it is necessary to engage “knaves” through market payments, and where demand is low, the payment should be enough to motivate a “knight”. This means that competition and profit should not be introduced in hierarchy mechanisms where the demand is low, and where the need for knightly motivation is high. Competition is normally associated with the market mechanism because the engagement of private providers introduces competition. However, we should distinguish between two scenarios; (i) competition between public and private providers; and (ii) competition among only private providers. Both are possible, depending on how the government devises the payment method and whether it treats public and private providers differently or the same. If the government wants to encourage competition between public and private providers, it should take two measures. The first measure is to apply a work-related payment mechanism, for example capitation, feefor-service, and cost per case or per diem payment. The payment method should apply to both public and private providers. The second measure is to limit any subsidy to public providers in order to create a level playing field for competition. If the government does not want to encourage competition, it should devise different payment mechanisms: non-workrelated payment for public providers and work-related payment for private providers. Private providers should be paid through a work-related method because a non-work-related method, such as salary or block contract, would transform the contract into an employment relationship. According to the above analysis, health care delivery can be organized as a hierarchy or market mechanism, and in each option, competition is possible. Therefore, we can establish four combinations: (1) hierarchy

05 Improving Health

115

8/24/11, 4:23 PM

Sopheap Ly

116

mechanism without competition, (2) hierarchy mechanism with competition, (3) market mechanism without competition and (4) market mechanism with competition. The meaning of the hierarchy mechanism without competition is that there are only public health care providers and they are not competing with one another; in the hierarchy mechanism with competition, they are competing. Market mechanism with competition means that there is competition between public and private health care providers, while the market mechanism without competition means that there are both public and private health care providers, but no competition between them occurs though there is competition among the private providers. From the theoretical discussion above, though the four mechanisms are possible in the health care sector, the hierarchy mechanism without competition is strongly suggested for two reasons. First, the hierarchy mechanism appears to be more appropriate for health care than the market mechanism because of the convenience of hierarchy in facilitating effective communication and influencing actors’ behaviour, and because health care is normally associated with high transaction costs. Second, competition is normally associated with profits; therefore, it is not suitable for public providers because those who work in health care in the public sector are not usually driven by profit. In addition, competition would supposedly undermine the ethics of health providers and the equity of users, thereby creating suspicion and antagonism among patients. Nevertheless, the three other mechanisms are also possible. The hierarchy mechanism with competition can be implemented where demands for health care are high and where there is a need to improve the efficiency of service provision. In order to introduce competition, it is important to design incentive systems appropriately; for instance, the provider payment methods should be work-related, such as capitation or fee for service. The market mechanism should be introduced where production costs outweigh transaction costs because unlimited integration or expansion of the public health sector would trigger management failure or loss of management control, bureaucratic costs, and delays. In addition, this mechanism could be introduced where the government has the capacity to control market failures and minimize transaction costs. Where the government is pro-market or pro-competition and advocates individual freedom, the market mechanism with competition should be

05 Improving Health

116

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

117

adopted. However, because the medical sector is well known for its high degree of information asymmetry, the government must have the ability to control this inequity by strengthening, maintaining, and monitoring the ethics and professional standards of health care providers. The government should adopt the mechanism of market without competition if it needs to supplement services from private providers and has limited ability to introduce competition between public and private providers. This section has discussed many theories on health care delivery. The next section examines experiences of structuring health care delivery in a number of developed and developing countries which have adopted a social health insurance system. The later section also includes a discussion on how the four health care delivery mechanisms analysed above could well fit the Cambodian social health insurance scheme.

4. DELIVERY MECHANISMS FOR SOCIAL HEALTH INSURANCE 4.1 Implications in Developed Countries Countries that have adopted social health insurance (SHI) have a system of public financing and private delivery (Besley et al. 1994). Japan, Germany, France, Belgium, Austria, and the Netherlands are some of the countries that have a SHI system. In these countries, private health care delivery is well established and citizens generally have choice of provider; yet the government plays a very important role in setting rules and making regulations which usually take two main forms. First, the government determines the fee schedules for providers’ reimbursement after centralized bargaining with providers’ associations. Second, the government has tight control over capacity expansion and technology adoption (Besley et al. 1994). In most of the SHI countries, the choice of provider is exercised by patients almost universally. The ability to choose the provider — both physician and hospital — is regarded as a central characteristic of the SHI system and is claimed to be a key aspect of why citizens in SHI countries have higher satisfaction levels on services provided than those in taxfinanced countries, in terms of responsiveness and short waiting time (Figueras et al. 2004). Therefore, citizens in SHI countries are willing to pay more for their health care systems in order to maintain provider choice since proponents often argue that choice of provider reflects an

05 Improving Health

117

8/24/11, 4:23 PM

Sopheap Ly

118

important social decision (Figueras et al. 2004). However, there are also a number of SHI countries which limit the choice of provider; for example, the Netherlands and Switzerland partly restrict the choice of providers (Busse, Saltman & Dubois 2004).9 Although the citizen’s choice is promoted in SHI countries, it is interesting to note that the majority of hospitals in those developed countries are public hospitals. For instance, 70 per cent of hospitals in Austria, 60 per cent in Belgium, and 55 per cent in Germany are public hospitals, while private-not-for-profit hospitals rank second with 25 per cent in Austria, 40 per cent in Belgium, and 40 per cent in Germany, and private-for-profit hospitals take third place (Busse, Saltman & Dubois 2004). In France, 65 per cent of hospital beds are public; private for-profits come second with 20 per cent, followed by private not-for-profits in third place with 15 per cent (Busse, Saltman & Dubois 2004). In most SHI countries, competition between providers is common. For example, in Japan, the service providers for the three insurance schemes (health insurance scheme, national health insurance scheme and health insurance for the elderly scheme) can be both public and private, operating under the supervision of the Ministry of Health, Labour and Welfare, and the prefectures (National Institute of Population and Social Security Research 2007). Generally, the users do not perceive any difference between them and they have freedom to choose “any medical service providers without constraint in terms of hospital type, location or other factors such as having referral or not” (National Institute of Population and Social Security Research 2007). However, because the fee schedule is determined centrally, providers cannot compete on prices, but on quality of services mainly. Also, in Japan, there is no competition between the insurers because the people are insured based on their location or residence and occupation and because the contract between purchasers and providers is in the form of a collective contract (National Institute of Population and Social Security Research 2007; WHO 2005).

4.2 Implications in Developing Countries The following section will examine health care delivery in five countries that were selected as case studies by Hsiao and Shaw (2007): Kenya, Ghana, the Philippines, Colombia, and Thailand. Their implementation of social health insurance differed, from Kenya, which had just begun initial design and legislation, to universal coverage in Thailand (Hsiao & Shaw 2007).

05 Improving Health

118

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

119

It is not yet clear whether Kenya will apply the market mechanism to health care delivery, but the government is expanding public capacity to provide health services (Fraker & Hsiao 2007). In the other four countries, there are both public and private providers; however, the adoption of hierarchy and market mechanisms varies and the degree of competition allowed is different (Ramachandra & Hsiao 2007; Jowett & Hsiao 2007; Pinto & Hsiao 2007; Hanvoravongchai & Hsiao 2007). In Ghana, where there are three separate health insurance schemes (district mutual health insurance scheme, private for–profit scheme, and community-based non-profit scheme), the hierarchy mechanism is applied only in the district scheme, and there is no promotion of competition between health care providers (Ramachandra & Hsiao 2007). In the Philippines, the market mechanism is applied and there is vibrant competition between providers (Jowett & Hsiao 2007). Among the five countries, only Colombia permits competition between insurers, in addition to competition between providers (Pinto & Hsiao 2007). The organization of health care delivery is different from one country to another, and within a country, it also varies between schemes. In Ghana, the only health care providers allowed under the public insurance scheme are public facilities. In this case, the delivery mechanism is hierarchy without competition. In the Philippines and Colombia, the delivery mechanism is the market with competition because there is competition between public and private providers. In Thailand, the social security scheme also applies the mechanism of market with competition. Thailand has distinct and interesting health insurance schemes; therefore let us discuss in detail the delivery mechanism for each. Before the implementation of universal health coverage in 2001, Thailand had four different schemes: (i) Medical Welfare funded by general taxation; (ii) Civil Servants Medical Benefits funded by general taxation as fringe benefits; (iii) Social Security funded by social security contributions and (iv) Health Card funded by voluntary premiums (Hanvoravongchai & Hsiao 2007). Health care delivery was arranged differently under each of the four schemes (Hanvoravongchai & Hsiao 2007). The Medical Welfare scheme used the mechanism of hierarchy without competition because the providers were only public and payment was non-work related (global budget). The Civil Servants Medical Benefits scheme used the mechanism of hierarchy with competition because the providers were public, but payment was work-related (fee for service). The Social

05 Improving Health

119

8/24/11, 4:23 PM

Sopheap Ly

120

Security scheme uses the mechanism of market with competition mechanism because the providers were both public and private and payment was work-related (capitation). The Health Card scheme used the mechanism of hierarchy with competition mechanism because the providers were public and payment was work-related (proportional reimbursement). The use of the hierarchy with competition mechanism in the voluntary integrated model (Health Card scheme) was an exception because this scheme was run by the government, which chose to contract with public providers only. Since the implementation of universal health insurance coverage, Thailand now has three schemes: (i) Universal Coverage scheme funded by general taxation; (ii) Civil Servants Medical Benefits scheme, and Social Security scheme (WHO 2005; Hsiao & Shaw 2007). The first and the third schemes use the market with competition mechanism, and the second scheme uses hierarchy with competition mechanism.

4.3 Variables for Delivery Mechanisms in Developing Countries Unlike developed countries, the situations in developing countries as seen above do not allow for the implementation of a single, nationwide health care delivery mechanism. Mechanisms must be implemented selectively and differently according to social, economic, and political conditions. The analysis below shows that there are three main variables that can lead to different arrangements for social health insurance: (i) levels of providers (primary and hospital care); (ii) location (urban and rural areas), and (iii) time frame (short- and long-term plans).

Levels of Providers This chapter adopts the concept of two levels of health providers as defined by OECD (1992). The first-level providers include general practitioners and pharmacists supplying over-the-counter medicines, while the second-level providers include hospitals and pharmacists who supply prescribed drugs. The OECD (1992) calls these two levels “primary care” and “hospital care” respectively. The distinction between levels of providers is mainly based on economies of scale and government monitoring capacity. In almost all

05 Improving Health

120

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

121

developed countries, the market mechanism is popular in primary care, and the hierarchy mechanism in hospital care (OECD 1992). The reason is that the market mechanism can encourage health care providers to perform effectively and competitively, while the hierarchy mechanism can contain costs more effectively. In addition, according to the World Health Organization (WHO 2000), two service configurations can be distinguished and are to be treated differently: “dispersed service” and “concentrated service” configurations. “Dispersed service configurations are usual for activities which do not benefit from economies of scale … such as primary care … Dispersed, competitive production by small producing units works well wherever markets are a satisfactory way to organise output … Concentrated service configurations are common for activities such as hospital care, central public health laboratories, and health education facilities, which do benefit from economies of scale …” (WHO 2000). The WHO proposes a hierarchy mechanism for concentrated service configurations (hospital care) because of the complexity: “These interventions are highly specialised and expensive, and require large teams of people with a wide range of skills … Most personnel can be employed as regular or part-time staff, rather than under the contractual relationships that appear to be better for dispersed activities.” (WHO 2000) Moreover, according to Mills (1995), contracts for management services and for primary care services with private providers are feasible in developing countries because the primary care services contract is easy to implement and the private sector is more capable of investing in primary care than in hospital services. Therefore, the discussion above shows that the market mechanism is appropriate for primary care, and the hierarchy mechanism for hospital care.

Urban and Rural Areas Mills (1995) also proposed that market mechanisms are more feasible in urban areas than in rural areas because of the difficulties in physically accessing private facilities in rural areas, where such private facilities are less concentrated. Also, Flood (2000) suggested that, before adopting the competition approach for health services, the degree of information asymmetry and asset specificity should be taken into account. In addition, in developing countries, information asymmetry is likely to be greater in

05 Improving Health

121

8/24/11, 4:23 PM

Sopheap Ly

122

rural than in urban areas because people in rural areas are less well educated and informed and more dispersed.

Short and Long-Term Plans A distinction should be made between short-term and long-term plans because of the ability of the government to negotiate and manage contracts, as well as monitor the performance and quality of services (Mills 1995). This ability may develop over time. Also, the government’s ability in securing financial and human resources to introduce competition and contract out services will increase with time if the economy is growing well. According to the life cycle implications outlined by Stigler (1968), the hierarchy mechanism would be feasible in young industries, and the market mechanism, when an industry grows. Stigler’s idea implies that the hierarchy mechanism is feasible when the government has just started to implement social health insurance scheme, and when this scheme grows and matures, the government should also apply the market mechanism. In summary, it is suggested that the hierarchy mechanism should be adopted for organizing health care delivery in the short-term, as well as hospital care in rural areas, while the market mechanism is more useful in the long-term and for primary care in urban areas. However, the application of either mechanism also depends on health care financing methods. For funding health care from general taxation, it is suggested that developing country governments should apply the hierarchy mechanism without competition because through this mechanism, the government can economize on administrative costs and ensure equity more effectively. This mechanism is also feasible for social assistance programmes for the poor that are funded from general taxation. For social health insurance scheme (compulsory health insurance), it is recommended that either the mechanism of hierarchy with competition or the mechanism of market with competition are feasible because, as discussed above, one of the main characteristics is a direct link between contributions and benefits, and the choice of provider is almost universally exercised by patients and regarded as a central characteristic of the system (Figueras et al. 2004). Although it is possible to implement a hierarchy mechanism without competition for social health insurance, its principle of choice of provider will be undermined by non-work-related provider payment methods that encourage perverse incentives for providers.

05 Improving Health

122

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

123

The OECD (1992) perceived that the non-work-related payment will lead to a consequence where “[E]fficient providers are rewarded by more work but not by increased resources. Inefficient providers are rewarded by a quiet life and idle resources.”

4.4 Implications for Cambodian Social Health Insurance Currently, the organization of health care delivery in Cambodia is through a hierarchy mechanism without competition. The public provision of health care is centralized under the supervision and direction of the Ministry of Health, which operates through vertical integrated branches at the provincial and district levels. Although there are three types of health providers, public, private for profit, and private not-for-profit, only public providers are financed by the government, and there is no contracting out10 of health services. Let us now come back to the purpose of this chapter, which is to suggest appropriate health care delivery mechanisms for the social health insurance scheme which the government plans to implement in 2015. The above sections have discussed various possibilities for structuring health care delivery from both theoretical and practical perspectives. In theory, it is recommended that transactions be internalized until the production costs inside an organization outweigh the transaction costs in the market. The imperative for forming an organization is that it is capable of attaining the highest degree of efficiency and facilitating communication and flow of information for correct decision making (Weber 1947; Simon 1997). In addition, in-house production can economize greatly on prices discovery, negotiating, and concluding separate contracts with suppliers, monitoring outcomes, and resolving disputes (Coase 1937; Williamson 1975). The nature of health care is characterized by high uncertainty and high information asymmetry, which, in this case can lead to opportunism, difficulty in designing long-term contracts with private health care providers, and eventually, to high transaction costs. This problem is more acute in developing countries because the capacity to cope, given such market failure issues, is very low (Mills 1995). Therefore, the hierarchy mechanism for health care delivery may be the most appropriate. However, the warning about the potential unlimited expansion of internalization within an organization which may eventually cost more

05 Improving Health

123

8/24/11, 4:23 PM

Sopheap Ly

124

than the transaction costs outside the organization must be heeded. Also, the new public management theory, which criticizes the command-andcontrol theory, supports the market mechanism approach and contractualism (contract making and enforcement) because it can achieve better efficiency. However, the new public management theory does not only refer to the fact that a service must be contracted out to the market (though this is one of the elements of this theory), it also refers to the adoption of private management style inside the bureaucracy, or in delivering public services. Although Mills (1995) does not favour applying market mechanism for health care delivery in developing countries, she proposes two things if the government wants to engage the private sector. First, private not-forprofit providers may be potential partners. Second, selective contracting with the private sector may be feasible, including contracts in urban areas, for non-clinical, management, and primary care services. The new public management theory also advocates choice and competition in accessing and providing a service. According to Le Grand (2007), allowing choice and competition has the best prospect of delivering a good service because it can ensure responsiveness, efficiency, high quality services, equity, and social justice. Le Grand (2007) clearly stated that introducing the element of competition into public services should not be confused with the privatization of those services. This means that the hierarchy mechanism can incorporate competition. However, introducing competition in developing countries is opposed by Mills and Broomberg (1994) because those countries have poor information systems and a low ability to create a competitive environment. They suggested that long-term relationships with either public or private health care providers should be developed instead. To sum up, it is found that there are four mechanisms which can be applied in health care delivery: (i) hierarchy mechanism without competition, (ii) hierarchy mechanism with competition, (iii) market mechanism without competition, and (iv) market mechanism with competition. To understand how these mechanisms are applied in practice, it is necessary to analyse the experiences of both developed and developing countries in structuring their health care delivery mechanisms for SHI system. Practice in most developed countries shows that choice of public and private providers is the norm and is a basic characteristic of the SHI

05 Improving Health

124

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

125

system. The main reason for this is the linkage between contributions and benefits, and the need to guarantee high satisfaction levels of services provided, in terms of responsiveness and short waiting time. Nevertheless, in most of those countries, the majority of hospitals are public hospitals and the private not-for-profit hospitals and private-for-profit hospitals rank second and third. Experiences of applying health care delivery mechanisms in developing countries are mixed. Ghana’s social health insurance scheme implements the hierarchy mechanism without competition. In the Philippines and Colombia, the market mechanism with competition is applied. In Thailand, the SHI scheme for civil servants applies the hierarchy mechanism without competition, while the SHI scheme for private workers applies the market mechanism with competition. The difference between these two schemes may stem from the fact that in the former scheme, the contributions are totally paid from general taxation (from the government) as fringe benefits for civil servants, whereas in the latter scheme, private workers have to share the contributions with their employers. From the above observation of experiences in both developed and developing countries, only two mechanisms are applied in their SHI schemes: hierarchy mechanism with competition and market mechanism with competition. In addition, in developing countries, the appropriate choice of a mechanism also depends on three main variables, including (i) levels of providers, (ii) location and (iii) time frame. For the Cambodian social health insurance scheme (civil servants and private workers),11 this chapter suggests that for the short-term plan (2015– 20), the hierarchy mechanism with competition should be applied for delivering hospital care in urban and rural areas, and primary care in the rural areas; and the market mechanism with competition should be applied only for primary care in the urban areas. For the long-term plan (post 2020), the hierarchy mechanism with competition should be applied in the rural areas for both primary and hospital care, and the market mechanism with competition should be applied in the urban areas for both primary and hospital care. The definition of primary and hospital care may vary. The reason this health care delivery should be structured in this way is explained below. For the Cambodian case, this chapter has identified primary care, that is, first-level health care providers as: health centres and operational district referral hospitals in the public sector, and general practitioners and clinics

05 Improving Health

125

8/24/11, 4:23 PM

Sopheap Ly

126

that provide comparable services in the private sector. Public facilities above operational district referral hospitals, including provincial and national hospitals and private facilities that provide comparable services, are identified as providing hospital care, that is, second-level providers. Private health care providers can be both for-profit and not-for-profit providers. Urban areas in Cambodia refer to the capital and provincial towns. The short-term time frame is from 2015 (when implementation of the social health insurance scheme is set to start) to 2020, and long-term is after 2020. The results from the combination of those variables are shown in Tables 5.1 and 5.2.

TABLE 5.1 Mechanisms for Social Health Insurance Scheme in Short-Term Plan (2015–20) Location

Urban areas

Rural areas

Primary care

Market mechanism with competition A

Hierarchy mechanism with competition C

Hospital care

Hierarchy mechanism with competition B

Hierarchy mechanism with competition D

Provider levels

TABLE 5.2 Mechanisms for Social Health Insurance Scheme in the Long-Term (post-2020) Location

Urban areas

Rural areas

Primary care

Market mechanism with competition E

Hierarchy mechanism with competition G

Hospital care

Market mechanism with competition F

Hierarchy mechanism with competition H

Provider levels

05 Improving Health

126

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

127

The only difference between models A and B is the level of providers. In model A the market mechanism with competition is proposed for primary care, and in B, the hierarchy mechanism with competition is proposed for hospital care. The rationale relates to economies of scale and the monitoring capacity of the government. Models A and C have one differing variable: location. In urban areas, the market mechanism with competition is proposed and in rural areas, the hierarchy mechanism with competition, for four reasons: (1) implementation of the SHI scheme will increase demand for health care dramatically; (2) the majority of employees are employed in manufacturing enterprises and the majority of civil servants live in urban areas; (3) in urban areas there are more private facilities than public facilities for primary care; and (4) it is unlikely that public facilities alone will have enough capacity to respond to high urban demands for health care. Hence private providers should be engaged in primary care in both short and long terms for urban areas (models A and E). They should also be engaged in hospital care in the long-term (F). Although one variable — the level of providers — differs, models C and D employ the same mechanism, the hierarchy mechanism with competition, for five reasons: (1) the location is rural areas, so the number of private providers is limited and the demands for health services from salaried workers are low; (2) there is information asymmetry between the government and providers, and between providers and patients; (3) opportunism of private providers; (4) low capacity of the government to monitor the performance of private providers; and (5) the government can contain costs more effectively through the hierarchy mechanism. These reasons are the same in models G and H. The reason for the adoption of the market mechanism with competition in model F is similar to the reason for adopting it in model A in that there will be a high demand for health services. This model is suggested for the long term because the government needs time to build its capacity to negotiate, conclude and monitor contracts with providers, and to deal with market failure. With regard to the implementation of the market mechanism with competition in hospital care, Cambodia should consider the experiences of Thailand (see WHO 2005; Hsiao & Shaw 2007). In Thailand, private health care providers are required to build their own network as a contractor with insurers and to register their clients with the network. As a result, those private providers receive capitation payments based on the number of registered clients on their list.

05 Improving Health

127

8/24/11, 4:23 PM

Sopheap Ly

128

This chapter has focused only on the question of structuring health care delivery for the SHI scheme; it does not deal with the issues of calculation and collection of contributions, benefit packages, payment methods, and the purchasing of health services. However, as a suggestion for the future implementation of the SHI scheme, it proposes a payment mechanism and the purchasing of health services as briefly outlined below. The provider payment mechanism suggested for both the hierarchy mechanism with competition and the market mechanism with competition is capitation payment because this allows the government to economize on administrative costs. However, because this type of payment encourages underproduction of health services, the government has to establish an effective monitoring system to assess and control the quality of services. In structuring the relationship between purchasers for health services for the insured and providers of health care services under the SHI scheme, the recent establishment and organization of Special Operating Agencies (SOAs) is very favourable.12 According to the draft manual, in order to manage SOAs, the commissioners or purchasers are distinct from health service provider SOAs. In the case where operational districts and provincial hospitals are health service provider SOAs, the provincial health departments play a role as commissioners (purchasers), which have to conclude “performance agreements” with the Ministry of Health to take charge of the role as commissioner. In the case where regional training centres and national centres are SOAs, the Ministry of Health is the commissioner. The draft manual on SOAs does not mention whether national hospitals can become SOAs. The contract between commissioner and SOAs is called a “service delivery management contract”. Being SOAs, the operational districts and provincial hospitals can subcontract with health centres or referral hospitals of the operational districts (MoH 2009a). The source of funding for those SOAs is from the government, health partners (service delivery grants), and other revenues such as user fees and revenue from the communitybased health insurance schemes (MoH 2009a). The SOAs final draft manual neither defines what the role of SAOs should play when the SHI scheme is implemented, nor specifies clearly if the contributions for the SHI scheme could be a source of revenue for SOAs. Also, only the Ministry of Health and Provincial Health Department can be the commissioners or purchasers. It is understood from the final draft of the Master Plan for Social Health Protection in 2009 that the SHI

05 Improving Health

128

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

129

schemes for civil servants will be managed by the Ministry of Social Affairs, and the SHI schemes for private workers will be managed by the Ministry of Labour. In such a case, the Ministry of Social Affairs and the Ministry of Labour can enter into a contract (performance agreements) with the commissioners (provincial health department). Having the provincial health department as purchaser will ease the implementation of the hierarchy mechanism with competition. However, for the implementation of the market mechanism in urban areas, the role of the provincial health department will be expanded to include contracting with private providers, which in this case would deviate from its original role as specified in the SAOs draft manual. In addition to contracting with national hospitals, the responsible ministries will have to be the purchasers. To facilitate the negotiation, conclusion, and monitoring of contracts with health care providers, the government should encourage the formation and establishment of providers’ associations.

CONCLUSION In developing countries, health care financing and health care delivery reforms have been initiated over the past two decades with technical and financial assistance from many international aid agencies, which mainly support adoption of the social health insurance system. For this reason, health financing reforms in most developing countries have focused on shifting from out-of-pocket payment to voluntary and compulsory health insurance. For health care delivery, the reforms mainly focus on strengthening the management and capacity of public health providers. However, given limited government management capacity and resources, this study suggests that developing countries should adopt an incremental reform approach to health care delivery in engaging the private sector, taking the level of providers, location, and time frame into consideration. When Cambodia implements its social health insurance system it will experience double movements: a gradual shift from the current tax-financed system and out-of-pocket payment towards prepayment, and from hierarchy without competition, to hierarchy with competition mechanisms and market mechanism with competition. This chapter strongly recommends that the government reconsider the potential of using private health care providers, especially private not-forprofit providers, in the case where there is no plan to include them as

05 Improving Health

129

8/24/11, 4:23 PM

Sopheap Ly

130

health care providers for the SHI scheme. To realize this, the government should speed up the establishment of an accreditation system for private health care providers.

Notes 1. The remaining 8 per cent do not seek any treatment (National Institute of Public Health 2006). 2. Ninety-two per cent of ill or injured household members sought at least one treatment, 27 per cent sought at least two treatments, and 10 per cent sought at least three treatments (National Institute of Public Health 2006). Although there is no big difference in the use of public facilities between the rich and the poor, the rich tend to use more costly public facilities than the poor (World Bank 2006). 3. According to Batt (1967), there are two important features of the employment contract that can be distinguished from an independent contract: first, “The servant must be under the duty of rendering personal services to the master or to others on behalf of the master, otherwise the contract is a contract for sale of goods or the like”; and second, “The master must have the right to control the servant’s work, either personally or by another servant or agent. It is this right of control or interference, of being entitled to tell the servant when to work (within the hours of service) and when not to work, and what work to do and how to do it (within the terms of such service) which is the dominant characteristic in this relation and marks off the servant from an independent contractor”. 4. Confusion over the term “social health insurance” used in the 2005 master plan for social health insurance in Cambodia is partly because in the context of developed countries, the term, “social health insurance”, also called “social insurance”, refers to the compulsory or mandatory health insurance which targets all the people, except the poor. Schemes offering health care for the poor or some groups of people such as the elderly or children are usually called “social assistance” or “social welfare”. For example, in Japan, the social security system comprises social insurance and social welfare (social assistance) (National Institute of Population and Social Security Research 2007). Also, the social protection system in Germany consists of “social insurance” and “social assistance” (Eichenhofer 2002). On the other hand, Hsiao and Shaw (2007) observe that the definition of “social health insurance” has been modified in developing countries and has thus deviated from its conventional meaning as used in developed countries. 5. For civil servants, the social security system has been provided in the Law on the Common Statute of Civil Servants in 1994, the Subdecree on the Regime of

05 Improving Health

130

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

6.

7.

8.

9.

10.

05 Improving Health

131

Retirement Pensions and Professional Unfitness of Civil Servants in 1997, the Royal Decree on the Social Security Benefits of the Civil Servants in January 2008, and the Subdecree on National Social Security Fund for civil servants in February 2008. The Law and the Subdecree regulate old age pension or retirement pension for retired civil servants. This subdecree was further implemented by a number of ministerial prakas which include prakas on Registration of Enterprise/Establishment and Employees in the National Social Security Fund, and Prakas on Determination of the Phase and Scope of Implementation of Occupational Risk Scheme issued on 11 February 2008 and Prakas on Determination of Contribution Rate, Prakas on Benefits of Workmen Compensation, and Prakas on Notification of Occupational Risks, issued on 16 June 2008. The other five principles are official jurisdictional areas, written rules, specialization, promotion based on competence, and performance-based salary (Weber 1978). Knight is defined as people who are predominantly public-spirited or altruistic, and pawn is defined as passive victims or passive recipients of services (Le Grand 2003). In contrast to knight and pawns, knave is defined as people who are motivated by their self-interest, and queen is defined as active agents, workers, or recipients (Le Grand 2003). The choice of provider is limited in the Netherlands because the Netherlands has introduced the managed competition model, according to which a sickness fund has its own list of providers with whom it contracts (see Flood 2000). Thus, if a person registers with that sickness fund, he/she can only get services from physicians or hospitals which contract with that sickness fund. The case of Switzerland is quite similar to that of the Netherlands. In Switzerland, there are Health Maintenance Organizations and Preferred Provider Organization which function like a sickness fund in the Netherlands (have a list of contracted physicians and hospitals) (Busse, Saltman & Dubois 2004). In addition, in principle, the people of Switzerland do not have free access to providers outside their canton, except in the case of an emergency (Busse, Saltman & Dubois 2004). However, in those countries (the Netherlands and Switzerland) people have the choice of insurer; thus, their choice of provider is followed by their choice of insurer. Cambodia piloted the contracting in and contracting out of health services from 1999–2002. However, the pure meaning of contracting out of health services here means when the government purchases a health service from an outside source (can be private for-profit or private not-for-profit health care providers) which uses its own workforce and resources as defined by Harding and Preker (2003). According to Harding and Preker (2003), it is called “contracting-in” when government hires “outside managers to come in and manage an internal workforce or service”. Therefore, both the contracting in

131

8/24/11, 4:23 PM

Sopheap Ly

132

and contracting out schemes which were piloted in Cambodia from 1999–2002 and which are different based on their degree of autonomy in using workforce and resources, may be best called “contracting for health service management” or “contracting-in”. 11. Both civil servants and private workers will have to share contributions for the social health insurance scheme with the government and private employers (MoH 2005). 12. The final draft of Social Operating Agencies Manual (MoH 2009) defines the SOAs as “new organisation in the health sector. They are part of the Ministry of Health but given operational autonomy, within an agreed mandate, to provide services under contract to the Ministry”.

References Annear, Peter Leslie. “Study of Financial Access to Health Services for the Poor in Cambodia, Phase 1: Scope, Design and Data Analysis”. Report for Ministry of Health of Cambodia, WHO, AusAID, and RMIT University, 2006. ———. “Developing a Strategy for Social Health Protection in Cambodia”. Paper presented at the UNESCAP “National Workshop on Promoting Sustainable Strategies to Improve Access to Health Care in the Greater Mekong Subregion”. held in Bangkok, on 23–25 July 2008. Arrow, Kenneth. “Uncertainty and the Welfare Economics of Medical Care”. American Economic Review 50, no. 5 (1963): 941–73. Bartlett, Will & Julian le Grand. “The Theory of Quasi-Markets”. In Quasi-markets and Social Policy, edited by Julian le Grand and Will Bartlett, pp. 13–34. Houndmills: Macmillan, 1993. Batt, Francis Raleigh. The Law of Master and Servant. London: Sir Isaac Pitman and Sons Ltd, 1967. Bautista, Maria Cristina G. Health Financing Schemes in Cambodia: Reaching the Poor with Quality Health Services. Phnom Penh: University Research Co., 2003. Besley, Timothy, Miguel Gouveia, and Jacques Dreze. “Alternative Systems of Health Care Provision”. Economic Policy 9, no. 19 (1994): 199–258. Broomberg, J. “Managing the Healthcare Market in Developing Countries: Prospects and Problems”. Health Policy and Planning 9, no. 3 (1994): 237–51. Busse, Reinhard, Richard Saltman, & Hands F. W. Dubois (eds.). “Organisation and Financing of Social Health Insurance Systems: Current Status and Recent Policy Development”. In Social Health Insurance Systems in Western Europe, edited by Saltman, Busse and Figueras. Maidenhead: Open University Press, 2004. Coase, R.H. “The Nature of the Firm”. Economica 4 (1937): 368–405.

05 Improving Health

132

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

133

Eichenhofer, Eberhard. “The Social Insurance in Germany: The Bismarck Model and its Challenge in the 21st Century”. In Reform and Perspectives on Social Insurance: Lessons from the East and West, edited by Ming-Cheng Kuo, Hans F. Zacher and Hou-Sheng Chan. Kluwer Law International, 2002. Figueras, Josep, Richard B. Saltman, Reinhard Busse, & Hands F.W. Dubois. “Patterns and Performance in Social Health Insurance Systems”. In Social Health Insurance Systems in Western Europe, edited by R.B. Saltman, R. Busse, & J. Figueras. Maidenhead: Open University Press, 2004. Flood, Colleen M. International Health Care Reform: A Legal, Economic and Political Analysis. London: Routledge, 2000. Fraker, Andrew & William C. Hsiao. “Kenya: Designing Social Health Insurance”. In Social Health Insurance for Developing Nations, edited by W.C. Hsiao & R.P. Shaw. Washington: World Bank, 2007. Frederickson, H. George. “Ethics and the New Managerialism”. Public Administration & Management: An Interactive Journal 4, no. 2 (1999): 299–324. Gottret, Pablo & George Schieber. Health Financing Revisited: A Practitioner’s Guide. Washington: World Bank, 2006. Hanvoravongchai, Piya & William C. Hsiao. “Thailand: Achieving Universal Coverage with Social Health Insurance”. In Social Health Insurance for Developing Nations, edited by W.C. Hsiao & R.P. Shaw. Washington: World Bank, 2007. Harding, April & Alexander S. Preker. Private Participation in Health Services. Washington: World Bank, 2003. Hirschman, Albert. Exit, Voice, and Loyalty. Cambridge: Harvard University Press, 1970. Hood, Christopher. “A Public Management for All Seasons?”. Public Administration 69, no. 1 (1991): 3–19. Hsiao, William C. & R. Paul Shaw (eds.). Social Health Insurance for Developing Nations. Washington: World Bank, 2007. Jowett, Matthew & William C. Hsiao. “The Philippines: Extending Coverage beyond the Formal Sector”. In Social Health Insurance for Developing Nations, edited by W.C. Hsiao & R.P. Shaw. Washington: World Bank, 2007. Lane, Jan-Erik. New Public Management. London: Routledge, 2000. Le Grand, Julian. Motivation, Agency and Public Policy: Of Knights and Knaves, Pawns and Queens. Oxford: Oxford University Press, 2003. Le Grand, Julian. The Other Invisible Hand: Delivering Public Services through Choice and Competition. Princeton University Press, 2007. Loevinsohn, Benjamin. Performance-based contracting for Health Services in Developing Countries: A Toolkit. Washington: World Bank, 2008. Marquand, David. Decline of the Public: The Hollowing Out of Citizenship. Cambridge: Polity Press, 2004.

05 Improving Health

133

8/24/11, 4:23 PM

Sopheap Ly

134

Mills, Anne. “Improving the Efficiency of Public Sector Health Services in Developing Countries: Bureaucratic Versus Market Approaches”. Health Economics and Financing Programme Working Paper 01/95, 1995. Available at . Mills, Anne, Sara Bennett & Steven Russell. Challenge of Health Sector Reform: What Must Government Do?. New York: Palgrave, 2001. Ministry of Health. Master Plan for Social Health Insurance in Cambodia. Phnom Penh: Ministry of Health, 2005. ———. Guideline for the Implementation of Community Based Health Insurance. Phnom Penh: Ministry of Health, 2006. ———. Health Strategic Plan 2008–2015. Phnom Penh: Ministry of Health, 2008a. ———. Strategic Framework for Health Financing 2008–2015. Phnom Penh, Ministry of Health, 2008b. ———. Joint Annual Performance Review 2008. Phnom Penh, Ministry of Health, 2008c. ———. “Special Operating Agency Manual”. Final draft. Phnom Penh: Ministry of Health, 2009a. ———. “Master Plan for Social Health Protection”. Unpublished final draft. Phnom Penh: Ministry of Health, 2009b. Mommsen, Wolfgang J. The Political and Social Theory of Max Weber. Cambridge: Polity Press, 1989. National Institute of Population and Social Security Research. “Social Security in Japan 2007”. Available at (accessed 1 December 2008). National Institute of Public Health, National Institute of Statistics [Cambodia] and ORC Macro. Cambodia Demographic and Health Survey 2005. Phnom Penh & Calverton, MD: National Institute of Public Health, National Institute of Statistics and ORC Macro, 2006. National Institute of Statistics. Cambodia Socio-Economic Survey 2004. Phnom Penh: National Institute of Statistics, 2004. Organization for Economic Cooperation and Development. The Reform of Health Care: A Comparative Analysis of Seven OECD Countries. Paris: OECD, 1992. Pinto, Diana & William C. Hsiao. “Colombia: Social Health Insurance with Managed Competition to Improve Health Care Delivery”. In Social Health Insurance for Developing Nations, edited by W.C. Hsiao & R.P. Shaw. Washington: World Bank, 2007. Ramachandra, Sreekanth & William C. Hsiao. “Ghana: Initiating Social Health Insurance”. In Social Health Insurance for Developing Nations, edited by W.C. Hsiao & R.P. Shaw. Washington: World Bank, 2007.

05 Improving Health

134

8/24/11, 4:23 PM

Social Health Insurance in Cambodia

135

Simon, Herbert A. Administrative Behaviour, a Study of Decision-Making Processes in Administrative Organizations. 4th ed. New York: Free Press, 1997. Simon, Herbert A. Models of Man: Social and Rational. New York: John Wiley & Sons, 1957. Stigler, George J. The Organization of Industry. Homewood, IL: Richard D. Irwin, Inc., 1968. Titmuss, Richard M. The Gift Relationship. New edition edited by A. Oakley & J. Ashton. London: London School of Economics, 1997. Weber, Max. Economy and Society: an Outline of Interpretive Sociology, edited by Guenther Roth & Claus Wittich, and translated by Ephraim Fischoff et al. Berkeley: University of California Press, 1978. ———. From Max Weber Essays in Sociology. Translated by H.H. Gerth & C. Wright Mills. Oxford: Oxford University Press, 1947. Williamson, O.E. Markets and Hierarchies: Analysis and Anti-trust Implications. New York: The Free Press, 1975. Williamson, O.E. “The Vertical Integration of Production: Market Failure Considerations”. The American Economic Review 61, no. 2 (1971): 112–23. World Bank. Cambodia: Halving Poverty by 2015? Poverty Assessment 2006. Washington: World Bank, 2006. World Health Organization. Social Health Insurance: Selected Case Studies from Asia and the Pacific. Geneva: WHO, 2005. ———. The World Health Report 2000, Health Systems: Improving Performance. Geneva: WHO, 2000. Yamamoto, Hiromi. “New Public Management: Japan’s Practice”. Institute for International Policies Studies Policy Paper 293E. Tokyo: Institute for International Policy Studies, 2003. Cambodian Laws and Regulations: Royal Kram No. 06/NS/94. Promulgating the Law on the Common Statute of Civil Servants, 26 October 1994. Royal Kram No. NS/RKM/0700/02. Promulgating Insurance Law. 25 July 2000. Royal Kram. Promulgating the Labour Law. 13 March 1997. Royal Kram No. 018/NS/RKM/0902. Promulgating Law on Social Security Scheme for Persons Defined by the Provisions of the Labour Law. 25 September 2002. Royal Decree on the Social Security Benefits of the Civil Servants. 22 January 2008. Subdecree No. 059/ANK/BK on the Regime of Retirement Pensions and Professional Unfitness of Civil Servants. 6 October1997. Subdecree on the Establishment of National Social Security Fund. 5 February 2007. Subdecree on the Establishment of National Social Security Fund for Civil Servant. 5 February 2008.

05 Improving Health

135

8/24/11, 4:23 PM

Toni Ashton and Mar ia C.G. Bautista

136

6 PURCHASING HEALTH SERVICES IN NEW ZEALAND Toni Ashton and Maria C.G. Bautista

INTRODUCTION New Zealand and Cambodia are both situated on the rim of the western Pacific Ocean. Beyond this, they appear to share few common features. Cambodia is a developing country with a GDP per capita of under US$600 and a population of more than fourteen million. New Zealand has a GDP per capita of over US$27,000 and a population of only four million. While Cambodia spent 5.9 per cent of its GDP on health care in 2005–06, New Zealand spent 9.4 per cent, slightly higher than the OECD average of 8.9 per cent. New Zealand compares favourably with the rest of the OECD countries for infant mortality and life expectancy at birth. Despite the contrast between the two nations, recent reforms in the public health system in New Zealand have some lessons for Cambodia, especially with respect to purchasing and contracting. New Zealand’s health system is predominantly publicly funded, with approximately 67 per cent of health expenditure coming from general taxation and a further 9 per cent from social insurance for accident-

06 Improving Health

136

8/24/11, 4:24 PM

Purchasing Health Services in New Zealand

137

related injuries. Out-of-pocket payments account for 16 per cent of total expenditure and private insurance for about 5 per cent (Ministry of Health 2008). User charges apply for general practice consultations and pharmaceuticals, but most other services, including maternity services, are provided free of charge. While personal health services are provided universally regardless of income, some social support and rehabilitation services are income tested. Service delivery is shared by public and private providers. All major population centres have at least one public hospital. Most larger centres also have one or more private hospitals that provide mostly surgical procedures on a fee-for-service basis, paid for either out of pocket or via private health insurance. Community-based services are generally delivered by private providers, including individual and group-owned practices, NGOs, and for-profit organizations. These arrangements have remained broadly the same since the public health system was first introduced in 1938 (Ashton et al. 2005). However, as developments in modern medicine have put increasing pressure on the system, successive governments have sought to restructure it in order to improve efficiency while maintaining universal access (Gauld 2001). This has resulted in a series of reforms, the particular design of which has been determined largely by the ideology of the government of the time. Each round of restructuring has focused primarily on how resources are allocated from the central government to providers, a key part being how to purchase services effectively. This focus makes New Zealand’s health reforms of special interest to Cambodia. In this chapter we describe these reforms and explore some of the substantive issues around the various purchasing arrangements and methods of contracting services.

A BRIEF HISTORY OF NEW ZEALAND’S REFORMS Prior to the 1990s, funding for services that were publicly provided — that is, public hospital services, public health services, and some communitybased services — was allocated via a population formula to fourteen area health boards that were responsible for planning services for their populations. Funding for any services of private providers, especially primary health care, remained the responsibility of departments within

06 Improving Health

137

8/24/11, 4:24 PM

Toni Ashton and Mar ia C.G. Bautista

138

the central government, particularly the Department of Health. The two different funding streams inhibited service planning by the area health boards and restricted integration between primary and secondary services. It also meant that, while the government had relatively tight control over area health board budgets, much of the funding to private providers was open ended and demand driven. In 1991, a new centre-right government decided that efficiency would be improved if the functions of purchaser and provider previously undertaken by the area health boards were performed by separate agencies. The idea was to establish a quasi-market in which both public and private providers would compete for contracts with regional purchasers. The area health boards were immediately abolished and, after two years of transition, the new regime was introduced on 1 July 1993. Most of the funding for primary and secondary care, as well as that for disability support services, was combined into a single funding envelope and distributed to four newly established regional purchasing agencies called Regional Health Authorities (RHAs). Treatments for accident-related injuries continued to be purchased separately by the Accident Compensation Commission. The RHAs negotiated funding agreements with the Ministry of Health, which indicated the proposed levels of service (broadly specified) and their budgets. The RHAs then purchased services via contracts from public hospitals — which were reconfigured as for-profit businesses called Crown Health Enterprises (CHEs) — and from private providers. A parallel system was set up for the purchase of public health services through a single agency, the Public Health Commission. However, while these arrangements were successful in raising the profile of public health services, this agency was abolished after eighteen months and its purchasing responsibilities were shared between the RHAs and Ministry of Health. Reasons given by the government included a lack of clarity about the roles of the Public Health Commission and the Ministry of Health, and a lack of any clear line of accountability from the Ministry of Health to public health services providers (Gauld 2001). In 1996, a new coalition government introduced further restructuring. This time, the purchasing function was centralized, the four RHAs being replaced by a single Health Funding Authority (HFA). While public hospitals were still required to act in a business-like manner, they were reconfigured as not-for-profit organizations in which “the principles of public service” were to replace the profit motive (Coalition Government

06 Improving Health

138

8/24/11, 4:24 PM

Purchasing Health Services in New Zealand

139

1996). The changes were in part a response to opposition by one of the coalition partners, which was ideologically opposed to a public health system being profit-oriented, but they were also designed to reduce transaction costs and potential inequities associated with contracting by multiple regional purchasers (Ashton et al. 2004). In 1999, a general election brought in a centre-left coalition government led by the Labour Party which claimed that the previous government had allowed the public health system to be “run down, privatized and commercialized” (New Zealand Labour Party 1999). It considered that basing a public health system on competitive tendering for contracts was inappropriate and that the system lacked democratic input. This stimulated another round of reform. The Health Funding Authority was abolished and most of its purchasing functions were devolved to twenty-one newly established district health boards (DHBs) funded on a population basis. (The Ministry of Health absorbed the responsibility for purchasing some services at the national level, including public health services.) As well as being responsible for purchasing all personal health services, the DHBs both own and operate the public hospitals. Thus, unlike the RHAs and the HFA, the DHBs are both purchasers and providers of services. In 2009, an incoming centre-right coalition government commissioned yet another review of the structure of the public health system (Ministerial Review Group 2009). The brief of the review group was to recommend ways of improving the quality and performance of the system without major changes to its basic structure. Based on these recommendations, the government decided to establish a new National Health Board within the Ministry of Health to recentralize some of the functions of the twentyone DHBs. These include the planning and funding of infrastructure and of specialist national services. The National Health Board will also be responsible for monitoring the DHBs, and a Shared Services Establishment Board has been appointed to consolidate some of the administrative functions of the DHBs, such as payroll and procurement (Minister of Health 2009). As was the case in 1996, this reversion to some centralized purchasing is in response to a perceived need to reduce duplication of decision making and the associated bureaucracy, inefficiency, and transaction costs. Approximately 20 per cent of the health budget (NZ$2.5 billion), previously held by the Ministry of Health, is now being handed over to the National Health Board, leaving the Ministry of Health responsible primarily for core policy and regulatory roles. About

06 Improving Health

139

8/24/11, 4:24 PM

Toni Ashton and Mar ia C.G. Bautista

140

500 civil service employees are expected to be shed following this reorganization. Table 6.1 summarizes the historical discussion of the New Zealand health system in terms of the purchasing organizations, the underlying political environment and ethos of reforms, and the institutional configurations and roles.

METHODS OF CONTRACTING FOR SERVICES Contracting for services was first introduced into the public health system in the early 1990s. At that time, area health boards, which accounted for about two-thirds of public funding, were responsible for both planning and providing services. From 1990, in return for their share of government funds, each board was required to sign a “contract” with the minister of health that specified the range of services it planned to provide, together with a set of performance indicators. This did not mean that expenditure was linked to a board’s output: while the contracts did include some rather crude measures of particular outputs, most service outputs remained unspecified and unmonitored, and quality measures were absent (Ashton 1999). Instead contracts were based on operating plans agreed with the minister that were consistent with the boards’ five-year strategic plans. Thus a key objective of this early contracting was to make explicit the planning, as opposed to the actual provision, of services. Each board’s performance was then measured against its agreed plan. The absence of any output measures, or of competitive pressures, meant that area health boards faced few incentives to operate efficiently. Payment mechanisms for private providers varied, with NGOs usually being paid via a block grant, and primary health care providers on a fee-for-service basis. Regardless of the method of payment, there was almost no monitoring of services and hence very little accountability from providers. In the case of block contracts, levels of provision were commonly specified in terms of inputs — usually the amount of time to be provided by different types of health professionals — rather than outputs. Thus wage rates were in effect set by the government as the purchasing agent. The introduction of the quasi-market in 1993 required that all services be purchased through contracts between the four RHAs and service providers. At this time, few people in New Zealand had much knowledge or experience of the practicalities associated with negotiating, specifying, and monitoring health services contracts. Moreover, the RHAs were not

06 Improving Health

140

8/24/11, 4:24 PM

06 Improving Health

Purchasing Organizations

141

Health Funding Authority (HFA)

21 District Health Boards (DHBs)

District Health Boards + National Health Board

1997–99

2000–10

2010

8/24/11, 4:24 PM

Conservative-led coalition

Labour-led coalition (1999–2008) Conservative-led coalition (from 2008)

Conservative-led coalition (1996–99)

4 Regional Health Conservative Authorities (RHAs) (1991–96)

Labour (1984–91)

Government

1993–96

1980s–91 14 Area Health Boards (AHBs)

Period

Stronger national leadership. Reduced bureaucracy. Increased regional collaboration.

Decentralization, community participation and collaboration. DHBs have elected boards.

Centralization; market model retained, but with more collaboration.

Market model; purchaserprovider split. Boards of directors of RHAs and CHEs appointed by the minister.

Social welfare model; AHBs have elected boards.

Underlying Ethos/ Policy Framework

Establishment of a National Health Board to undertake planning and funding of ser vices that are provided at a national le vel, monitor DHBs, and to oversee whole-of-system performance.

Funding devolved to 21 DHBs, which pro vide hospital services (and some community-based services) and purchase other ser vices from NGOs and other private providers. Primary care purchased through primary health organizations (PHOs), which are non-profit networks of general practitioners and other primary health providers. PHOs paid per capita subsidies for their enrolled populations.

Single national purchasing agency; CHEs reconfigured as non-profit and renamed Hospital and Health Services; some services shed to community trusts or NGOs; GPs negotiate for contracts through independent practitioner associations (IPA)

All primary and secondary care purchased by RHAs from competing providers. Public hospitals (now called Crown Health Enterprises–CHEs) structured as for-profit businesses. Public health services purchased by the Public Health Commission until 1995.

AHBs provide hospital and public health services. Primary care funded separately by Department of Health. Increasing emphasis on ser vice delivery, management, and accountability. Shift from cure to prevention.

Institutional configurations and roles

TABLE 6.1 Institutional Arrangements in New Zealand since the 1980s Purchasing Health Services in New Zealand

141

Toni Ashton and Mar ia C.G. Bautista

142

given any guidance about purchasing strategies, such as how to specify services or set prices. Because the contracts were legally binding, the RHAs initially looked towards commercial contract models for guidance. This resulted in prolonged — and often adversarial — negotiations and very detailed and lengthy contracts (Ashton et al. 2004). Many service providers, who were used to working more collaboratively, felt quite alienated by this new competitive environment, and a number of senior managers resigned (Malcolm et al. 1996). For some services — especially hospital-based services — there was only one service provider in the area, so a bilateral monopoly existed rather than a competitive market. While the RHAs did try opening up some services to competitive tendering, the vast majority of contracts went to incumbent providers. Nevertheless, all contracts were contestable, and the threat of the possible loss of a contract encouraged even longstanding monopoly providers to focus more on both the cost and the quality of their services (Ashton et al. 2004). Unfortunately, the focus on costs did not appear to lead to any improvement in the efficiency of service provision. Indeed, after three years of these quasi-market arrangements, the agency responsible for monitoring the RHAs reported: “… the pace of performance seems, if anything, to have declined since the advent of the reforms” (Crown Company Monitoring and Advisory Unit 1996). Possible reasons for this include: the absence of competition for many services, the fact that providers had not had sufficient time to make the changes necessary to secure efficiency gains, and increased transaction costs associated with contracting (Gauld 2001; Ashton et al. 2004; Cumming & Mays 2002). Over time, and especially following the replacement of the four RHAs by a single purchasing authority, the style of contracting changed. The HFA worked more closely with service providers in planning and developing services, so that more relational styles of contracting began to replace the original arm’s length commercial approach, and the duration of contracts increased. The HFA also worked with service providers to develop a national framework that provided a standardized template for contracts. The framework specified a common set of terms and conditions, defined service specifications broadly and set a national pricing schedule for hospital services, based on estimates of the cost of efficient provision. All of these changes reduced the costs of contracting and generally improved purchaser-provider relationships. At the same time, the continued

06 Improving Health

142

8/24/11, 4:24 PM

Purchasing Health Services in New Zealand

143

use of formal written agreements ensured that providers maintained a focus on both costs and quality of services. The establishment of DHBs in 2000 and the reintroduction of democratically elected boards was expected to bring about a tidal change to purchasing, with any remnants of competitive contracting being replaced by a more collaborative approach between the DHBs and providers. Contracts now became known as “service agreements”, in line with this new orientation. In the case of hospital services, the fact that DHBs own the public hospitals means that their monopoly status is now reinforced, leaving little space for non-government providers to win contracts, even if they could provide services more efficiently. And where some competition does exist, the development of stronger partnerships between the DHBs and NGOs has been inhibited by the fact that multiple providers are still competing for limited funds. Not surprisingly, the DHBs remain unwilling to enter into longer term contracts with NGOs in circumstances where competition could potentially put downward pressure on prices and/or encourage higher quality (Ashton et al. 2008). Overall, the shifts that have occurred in New Zealand suggest that a more relational style of contracting, in which purchasers and providers work together to plan services, reduces transaction costs and improves stability of supply. For some services, this style of contracting can also reduce prices compared with short-term spot contracting if providers build in some premium to cover their investment risk. However, where real competition exists — as it does, for example, in long-term residential services for the elderly — the threat of the loss of a contract can provide an important incentive to maintain service quality. In the absence of such an incentive, purchasers may need to monitor providers more closely, and this inevitably increases transaction costs.

DISCUSSION While it is difficult to assess systematically the impact of these layers of structural changes on health sector performance, a cursory glance at total health care expenditures per capita reveals some interesting trends. Using OECD-reported expenditure data from 1980 to 2007, the annual rates of change of per capita spending before the 1990s averaged around 7.0 per cent, while between 1991 and 2000 the average rate fell to just under 5.0 per cent. Subsequently, the rate inched up again to slightly above 6 per

06 Improving Health

143

8/24/11, 4:24 PM

Toni Ashton and Mar ia C.G. Bautista

144

cent for the period 2001–07. Whether these differences are significant remains the subject of further analysis, especially one which controls for simultaneous developments occurring in the health sector and the broader economy. However, initial indications are that purchasing arrangements and/or the political ethos of the government appear to have some effect on expenditure trends. The developments in purchasing arrangements in New Zealand came with some institutional reconfigurations that have system-wide effects and thereby affect health care spending patterns. One enduring institutional by-product of the 1990s reform is the Pharmaceutical Management Agency (PHARMAC), the single purchasing agency for community medicines. Established jointly in 1993 by the four RHAs, PHARMAC is a crown agency that operates independently, with the primary remit of managing the pharmaceuticals subsidy. This involves negotiating prices with pharmaceutical companies as well as deciding which drugs to subsidize. Through the use of techniques and tools widely adapted internationally — the use of formulary, reference pricing, tenders, and price volume contracts, among others — PHARMAC has managed to increase volumes of purchases with costs contained to just under 3 per cent per year (Pharmaceutical Management Agency 2009). It has been so successful that there are legislative proposals to expand its remit to other medical services. There have been other interesting unplanned responses to each round of reforms. For example, in the 1990s the introduction of the quasi-market encouraged general practitioners (who had previously worked mostly independently as self-operated businesses) to form provider networks called independent practitioner associations (IPAs). The formation of IPAs was largely a response to the imbalance of power between the new regional purchasing agents (the RHAs) and individual providers. IPAs reduced competition between providers and thus strengthened their bargaining power. At the same time, IPAs very quickly established a role in improving the quality of primary care, for example, by assisting practices to improve the quality and use of patient data and by providing management support to individual practices. They also enhanced collaboration among individual providers and encouraged an organized approach to general practice that had previously been absent. Further modifications to the primary care sector were put in place during the first half of the 2000s. While previous reorganizations had

06 Improving Health

144

8/24/11, 4:24 PM

Purchasing Health Services in New Zealand

145

largely centred on the purchasing of hospital services, reforms in the post2001 period focused on the primary care sector, particularly the formation of primary health organizations (PHOs) as not-for-profit networks of primary health providers. PHOs were designed to be multidisciplinary, an agglomeration of GPs, plus other primary providers, with one of the key aims being to provide comprehensive, coordinated, and continuous care. Health promotion and other public health programmes were included in their remit. Higher, capitation-based subsidies for providers who belonged to a PHO provided an incentive for PHO formation, and some PHOs received extra funding to improve “access” for underserved population groups. In 2009, 95 per cent of the population belonged to one of around eighty-two PHOs (Ministry of Health 2009). Good progress has been made in improving access, both through reductions in user charges (Cumming & Gribben 2007) and through PHOs that cater primarily for Mäori and Pacific people. Recent reforms are directed at making PHOs more efficient. There are proposals for DHBs to be more proactive in working with PHOs to develop new models of care — including shifting some hospital-based services to the community — and to cut back on management subsidies to smaller PHOs to encourage mergers (Ministerial Review Group 2009). It remains to be seen whether considerations of efficiency, by pushing for mergers of smaller PHOs, will reverse some of the equity gains in primary care. One common feature of all the institutional models implemented has been the maintenance of strong guidance from the central government. Even under the market model that prevailed during the 1990s, the government or its agencies maintained control over broad funding decisions (Ashton et al. 2005). Similarly, while funding decisions have now been devolved to the DHBs in an effort to strengthen community participation, the DHBs must still adhere to a set of national priorities and operational requirements, and negotiate annual service coverage schedules with the minister of health. Thus, regardless of the prevailing institutional arrangements, strong central control has remained a feature. Another lesson that can be learned from the New Zealand experience is the importance of taking into consideration the norms and values of those working in the health system. During the 1990s, the alienation that many health professionals felt working in a competitive environment undermined the culture of collaboration that had previously existed in the public hospitals. Ideas and innovations were no longer shared, and some

06 Improving Health

145

8/24/11, 4:24 PM

Toni Ashton and Mar ia C.G. Bautista

146

well established teams were torn apart. Many senior personnel resigned and much institutional memory was lost. The market model also had implications for health service research because data that were previously publicly available were now often protected on the grounds that they were “commercially sensitive”. This made it difficult for researchers to undertake independent evaluations of the quasi-market reforms.

CONCLUDING REMARKS The introduction in New Zealand of a series of different institutional arrangements over the past two decades has resulted largely from shifts in the political ethos of governments. While right-of-centre governments have sought greater reliance on market mechanisms, left-of-centre governments have preferred to encourage collaboration and strengthen democratic processes. None of New Zealand’s alternative institutional arrangements has emerged as being clearly superior to the others: each has its own strengths and weaknesses. But importantly, new systems and practices stimulated by the reforms have often endured despite subsequent structural reform. Arguably, these changes, many of which have been unplanned, may have had more effect on overall system performance than structural reform (Ashton et al. 2005). For Cambodia, there appears to be a decade or more of experience with different schemes to improve access and service delivery (Annear 2006). Cambodia leads the developing world with health protection schemes (health equity funds) and service contracting. The extent to which these schemes embed new systems or practices that can endure without structural reform, particularly with respect to more sustainable funding, remains to be seen. Cambodia’s multitiered health system is organized around facilities, with health centres and hospitals as focal points. The significant number of players, from donors to international and local non-government organizations, presents some planning and coordination challenges. With health human resource capacities and numbers strained, the strengthening of local capacities to assess needs, identify priorities, plan and bring health decision making closer to communities, would be a default option. The risk of fragmentation of the system can be avoided by a strong centre or mediating institutions that can respond to the needs, and harness resources below with funding and guidance from the top. The New Zealand health system presents, for countries looking at models in health care purchasing and health reforms in general, a specific

06 Improving Health

146

8/24/11, 4:24 PM

Purchasing Health Services in New Zealand

147

context. The relative ease with which purchasing arrangements have been changed several times in just two decades attests to the resilience of the health system, but has also contributed to its complexity. The broader institutional changes reported here have been accompanied by many innovations at the grass roots. While some of these innovations would probably have occurred in the absence of any institutional change, the culture of health sector reform has itself been an important stimulus for innovation at the service level. Either way, change is likely to remain an enduring quality of the health system in New Zealand, its strong democratic ethos having created a high tolerance for reform.

References Annear, P.L. A Study of Financial Access to Health Services for the Poor in Cambodia, Phase 1 and 2. Canberra: AusAID, 2006. Ashton, T. “Quasi-Markets and Contracting for Health Services”. PhD thesis, University of Auckland, 1999. Ashton, T., J. Cumming & J. McLean. “Contracting for Health Services in a Public Health System: The New Zealand Experience”. Health Policy 69 (2004): 21–31. Ashton, T., N. Mays & N. Devlin. “Continuity through Change; the Rhetoric and Reality of Health Reform in New Zealand”. Social Science and Medicine 61 (2005): 253–62. Ashton, T., T. Tenbensel, J. Cumming, & P. Barnett. “Decentralizing Resource Allocation: Early Experiences with District Health Boards in New Zealand”. Journal of Health Services & Research Policy 13, no. 2 (2008): 109–115. Coalition Government. “Policy Area: Health”. Coalition Government Policy Document, 1996. Crown Company Monitoring and Advisory Unit. “Crown Health Enterprises: Briefing to the Incoming Minister”. Wellington: Crown Company Monitoring and Advisory Unit, 1996. Cumming, J. & B. Gribben. Evaluation of the Primary Health Care Strategy: Practice Data Analysis 2001–2005. Wellington: Health Services Research Centre, 2007. Cumming, J. & N. Mays. “Reform and Counter Reform: How Sustainable is New Zealand’s Latest Health System Restructuring?”. Journal of Health Services & Research Policy 7 (2002): Suppl 1, S46–55. Gauld, R. Revolving Doors: New Zealand’s Health Reforms. Wellington: Institute of Policy Studies, 2001. Malcolm, L., P. Barnett, & J. Nuthall. “Lost in the Market? A Survey of Senior Public Health Service Managers in New Zealand’s Reforming Health System”. Australian and New Zealand Journal of Public Health 20, no. 6 (1996): 567–73.

06 Improving Health

147

8/24/11, 4:24 PM

Toni Ashton and Mar ia C.G. Bautista

148

Minister of Health. “The Government’s Response to Ministerial Review Group’s Report ‘Meeting the Challenge’ ”. Wellington, 2009. Ministerial Review Group. “Meeting the Challenge: Enhancing Sustainability and the Patient and Consumer Experience within the Current Legislative Framework for Health and Disability Services in New Zealand”. Wellington, 2009. Ministry of Health. Health Expenditure Trends in New Zealand 1996–2006. Wellington: Ministry of Health, 2008. ———. “Primary Health Care: Frequently Asked Questions”. Wellington: Ministry of Health, 2009. New Zealand Labour Party. Focus on Patients: Labour on Health. Wellington: New Zealand Labour Party, 1999. Pharmaceutical Management Agency. “Annual Review 2009”. PHARMAC, 2009.

06 Improving Health

148

8/24/11, 4:24 PM

Primary Health Care in Cur itiba, Brazil

149

III Optimal Health Workers Contracts

07 Improving Health

149

8/24/11, 4:24 PM

Geoffrey Shepherd

150

07 Improving Health

150

8/24/11, 4:24 PM

Primary Health Care in Cur itiba, Brazil

151

7 A CIVIL SERVICE THAT PERFORMS Primary Health Care in Curitiba, Brazil1 Geoffrey Shepherd

INTRODUCTION “Curitiba is a city of 1.9 million people. It is the capital of Paraná, one of Brazil’s more developed states, located in the south of the country, and one of Brazil’s wealthiest cities. Curitiba is known for its urban innovations and the quality of its life. It has an innovative rapid-transport bus system. It has more green space than any other Brazilian city. Seventy per cent of the city’s garbage is recycled. Ninety-nine per cent of Curitibans say that they would not want to live anywhere else”. (McKibben 1995). Curitiba’s public health care system is less well known. but is equally impressive. Over four decades of steady development, the city has emerged as an exemplary practitioner of the Brazilian model of public primary health care. This chapter looks at how Curitiba’s management of its human resources has contributed to this achievement. Curitiba has a well qualified and highly motivated workforce consisting largely of tenured civil servants. But Curitiba is not a typical

07 Improving Health

151

8/24/11, 4:24 PM

Geoffrey Shepherd

152

Brazilian story. Problems of delivering health services in Brazil’s subnational governments are widespread, and a poorly performing workforce — in most cases with the same traditional civil service regime — is an important part of the explanation. The case of Curitiba suggests that the interests of workers can be made to align with public objectives, even within such a regime. This case study goes beyond the most formal or most visible aspects of the labour contract and places the management of human resources within the overall management system. It also looks at the external factors of history, national policies, and funding that influence the management system. But looking in-depth at one case and casting a broad net to understand it do not lead to unambiguous conclusions. The lessons from this case and their applicability to other health systems have an element of speculation.

PRIMARY HEALTH CARE IN CURITIBA The Unified Health System Brazil’s 1988 constitution mandated a decentralized, universal, and free health service financed from social welfare funds. This reform was largely part the product of the Brazilian health care reform movement (Movimento da Reforma Sanitária). This was a remarkable revolution, in the context of Brazil’s return to democracy, in which a professional group formulated a radically alternative policy and then gained key political and bureaucratic posts to implement it (Weyland 1996, Chapter 7). The new constitution led to a series of health policy reforms under the Unified Health System (SUS in its Portuguese acronym) that have transformed the organization, financing, and provision of health services. The federal government’s role began to shift from service provider to financer, promoter, and regulator. In the late 1980s Brazil converted its federal public health financing system to a single national fund. In the mid-1990s, it moved away from supply-driven financing of health expenditures (based on fee per service) towards needs-driven funding, instituting a per capita payment for primary care services distributed directly to municipalities. This simple reform caused a vast improvement in the equity of the system. Poor municipalities suddenly had funds for primary health services on a scale they had not seen before. More recently,

07 Improving Health

152

8/24/11, 4:24 PM

Primary Health Care in Cur itiba, Brazil

153

this capitation system was enhanced by various incentive-based programmes, notably the Family Health Programme (PSF), through which the federal government has transferred additional funds to municipalities that agree to implement a particular programme. The most salient shift in health care policy was a declared emphasis on primary and preventive care rather than on secondary and tertiary care (specialized clinics and hospitals) and curative medicine. The PSF has been a centrepiece of that primary health care model. Under the PSF, first introduced in 1994, municipal health care teams proactively provide an integrated package of health services to families in a defined community.2 These standardized teams work with a uniform set of equipment and procedures. Each team is responsible for providing care to a defined set of 600–1000 families. This is one of Brazil’s largest federally funded health programmes, reaching 47 per cent of the population by the end of 2007. The non-government sector provides only 5 per cent of primary health care services. In comparison to primary care, federal policy towards hospitals remains considerably more “passive”. Still, federal spending on curative services far outweighs spending on preventive and public health, even though the share is declining. Under the SUS, the provision of services is decentralized as far as possible to states and municipalities and jointly financed by all three levels of government. Curitiba has full responsibility for all levels of health service. In 2007 its services were almost 60 per cent federally financed, with almost all the balance coming from the municipality.

The Evolution of Curitiba’s Primary Health System The SUS had a substantial impact on Curitiban practice from the 1990s, but Curitiba started its health reforms earlier, partly under the same influence of the health care reform movement. The city created a Health Department in 1979 and initiated a health care model based on four principles: democratization, extension of coverage and prioritization by level of care, integrated medicine, and community participation (Ducci et al. 2001, p. 16). During the 1980s the city opened a number of largely standardized health units for primary care. A body of health workers developed in parallel, and many of these remain today. The transfer of responsibility for health to Curitiba began in 1992 and was completed in 1998. Primary health care expanded through the opening

07 Improving Health

153

8/24/11, 4:24 PM

Geoffrey Shepherd

154

of PSF Health Units. In the early 1990s, as the health system expanded, Curitiba began to recognize its growing management problems, actual and prospective. The system’s growth was challenging the way that people, money and information were managed. This recognition — the fruit of a strategic planning mindset — ushered in a new, more “managerialist” era. In 1992, the system was decentralized. Several initiatives affected human resources more directly: the first of several staff bonus schemes was introduced in 1994; there was a large management training initiative in 1995; medical training was put on a more systematic basis from 1995; new staff (community health agents) were employed under private labour law from 1999; and management contracting was introduced in 2002.

The Primary Health Care System Today Current primary health care reflects a philosophy built up over three decades, sowing its character to the health care reform movement and the SUS, but also to more local developments. Based on descriptions of the system given by practitioners and from observations of how the system works, the following appear to be central elements of the primary health care philosophy: • Care is based on a humanized and socially oriented medicine. This approach seeks to prevent and cure through greater attention to the particular needs and situation of an individual patient and a greater understanding of his/her environment. This can change the typical role of the doctor from a leader of a team to a specialist within a team. • Primary health care is organized to focus on local problems. Specific areas covered by different units (and micro-areas within them) face specific health problems that can be attacked systematically. For instance, an area populated by industrial workers and one populated by poor rural immigrants face different problems. This local focus is realized through a decentralized primary care network of nine health districts, each administering around ten health units for primary care. • Primary health care emphasizes preventive practice. Curitiba does this mostly through specific programmes, for instance, for safe pregnancies or dental health (Ducci et al. 2001).

07 Improving Health

154

8/24/11, 4:24 PM

Primary Health Care in Cur itiba, Brazil

155

• More recently, Curitiba has emphasized evidence-based medicine, medical practice based on scientific evidence, distilled where possible into standard operational processes. A dominant part of the city’s primary health care (general medicine) and a substantial amount of secondary health care (specialized medicine) are in public hands. There are more than 100 health units, in which 70 per cent of the city’s population is registered (Marty 2010). These units provide most of the city’s primary health care and a large amount of the secondary. (Thirty per cent of the population is privately insured, but makes partial use of the public system — for vaccinations, for instance.) There are about 150 non-profit or private primary and secondary health facilities. Most tertiary care (hospitals) is in private or non-profit hands. The city pays these hospitals on a fee-per-service basis. PSF teams cover an average population size of 3,450 (in 2005). The federal government, which provides financial support to municipal PSFs, mandates a standard structure. Each PSF Health Unit has a manager and three to four teams operating in different geographical areas. Each team has ten public employees (one family doctor, one nurse, four auxiliary nurses, a dentist, two dental auxiliaries, a dental hygiene technician), and four community health agents (Secretaria Municipal da Saúde de Curitiba 2010). These agents are employed under private labour law.3 Contracted on the basis of a test and receiving one to two years’ training, they make community visits and help ensure that all families in the catchment area are identified. PSF doctors (like dentists) have a 40-hour week and have often made a career choice to enter family medicine. Some health units have been converted to PSF units, but the remaining units (called Basic Health Units) still account for about half of Curitiba’s health units. (Shortages in the supply of family doctors have prevented a faster expansion of PSF Health Units.) Basic Units are comparable in size to PSF Units, but are not constituted in teams. With a clinical staff of more than thirty, plus some community health agents, they serve larger populations and are less able to go out into the community. They have the equivalent of about three full-time doctors (general practitioner, paediatrician and gynaecologist working a 20-hour week). About ten Basic Health Units have a specialization (for instance in HIV/AIDS) and serve clients beyond their area.

07 Improving Health

155

8/24/11, 4:24 PM

Geoffrey Shepherd

156

Health Outcomes The available evidence does not allow us to relate results rigorously to administrative practices, but there is plenty of evidence indicating good health outcomes. For instance, the infant mortality rate fell from 42.7 per 10,000 of population in 1980, to 8.9 in 2009; adolescent pregnancies from 19.3 per cent in 1999 to 14.6 per cent in 2009, leprosy cases from 0.70 per 10,000 of population in 2000 to 0.14 in 2009 (Secretaria Municipal da Saúde de Curitiba 2010). Curitiba performs well for selected health indicators compared with other municipalities with a similar epidemiological and demographic profile. In 2004, Curitiba ranked either first or near the top among comparable cities in a selection of indicators reported on by the World Bank (2006). For instance, Curitiba had one of the lowest infant mortality rates in Brazil, with 11.2 deaths per 1,000 births. This can be ascribed to intensive prenatal care, since 80 per cent of births are from mothers who attended seven or more prenatal consultancies. It also performed at the top for non-foetal deaths by non-defined causes, with 1.1 per cent. The city had a high average of 1.6 annual medical visits per person in basic care.4

Management of Human Resources — the Formal Labour Contract The Municipal Health Secretariat (SMS) has about 8,000 employees (one quarter of the municipal total) and accounts for one fifth of the municipal budget: • There are almost 7,000 civil servants, four times the number than in the early 1990s. More than 40 per cent of them are professionals, with a similar number at the intermediate level, and the rest at basic level. • About 200 of the career professionals are managers who occupy, for the duration of their appointment, a politically nominated post. • Almost 1,200 people are employed under private labour law as community health agents.

Careers of Health Care Civil Servants The rules for civil servants at all levels of government are set out in a federal law passed in 1990, following the requirements of the 1988

07 Improving Health

156

8/24/11, 4:24 PM

Primary Health Care in Cur itiba, Brazil

157

constitution. Entry is based on academic qualifications and a written public examination. The top candidates assume an office defined by a narrow occupational group: doctors, dentists, and nurses are at the superior level, and auxiliary nurses at the intermediate level. After a three-year probationary period, civil servants receive legal protection against dismissal.5 New entrants start at the bottom and regularly and automatically advance within and between grades. Lateral movement outside the given office is restricted. An auxiliary nurse hoping to move up to the higher level for nurse, for instance, would typically go to night school and then take the public examination. Health jobs are professions in themselves: people typically become doctors, dentists, and nurses for life. In the public sector, the most obvious career development path typically available is to gain, temporarily or permanently, a management post (a political appointment). Within the SMS, there is one career move that has aspects of career development: the move from a Basic Health Unit to a PSF Unit results in more money and a different (more team- and community-oriented) job.6 There is a strong demand — and a weaker supply — for doctors in Basic Units to pass to PSF Units. For most health professionals, satisfaction seems to come from the job itself, not advancement prospects. In the SMS, part of this satisfaction also comes from the frequent opportunities for training and professional development. Most health professionals have the choice to work in the public or private sectors. For doctors and nurses in particular, public and private sector jobs may be quite different because of the public sector’s emphasis on primary health care and preventive and family medicine. In practice, nurses, auxiliary nurses, and dentists are in adequate supply, and the public sector pays them competitively with the private sector. Some of these professionals typically have made a career choice based on their preference for what the public sector does in health care. As a result, there has been a reasonable stability among public sector health employees, which has probably helped foster teamwork and increase experience in health units. Doctors are different. First, the profession exercises control of entry into medical schools. This has raised the price of doctors and lowered their numbers on the labour market. Second — and despite the health care reform movement — the preference of most doctors (reflected in the availability of university training courses) is to specialize rather than practise general or family medicine. However, the PSF has since 1996

07 Improving Health

157

8/24/11, 4:24 PM

Geoffrey Shepherd

158

developed a new labour market for generalist doctors: doctors who take up posts in PSF Health Units have a greater tendency to stay in the public sector than doctors in Basic Units. And in recent years, federal funds have been used to press universities to begin to adopt medical and nursing curricula. The scarcity of doctors has implications for the turnover of doctors and the way they are employed. Thirty per cent of doctors appointed to the SMS leave in the first year, usually for the private sector.7 Curitiba’s doctors (like doctors elsewhere in Brazil and in many poorer countries) hold multiple jobs — an average of two to three jobs (and up to five), each formally requiring twenty hours of work a week. Often, this multiple-job strategy seeks to combine the benefits of higher salaries in the private sector, with the higher prestige, greater learning opportunities, and benefits (particularly pensions) of the public sector. Doctors seem to be the weakest link in health unit teams because of their high turnover and multiple jobs.

Managers All managers in the SMS (around 200 posts) are political appointees serving at the pleasure of elected officials. There are four levels: the municipal secretary of health, then directors, then district health supervisors and coordinators and, finally, health unit managers. People occupying these positions get small increments to their salaries as civil servants. They tend to come more from the ranks of nurses and dentists than of doctors. By tradition, managers, including the secretary, are also civil servants from the SMS. This tradition, reflecting the limited influence of party politics on health care in Curitiba, also means that managers, though politically appointed, share the same culture as their non-manager colleagues. When people no longer have a political appointment, they are free to resume normal office within the secretariat. This makes the structure of management very flexible.

Bonus Schemes The SMS has had several systems for bonuses on top of basic salaries. In several cases, the purpose has simply been to raise the level of remuneration where the labour market was supplying too few qualified people. All civil servants in PSF Health Units are eligible for a bonus funded by the federal

07 Improving Health

158

8/24/11, 4:24 PM

Primary Health Care in Cur itiba, Brazil

159

government: doctors and dentists receive 80 per cent of their basic salary, nurses 60 per cent, and auxiliary nurses 50 per cent. In 2002, the SMS introduced a special bonus for doctors equal to 20 to 40 per cent of the basic salary. From 2002, the turnover of doctors fell, and it is thought that this incentive was largely responsible for this. The SMS has also experimented with two schemes to encourage better performance. The first was launched in 1995, when the SMS felt the need to create an incentive to support the drive for greater quality in health services. The Quality Incentive Plan (PIQ) established a bonus for all SMS units, with the exception of PSF Units. Based on performance according to a small set of indicators (health outcomes, and health and management outputs), the staff in about 10 per cent of the units were awarded a 30 per cent bonus. This scheme turned out to be problematic. The focus of teams became one of simple compliance with a checklist (sometimes by “gaming” the system). Moreover, senior managers in the SMS perceived that the team-based incentive provoked rivalry among teams and undermined the unity of the system as a whole. The PIQ was terminated in 1997. The current scheme, the Incentive Programme for Quality Development (IDQ), was created in 2000, when there was a rapid expansion of PSF Health Units. Each quarter, SMS workers are evaluated to see whether they qualify for a bonus of 20 per cent of their monthly base salary (in a few units, where working conditions are more difficult, the bonus can be as high as 50 per cent). All SMS employees, except some political appointees, are eligible. The IDQ evaluation is based on four elements: the supervisor’s evaluation of the employee, self-evaluation, SMS evaluation of the unit, and community evaluation of the unit.8 Typically, more than 95 per cent of employees win the bonus each quarter. Most of the others are disqualified because of “excluding factors” — leave, absence, lateness, etc. (Figure 7.1). Leave is the most important cause. That the rate of disqualifications under these factors fell markedly after 2005 suggests that staff have become more attentive about being absent from the job. Those disqualified from the bonus for poor performance account for a very low share. This was at first well below 1 per cent, then rose to almost 3 per cent by the end of 2005, but fell again in 2006; in 2007–09 it averaged less than one-half per cent.9 SMS reported that low performers are identified and are usually the same people. Their managers try to assign these employees to other jobs, to which they might be better suited.

07 Improving Health

159

8/24/11, 4:24 PM

07 Improving Health

160

2004 2005 2006 2007 2008

Source: Calculations based on data from Municipal Health Secretar iat (SMS)

0.0%

2.0%

4.0%

6.0%

8.0%

10.0%

2009

Other reasons

Performance

FIGURE 7.1 Incentive Programme for Quality Development: Share of Participants not Receiving Bonus, 2004–09

160 Geoffrey Shepherd

8/24/11, 4:24 PM

Primary Health Care in Cur itiba, Brazil

161

Where so few cases of disqualification for the bonus occur, it is reasonable to suppose that the scheme is catching only very extreme cases of poor performance — moreover, ones that are more easily measured, such as lateness — and that the IDQ is not stimulating better performance in general.10 But the results of a World Bank survey seem, at first sight at least, to contradict this.11 Ninety per cent of respondents thought that IDQ had an influence on behaviour when it was introduced. Consistent with the pattern of qualification and disqualification in Figure 7.1, this influence was considered greatest in the area of workplace behaviour. But it was also perceived to have improved motivation and, to a lesser extent, work organization and productivity. The survey also suggested that respondents felt the impact had fallen over time: only 61 per cent believed that IDQ continued, in 2005, to provide an important influence on behaviour. Sometimes, management innovations have a short-lived effect (the so-called Hawthorne effect): they enhance performance when they are still a novelty and provide a spotlight on workers, but as they lose their novelty, workers adjust their effort back to the norm. The IDQ may be one such innovation. But according to the World Bank survey, the IDQ may also have had a more valuable side effect that, it is hoped, will last longer. It seems that the IDQ may act as a motivator through the evaluation itself. This process was perceived to be fair and coherent and to contribute most by promoting employee-supervisor dialogue (rather than by direct reward/punishment, except punishment because of “excluding factors”). For instance, it was perceived to help solve team conflicts. Collectively passing the IDQ test was a major preoccupation of teams (72 per cent said this is “always” so and 25 per cent “frequently” so). When a problem was perceived with a staff member, the team acted to solve it. Thus, it may be that the enthusiasm (if waning) that staff showed for the bonus as a management tool related to evaluation as a tool for identifying problems, rather than the bonus as a performanceenhancing reward.

Strategic Management — Informal Aspects of the Labour Contract The Municipal Health Secretariat has some pronounced, often innovative, management features. Several of these contribute to informal aspects of the labour contract that underpin good performance.

07 Improving Health

161

8/24/11, 4:24 PM

Geoffrey Shepherd

162

Client Orientation There are several dimensions to client-oriented management. In 1991 the SMS was decentralized into nine health districts, the facilities of which were designed with accessibility of public transport in mind. The Family Health Programme, later reinforced by community health agents, created a more permanent link between the health provider and the patient and her/his family. Tripartite health councils represent the SMS, employees and clients at the municipal, district, and local levels. They meet monthly, as well as every two years in conference. Local populations consider them an effective channel for complaints and suggestions. In the late 1990s, the SMS began to reorganize the work (in particular nursing) in order to make access to services easier, create a more clientoriented culture, and dispose of cases more effectively. Changes in behaviour and processes have reduced queues, improved the rate at which problems are resolved, and reduced complaints. Curitiba’s health care clients have also been individually empowered by a telephone complaints and information system for city services (established in 1993) and, more recently, by twice-yearly telephone community evaluations, specifically on health units. Surveys about individual units provide community evaluations that form part of the system of personnel evaluation (which in turn has consequences for the IDQ bonus). In this way, for instance, managers get to know whether enough staff are available to serve clients (Ministério da Saúde 2007, p. 51). The World Bank’s survey (2006) indicated that health unit managers made use of these evaluations to improve client services. Empowering the community and individual clients allows the municipality to benefit from local knowledge and preferences on health issues and services. It also allows the municipality, the monopoly supplier of a free service, to gauge whether it is providing the right service. Thus information asymmetry is reduced.

Managing Information Both clinically related knowledge (social and medical) and management routines need to be managed. The health units depend on better social knowledge. Hence, the expansion of primary health care services has been

07 Improving Health

162

8/24/11, 4:24 PM

Primary Health Care in Cur itiba, Brazil

163

planned on the basis of area-by-area socio-economic analyses about clients and the environment they live in. The collection and application of local and worldwide medical and clinical knowledge are constant endeavours for any health authority. The SMS carries out systematic epidemiological activities and programmes linked to universities and training institutes, such as an intensive training relationship with the University of Toronto. One important result of Curitiba’s emphasis on evidence-based medicine has been the development of integrated protocols (Marty 2010). These are manuals for clinical and preventive procedures in various areas. There are almost twenty, covering such areas as leprosy, adolescent health, HIV, obstetric emergencies, and mental health.12 The protocols also help integrate primary health care with specialized medicine and hospital services. In part, these were originally developed (from 1999 onwards) to prioritize cases as clients entered the unit in a way that would minimize queuing. Protocols are part of the SMS management information system. A doctor who attends to a pregnant woman, for instance, knows all the exams the patient has already undergone and what others she still needs. In addition to promoting good health practices, the protocols appear to have important economic effects. It is not clear how far all these effects were intended or unintended: • According to the SMS, the negotiation of the protocols (between technical staff, directors and scientific societies) provided consensus among teams.13 • By standardizing clinical practice, protocols (in theory at least) permit greater economies of scale in treatment and procurement. Treatments become standardized.14 Protocols mean that a predetermined set of medicines is purchased for the entire network. The SMS also believes that protocols save public funds that could be lost in lawsuits against the state. • By standardizing clinical practice, the protocols also make it easier for managers to measure the performance of clinicians.15 The protocol’s indicators are monitored and then fed into the reports about the accomplishment of management targets. The SMS has a very advanced information system (under construction since 1988), built on computerized databases that communicate with one another, with other municipal systems (citizen identification, human

07 Improving Health

163

8/24/11, 4:24 PM

Geoffrey Shepherd

164

resources, etc.) and, in some cases, with private providers. At the heart of the system is the computerized patient record, with medical, dental, and nursing modules, established in 1999 (Xavier & Shimazaki [n.d].). Each person registered in the system has a magnetic card that allows them access to their records and to health services. The patient record is an instrument in clinical management because it is compatible with the integrated protocols. According to those working in the sector, the computerized patient record has had clear health and managerial outcomes (Xavier & Shimazaki [n.d]; Marty 2010): it has systematized nursing practice and led to changes in the way dentistry is done (for instance, requiring the development of dental health protocol training). By facilitating the monitoring of individuals, families, and populations at risk, the system has enabled coordination of health actions; epidemiological information is more timely; laboratory results are returned more quickly; the system has integrated different services, so that clients find that their appointments are set more efficiently (and they have more choice); hospital bed allocations are better managed. Other IT systems allocate hospital beds for specialized medicine and hospital resources financed by the federal government, set appointments for consultations, examinations, and therapies, provide for obligatory notification of disease information (which is then transmitted to federal Ministry of Health systems), administer the municipal laboratory, and administer the pharmacy. Finally, a management information system monitors administrative performance. By permitting greater standardization of medical procedures, the integrated protocols stand out as a tool for reducing information asymmetry between clinicians on the one hand and managers and clients on the other: the more standardized the procedures that clinicians perform, the more managers can monitor and measure them. The computerized patient record plays an important part in making standardization and its monitoring a reality.

Specialization and Coordination The SMS has concentrated its resources (and its most important innovations) in the area of primary and preventive medicine, rather than hospitals. It has further sought to reap the benefits of specialization by emphasizing

07 Improving Health

164

8/24/11, 4:24 PM

Primary Health Care in Cur itiba, Brazil

165

problem-specific programmes (diabetes, for example) and by standardizing medical and management processes (notably PSF Health Units and the protocols). It has also invested heavily in systems to coordinate primary, secondary and tertiary health services, and the public and private units providing them, and to make the primary health care network the gatekeeper for these services.

Managing Performance The Curitiba experience suggests three related aspects of management practice designed to affect performance: strategic planning, a process of continuous learning that anticipates changes the organization is likely to face and defines goals to accommodate them; management contracting, a process that seeks to align management objectives with employee objectives; and promoting a performance culture, another approach to aligning management and employee objectives. Strategic planning and systems thinking have been a central habit of mind among health managers since a Health Department was created in 1979. Among the foremost formal strategic planning exercises was a management course for leaders of health units, initiated in 1995 with the aim of setting a track for reorganizing basic health care. After a year and a half of reflection, the course generated a number of proposals that substantially shaped the managerial innovations of the 1990s, including the development of the information system and performance targeting. Many aspects of Curitiba’s health care management system are informed by systems thinking: for instance, information systems have consciously changed management methods (and vice versa); integrated protocols have similarly contributed to changes in management methods. Strategic planning has also become a more formal, routinized activity that has paved the way for management contracting. Health planning in Curitiba is mandated by federal legislation. Curitiba’s Municipal Health Council agrees on four-year municipal health plans that set out, in very general, unquantified terms, how health services are to be delivered and managed, and how major medical problems are to be tackled. This plan is then translated into more specific actions in the annual operating plan (POA). A number of quantitative indicators have been developed since 2000, based on the standardized parameters of the protocols, to help make the strategic POA targets more operational. (There are other inputs to the

07 Improving Health

165

8/24/11, 4:24 PM

Geoffrey Shepherd

166

POAs, including agreements with the federal government and local health planning.) Based on these POAs, annual management contracts have been signed since 2003. (Ministério da Saúde, Política Nacional de Humanização 2006; Ministério da Saúde 2007; Marty 2010.) The management contracts appear to provide an effective means of reinforcing the performance orientation of staff in primary health care. This is quite a rarity for Brazil or other countries. The management contracts are simple and short: they summarize the general mutual obligations of principal and agent. The POA, annexed to the contract, sets specific performance targets. The SMS contracts with the nine health districts and the health districts contract with their health units. No rewards or punishments are directly associated with performance under the contracts.16 However, 35 per cent of the points awarded individuals under the Incentive Program for Quality Development depend on the unit’s performance under the management contract. In 2009, the POAs covered eighty-one indicators, thirteen of which provided the basis for IDQ evaluations (Marty 2010). Targets are agreed in discussions between the SMS and the health districts, and between the latter and the health units. Some of the indicators are agreed on with the federal Ministry of Health, regarding targets of coverage, visits, and procedures. Target setting takes into account differences between areas (for instance, the more middle-class the neighbourhood, the more inhabitants are likely to use private health services), and differences of installed capacity. According to World Bank interviews and survey results, targets are fixed between health district and health unit with a mix of negotiation and imposition. But the health district is open to renegotiation when there are difficulties in fulfilment. This would suggest that there is a reasonable balance between the “topdown” direction and “bottom-up” participation that many people consider necessary for good management. The effectiveness of the contracting system depends on the quality of the indicators, so participants have less opportunity to “game” the system. Two information systems provide the key to good information: the computerized patient record, allowing faster and more accurate patient information; and the protocols, allowing more standardized definitions of medical services. Every month the management information system transmits performance data from the health units to their districts. Every quarter the data are sent from the districts to the SMS, which carries out

07 Improving Health

166

8/24/11, 4:24 PM

Primary Health Care in Cur itiba, Brazil

167

a comparative analysis and reports to the city council and the Municipal Health Council on the progress of the POA. For each indicator, objectives are defined in terms of the percentage coverage of the target population. In practice, it appears that some indicators, such as those relating to safe pregnancies and infant deaths, are more closely watched than others. For all targets, there is a wide range of overfulfilment and underfulfilment. This may be because it is difficult to set accurate indicators, especially as many non-Curitibans use the city’s medical facilities, but it also clearly suggests that monitoring performance is not a pro forma activity. There has been a vigorous learning process with the indicators, which have been dropped, added, and fine-tuned. Their number has grown over time. In turn, the measuring of indicators has fed back into clinical practice. For instance, measurement indicated a need to improve the rate at which pregnant women were examined (Marty 2010). An overwhelming number of respondents to the World Bank survey (both managers and health workers) found management contracting useful (and not excessively costly to them): • The negotiating of contractual targets helped fix staff expectations: 85 per cent of survey respondents found the contracts substantially or entirely useful in establishing priorities. • The monitoring and discussing of performance under the contracts feed back into the organization of work (that is, as an instrument of local strategic planning). The districts regularly consult with the units on their performance. The SMS, in turn, meets with the districts. These meetings examine indicators, identify poorly performing health units and identify problems. Following these meetings, the SMS and the district provide help to the health units through training and exchange of ideas. • Management contracting is part of a broader set of management instruments that encourage teamwork, both in everyday operations and in problem solving. There was a strong perception — 92 per cent of survey respondents — that teams worked well. The survey emphasized that the IDQ is a vehicle for staff evaluation and the identification and solution of problems related to team performance. According to the survey, ensuring that staff gain the IDQ bonus was a major preoccupation of teams: 50 per cent of an individual’s

07 Improving Health

167

8/24/11, 4:24 PM

Geoffrey Shepherd

168

evaluation depends on team performance. Teams generally meet every week. When they perceive a problem, they most often talk to an underperforming colleague or adjust work processes. • Health unit managers are reasonably aware of how other units are performing. But the survey suggested that staff generally hesitated to openly judge the performance of other units openly (and, if they judged them, hesitated to find them better or worse). It can be inferred from this that there is no strong sense of rivalry between health units and that trying to foster interunit competition as a motivation for better performance may be misplaced, at least within the culture of Curitiba’s public administration. These observations are echoed in comments that SMS managers and staff have themselves made. We learned from interviews that the trust established between district and unit was high. District supervisors have all worked in units and may go back to them. When a problem arises in a unit, trust paves the way to cooperation, not defensiveness. Various approaches to solving work problems have been developed: visits to units, group discussions among nurses, and so on. When one unit told the district that it could not achieve a particular target related to the number of cases of hypertension registered, the district suggested novel ways to do detective work to find the clients or, if necessary, to explain the shortfall. Performance management is associated with a thriving performance culture among health workers in Curitiba’s public sector. It appears that staff actively support the values and objectives of the SMS. Extrinsic motivation — “I do it because I’m materially rewarded” — is always important, but does not appear to dominate in Curitiba. According to the survey, respondents reported that they were motivated by extrinsic factors (job stability and nearness to home), as well as intrinsic factors (job content, learning opportunities, training opportunities). They reported that salary levels and flexible hours were less important. The general enthusiasm of professional SMS staff for their work came through in interviews. It is also expressed in the astounding amount of effort staff voluntarily put into preparing professional papers for the biennial meetings on public health that SMS organizes for primary health care staff.17 It is by no means clear that SMS management practices alone created this thriving performance culture. However, it is clear that a performance culture is in the air of Curitiba.

07 Improving Health

168

8/24/11, 4:24 PM

Primary Health Care in Cur itiba, Brazil

169

Where Did Curitiba’s Performance Ethos Come From? Curitiba’s performance in primary health care is in many ways extraordinary. Where did this come from? It can be argued (without resounding proof) that three different effects have acted in combination: national health care policy, the network of health care professionals, and the traditions of government specific to Curitiba.

National Health Care Policy There are two aspects to national health care policies: the substance of these policies and arrangements for health financing. First, the influence of the SUS on Curitiba’s primary health care is clear. Beginning in the early 1990s, the government, through the SUS, provided a workable set of policies. The policies initially stemmed from the ideas of the health care reform movement and were articulated and refined by the federal government under the SUS. The SUS health philosophy emphasizes humanized and socially oriented medicine, a focus on local problems, and preventive medicine. Its management philosophy emphasizes, among other things, highly structured teams using standard procedures (notably the PSF), decentralization within cities, formal health planning, institutions of social control, and telephone hotlines. Second, the level and stability of financial resources have been guaranteed under the SUS. The level of public spending on health in Brazil is now mandated by law. A constitutional amendment in 2000 required that, by 2004, levels of health funding would be guaranteed and stabilized at all three levels of government.18 In Curitiba public spending on health has been rosy. Spending on primary health care per head of population at constant prices rose steadily from 2001 to 2005 (Figure 7.2, left-hand axis). Payroll expenses per employee at constant prices rose steadily — particularly after 2004 — averaging seven per cent a year (Figure 7.2, right-hand axis). A guaranteed and stable flow of financing greatly facilitates strategic planning. But at the same time it may also undermine incentives for greater efficiency.

Networks of Health Professionals Curitiba’s health care sector has been affected by two overlapping groups that associate to pursue a collective objective. The first network is composed of doctors and other health care workers. These are among the most

07 Improving Health

169

8/24/11, 4:24 PM

07 Improving Health

170

2002 2003 2004 2005 2006

2007

2008

2009

8/24/11, 4:24 PM

Source: Annex Tables 7.1 and 7.2.

Health payroll expenditures per employee, R$000 (right axis)

Primary health expenditures per caput of population, R$ mn. (left axis)

0

R$ 0 2001

5

10

15

20

25

R$ 20

R$ 40

R$ 60

R$ 80

R$ 100

R$ 120

30

35

R$ 160

R$ 140

40

R$ 180

FIGURE 7.2 Curitiba — Primary Health and Health Payroll Expenditures, 2001–09, in Constant Prices of 2009

170 Geoffrey Shepherd

Primary Health Care in Cur itiba, Brazil

171

powerful professions in Brazil. Their power has given them a strong presence in government; this is a major reason health spending is constitutionally guaranteed. The objective of health care professional groups is to serve their members’ interests, in particular, by maintaining a selfregulating monopoly. The standardization that has been occurring through the PSF (and even more markedly in Curitiba) threatens to weaken this monopoly by increasing the information available to politicians and the public. It is, therefore, perhaps no coincidence that doctors, the most highly organized of the health care professionals, have been less enthusiastic than other professionals about primary health care reform. The second network, the sanitaristas (health workers), is not a formal group, but brings together, more or less, the primary health care workers of the public sector (and cuts across the health care professions). Apart from a common interest in securing more resources for the sector, the objectives and impact of this network have not been the same as those of the professionals, particularly doctors. Born in the 1960s and 1970s, before Brazil’s return to democracy, the health care reform movement adopted a very political view that linked health to social conditions and democracy, and took to political action to pursue its beliefs. This movement — in the absence of an articulated popular demand for reform — dictated the health clauses of the 1988 constitution, which led in turn to the SUS. With this objective largely achieved, the network of sanitaristas that grew out of the health care reform movement also became a “community of practice”, a network of people who collaborate to share ideas about, and find solutions to, common problems. The sanitaristas circulate within a large, mobile, public health labour market at all levels of government. There are substantial interchanges of people and information through this market, through academic faculties, and through formal organizations, such as the Association of Family Health, the National Council of Municipal Health Secretaries, and the Brazilian Association of Graduates in Collective Health. The network of sanitaristas, as a community of practice, brings a common experience and understanding of problems and common tools. Most of today’s senior managers were products of this network and lived through Curitiba’s health care reforms. It is reasonable to infer that this movement has been helpful in imparting specialized knowledge to Curitiba’s health care workers, supporting a strong working ethos that motivates performance and the sharing of knowledge, and facilitates teamwork.

07 Improving Health

171

8/24/11, 4:24 PM

Geoffrey Shepherd

172

City Government Curitiba has an outstanding city management that has acquired a worldwide reputation for its innovative solutions to problems of urban services. Curitiba’s period of intensive urban innovation began in the 1960s with the creation of a city planning agency, the Research and Urban Planning Institute of Curitiba (IPPUC) (Campbell 2006). IPPUC produces a master plan for the city, which it monitors and updates constantly. But Curitiba’s strategic planning can also be characterized as a fairly “soft” approach to management, emphasizing a culture or mindset — about imagining a future and seeing how to organize and coordinate resources to get there — rather than the “hard” rules of systems for human resource management or financial management (New Zealand Parliamentary Commissioner for the Environment 2002; Gnatek 2003). In addition, the city has been good at integrating policies across sectors, adopting pragmatic, flexible, and experimental approaches to new problems, using research and information to inform decisions, and listening to its citizens (who are keen to report problems). Curitiba has also invested significantly in human resources, placing strong emphasis on meritocratic entry of new civil servants, educational qualifications, leadership, and training. Curitiba’s achievements have been secured in a singular political environment. There has been remarkable political continuity: technocrats (particularly urban planners, but also health care professionals) have played a powerful role as political entrepreneurs, and they have been able to provide consistent, coherent leadership. After 1971, Curitiba’s two longest serving mayors, both urban planners from IPPUC, spent more than twenty years between them in that office. With the experience of several decades of improving city services, citizens have a stake in the system and have voted for political continuity.19 The management style of the Municipal Health Secretariat clearly bears much in common with other parts of the city’s administration. The same strategic planning approach is central. Curitiba’s managers appear to be adept at managing complexity. Policies are applied flexibly and progressively. The interconnectedness of the different systems that combine to produce effective health services is well understood. Good information is emphasized. Client feedback is important. Human resources are nurtured. And like city leadership, the leadership of the SMS has been technocratic and has benefited from continuity. It is plausible to argue that public health care has benefited greatly from Curitiba’s management traditions.20

07 Improving Health

172

8/24/11, 4:24 PM

Primary Health Care in Cur itiba, Brazil

173

Contract and Performance: An Interpretation Curitiba has performed above the Brazilian norm. Its very good primary health outcomes strongly indicate that the city has a health care workforce substantially oriented to high performance. This performance owes something to national factors, in particular, a robust national policy framework for primary health care, a good national health information system, substantial budgetary support from the federal government, and a very dynamic network of health care workers dedicated to making national policies work. Despite these national advantages, Brazil has many municipalities that perform poorly. A World Bank study of governance in Brazil’s health system (World Bank 2007, Chapter 4) had some alarming findings about the state of human resources.21 This suggests that additional explanations of Curitiba’s performance have to be sought. The formal labour contract does not appear to be conducive to staff performance in Curitiba. It would be the consensus of many Brazilians that, whether in health or in other sectors, Brazil’s civil service rules do not enhance performance.22 The dismissal of Brazilian civil servants on performance, or any other, grounds is rare. Remuneration dependent on performance is rare. Career promotion is largely automatic. On the other hand, Brazil’s federal government and some subnational governments have a better civil service — more effective and less corrupt — than is to be found in most other parts of Latin America. Even so, there were — beyond the emphasis on competitive, merit-based procedures for recruitment of public servants — no particular indicators from the Curitiba case study that the rules of the civil service statute contributed to staff performance. However, shortcomings in the formal labour contract have been offset by managerial innovations that make information more symmetric on the one hand, and emphasize intrinsic incentives on the other. These measures align the interests of health workers with those of management especially, but also those of clients. Information asymmetries have been offset through two principal channels: standardization and information exchange. First, the combination of standardized clinical procedures, particularly through medical protocols, and sophisticated information systems, means that managers have much more meaningful information on what clinicians are doing. (Clinicians, of course, can also perform better if they have better information.) Second, negotiation and reporting under the results contracting system, and staff

07 Improving Health

173

8/24/11, 4:24 PM

Geoffrey Shepherd

174

evaluations under the Incentive Programme for Quality Development, bring problems and constraints to the surface and give managers and staff a common platform of information, thus helping them to align their knowledge and expectations. The IDQ and the evaluation process have also proved important in improving information and aligning expectations within teams, thereby promoting teamwork. Intrinsic incentives are harder to pin down. The current performance bonus looks superficially like an extrinsic incentive to individuals and also to teams within the health units. But since very few people (1–2 per cent) fail to qualify for the bonus, the IDQ does not seem to provide a strong extrinsic incentive to individual effort (beyond avoiding lateness and absenteeism). By the same token, the IDQ should provide little incentive for teamwork. Yet our results indicated that passing the IDQ test was a major preoccupation of teams which devote substantial attention to solving problems that arise with individual team members. There is something of a paradox here. A plausible resolution may be that the incentive is largely intrinsic: the satisfaction of passing the “test” of the management contract and community evaluation. The IDQ bonus does not involve rivalry with other units because it does not function as a zero-sum game. Rewards are modest. And the scheme is based on trustworthy information. Each of these features contributes to a greater perception of fairness on the part of employees. The IDQ experience also suggests that what has most clearly driven team performance in Curitiba are the intrinsic incentives and attitudes of trust and cooperation that follow from the health sector culture and the problem solving culture (strategic planning) in Curitiba. Managers did not, perhaps, invent the strong work culture that characterizes Curitiba’s health workforce, but they actively support it (through training policies, for instance). In short, mechanistic applications of extrinsic incentives have not been the main drivers of performance improvements. Rather, this case highlights the role of organizations as organic environments within which managers and staff must find common goals (through carefully devised organizational structures and management processes) and seek ways to cooperate with one another through a mixture of diverse managerial and information tools. These conclusions are similar to those of Miller (1992, pp. 237–38) in his assessment of how inherent “managerial dilemmas” between individual employees’ self-interests and the corporate interests of the principal may

07 Improving Health

174

8/24/11, 4:24 PM

Primary Health Care in Cur itiba, Brazil

175

(sometimes) be reconciled. He argues that economic incentives are insufficient for managing complex organizations because agents can always find ways to cheat their principals (that is, information is asymmetric); thus, “political” mechanisms to improve credible commitment, thereby cooperation, become important. Using the game theory, he points out that it can be in the mutual interest of agents to cooperate if they believe that they have a long-term future in the organization (which makes cooperation and its fruits a “repeated game”), if they engage in activities that create a shared confidence that others will cooperate, and if they have some reciprocal ability to punish non-cooperators. Horizontal cooperation in smaller teams, he argues, can be reinforced by informal social rituals of reward and punishment, while vertical cooperation between subordinate and superior can be reinforced by creating enough trust to allow each side to reveal more information. Inculcating a sense of common identity and purpose is central to this.

Speculating on Policy Implications It is hardly appropriate to speak of replication of the Curitiba experience, especially in other countries.23 First, Curitiba’s story is one of complexity and nuance: a set of interacting systems developed over three decades and underpinned by a strong and enduring performance culture, which is driven as much by informal values as by formal design. Second, a department with only 7,000 staff, is by many standards, modest in size, hence more easily managed. Third, Curitiba’s experience occurred within a particular set of historical conditions, including the city’s broader reform experience, and the Brazilian health care revolution. With that caution in mind, it is nonetheless useful to understand how Curitiba has applied strategic planning. It has applied it more as a methodology (or work habit) than as a formal process — a methodology that has pervaded the organization from top managers to the operational level. It has involved a mindset that thinks about the future, looks for and solves problems, understands systems’ complexity, and sees that problems create opportunities. It is empirical and experimental and takes risks. The formal tools of performance management are an adjunct to, not a driver of, strategic planning. Our case study identified two types of management tools to align the expectations of principal and agent: those that offset information

07 Improving Health

175

8/24/11, 4:24 PM

Geoffrey Shepherd

176

asymmetry; and those that provide direct incentives to tie an employee’s behaviour to performance outputs.

Reducing Information Asymmetry This study suggested two particular ways for reform-minded governments to align expectations and incentives between principals and agents better: • Invest in strategic planning by clarifying expectations and establishing regular feedback on performance. These processes work better when operators — the frontline troops — are involved in the planning and evaluation cycle. • Invest in better strategic management of information by standardizing processes and definitions; ensuring the quality of data; tapping information from the community; and providing IT systems to manage this information.

Direct Performance Incentives The lessons from this study that relate to performance incentives are more difficult to interpret. The SMS in Curitiba places great emphasis on intrinsic staff motivations (investing in esprit de corps, using staff appraisals to identify and solve problems) and maintains a hierarchical management format. Curitiba has tried to use performance bonus schemes, but the current scheme appears to have been effective more for its compatibility with intrinsic incentives, than for its extrinsic effect. On the other hand, it is easy to show cases in the health sector where reliance on extrinsic incentives is more appropriate. The other case handled in the World Bank’s study (2007) from which the Curitiba case is drawn, provides an example: non-profit hospitals contracting with the state of São Paulo rely primarily on extrinsic incentives — the freedom and incentives to manage for managers, and the discipline of external labour markets for staff. (On the other hand, they do not use performance bonus schemes.) These hospitals have proven more efficient than São Paulo’s public hospitals, whose staff fall under the Brazilian civil service statute. The two cases, taken together, may suggest that extrinsic personnel incentives are more compatible with the use of external labour markets, while intrinsic incentives may be more compatible with internal labour markets, such as

07 Improving Health

176

8/24/11, 4:24 PM

Primary Health Care in Cur itiba, Brazil

177

the market created by the Brazilian civil service statute, where entry and exit are limited (World Bank 2006, pp. 80–81).24 Of the two types of management tool, steps to reduce information asymmetry are generally less sophisticated and entail fewer risks of “unintended consequences” than direct performance incentives. Thus, most health authorities in developing countries may find it appropriate to focus first on basic statistics and management information, as well as minimum administrative capacities (for example, simple planning, logistics management). Authorities must have a clear policy objective that can be translated into meaningful performance targets or expectations at a sectoral/operational level. At the beginning, when the government’s experience with designing proper performance indicators is limited and its capacity for information management is weak, it is preferable to focus on a limited set of performance indicators: far better to have a few indicators that can be reliably tracked and that serve as references for the government and service providers in their regular performance review discussions, than to propose myriad indicators without the ability to track them on a timely and reliable basis. There are gradients of sophistication in each set of tools. Consider strategic planning, for example. Clarifying expectations can be tackled at a minimum of four different levels: (1) informal or formal discussions; (2) formal strategic or indicative planning; (3) relational contracting; and (4) enforceable contracts. Similarly, feedback can take place through simple discussions (informal or formal), or through a formal evaluation. Once again, what really matters is the political commitment behind the expressed performance expectations. The choice of instruments is secondary. Governments with limited institutional capacity are well advised to be cautious about adopting formal contractual tools, as it will prove particularly difficult to set appropriate performance targets ex ante, which may result in frequent contract revisions (reducing the credibility of contracts as an enforceable tool), or the complete disregard of contracts. The advice that the managers of the SMS themselves provide on management contracting emphasizes realism, good communication, careful monitoring, and listening.25 Similarly, mechanistic applications of performance incentive tools, especially performance-related pay, should be approached with caution by governments with weak administrative capacities. The conditions under which performance-related pay can be effective are rather stringent:

07 Improving Health

177

8/24/11, 4:24 PM

Geoffrey Shepherd

178

performance itself must be reliably monitored, managers must have the courage and power to apply both positive and negative incentives, and so on. When such conditions are not met, performance-related pay schemes become a formality without real impact on staff behaviour, or may even produce negative consequences. On the other hand, the development of professionalism and esprit de corps is a longer-term endeavour that is not within easy reach of any particular government administration. Conscious efforts to strengthen professionalism through careful recruitment strategies will pay off in the medium to long run.

ANNEX TABLE 7.1 Health Financing in Curitiba, 2001–09 Municipal expenditure Primary health

Health payroll

Federal Govt. health transfers

R$110 R$131 R$157 R$193 R$237 R$241 R$261 R$225 R$276

n.a. R$77 R$84 R$99 R$128 R$163 R$193 R$222 R$255

n.a. R$222 R$243 R$293 R$272 R$330 R$284 R$375 R$480

In constant prices of 2009 (R$millions): 2001 R$2,293 R$464 R$171 2002 R$2,395 R$474 R$185 2003 R$2,464 R$490 R$206 2004 R$2,541 R$510 R$237 2005 R$2,552 R$582 R$279 2006 R$2,902 R$623 R$277 2007 R$3,094 R$614 R$287 2008 R$3,274 R$656 R$234 2009 R$3,468 R$754 R$276

n.a. R$109 R$110 R$122 R$151 R$187 R$213 R$230 R$255

n.a. R$315 R$319 R$361 R$321 R$379 R$313 R$389 R$480

Total

Health

In current prices (R$millions): 2001 R$1,470 R$298 2002 R$1,688 R$334 2003 R$1,878 R$373 2004 R$2,064 R$415 2005 R$2,167 R$494 2006 R$2,526 R$542 2007 R$2,811 R$558 2008 R$3,159 R$633 2009 R$3,468 R$754

Consumer price index (FIPE, São Paulo)

2005 2204 2384 2541 2656 2723 2842 3017 3128

Sources: Expenditure and tr ansfers: Curitiba Municipal Health Secretar iat (SMS) at Consumer price index: FIPE (F oundation Institute of Economic Studies , University of São Paulo) at

07 Improving Health

178

8/24/11, 4:24 PM

Primary Health Care in Cur itiba, Brazil

179

ANNEX TABLE 7.2 Population and Public Health Employees in Curitiba, 2001–09 Population (000) 2001 2002 2003 2004 2005 2006 2007 2008 2009

1620.2 1644.6 1671.2 1727.0 1757.9 1788.6 1808.2 1828.1 1851.2

Number of health employees (December) 4,843 4,734 4,726 5,027 5,180 5,723 6,237 6,394 6,891

Source: Population: IBGE (Br azilian Institute of Geog raphy and Statistics) at Employees: Curitiba Municipal Health Secretariat (SMS) at

Notes 1. This chapter draws largely on a World Bank report on enhancing performance in Brazil’s health sector (World Bank 2006). The report, presented two case studies: primary health care in Curitiba, and non-profit hospitals in the state of São Paulo. The Curitiba case was prepared by the present author with Juliana Wenceslau and Samuel Jorge MoysÈs and benefited from the comments of Jerry La Forgia, Chico Gaetani, April Harding, and Yasuhiko Matsuda. The work was based on interviews and a field survey carried out in 2005 (see Chapter 3 and Appendix D of World Bank 2007). Information has been updated in the present chapter, thanks to the help of Curitiba’s Municipal Health Secretariat, in particular Eliane Chomatas, Curitiba’s secretary of health, and InÍs Kultchek Marty, adviser to the Superintendency of Management. Since 2005, Curitiba’s primary health care system has continued to expand vigorously and management contracting has become a central feature of health care management, but the organizational features of the system have not changed significantly otherwise (Marty 2010). The findings, interpretations, and conclusions are the author’s own and should not be attributed to the World Bank, its Executive Board of Directors, or any of its member countries. 2. The PSF English-language website is . 3. In many Brazilian municipalities, other PSF staff are employed under private labour law.

07 Improving Health

179

8/24/11, 4:24 PM

Geoffrey Shepherd

180

4. See also IMAP (2005), which pointed to some typical achievements of the health system, and Sant’Ana et al. (2002), which studied the impact of the first PSF unit, opened in 1993. 5. Following a 1998 constitutional amendment, civil servants can, in theory, lose their jobs for poor performance; in practice, this is hardly known. 6. The PSF units are staffed through an internal competition (written exam plus interview) open to qualified SMS staff. This system was introduced to provide a legitimate appointment procedure that helps to counteract any disincentive that basic unit staff feel as a result of getting lower pay for similar work. Our survey suggested that, surprisingly perhaps, there was little sense of a difference in effort and performance between PSF unit and basic unit staff. 7. A figure based on the years up to 2005. 8. In a separate scheme, community health agents can also earn a 20 per cent bonus for having reached the work targets set for them. 9. In 2005, consistently with the peak in poor performance cases in 2005–06, the SMS reported that it was trying to raise standards of performance over time. 10. It is often observed that evaluators in Brazil (as elsewhere in Latin America) are reluctant to give a low performance evaluation. This may be because the prevailing culture values equal treatment more than rewarding merit; or it may reflect a supervisor’s fear that low evaluations will have adverse consequences for them. 11. In July–August 2005, as part of its study of the management of primary health care in Curitiba, the World Bank, with the help of Curitiba’s Municipal Health Secretariat, carried out a confidential survey of the perceptions of professional staff and managers about the performance incentives they faced (Appendix D of World Bank 2006). Of 372 eligible staff in a representative sample of thirtyone health units, 254 responded. 12. See . 13. At the beginning, the doctors were resistant, but after the entry of scientific societies into the discussion, they changed their position. Indeed, they came to see protocols, with their support from scientific societies, as a protection against claims of medical malpractice. 14. The Aravind Eye Hospital in India is a well known example of economies of scale through specialization — in this case in cataract surgery (Rangan 1994). With less than 1 per cent of the country’s ophthalmic labour power, Aravind performs about 5 per cent of all cataract surgeries in India. (). 15. Sometimes doctors do not fill in the information properly or do not follow the agreed procedures. To check on this, designated doctors (who are allowed access to confidential information) monitor the use of protocols in a sample of

07 Improving Health

180

8/24/11, 4:24 PM

Primary Health Care in Cur itiba, Brazil

16.

17. 18.

19.

20.

21.

07 Improving Health

181

health units. The intention is for each health district to have one doctor specialized in monitoring protocols. Thus, they have the characteristics of a relational contract, a quasi-contractual agreement getting its strength not from formal sanctions, but from the shared needs of principal and agent to do business together over the longer term (Marty 2010). Unlike some relational contracts, the SMS management contracts are, through the agreed targets, highly specific about some aspects of expected performance. In the 2008 meeting, 1,821 public servants participated and between them made 254 professional presentations (Secretaria Municipal da Saúde 2007). Constitutional Amendment No. 29: the minimum federal health spending was required to rise as a share of the budget from 2000 to 2004, then stay at a fixed ratio to nominal GDP. By 2004 the floor on state and municipal spending was to become 12 and 15 per cent, respectively, of revenue collections. These were transitional rules but, in the absence since 2000 of an anticipated complementary law, these arrangements have remained in force. Evans (1997) suggests that Curitiba’s benign form of constructive, democratic politics benefited from two factors. First, Curitiba did not have the traditional dominant oligarchy that typified much of Brazil. The power of traditional elites was further diluted by new immigrants arriving in the later nineteenth century to become independent farmers. The lack of dominant elites encouraged the growth of collective institutions and the provision of public goods. Second, political entrepreneurship, which legitimized planning, was also a necessary part of the explanation. It can also be argued that Curitiba has good political reasons to emphasize the provision of social services. Evans (1997) has pointed out that Curitiba’s politicians have long had to manage the challenge of conflicts resulting from immigration into the city. Curitiba has seen a very high rate of immigration — at times the highest in Brazil — some of which is, of course, the consequence of its success. Hence, the city has had to find ways of winning over the newly arrived poor in the urban periphery. In this light, the effectiveness and expansion of health services can be seen as a political imperative in the struggle to acquire the political support of new immigrants. The report identified a range of problems across many municipal and state health authorities. Poor work incentives were reflected in absenteeism, incomplete work hours, excessive leave, and low salaries for qualified personnel. Facility managers had little autonomy in managing their personnel, while managers at the centre were unable to control staff allocation. Employees working in one institution were often accountable to another. There was little performance evaluation. There was an excess of low-qualified personnel and a deficit of properly qualified personnel, especially in management.

181

8/24/11, 4:24 PM

Geoffrey Shepherd

182

22. Perhaps the best known champion of this view is Carlos Bresser Pereira, who, as minister of federal administration and reform of the state (1995–98), was the architect of various initiatives to introduce performance-oriented management into government. 23. This section draws heavily on the conclusions of the original study (World Bank 2006, pp. 79–82). 24. One case study shows subcontracting working well; the other shows the direct provision of health services working well. So, taken together, the two studies have little to say about the relative merits of either mode. 25. Ministério da Saúde, PolÌtica Nacional de Humanização 2006, p. 15 (our translation).

References Campbell, Tim. “IPPUC: The Untold Secret of Curitiba: In-House Technical Capacity for Sustainable Environmental Planning”. Draft, Urban Age Institute, 2006. Ducci, Luciano, Maria Alice Pedotti, Mariangela Galvão Simão & Samuel Jorge Moysés (eds.). Curitiba: A Saúde de BraÁos Abertos. Rio de Janeiro: Centro Brasileiro de Estudos de Saúde, 2001. Evans, Peter. “Sustainability, Degradation and Livelihood in Third World Cities: Possibilities for State-Society Synergy”. Paper prepared under the auspices of the Research Group on States and Sovereignty for UN21 Project, Conference on the Global Environment, New York, 15 November 1997. Gnatek, Tim. “Curitiba’s Urban Experiment”. Available at . IMAP (Instituto Municipal de Administração Pública). “Avaliação das Políticas Públicas Municipais de Curitiba — 1997 a 2004”. Available at . Marty, Inês Kultchek. “Primeiras Experiências de Contratos de Gestão em Curitiba: Secretaria Municipal da Saúde”. In Gestão para Resultados em Curitiba — A Experiência de Contratualização, edited by Homero Giacomini. I Edição, Curitiba: Instituto Municipal de Administração Publica, 2010. McKibben, Bill. Hope, Human and Wild. Boston: Little Brown, 1995 (excerpted in Yes! magazine, summer 1999). Miller, Gary J. Managerial Dilemmas. Cambridge: Cambridge University Press, 1992. Ministério da Saúde, Política Nacional de Humanização. “Boas práticas de humanização na atenção e gestão do Sistema Único de Saúde — SUS: contrato interno de gestão na Secretaria Municipal de Saúde de Curitiba”. Secretaria de Atenção à Saúde, 2006. .

07 Improving Health

182

8/24/11, 4:24 PM

Primary Health Care in Cur itiba, Brazil

183

Ministério da Saúde. “Curitiba firma Contrato de Gestão com metas para a saúde”. Revista Brasileira Saúde da Família, Ministério da Saúde. Departamento de Atenção Básica — DAB, No. 14 (2007): 48–57. Available at . New Zealand Parliamentary Commissioner for the Environment. “Showing the Way: Curitiba: Citizen City”. Wellington: Parliamentary Commissioner for the Environment, 2002. Available at . Rangan, V. Kasturi. “The Aravind Eye Hospital, Madurai, India: In Service for Sight”. Harvard Business School Case 593-098, 1994. Sant’Ana, Ana Maria, Walter W. Rosser, & Yves Talbot. “Five Years of Family Health Care in São José”. Family Practice 19 (2002): 410–15. Secretaria Municipal de Saúde de Curitiba. “Caderno de Resumos: 8° Encontro de Saúde Coletiva de 13 a 16 de Julho de 2004”. 2004. Secretaria Municipal da Saúde de Curitiba. “Forma de Pagamento e Resultados dos Serviços Públicos em Curitiba”. PowerPoint presentation at 5° Seminário Internacional de Atenção Primária à Saúde. Xavier, Elziane Cazura & Maria Emi Shimazaki (n.d.). “A Experiência de Curitiba com o Prontuário Eletrônico — A Ousadia em Inovar”. (n.d.). Available at . Weyland, Kurt. Democracy Without Equity: Failures of Reform in Brazil. Pittsburgh: University of Pittsburgh Press, 1996. World Bank. “Brazil: Enhancing Performance in Brazil’s Health Sector: Lessons From Innovations in the State of São Paulo and the City of Curitiba”. Report No. 35691-BR, 27 October 2006. Available at . World Bank. “Brazil: Governance in Brazil’s Unified Health System (SUS): Raising the Quality of Public Spending and Resource Management”. Report No 36601BR, 15 February 2007. Available at .

07 Improving Health

183

8/24/11, 4:24 PM

Che Katz and Thoai D. Ngo

184

8 INCREASING UPTAKE OF REPRODUCTIVE HEALTH SERVICES USING INNOVATIVE FINANCING MODELS Experiences of Marie Stopes International Che Katz and Thoai D. Ngo

CLIENT CHOICE THROUGH DEMAND-SIDE FINANCING Marie Stopes International (MSI) is globally recognized as one of the leading family planning organizations. The organization uses demandside financing to improve access to family planning, sexually transmitted infection (STI) services, and safe motherhood services, using a range of innovative financing models. MSI is currently running voucher schemes and outputs-based aid (OBA) financing in several countries, including Kenya, Uganda, Tanzania, Cambodia, and Yemen. This chapter explores MSI’s global experiences with demand-side financing for sexual and reproductive health services.

08 Improving Health

184

8/24/11, 4:25 PM

Experiences of Marie Stopes Inter national

185

WHY DEMAND-SIDE FINANCING? In some settings, limited government capacity challenges the ability to design and implement effective public health service delivery mechanisms (Eichler 2006). Historically, financial input into health systems has focused on the supply side (infrastructure, staff, salaries, etc.), often, though not exclusively, through the government/public sector. This approach is failing to impact significantly on health outcomes as a result of inefficiencies, inflexibility, and monopolies inherent in many public sector systems (Sandiford et al. 2005; Bhatia & Gorter 2007; Bhatia et al. 2006). A large proportion of health care in developing countries is delivered through the non-state sector. In Pakistan, for example, 80 per cent of total health expenditure is out-of-pocket payments by individuals for curative services provided by independent, private health care providers (Pakistan Medical Research Council 1998). The scale of this private sector reflects on public services — both the shortage of public health care facilities in rural areas and urban slums, and the perceived low quality of their services by clients. The existence of a large private sector in developing countries presents an opportunity to increase the poor’s access to key health services. However, harnessing the private sector to provide services to low-income groups and thereby contribute to public health objectives presents a number of challenges. The quality of private service provision is often poorly regulated by the state and hence can be low or varied. The services may be biased towards those that maximize revenue rather than those that offer the best public health outcomes, such as long-term family planning methods, bed nets, and other preventive services.

SEXUAL AND REPRODUCTIVE HEALTH Sexual and reproductive health is underresourced in many developing countries. Evidence indicates that upfront investment in preventive services, such as family planning, contributes directly to poverty alleviation, and cost savings for both the health system and consumers due to prevention of complications of unplanned deliveries, post-abortion care, and maternal and infant mortality (Goldie et al. 2010). Nevertheless, with the exception of safe deliveries, donors have largely ignored most aspects of reproductive health, such as family planning and safe abortion.

08 Improving Health

185

8/24/11, 4:25 PM

Che Katz and Thoai D. Ngo

186

In many developing countries, investments in vertical HIV/AIDS programming have also skewed the distribution of health resources, resulting in reproductive health services being underresourced and underutilized. In some countries, the over-incentivizing of HIV/AIDS services has resulted in providers being demotivated towards providing services that are not tied to incentives. Providers are further demotivated from offering family planning by the high fees health providers can charge for illegal abortions, which can directly contribute to high maternal mortality rates. Offering service providers incentives for safe delivery services also deters providers from promoting family planning and preventing unplanned pregnancies, particularly the more reliable longterm and permanent methods where there is high demand, but lack of supply. Furthermore, claim rates on existing equity funds for reproductive health and family planning services are typically low because recipients think that these funds are available only for curative or chronic health problems. Many of these funds also reimburse only at tertiary health facilities and not at the lower levels of the health system, where most people have their family planning needs met.

HOW DOES DEMAND-SIDE FINANCING WORK? An alternative approach is the provision of financial support to the demand side, based on outputs. The premise of demand-side financing (DSF) is to focus subsidies on specific services for specific populations. It can be used within both the public and private sectors. DSF enables output-based remuneration, increases the focus on evidence-based practices, and, perhaps most importantly, shifts the focus and subsequently the purchasing power from the provider (health care professional) to the patient. Such an approach is arguably more effective in improving service delivery (Eichler et al. 2001) and more beneficial to the poor and vulnerable by targeting resources directly to this group (Sandiford et al. 2005). In addition, by providing choice to the client, such schemes can stimulate competition between providers, leading to positive, client-focused changes in service provision (Bhatia & Gorter 2007). Various types of demand-side financing tools, such as vouchers, contracting out, social insurance schemes, and output-based aid, can be used to attract clients, reduce out-of-pocket expenses, increase equitable access to sexual and reproductive health services, stimulate patient choice, and increase efficiency (Bellows et al. 2009).

08 Improving Health

186

8/24/11, 4:25 PM

Experiences of Marie Stopes Inter national

187

WHY VOUCHERS? Vouchers have been increasingly used to promote social gain, in both the education and health sectors, through both public and private providers. Voucher schemes have the advantage of being able to target specific groups according to geographic, demographic, or socio-economic criteria. Thus, they offer a means of ensuring that the primary beneficiaries of a programme are those who need the services most, or who have the most difficulty in accessing public services. Furthermore, the benefit to the voucher user is not just the reduced cost of the service, but the fact that the price of the voucher is fixed. Thus, a pregnant woman seeking clinical services for safe delivery knows that she will not face higher charges in the event of a complicated delivery or Caesarean section. Similarly, family planning vouchers enable users to choose the method of contraception that suits them rather than the cheapest option. The World Bank Group has published a guide to competitive vouchers in health (Sandiford et al. 2005), and the document clearly describes the logistics of a typical voucher scheme, though the administration can vary from one programme to another. Figure 8.1 provides an overview of the basic design. Despite the advantages, donor support for voucher schemes has been criticized in some quarters for diverting funds that could be invested in public health services. It is true that in most developing countries, the public health sector suffers from chronic under-investment, and remedying this will remain a priority for international development efforts for the foreseeable future. However, private sector providers are likely to remain a significant source of health care, so it is important that they are brought within the government regulatory framework as a means of improving their service quality and safety. Voucher schemes enable governments and private providers to work together, which develops the state’s capacity for contracting, regulating, and monitoring non-state health providers, and hence improving the quality of care. In this way, voucher schemes can contribute to long-term health system strengthening.

CASE STUDIES OF SUPPLY AND DEMAND FINANCING The high unmet demand for reproductive health services, particularly long-term and permanent family planning methods and safe abortion

08 Improving Health

187

8/24/11, 4:25 PM

Che Katz and Thoai D. Ngo

188

FIGURE 8.1 Basic Design for a Voucher Scheme

DONOR FUNDS

STEP 1

STEP 6

VOUCHER MANAGEMENT AGENCY STEP 5

STEP 2

CONTRACTED HEALTH CARE PROVIDER

VOUCHER DISTRIBUTERS

STEP 3

STEP 4 CLIENT

Source: Boler & Harris (2010).

Step 1: A voucher management agency receives funds to implement the programme Step 2: The voucher management agency provides vouchers to an appointed distribution agency Step 3: Clients in the targeted populations receive the vouchers from the distribution agency Step 4: The clients then redeem the vouchers at participating health facilities for specified services Step 5: The provider then sends the voucher back to the management agency and is reimbursed the cost for the treatment/service. Step 6: The voucher management agency reports back to the donor.

services, has led MSI to explore alternative financing mechanisms to ensure wider access to reproductive health services. MSI has adopted a total market approach through public, NGOs and private providers. These

08 Improving Health

188

8/24/11, 4:25 PM

Experiences of Marie Stopes Inter national

189

innovative models are being tested in most of the forty countries where MSI works. While this work is relatively new, MSI has seen encouraging signs that include: improved uptake of services; better reach, especially to more vulnerable and underserved populations; and increased competition among providers, which has resulted in improved quality and higher client satisfaction.

Marie Stopes International Cambodia In 2007, using MSI discretionary funds, Marie Stopes International Cambodia (MSIC) began a small, outputs-based assistance (OBA) project focusing on supporting trained public sector providers to offer long-term family planning methods (IUD) and permanent methods (tubal ligation) that were previously difficult to access in Cambodia. The project involved paying service providers the user-fee for each IUD or tubal ligation performed, thus making these services more accessible and affordable to consumers. Unfortunately only five providers in the country were qualified to provide tubal ligation services. This project is supplemented by MSIC’s own clinical outreach team, which covers more than half the referral hospitals with regular, voluntary surgical contraception services (tubal ligation and vasectomy). Through this OBA programme, the number of IUD insertions increased from thirty-two in 2007 to 1,852 in 2009 (Figure 8.2). Additionally, the number of tubal ligation services provided directly by government providers increased from 286 between January and July 2009, to 437 between July and December 2009 (Figure 8.3). Based on the experience of the IUD project, AusAID has supported MSIC to conduct a pilot project in early 2010 using an IUD voucher scheme, coupled with an OBA user-fee for service providers. Through this project, MSIC will distribute IUD vouchers to poor households in Koh Kong and Svay Rieng districts, and pre-assessed providers will be reimbursed based on redeemed vouchers. MSIC’s public private partnership, conducted by MSIC’s clinical outreach team, works with the referral hospital counterparts to build their capacity in provision of tubal ligations. This project has increased its provision of tubal ligations from 758 in the first six months of operation (July to December 2007), to 2,016 (July to December 2009) (Figure 8.4).

08 Improving Health

189

8/24/11, 4:25 PM

08 Improving Health

190

8/24/11, 4:25 PM

Number of IUD Insertion

July-December 2007

32 January-June 2008

235

July-December 2008

947

January-June 2009

1631

Source: Marie Stopes International Cambodia, Health Management Inf ormation Data, 2010.

0

200

400

600

800

1000

1200

1400

1600

1800

2000

July-December 2009

1852

FIGURE 8.2 MSI Outputs-Based Assistance (OBA) for IUD Insertions in Cambodia

190 Che Katz and Thoai D. Ngo

08 Improving Health

191

8/24/11, 4:25 PM

Number of Tubal Ligations

January-June 2009

286

July-December 2009

437

Source: Marie Stopes International Cambodia, Health Management Inf ormation Data, 2010.

10

50

100

150

200

250

300

350

400

450

FIGURE 8.3 MSI Outputs-Based Assistance (OBA) for Tubal Ligations in Cambodia

Experiences of Marie Stopes Inter national

191

08 Improving Health

192

0

500

1000

1500

2000

2500

776

909

2016

July-December January-June July-December January-June July-December 2007 2008 2008 2009 2009

758

1960

Source: Marie Stopes International Cambodia, Health Management Inf ormation Data, 2010.

Number of Tubal Ligation

FIGURE 8.4 MSIC Public-Private Partnership Voluntary Surgical Contraceptive in Mobile Outreach Programme

192 Che Katz and Thoai D. Ngo

8/24/11, 4:25 PM

Experiences of Marie Stopes Inter national

193

Marie Stopes International Uganda The aim of the Marie Stopes International Uganda project is to: reduce the number of mothers and children dying (or being disabled) due to absence or underutilization of skilled medical providers; and to increase access to safe childbirth for poor mothers, including antenatal and post-natal visits and basic emergency obstetric care. The project also aims to contribute to the reduction of HIV/STI incidence in Uganda. The project includes: (1) delivering high quality STD diagnosis and treatment through established qualified health service providers, using a voucher system, (2) reducing STD transmission by increasing access to the diagnosis and treatment of STDs, and (3) managing 35,000 STD cases and providing 110,000 safe deliveries. This project targeted the poor and individuals who were at high risk for STDs infections in twenty-six districts populated by 10.2 million people. The results of the project include 5,541 STD cases treated, and 2,567 babies delivered from February to December 2009.

GLOBAL EXPERIENCE IN DEMAND-SIDE FINANCING While the existing literature on the effect of voucher schemes on health outcomes is limited, there are examples of vouchers being used effectively to increase demand, access, equity, and use of services and products (for example, voucher schemes to increase uptake of malaria control in Tanzania [Mushi et al. 2003] and HIV/STI interventions in Nicaragua [Borghi et al. 2005] and Uganda [Bellows et al. 2008]). In Tanzania, the evaluation found that the voucher return rate was extremely high (97 per cent, [7,720 out of 8,000]), but awareness about the scheme declined over time. The researchers concluded that while a discount voucher scheme was a feasible system for targeting subsidies, a substantial amount of time and effort was needed to achieve high awareness and uptake. In Nicaragua, 4,815 vouchers were distributed to groups at high risk of STI; 1,543 patients were tested for STIs, and 528 STIs were cured in this period. The evaluation found that the voucher scheme offered an effective and efficient means of targeting and curing STI in high-risk groups, as well as encouraging quality care practices.

08 Improving Health

193

8/24/11, 4:25 PM

Che Katz and Thoai D. Ngo

194

In Uganda, a voucher programme was used to increase uptake of STI treatment for those most in need. DSF (voucher scheme) was chosen because low socio-economic status and risky behaviours were shown to be factors that influence the incidence of STIs and, more importantly, that cost was a significant barrier to accessing treatment. A quasi-experimental study with a two-panel design was designed, and a household survey was conducted to assess if the poor were disproportionally infected with STIs, and if treatment seeking behaviour was significantly different between areas with vouchers and areas with no vouchers. Among the seven clinics surveyed, non-voucher client visits for STI-related laboratory tests increased by an average of 32 per cent in the first year of vouchers compared with the year before. Some clinics saw considerable increases while others recorded a decrease in non-voucher client numbers. Total client visits (voucher and non-voucher) increased by 226 per cent for the same period, with all clinics experiencing an increase from the first year to the second. This indicates that voucher-bearing clients contribute significantly to the client volumes. Similar advances have taken place in the reproductive health field in order to improve access, uptake, and improvements in sexual and reproductive care to poor women (Mati 2008) and adolescents (Meuwissen et al. 2006a; Meuwissen et al. 2006b). A quasi-experimental intervention study in Nicaragua found that voucher receipt among adolescents was the main cause of the improved knowledge, use of services, contraception use, and condom use (Meuwissen et al. 2006b). Qualitative research suggested that vouchers removed various barriers (financial, logistical) and guaranteed confidentiality and choice. However, there were a number of constraints: the study relied on reported behaviour of young adolescents, and no triangulation with clinical data was presented. Furthermore, while participating clinics were selected based on suitability (defined by the services offered and location), the study did not account for inherent differences likely to be present in these services. An evaluation of a USAID-funded voucher programme in Zambia highlighted challenges in targeting those most in need (Kafue District CHEWS Team 2006). The project team concluded that the inability or unwillingness of voucher agents to use the targeting criteria could be due to the complexity of the matrix that the programme had established to identify and select the poorest in a generally poor population, and the lack of resources for support or supervision. They recommended

08 Improving Health

194

8/24/11, 4:25 PM

Experiences of Marie Stopes Inter national

195

careful review and field testing of the criteria used to select people, as well as ensuring there is adequate investment to implement such programmes. In an unpublished review, Bellows et al. (2009) illustrated that DSF schemes using vouchers are effective in improving equity (by targeting those most in need), as well as improving overall population health (by targeting those most at risk). In addition to targeting, vouchers were deemed advantageous in stimulating service utilization by breaking down some barriers to uptake, including lack of information and high out-ofpocket expenses. From the supply side, vouchers were reported to motivate quality improvements and encourage competition, thereby controlling costs. The authors concluded their review by highlighting a number of key unanswered questions: the dearth of quantitative assessments of the targeting process, and the lack of evidence of quality improvements that such schemes could stimulate. In another review article, Bhatia and Gorter (2007) recommend the increased uptake of DSF through competitive voucher schemes. The authors argue that implementing such schemes through the private sector is likely to be more effective and acceptable to the very poor who currently have limited access to formal health systems. These authors also highlighted that the limited available evaluation of the effectiveness of voucher schemes makes evidence-based policy making impossible for both governments in developing countries and international donors. They strongly recommend more studies to test competitive voucher schemes.

MSI’S FRAMEWORK FOR MEASURING IMPACT OF DSF MSI uses a conceptual framework for measuring the impact of demandside financing. This framework covers five elements: effectiveness, efficiency, equity, access, and acceptability (see Annex 1 for more details). Voucher programmes often relate directly to effectiveness. However, it is acknowledged that factors such as equity, access, acceptability, and efficiency must be addressed in order to achieve programme effectiveness. This framework represents a tool for ongoing assessment, and periodic evaluation, of voucher schemes. It will be used by MSI for future sexual and reproductive health voucher programmes in Sierra Leone and the Philippines, supported by the World Bank.

08 Improving Health

195

8/24/11, 4:25 PM

Che Katz and Thoai D. Ngo

196

CONCLUSION AND LESSONS While there is limited literature on voucher programmes and their evaluations, MSI’s experiences indicate that: • Public and private sector approaches can be linked through voucher schemes to address sexual and reproductive health issues. • When implemented in the right context, voucher schemes appear to be effective and efficient. • Demand-driven voucher schemes help to create competition, thereby raising the quality of care across the market. • Voucher schemes provide a more targeted approach for services that are underutilized. • Voucher schemes have the ability to target populations with high unmet demand better. However, there is a dearth of rigorous evaluations of the impact of voucher schemes on multiple outcomes (quality, access, efficiency and effectiveness of service delivery). The few studies that have sought to evaluate the effect of voucher schemes are generally conducted on a small scale in discrete populations, making generalization of the results difficult.

08 Improving Health

196

8/24/11, 4:25 PM

Experiences of Marie Stopes Inter national

197

Annex 1 Marie Stopes International Conceptual Framework for Measuring Demand-Side Financing Effectiveness

Efficiency

Equity

Access

Acceptability

Effectiveness is the extent to which the voucher intervention achieves its intended outcomes; that is, reducing financial barriers for the poor and increasing their uptake of targeted family planning services (WHO 2009). Efficiency refers to producing the maximum possible output from a given set of inputs. In the context of the current study, efficiency pertains to the performance of the management entity in terms of accreditation of service providers to participate in the voucher programme, distribution of vouchers to participating service providers, and other operational issues. Equity means that every potential client has a fair opportunity to obtain services, and no one should be unfairly disadvantaged from accessing a service. According to the WHO, there are two kinds of equity: horizontal (all those in similar circumstances should be treated the same), and vertical (those in different circumstances with respect to a characteristic of concern for equity should be treated differently) (European Observatory on Health Systems and Policies 2008). The focus here is to ensure that women who cannot afford to pay for targeted family planning will still be able to access those services. There are various types of access or issues concerning accessibility of services: financial, geographic, and cultural (European Observatory on Health Systems and Policies 2008). Financial accessibility measures the extent to which people are able to pay for services. Geographic accessibility measures the extent to which services are a) available and b) reachable by the population. In addition to physical distance, other factors such as the availability of local modes of transportation can influence geographic accessibility. Finally, cultural accessibility pertains to whether access to health services is inhibited by cultural norms and taboos, for example, whether a woman can use longterm methods of family planning without the consent of her husband or partner. Acceptability refers to the willingness of those involved (e.g. Marie Stopes Sierra Leone, OBA management agency, health service providers) to implement the voucher programme, and of the poor to use the vouchers to obtain family planning services (WHO 2009).

Source: Research & Metrics Team, Health System Department, Marie Stopes International, London, UK, 2010.

08 Improving Health

197

8/24/11, 4:25 PM

Che Katz and Thoai D. Ngo

198 Key Indicators Thematic area

Indicators

Method of data collection

Indicators Pertaining to the Hypothesis (“Implementation of an output-based aid voucher programme will be associated with higher uptake of long-term family planning methods by the poor in intervention sites compared to non-intervention sites”) Effectiveness

1.1 Number of acceptors new to modern contraception 1.2 % of new family planning acceptors who meet the programme criteria of “poor” 1.3 % of all family planning clients over the past year who redeemed vouchers

Clinic records Clinic records; exit interviews Clinic records

Indicators Pertaining to Research Question 1 (“How effective is the OBA voucher programme in reducing out-of-pocket family planning expenses and increasing use of family planning services among its beneficiaries?”) Effectiveness

Equity

Access

08 Improving Health

1. Number of acceptors new to modern contraception 2. % of new family planning acceptors that meet the programme criteria of “poor” 3. % of all family planning clients over the past year who redeemed vouchers 4. % of the target population receiving vouchers 5. Out-of-pocket expense (cost) incurred for contraceptive supplies and/or services, as a percentage of one month’s household income 6. Average out-of-pocket expense (cost) incurred by family planning clients 7. % of distributed vouchers that are redeemed for family planning services during the project period

198

Clinic records Clinic records; exit interviews Clinic records Voucher tracking and verification system Exit interviews

Exit interviews Voucher tracking and verification system

8/24/11, 4:25 PM

Experiences of Marie Stopes Inter national Acceptability

199

No quantitative indicators will be assigned to this dimension. However, qualitative data will be analysed to explore themes such as the following: • Among target beneficiaries, perceived quality of care of service providers that accept vouchers from clients • Stigma attached to clients who redeem vouchers for services • Among service providers, costs and benefits of participating in a voucher programme • Among beneficiaries, costs and benefits of participating in a voucher programme.

Exit interviews Provider interviews (quantitative) In-depth interviews with selected providers (qualitative)

Indicators Pertaining to Research Question 2 (“How efficient is the output-based aid voucher programme as a mechanism for facilitating access to family planning services among the poor?”) Efficiency

Accreditation process: 8. Average duration (in days) for accreditation of a service provider by OBA management agency Voucher distribution: 9. Average number of vouchers distributed per month, by community Reimbursement/claims processing: 10. Voucher return rate (number of voucher claims submitted by voucher service providers to the OBA management agency, divided by the total number of voucher clients over the past 12 months) 11. Average lag (in weeks) between claims submission and reimbursement Procurement/logistics management: 12. Number of incidents of empty stock of contraceptives (by type and frequency, duration) in the past 12 months

Voucher tracking and verification system Voucher tracking and verification system Voucher tracking and verification system In-depth interviews

Source: Research & Metrics Team, Health System Department, Marie Stopes International, London, UK, 2010.

08 Improving Health

199

8/24/11, 4:25 PM

Che Katz and Thoai D. Ngo

200

References Bellows, B., E. Mulogo, and F. Bagenda. “Evaluation of Output-based Aid (OBA) in Uganda: Impact of Contracted Facilities and Social Marketed Vouchers on Knowledge, Utilization, and Prevalence of Sexually Transmitted Infections (STIs) 2006–2007”. Berkeley, California: Venture Strategies for Health and Development, 2008. Bellows, B., D. Griffith, C. Grainger, & A. Gorter. “Literature Review: Vouchers for Reproductive Health”. Draft (unpublished), 2009. Bhatia, M.R. & A.C. Gorter. “Improving Access to Reproductive and Child Health Services in Developing Countries: Are Competitive Voucher Schemes an Option?”. International Journal of International Development 19, no. 7 (2007): 975–81. Bhatia, M.R., C.A.K. Yesudian, A. Gorter, & K.R. Thankappan. “Demand Side Financing for Reproductive and Child Health Services in India”. Economic and Political Weekly 41, no. 3 (2006): 279–84. Boler, T. & Harris, L. “Reproductive Health Vouchers: From Promise to Practice”. London: Marie Stopes International, 2010. Borghi, J., A. Gorter, P. Sandiford, & Z. Segura. “The Cost-effectiveness of a Competitive Voucher Scheme to Reduce Sexually Transmitted Infections in High-risk Groups in Nicaragua”. Health Policy and Planning 20, no. 4 (2005): 222–31. Eichler, R. “Can ‘Pay for Performance’ Increase Utilization by the Poor and Improve the Quality of Health Services?”. Discussion paper for the first meeting of the Working Group on Performance-Based Incentives, Center for Global Development, 2006. Available at . Eichler, R., P. Auxila, & J. Pollock. “Output-Based Health Care: Paying for Performance in Haiti”. The World Bank Group, Private Sector and Infastructure Network, Note Number 236, August 2001. Available at . European Observatory on Health Systems and Policies. “Glossary”. Available at . Goldie, S.J., S. Sweet, N. Carvalho, U.C. Natchu, and D. Hu. “Alternative Strategies to Reduce Maternal Mortality in India: A Cost-effectiveness Analysis”. PLoS Med. 7, no. 4 (2010): 1000264. Kafue District CHEWS Team. “Zambia Community Health Waiver Scheme: Final Evaluation”. Partners for Health Reformplus Project. Bethesda MD: Abt Associates, 2006. Mati, J.K.G. “Report of the Mid-term Review of the Reproductive Health Output Based Approach Project in Kisumu, Kitui, Kiambu, Korogocho and Viwandani”. Submitted to the National Coordinating Agency for Population and

08 Improving Health

200

8/24/11, 4:25 PM

Experiences of Marie Stopes Inter national

201

Development (NCAPD), 2008. Available at . Meuwissen, L.E., A.C. Gorter, A.D. Kester, & J.A. Knottnerus. “Does a Competitive Voucher Program for Adolescents Improve the Quality of Reproductive Healthcare? A Simulated Patient Study in Nicaragua”. BMC Public Health, 6(204), 2006a. Meuwissen, L.E., A.C. Gorter, & A.J. Knottnerus. “Impact of Accessible Sexual and Reproductive Healthcare on Poor and Underserved Adolescents in Managua, Nicaragua: A Quasi-experimental Intervention Study”. Journal of Adolescent Health 38, no. 1 (2006b): 56.e1-56. Mushi, A.K., J.R. Schellenberg, H. Mponda, & C. Lengeler. “Targeted Subsidy for Malaria Control with Treated Nets Using a Discount Voucher System in Tanzania”. Health Policy and Planning 18, no. 2 (2003): 163–71. Pakistan Medical Research Council. National Health Survey of Pakistan: Health Profile of the People of Pakistan. Islamabad, 1998. Sandiford, P., A. Gorter, Z. Rojas, & M. Salvetto. “A Guide to Competitive Vouchers in Health”. Washingston DC: Private Sector Advisory Unit, World Bank, 2005. World Health Organization. “Definitions”. 2009. Available at .

08 Improving Health

201

8/24/11, 4:25 PM

202

Seng Bundeth, Net Neath, P agaiya Nonglak and Sok Sethea

9 UNDERSTANDING RURAL HEALTH SERVICE IN CAMBODIA Results of a Discrete Choice Experiment Seng Bundeth, Net Neath, Pagaiya Nonglak and Sok Sethea

BACKGROUND Brief Overview of Cambodia’s Health Service System In Cambodia, health centres and health posts, mainly located in rural areas and providing the minimum package of activities (MPA), are the first point of patient contact. These facilities are staffed by nurses and midwives and do not normally include doctors, except in the case of former district hospitals. The health post is the lowest level of facility and is usually served by one or two staff. The next facility level are the referral hospitals, which are classified according to their capacity to provide the Complementary Package of Activities (CPA). Although their staffing includes full-time doctors, CPA 1

09 Improving Health

202

8/24/11, 4:25 PM

Understanding the Rural Health Ser vice in Cambodia

203

hospitals do not have general anaesthesia surgery. CPA 2 hospitals have general anaesthesia surgery, and CPA 3 hospitals, the highest level, usually located in provincial centres, offer additional specialized services. In 2003, fifteen referral hospitals were classed as CPA 3 (MoH 2006a). The five MoH national hospitals top the hierarchy of government health care facilities. All are located in Phnom Penh and offer a range of specialized tertiary services. Three private non-profit hospitals, which are members of the Kantha Bopha Group, are also included in official statistics as “national hospitals”. Of those, two are in Phnom Penh and another one is in Siem Riep. The Ministry of Health plans to increase the number of operational districts and health facilities to enhance the accessibility of care (MoH 2008). The private health sector has grown substantially since 1993, and there are now more for-profit than public facilities in Cambodia. Recently, two private hospitals have been established in Phnom Penh. In addition, there are many different types of agencies within the private health sector, ranging from medical clinics (without beds) to polyclinics equipped with high-technology equipment. There are 2,457 medical clinics without beds and 274 with beds. Accounting for only 9.5 per cent of the country’s population, Phnom Penh harbours a disproportionate share of private facilities (Table 9.1). TABLE 9.1 Licensed and Unlicensed Private Sector Clinics in 2003 Population

Maternity facilities

Medical clinics without beds

Medical clinics with beds

Dental clinics without beds

Dental clinics with beds

1.3 million 9.5%

2 7.4%

573 23.3%

205 74.8%

213 48.5%

35 100%

Rest of country

12.0 million 90.5%

26 92.6%

1,884 76.7%

69 25.2%

226 51.5%

0 0%

Total

13.3 million

28

2,457

274

439

Phnom Penh

35

Source: Office of Ethics and Regulation, Hospital Ser vices Department, MoH, 2004 (quoted in MoH 2006a).

09 Improving Health

203

8/24/11, 4:25 PM

204

Seng Bundeth, Net Neath, P agaiya Nonglak and Sok Sethea

Four paediatric, private, non-profit hospitals are managed and funded outside the government sector. Additionally, the Centre for Hope Hospital operates virtually entirely outside the Ministry of Health. Some NGOs, such as the Reproductive Health Association of Cambodia and Reproductive and Child Health Alliance, also run health clinics, particularly in reproductive health care. However, the boundary between this sector and the government is blurred because some non-government agencies and international development agencies work in partnership with the Ministry of Health and contribute substantial management and financial support to the operation of government hospitals and health centres.

SITUATION AND TREND OF HEALTH WORKFORCE Even though the private, for-profit sector boasts the most facilities, the majority of Cambodian health workers are employed in the public sector. Table 9.2 presents the existing public health workforce for the most important clinical cadres. These figures do not include workers who work exclusively in the private sector, though many in the public sector offer their services in private facilities as well. Whereas the public health workforce headcount expanded by 9 per cent between 2004 and 2008, the population grew by only 6.5 per cent. Combined with the growth in private for-profit facilities, these figures suggest that the ratio of health workers to population has been steadily TABLE 9.2 Public Health Workforce, 2008 Category Medical Doctor Medical Assistant Nurse, Secondary Nurse, Primary Midwife, Secondary Midwife, Primary Other Total Population1

2004

2008

2,120 1,340 4,498 3,356 1,756 1,063 2,931 17,064 12,571,351

2,196 1,258 5,186 3,534 1,844 1,478 3,054 18,592 13,388,910

% Increase 4 –6 15 5 5 39 4 9 6.5

Note: Population census 2008, and o wn calculations assuming a constant population gro wth rate of 1.58 per cent per y ear between 1998 and 2008. Source: National Health Statistics 2004 & 2008 (quoted in MoH, 2006 c and 2008).

09 Improving Health

204

8/24/11, 4:25 PM

Understanding the Rural Health Ser vice in Cambodia

205

increasing. Growth rates are, however, disproportionate, with the highest growth being in primary midwives and secondary nurses (15 per cent), while the numbers of medical assistants declined. Furthermore, the number of public medical doctors and primary nurses has not kept pace with population growth. Although the number of health workers has increased, their geographical allocation remains unequal and inequitable. In 2003, the population to doctor ratio in Phnom Penh was 1,200:1, which was 1/10 the ratio in the rest of Cambodia. The factor differences for midwives and nurses were much better, at 2.3:1 and 1.6:1 in 2003, but there was still a discrepancy in access between urban and rural areas. The factor differences between Phnom Penh and the rest of Cambodia only marginally improved from 2003 to 2008 (Table 9.3). Further details on geographical allocation are shown in Table 9.4, where twenty-four provinces have been grouped into four regions according to the coverage areas of Ministry of Health regional training centres (RTCs), plus Phnom Penh. The picture that emerges confirms a high concentration of health workers, particularly doctors, in Phnom Penh compared with the other regions. This picture is somewhat offset by the location of private for-profit facilities. As shown in Table 9.1, the supply of services such as maternity and medical clinics without beds is heavily concentrated in rural areas. However, whether this leads to improved access, especially for poorer segments of the population, depends on the fees charged by these providers.

TABLE 9.3 Population to Doctor, Nurse, and Midwife Ratios in Phnom Penh and the Rest of Cambodia, 2003 and 2008 2003

2008

Category

Phnom Penh

Rest of Cambodia

Factor difference

Phnom Penh

Rest of Cambodia

Factor difference

Doctors Nurses Midwives

1,200 850 3,200

12,300 2,000 5,200

10.2 2.3 1.6

1,155 787 3,339

11,511 1,714 4,124

10.0 2.2 1.2

Source: National Health Statistics 2004 & 2008 (quoted in MoH 2006 c & 2008).

09 Improving Health

205

8/24/11, 4:25 PM

Seng Bundeth, Net Neath, P agaiya Nonglak and Sok Sethea

206

TABLE 9.4 Population to Doctor, Nurse, and Midwife Ratios, by Region, 2008 Population per

Region Phnom Penh Battambang Kompong Cham Kampot Stung Treng Average

Factor differences

Doctor

Nurse

Midwife

Doctor

Nurse

Midwife

1,155 12,169 12,971 8,693 12,302 6,097

787 1,598 2,120 1,562 1,114 1,535

3,339 3,654 5,052 4,677 2,159 4,030

1.0 10.5 11.2 7.5 10.7 5.3

1.0 1.4 1.8 1.4 1.0 1.3

1.0 3.2 4.4 4.0 1.9 3.5

Source: National Health Statistics 2004 & 2008 (quoted in MoH 2006 c & 2008).

Based on the staffing norms put forward by the Ministry of Health (MoH 2006c) — 3,800 persons to one doctor, 1,750 persons to one nurse, and 4,700 persons to one midwife — there would be a need to produce annually 164–194 new doctors, 353–422 new nurses, and 133–157 new midwives until 2015 (Table 9.5). There is one government and one private medical school in Phnom Penh. The University of Health Sciences (government) produces about 150 doctors and 100 diploma nurses each year, with the exception of 2010, when only eighty-two medical students and 111 nurses will graduate. The International University (private), whose first intake of medical students

TABLE 9.5 MoH Projection of Doctor, Nurse and Midwife Intake Requirements 2006 2007 2008 2009 2010 2011 2012 2013 2014 Population (millions) Doctors required Annual intake Nurses required Annual intake Midwives required Annual intake

14 14.3 14.6 14.9 15.2 15.5 15.8 16.1 16.4 16.7 3,700 3,754 3,833 3,910 3,988 4,068 4,149 4,232 4,317 4403 164 167 169 172 174 179 183 187 190 194 8,000 8,515 8,693 8,876 9,062 9,253 9,447 9,645 9,848 10,055 353 360 367 375 382 390 397 405 414 422 3,000 3,038 3,099 3,161 3,224 3,289 3,355 3,422 3,490 3,560 133 135 137 139 142 145 148 151 154 157

Source: Health Workforce Development Plan (2006–10).

09 Improving Health

2015

206

8/24/11, 4:25 PM

Understanding the Rural Health Ser vice in Cambodia

207

was in 2002, expects to see about 170 doctors graduate in 2009. Students at the International University pay fees of about US$1,500 per year, while students at the University of Health Sciences pay about US$850 per year. The International University also trains nurses to diploma and graduate levels, but none will finish their training in 2009. Besides Phnom Penh, there are government regional training centres in Battambang, Kompong Cham, Kampot, and Stung Treng, which all train midwives and nurses. All students from RTCs received government scholarships and were required to work for the government from three to five years during 2002–07. However, after that the government has decreased the number of students, except for the midwives. Table 9.6 shows the number of medical, nurse, and midwifery students who will graduate in 2010. These numbers suggest that the production of health workers is on track to meet the Ministry of Health standards. However, if historic allocation patterns apply, the majority of the new doctors, nurses, and midwives will find employment in Phnom Penh, which means that the inequitable distribution of the health workforce will persist. Like many countries, Cambodia has difficulty in posting and retaining doctors, nurses, and midwives in disadvantaged rural areas. Anecdotal evidence for Cambodia suggests this is mainly because of the poor financial and non-financial attributes associated with rural jobs, such as low public sector pay, limited opportunities for private practice, poor educational TABLE 9.6 Medical, Nursing and Midwifery Students Graduating in 2010 Phnom Battambang Penh

Kompong Cham

Kampot

Stung Treng

Total

Medical students Diploma of nursing Post-basic midwife Primary midwife Secondary nurse Primary nurse Secondary midwife

82 111 0 0 0 0 0

0 0 32 63 92 32 0

0 0 36 30 52 0 0

0 0 0 61 105 26 0

0 0 0 30 62 0 0

82 111 68 184 311 58 0

Total

193

219

118

192

62

814

Source: Compiled by CDRI team based on lists giv en by all training centres.

09 Improving Health

207

8/24/11, 4:25 PM

208

Seng Bundeth, Net Neath, P agaiya Nonglak and Sok Sethea

services, housing and other amenities, weak managerial regimes, limited professional support, and distance from family. The Ministry of Health has designed three strategies to tackle allocation and retention of health workers in rural areas (MoH 2006c). The first is to plan the health workforce in line with health sector planning and the health coverage plan. The second is to develop and implement human resource management to deploy staff in underserved areas through “contracting of health service deliveries”. The third strategy is to increase and retain midwives in public sector facilities in rural areas by offering them an attractive remuneration and benefit package for selected locations. Moreover, the ministry described three complementary ways in which financial incentives for rural service can be provided: by continuing to promote better remuneration and salary through the civil servant reform scheme of the government; by improving the facility-managed salary supplementation using financial resources from user fees, the Health Equity Fund, service delivery grants, contracting, the Special Operation Agency, and community-based health insurance; and by expanding the Priority Mission Group and compatible incentive mechanisms. These financial incentives, which amount to salary top-ups for rural service, have been trialled in selected areas. However, their effects on health workers’ motivation, retention, and productivity have not been adequately assessed. Meanwhile, evidence from other countries suggests that health worker motivation and rural service uptake are complex and driven by a combination of “extrinsic” factors, or incentives (including but not limited to financial factors only), and “intrinsic” incentives arising from, for example, the wish to help people and the satisfaction derived from the delivery of health care. The structure of these incentives and their relative strength is likely to vary between individuals, cultures, and social settings. The empirical results from other countries may therefore not apply directly to Cambodia, and incentive regimes designed in other countries may not have the intended effect on Cambodian health workers. It is therefore essential to increase understanding of the financial and non-financial incentives required to post different classes of new clinical graduates in remote areas and retain them there. This study aims to contribute to that understanding. More specifically, it should inform the policy debate in at least the following ways:

09 Improving Health

208

8/24/11, 4:25 PM

Understanding the Rural Health Ser vice in Cambodia

1.

2.

3.

209

What salary (reservation wages) would be required to attract and retain different categories of clinical staff to urban and to rural areas? What intrinsic motivations make staff more inclined to work in rural areas, and how could these attributes be used to select applicants for clinical training? What value do graduate clinicians attach to opportunities for private practice, schooling, and other amenities in choosing where they will work?

STUDY DESIGN, METHODOLOGY, AND IMPLEMENTATION This section describes the purpose and the design of a cross-sectional survey of final year medical and nursing students in Cambodia. It sets out the research design, tool development, data collection, and data analysis.

Literature Review There is rapidly growing literature on the motivations that drive public sector worker performance. This study is not concerned with worker motivations in relation to effort, productivity, or service quality, but only with the incentives required to attract and retain health workers in geographical locations that may not be preferred. In this case, economic theory would suggest that financial compensation is required to offset the loss of utility of working in a non-preferred location. However, there is a considerable body of evidence to suggest that public sector workers are also motivated by non-financial considerations, including public sector ethos, professional reputation, job satisfaction, recognition and altruism (Burgess & Metcalfe 1999; Dieleman et al. 2003; Kitange & Hanoi School of Public Health 2001; Le Grand 2003; RoseAckerman 1996; Stilwell 2001; Williams & Windebank 2001; Jolley 2008; Karlsson 2008; Myers 2008). The non-financial incentives motivating behaviour has been classified in different ways by different authors (for example, Dixit 2002; Dixit 2004; Kirton 2001; Jacobson 1995; Williamson 1998) and for government health workers by Bennett & Franco (1999) and Franco & Bennett (2000). Synthesizing these classifications generates a hierarchy of non-financial

09 Improving Health

209

8/24/11, 4:25 PM

210

Seng Bundeth, Net Neath, P agaiya Nonglak and Sok Sethea

incentives that might modify a worker’s reservation wage in a nonpreferred location. Non-financial incentives fall broadly into two groups: extrinsic incentives and intrinsic incentives (Table 9.7). Extrinsic incentives might include: (1) political (workers might be persuaded to work in non-preferred locations if asked by a high political authority, possibly in the hope of recognition or promotion); (2) organizational (workers might agree to work in a well managed, but otherwise non-preferred location because their contribution is valued or out of a sense of “organizational citizenship” or team spirit; (3) managerial (workers might agree to work in a non-preferred location because management sets targets for and rewards such service by recognition, honours, training opportunities, or promotion — or because their work would be less [or more] well observed); (4) social (workers might prefer to work in a location where they have strong social obligations to their family or community). Intrinsic or “embedded” incentives might include reputation (workers may gain peer group approval), vocation (because they believe that their work has an intrinsic value), and altruism (because they gain satisfaction by helping people). Although it may be that choice of job will be influenced by an organization’s reputation, it is likely to be difficult for a student to place a

TABLE 9.7 Incentives and Public Sector Motivators Extrinsic incentives Financial

Non-financial

Fees for service (piece rates) Salaries Prizes Gifts Political Organizational Managerial Social

Intrinsic incentives Vocation Altruism Reputation Obedience

09 Improving Health

210

8/24/11, 4:25 PM

Understanding the Rural Health Ser vice in Cambodia

211

value on working in a particular organizational or management regime without prior experience. As a result, the non-financial incentives of most interest to this study are: (1) political (as political persuasion has been the main policy instrument to date); (2) organizational (at least in terms of the preference of working for the government or another health provider); (3) social (reservation wage rates may be lower when health workers are posted to their own communities); and (4) “intrinsic” incentives of the kind listed above. A key assumption of this study is that non-financial incentives reduce the financial compensation of workers who are posted to otherwise nonpreferred locations. However, Dixit (2004) suggests that financial and nonfinancial incentives may interact in much more complex ways. For example, it may be that stronger financial incentives undermine the strength of social and intrinsic incentives. As a result, it may be that this assumption is subject to further enquiry. Several studies have assessed health workforce recruitment and retention in rural areas. Daniels et al. (2007) studied rural retention of a variety of health workforces in the southern United States and found that rural background and preference for a small community were associated with rural recruitment and retention. Having enough work available, income potential, professional opportunity, and serving community health needs were important for their job choices. The evidence that rural background was a predictor of rural employment was at the basis of the Richards et al. (2005) study carried out in the Scottish Highlands. In a developing country (Vietnam), Dieleman et al. (2003) found that rural health workers’ job motivation was influenced by both financial and nonfinancial incentives. The main motivating factors were appreciation by managers, colleagues, and the community, a stable job, income, and training. The main discouraging factors were low salaries and difficult working conditions. Stenger et al. (2008) revealed that the extent to which doctors in rural areas in the United States were satisfied was related to their spouse/ partner’s happiness with the workplace location. Other elements that satisfied the majority of respondents were call group, quality of local referral specialists, and degree of intraspecialty and interspecialty collegiality. Although the majority of respondents anticipated remaining in their present practice, those who intended to leave raised dissatisfaction with low pay and a high workload.

09 Improving Health

211

8/24/11, 4:25 PM

212

Seng Bundeth, Net Neath, P agaiya Nonglak and Sok Sethea

A study of job satisfaction of nurses found that lifestyle and personal life issues were important to job choice (Molinari & Monserud 2008). Features such as time away from work, rural lifestyle, recreation opportunities, climate, and social activities were necessary for retention. Rural nurses, who were mostly satisfied and preferred to live in rural areas, generally possessed rural backgrounds. Satisfied nurses lived close to family, friends, and spousal employment. Therefore, in order to increase retention, hospitals may consider marketing rural lifestyle opportunities and interviewing applicants about their rural backgrounds and connections. Participants provided emphatic thoughts about the most and least satisfying aspects of their jobs. The insights from the literature review were used in the design of the study and questionnaires.

Discrete Choice Experiment A discrete choice experiment (DCE), and conjoint analysis, are quantitative methodologies for evaluating the relative importance of the different product attributes that influence consumer choice behaviour (Louviere et al. 2000). In health, conjoint analysis has increasingly been used to assess patient preferences for health care service delivery (Ryan 2004; Ryan & Farrar 2000), but a few studies have also utilized conjoint analysis on the preferences and choices of health care worker’s placements (Scott 2001). In this study, conjoint analysis, in the form of a discrete choice experiment, will be used to evaluate the relative importance of different job characteristics that may be used to attract and retain health workers in rural areas. Lagarde and Blaauw (2009), based on a systematic review of studies using DCE to elicit health worker job choices, found that non-pecuniary incentives are significant, sometimes more powerful than financial ones, in explaining job choices, and that different groups of health workers may have different preferences for job attributes. Most of the studies reviewed by Lagarde and Blaauw (2009) were conducted in developed countries and produce insights that are often specific to health labour markets in such countries. Recently Günther et al. (2010) studied job preferences of young doctors in Germany and the role of six attributes in job choice: professional cooperation, income, career opportunities for the partner, availability of childcare, leisure activities, and on-call duties. The results indicated that a change in

09 Improving Health

212

8/24/11, 4:25 PM

Understanding the Rural Health Ser vice in Cambodia

213

income led to the largest utility change, yet non-monetary attributes such as on-site availability of childcare and fewer on-call duties would decrease the additional income required to compensate the disutility of a rural practice. A few studies of doctors’ job preference in developing countries highlight a different set of insights. Chomitz et al. (1998) found significant differences in location preferences across different groups of medical students in Indonesia. Some of the factors that increased willingness to work in rural areas were rural background, being women, and not having studied in a private school. Hanson and Jack (2008) showed that the most important job characteristic for doctors in Ethiopia was the possibility of working in the private sector (which was not allowed for public doctors at the time of the study). A pay increase was the next most valued aspect, followed by improved housing, being posted in Addis Ababa (compared with a regional city), and better equipment. Compulsory service in the public sector in exchange for training was the least important preference. A study by Kolstad (2010) showed that different measures were likely to attract young doctors to rural areas. Alongside increased salary and hardship allowance, opportunities for further training were found to be among the most powerful attributes to attract doctors into rural areas in Tanzania. Other measures included housing, health infrastructure improvement, and adequate facility health equipment. Individual factors associated with rural choice were high degrees of willingness to help others. For nurses in Ethiopia, Hanson and Jack (2008) reported that the most powerful attraction was increased salary, closely followed by the possibility of being posted in a regional capital. Also, contrary to doctors’ preferences, nurses valued better equipment more than better housing, and the opportunity to work in the private sector came ahead of avoiding paying back years of training with years of work in the public sector. In another study of nurses in South Africa, Penn-Kekana et al. (2005) found that financial incentives were twice as powerful as the next job attributes, better facility management and improved equipment. Facilities that were well staffed and having social amenities were the least important determinants of nurses’ job choices. The study also found that younger nurses and those who worked in hospitals were more sensitive to salary levels, while nurses in rural areas were more concerned about facility management.

09 Improving Health

213

8/24/11, 4:25 PM

214

Seng Bundeth, Net Neath, P agaiya Nonglak and Sok Sethea

A study of job choices of nurses working in the public sector in Malawi by Mangham (2007) showed that graduate nurses appreciated higher pay, but also highly valued the opportunity to upgrade their qualifications quickly, as well as the provision of housing. Interestingly, nurses preferred jobs located in district towns more than cities. Faster promotion was valued more by younger nurses. The above studies indicate that although financial incentives often prove the most powerful, a wide range of non-financial incentives also play an important role in influencing job location choice. Moreover, nonfinancial incentives decrease the additional income required to compensate for the disutility of rural practice. They also show that the financial and non-financial job attributes explaining rural job choice vary considerably between countries. The studies also underline the importance of understanding local context in order to design effective policies to attract and retain health workers in rural areas.

Study Purpose • To determine the attitudes of health graduates towards working in rural areas and the values that influence these attitudes; • To evaluate health graduates’ preferences for various policy interventions, including differential salary packages, which may be used to improve the recruitment and retention of health professionals in rural areas.

Overall Study Methodology Since we could not observe real choice behaviour for rural service, we applied a stated preference method, also known as contingent valuation. Essentially this means we asked respondents to make choices between hypothetical job offers. Each offer varied in a number of attributes. From their replies, we elicited preferences. We used a discrete choice experiment (DCE) to investigate the relative importance health students attached to different policy interventions or jobs with different characteristics that may be used to attract them to rural areas. A self-administered questionnaire was given to each participant to collect basic individual characteristics. Questions on educational

09 Improving Health

214

8/24/11, 4:25 PM

Understanding the Rural Health Ser vice in Cambodia

215

background, attitudes towards living and working in rural areas, and reasons for the choice of a health career were also asked. Attitudes towards intrinsic incentives and government policies were included.

Sampling Strategy Sample size considerations in discrete choice experiments have generally been based on empirical experience rather than mathematical calculation (Louviere et al. 2000). The usual requirements are 100–150 respondents per subgroup of interest (Scott 2001). The sample unit was a final year of medical and nursing students. Lists of final year students from each school were used to develop the sample frame. A sample stratified per school was proposed. If non-equal sample fractions are chosen, then the results must be weighted to obtain representative results. When equal sample fractions are chosen, then the design is self-weighted. There are two medical schools and four regional training centres for nurses. They are located in five different regions, namely Phnom Penh, Kampot, Kompong Cham, Battambang and Stung Treng. Of the medical schools, only the University of Health Sciences agreed to collaborate in the study. All four RTCs took part. The sample frame is shown in Table 9.6. There were only eighty-two final year medical students due to graduate in 2010 from the University of Health Sciences, and therefore all students were included. In the nursing group, we excluded primary nurse students because they have had only one year of training, as opposed to three years for the other students. For the nurse student group, we applied a two-stage sample strategy. Firstly, we distinguished between the university and the four RTCs taken together. We applied equal sample fractions to obtain a self-weighted sample (Table 9.8). The sample size of secondary nurses and diploma nurses was 170 and sufficiently large. Students in each sample were randomly selected from pooled nurse student lists. In the midwife group, we decided to exclude the post-basic midwives to increase homogeneity and because this category may be phased out over time, and the secondary midwives were not in their final year at the time of the study. The sample frame of midwives was 184 and thus sufficiently large. For midwifery, all final year students were included.

09 Improving Health

215

8/24/11, 4:25 PM

216

Seng Bundeth, Net Neath, P agaiya Nonglak and Sok Sethea TABLE 9.8 Sample Size

Sample Frame

University of Battambang Health Science

Medical students Diploma of nursing Primary midwife Secondary nurse

Kompong Kampot Stung Total Cham Treng

82 45

Total

127

63 37

30 21

61 42

30 25

82 45 184 125

100

51

103

55

436

Discrete Choice Experiment Development The first step was to decide which job attributes to include in the experiment and to define the appropriate levels for each attribute. The focus of our DCE was on different financial and non-financial incentives that might be used to attract and retain health workers in rural areas in Cambodia. The selection of policy options to include in the DCE was based on: • A literature review of interventions that have been tried and have been successful in attracting and retaining health workers in rural areas in other developing and developed countries; • A review of human resource policy documents and key informant interviews with policymakers to identify feasible local strategies; and • Focus group discussions (FGDs) with medical and nurse students, and with doctors and nurses currently working in rural areas to obtain their suggestions and preferences. The qualitative pre-research consisted essentially of conducting FGDs with final year medical, nursing, and midwife students; and with doctors, nurses and midwives who currently work in rural facilities in order to get the views about important job attributes from those who experienced rural jobs and those who did not. An FGD guideline was developed based on a study used in Thailand (Pagaiya et al. 2009). It consisted of questions in relation to job characteristics

09 Improving Health

216

8/24/11, 4:25 PM

Understanding the Rural Health Ser vice in Cambodia

217

of interest, respondents’ job of choice, and measures to attract health workers to rural areas. Since the focus of this study was rural-urban career choice, it was important to conduct FGDs with students in both Phnom Penh and regional training centres, as location of the training institution is believed to impact on choice behaviour. An equal number of FGDs were administered in Phnom Penh and outside Phnom Penh. In Phnom Penh, two FGDs were conducted: one with medical students and another with nursing students. Outside Phnom Penh, two FGDs were conducted in Kampot regional training centre (RTC): one with midwifery students and another with nursing students. Semi-structured interviews were conducted with five doctors, five nurses, and five midwives currently working in rural areas (Kampong Trach district, Kampot province).

Determining Job Attributes for the Nurse/Midwife Groups The FGD conducted with twenty-two nurses and midwives showed that they would be interested in jobs that have characteristics such as proximity to their hometown, security, appropriate infrastructure, including water supply and electricity, higher salary, housing provision, adequate medical equipment, increased patient load, faster promotion, transportation allowance, and a good working environment. It was suggested that salary be increased by 50 to 100 per cent of the current level. The in-depth interviews with five nurses and midwives working in rural facilities obtained similar results. Job characteristics or incentives that can retain nurses and midwives working in rural areas included: higher salary, transportation allowance, safety at workplace, adequate equipment, adequate electricity and water supply, closeness to family, opportunity for training, housing provision, and good management. A short checklist filled in by final year nurse and midwife students confirmed the above results. As only about seven attributes can be taken into account in a DCE, a choice had to be made from within the attributes suggested. This was done by the researchers depending on the weight respondents generally placed on the attributes and also taking into account which policy interventions are feasible. For example, one may argue that it is not sensible to include housing provision as an attribute of a hypothetical rural job when current health policy doesn’t offer housing to workers, even if workers value it.

09 Improving Health

217

8/24/11, 4:25 PM

Seng Bundeth, Net Neath, P agaiya Nonglak and Sok Sethea

218

It was suggested that salary be increased by US$25 increments up to a maximum total of US$120. “Equipment and supplies” took the values “usually inadequate” or “usually adequate”. Housing provision, transportation allowance, and closeness to hometown took the values “yes” or “no”. The time after which promotion is obtained was set at “one” or “two” years; and training opportunities were either “none” or “two weeks per year” (Table 9.9). However, some important attributes were dropped: medical benefit package, security at workplace, and management system. Because security at workplace needs to involve more actors, and the implementation of a medical benefit package is against the constitution, these attributes were excluded. Management system was very difficult to describe in a uniform way, and was therefore excluded.

Determining Job Attributes for Doctors As in the case for nurses, the job attributes of interest to doctors were derived from a FGD with students, in-depth interviews with doctors

TABLE 9.9 Proposed Job Characteristics of Nurses and Midwives Job attribute Salary

Equipment and supply Housing provision Transportation allowance Time after which promotion is obtained Opportunity for training per year Close to hometown

09 Improving Health

218

Rural facilities

Urban facilities

Same (US$45) US$70 US$95 US$120 Usually inadequate Usually adequate No Yes No Yes 1 year 2 years 2 weeks 4 weeks Yes No

Same (US$45)

Usually inadequate Usually adequate No Yes No 1 year 2 years 0 weeks 2 weeks Yes No

8/24/11, 4:25 PM

Understanding the Rural Health Ser vice in Cambodia

219

working in a rural facility, and a checklist completed by final year medical students. The FGD suggested that the most important attribute in their choice of job location was salary. Students suggested that salary should be increased by 50 per cent, 150 per cent, and 250 per cent. Other job characteristics of interest to them were security in the workplace, adequate equipment, transportation, and accommodation. The in-depth interview with five rural doctors had slightly different results. Doctors would be interested in working in rural areas if the salary was good. They suggested that the salary should be increased to US$150– 300 per month. The other attributes of interest were accommodation, opportunity for career advancement, opportunity for specialty training, transportation, adequate equipment, good management, teamwork, and friendly communities. The short questionnaire filled in by the students and doctors did not suggest attributes other than those discussed above. Again the researchers selected seven attributes and set their levels. Salary increase was suggested to reach a maximum of US$300. As in the case for nurses, equipment and supplies were either “usually inadequate” or “usually adequate”. Housing provision, workplace close to hometown, opportunity for specialty training, and allowing time to work in the private sector took the values “yes” or “no”. Time after which promotion is obtained was set at either “one” or “two” years (Table 9.10). However, some important attributes were dropped, such as medical benefits package, security at workplace, community recognition, and management system. As security at work place needs to involve others, it was excluded. Management and social recognition were found to be important, but implementation in this study was complicated, so these attributes were excluded. For the DCE development, the job attributes listed above were combined into a set of hypothetical jobs between which participants were asked to choose. We used Sloane’s orthogonal array to develop the most efficient study design and select the hypothetical jobs to be used in the DCE. The selected combinations were organized into a series of sixteen choice pairs and respondents selected the best job in each pair. This study required two questionnaire instruments: a DCE questionnaire and a general questionnaire. The general questionnaire comprised questions on respondent characteristics, their educational background, and their

09 Improving Health

219

8/24/11, 4:25 PM

220

Seng Bundeth, Net Neath, P agaiya Nonglak and Sok Sethea TABLE 9.10 Proposed Job Characteristics of Doctors

Job attribute Salary

Equipment and supply Housing provision Workplace location Time after which promotion is obtained Opportunity for specialty training Allowing time to work in private sector

Rural facilities

Urban facilities

Same (US$50) US$100 US$200 US$300 Usually inadequate Usually adequate No Yes Far from hometown Close to hometown 1 years 2 years Yes No Yes No

Same (US$50)

Usually inadequate Usually adequate No Yes Far from hometown Close to hometown 1 years 2 years Yes No Yes No

attitudes in relation to altruism, professionalism, working and living in rural areas, and policy interventions.

Piloting The objective of piloting the general questionnaire was to make sure respondents understood and were able to answer the questions. The objective of the pilot DCE questionnaire was to make sure the respondents understood what they were asked to do and understood the attributes in a sufficiently similar way. After a pilot group filled in the questionnaire, the students were provided with a short questionnaire to reflect on the DCE and questionnaire. Piloting was carried out with thirty medical and thirty nursing students from the University of Health Science and the Technical School of Medical Care in Phnom Penh. Following the pilot, “transportation allowance” in the doctor DCE questionnaire was changed to “workplace location”. The pilot also showed

09 Improving Health

220

8/24/11, 4:25 PM

Understanding the Rural Health Ser vice in Cambodia

221

that the self-administered questionnaire was too long, so factor analysis was used to eliminate some questions.

Data Collection and Analysis Three teams carried out the data collection. Each comprised a team leader and an assistant. In total, six institutions — two medical schools and four nursing colleges — were visited, and 406 final year students were surveyed. The respondents comprised seventy-one students from the University of Health Science, fifty-two students from the Technical School of Medical Care, 102 students from the RTC of Kampot, eighty-six from the RTC of Battambang, forty-one from the RTC of Kompong Cham, and fifty-four from the RTC of Stung Treng. Four enumerators entered the data into Excel spreadsheets. Double data entry was used for high accuracy and reduced time for data cleaning. Data was cleaned and checked using Excel and SPSS. The data was transferred to SPSS and STATA for the analysis. Data from DCE questionnaires were entered, cleaned, and managed in STATA. DCE questionnaire analysis used regression techniques to model respondents’ choices as a function of the different package components. To analyse the labelled choice experiment, job characteristics were dummy-coded and then interacted with the label. Using a conditional logit model, the odd ratios of the independent variables can be interpreted as the effect of the job characteristic on the likelihood of choosing a rural position, or the likelihood of choosing an urban position. In addition, the prediction for rural choice was computed to demonstrate the relative “market share” of each label under different configurations and under the simulated scenarios. For descriptive purposes, we analysed the respondents’ reservation wages against their socio-economic and demographic backgrounds. For example, we can determine how reservation wages differ for students from a rich versus a poor background, or for students from a rural background versus an urban background. We measured the respondents’ medical knowledge, intrinsic motivation, and job aspirations to determine if, and how, these drive reservation wage choices. We also applied multivariate analysis to explain the dependent variable — reservation wages for rural service — with the independent variables described above.

09 Improving Health

221

8/24/11, 4:25 PM

222

Seng Bundeth, Net Neath, P agaiya Nonglak and Sok Sethea

RESULTS Respondent Students’ Characteristics The response rate to the DCE questionnaire by midwifery students was 89 per cent, by medical students 87 per cent, and by nursing students 100 per cent. Therefore, the non-responses should not have distorted the results. Table 9.11 displays the characteristics of the final year students surveyed. Most nursing students (69.6 per cent) were selected from RTCs, while the remainder were from the Technical School of Medical Care. All midwifery students were selected from RTCs and all medical students from the University of Medical Science. Two thirds of nursing students were women; all midwifery students were women; and two thirds of medical students were men. Almost all students are Buddhists. The extent to which religion is important to students varied considerably: 77 per cent of midwifery students but only 21 per

TABLE 9.11 Respondent Students’ Characteristics Characteristics Training Location Phnom Penh Regional Sex Male Female Religion Buddhism Islam Christianity Importance of Religion Unimportant Neither Important Childhood Background Rural Urban

09 Improving Health

222

Nurse N (%)

Midwife N (%)

Medical N (%)

52 (30.4) 119 (69.6)

0 164 (100)

71 (100) 0

57 (33.3) 114 (66.7)

0 164 (100)

47 (66.2) 24 (33.8)

169 (98.8) 2 (1.2) 0

162 (98.8) 0 2 (1.2)

69 (97.2) 1 (1.4) 1 (1.4)

1 (0.6) 46 (26.9) 124 (72.5)

0 38 (23.2) 126 (76.8)

24 (33.8) 32 (45.0) 15 (21.2)

84 (49.1) 87 (50.9)

128 (78.0) 36 (22.0)

17 (23.9) 54 (76.1)

8/24/11, 4:25 PM

Understanding the Rural Health Ser vice in Cambodia

223

cent of medical students said that religion was important in their lives. Nursing students came from rural as well as urban backgrounds, midwifery students were predominantly from rural backgrounds, while medical students were mostly from an urban background. In as much as childhood background is important for job location, this suggests that future cohorts of doctors will continue to prefer urban locations. Table 9.12 presents respondent students’ social activities and experience. Most were not actively involved in social organizations, such as religious organizations, sports, recreational organizations, looking after sick relatives, or savings clubs at the time of the study. Seventy-seven per cent of medical students had not had any experience working in rural facilities. This reflects the evidence that rural hospitals and health centres lack capacity to allow medical students to do their internship there. It seems like a missed opportunity not to give medical students work experience in rural settings during their education. However, the study found that most medical students reported that they had attended a course on community health or primary health care.

TABLE 9.12 Social Activities and Health Experience of Respondent Students Characteristics Membership in social groups as student Non-member Inactive member Active member Have attended a course on comm unity health or primary healthcare Yes No Have spent some time in rural hospitals or health centres Yes No Have work experience in health sector Yes No

09 Improving Health

223

Nurse N (%)

Midwife N (%)

Medical N (%)

110 (64.7) 47 (27.6) 13 (7.6)

114 (69.1) 40 (24.2) 11 (6.7)

57 (80.3) 11 (15.5) 3 (4.2)

155 (90.6) 16 (9.4)

136 (82.9) 28 (17.1)

51 (71.8) 20 (28.2)

131 (76.6) 40 (23.4)

103 (62.8) 61 (37.2)

16 (22.5) 55 (77.5)

9 (5.3) 162 (94.7)

2 (1.2) 162 (98.8)

18 (25.3) 53 (74.7)

8/24/11, 4:25 PM

224

Seng Bundeth, Net Neath, P agaiya Nonglak and Sok Sethea

On the other hand, the majority of nursing and midwifery students said that they had spent some time in hospitals or health centres in the rural areas. The majority of all student groups reported that they had no work experience in the health sector before their study. Table 9.13 shows the education and occupations of students’ parents. Most fathers completed secondary school, while only a small proportion of fathers completed a bachelor or higher degree. More fathers of medical students than fathers of nursing and midwife students had completed post-secondary education. Mother’s education generally seemed to lag one educational cycle behind father’s education. Approximately three quarters of the mothers of midwifery and nursing students had not been educated higher than secondary school level. Mothers of medical students were better educated, with about half achieving higher than secondary level education. A high proportion of students’ parents were self-employed in farming or private business, with the exception of mothers of medical students (49 per cent). Only a minority of parents were employed in the public sector. While acknowledging that self-reported wealth levels are notoriously unreliable, we nevertheless obtained results suggesting that, going by the poorest quintiles only, midwifery students are from the poorest backgrounds, followed by nursing students, then medical students. Table 9.14 reports on how the students financed their studies as well as obligations to the source of finance. The majority of all students had financial support from their parents (93 per cent of nursing students, 96 per cent of midwifery students, and 80 per cent medical students). However, midwifery and nursing students also had financial support from government scholarships. On the other hand, medical students had additional financial support from salaries or study leave allowances. Nearly all midwifery students reported having an obligation to the financing source, whereas only a minority of medical and nursing students did. The same dichotomy was observed with respect to whether students thought they had a say over the locality of their future jobs — midwifery students generally thought they did, in contrast to nursing and medical students. Table 9.15 presents the students’ rating on a six-point Likert scale about their attitudes towards altruism and working and living in rural

09 Improving Health

224

8/24/11, 4:25 PM

Understanding the Rural Health Ser vice in Cambodia

225

TABLE 9.13 Parents’ Background Education Level

Nurse N (%)

Midwife N (%)

Medical N (%)

Father’s education Illiterate No formal education Primary school Secondary school High school Post-high school (Bachelor or higher) Vocational training

5 (2.9) 7 (4.1) 31 (18.1) 59 (34.5) 40 (23.4) 17 (9.9) 12 (7)

13 (7.9) 16 (9.8) 48 (29.3) 37 (22.6) 39 (23.8) 6 (3.7) 4 (2.4)

2 (2.8) 1 (1.4) 6 (8.5) 11 (15.5) 18 (25.4) 30 (42.3) 3 (4.1)

Mother’s education Illiterate No formal education Primary school Secondary school High school Post-high school (Bachelor or higher) Vocational training

57.9 15 (8.8) 12 (7.0) 64 (37.4) 52 (30.4) 18 (10.5) 7 (4.1) 3 (1.8)

46.4 27 (16.5) 16 (9.8) 72 (43.9) 36 (22.0) 9 (5.5) 1 (0.6) 3 (1.8)

4 3 12 16 16 19 1

40.9 (5.6) (4.3) (16.9) (22.5) (22.5) (26.8) (1.4)

Father’s occupation Has work Private sector Public sector Has no job NA

151 84 67 1 19

(88.3) (49.1) (39.2) (0.6) (11.1)

144 103 41 1 19

(88.4) (63.2) (25.2) (0.6) (11.4)

57 30 27 1 13

(80.3) (41.6) (38.0) (1.4) (18.3)

Mother’s occupation Has work Private sector Public sector Has no job NA

147 127 20 2 22

(86.0) (74.3) (11.7) (1.1) (12.9)

146 (89.0) 138 (84.1) 8 (4.9) 1 (0.6) 17 (10.4)

43 35 8 1 27

(60.6) (49.3) (11.3) (1.4) (38.0)

14 28 33 46 50

(8.2) (16.4) (19.3) (26.9) (29.2)

15 14 14 14 14

(21.1) (19.7) (19.7) (19.7) (19.7)

Wealth Index First quintile (poorest) Second quintile Third quintile Fourth quintile Fifth quintile (richest)

09 Improving Health

225

53 39 34 21 17

(32.3) (23.8) (20.7) (12.8) (10.4)

8/24/11, 4:25 PM

226

Seng Bundeth, Net Neath, P agaiya Nonglak and Sok Sethea TABLE 9.14 Financial Sources and Obligations

Variable

Nurse N (%)

Financial sources Parents 159 (93) Salary/study leave 5 (2.9) Loan 4 (2.3) Scholarship (public) 128 (74.9) Scholarship (private) 3 (1.8) Obligations towards financing source Yes 3 (2.3) No 125 (97.7) Ability to express preference for rural or urban areas Yes 168 (98.2) No 3 (1.8) Hope to fulfil the obligations Yes 171 (100) No 0

Midwife N (%) 158 4 21 164 8

Medical N (%)

(96.3) (2.4) (12.8) (100) (4.9)

57 20 4 8 3

(80.3) (28.2) (5.6) (11.3) (4.2)

162 (98.8) 2 (1.2)

3 12

(20) (80)

21 (12.8) 143 (87.2)

70 (98.6) 1 (1.4)

162 (100) 0

71 (100) 0

TABLE 9.15 Students’ Attitudes towards Altruism and Rural Service Attitudes

Nurse Mean (SD)

Midwife Mean (SD)

Medical Mean (SD)

Towards altruism Towards working and living in r ural areas

4.11 (0.54) 3.89 (0.65)

4.20 (0.61) 4.28 (0.59)

4.04 (0.65) 3.50 (0.64)

SD = standard deviation

areas. Each dimension of attitude was probed into with a series of questions. The table reports on the average score for all the questions on a dimension. Altruism levels in students seemed to be generally high, with midwives scoring highest, and medical students lowest. All three groups of students scored working and living in rural areas low. Medical students had the lowest score and midwifery students the highest. This may reflect the background of the students, as most midwifery

09 Improving Health

226

8/24/11, 4:25 PM

Understanding the Rural Health Ser vice in Cambodia

227

students grew up in rural areas, while most medical students grew up in cities (Table 9.11).

Respondents’ Career Plans and Job Expectations Table 9.16 shows the students’ answers regarding their career plans. Most medical students (93 per cent) and nursing students (77 per cent) stated that they would prefer to work in urban areas after graduation, TABLE 9.16 Career Plans Variable Work location preference after graduation Rural Urban Health facility preference after graduation Health post Private individual practitioner/self employed Health centre Private clinics Private poly-clinics Public hospital Work location preference after 5 years working Rural Urban Health facility preference after 5 years working Health post Private individual practitioner/ self employed Health centre Private clinics Private poly-clinics Public hospital Sector preference after 5 years working Public sector Private sector Private not-for profit sector (NGO)

09 Improving Health

227

Nurse N (%)

Midwife N (%)

Medical N (%)

40 (23.4) 131 (76.6)

107 (65.2) 57 (34.8)

5 (7.0) 66 (93.0)

1 1 44 3 6 116

5 0 133 3 0 23

2 3 17 8 1 40

(0.6) (0.6) (25.7) (1.8) (3.5) (67.8)

(3.0) (81.1) (1.8) (14.0)

68 (39.8) 103 (60.2)

96 (58.5) 68 (41.5)

1 (0.6)

1 (0.6)

0 35 3 9 123

(20.5) (1.8) (5.3) (71.9)

123 (71.9) 9 (5.3) 39 (22.8)

1 59 10 0 93

(0.6) (36.0) (6.1) (56.7)

136 (82.9) 15 (9.1) 13 (7.9)

8/24/11, 4:25 PM

(2.8) (4.2) (23.9) (11.3) (1.4) (56.3)

20 (28.2) 51 (71.8)

0 0 8 8 4 51

(11.3) (11.3) (5.6) (71.8)

48 (67.6) 11 (15.5) 12 (16.9)

228

Seng Bundeth, Net Neath, P agaiya Nonglak and Sok Sethea

in contrast to midwifery students (35 per cent prefer urban), who preferred rural areas. Interestingly, nursing and medical students seem to be more willing to work in rural areas as time passes after their graduation. Five years after graduation, significantly fewer medical students (71 per cent) preferred to work in urban areas than directly after graduation (93 per cent), which is statistically significantly different with a standard deviation of 12.8 per cent. The same trend was observed in nursing students (60 per cent from 77 per cent), but this is statistically not significantly different. The shift in preferences for midwives was not great. These results suggest that preferences for job location change over time and differ from one group to another. The majority of medical students and nursing students reported that they would like to work in public hospitals after graduation, however approximately 25 per cent in each group reported that they would prefer to work in health centres. Five years later, these proportions rose slightly for both the nursing group (5 per cent increase for public hospitals) and the medical student group (16 per cent increase), indicating that public hospitals become a more desirable workplace as careers progressed. A marked shift in preferences for facility level was observed in midwives, the majority of whom wished to start a career in health centres (81 per cent) but who wished to move to public hospitals after five years (14 per cent post-graduation vs 57 per cent five years after graduation). The public sector was the most popular sector location for employment, especially for midwives. While it isn’t fully clear why students preferred public hospitals to private hospitals, it was noted during FGDs that jobs at public hospitals were thought to be more secure, with fewer responsibilities, better opportunities to gain experience, and more freedom. Table 9.17 displays the relationships between students’ workplace preferences and their socio-economic characteristics. The table reports associations between variables where preference for location is explained by a range of socio-economic characteristics, as well as attitudes, career preferences, and so on, for each student group. We report only the relevant and statistically significant results. Gender was not an important driver of location preference. The results were not statistically significant for nursing students or for medical students (all midwifery students were women). Father’s employment status was likely to have an influence on the chance the nursing students and medical students choosing to work in

09 Improving Health

228

8/24/11, 4:25 PM

09 Improving Health

Sex Male Female Location Phnom Penh Regional Father’s employment status Employed Unemployed Mother’s employment status Employed Unemployed Childhood Background Rural Urban Social activities as students Participant Non-participant Wealth Index First quintile (poorest) Second quintile Third quintile Fourth quintile Fifth quintile (richest)

Rural

229

1.527

2.275

–0.67

0.175

–0.45

–1.26

112(74.2) 39(25.8) 19(95) 1(5.0) 112(76.2) 35(23.8) 19(79.2) 5(20.8) 59(70.2) 25(29.8) 72(82.8) 15(17.2) 115(77.2) 34(22.8) 16(72.7) 6(27.3) 7(50.0) 7(50.0) 20(71.4) 8(28.6) 23(69.7) 10(30.3) 41(89.1) 5(10.9) 40(80.0) 10(20.0 )

–1.7

t

46(88.5) 6(11.5) 85(71.4) 34(28.6)

39(68.4) 18(31.6) 92(80.7) 22(19.3)

Urban

Nurse

0.209

0.651

0.861

0.502

0.024

0.129

0.092

P-value

14(26.4) 18(46.2) 11(32.4) 8(38.1) 6(35.3)

42(30.9) 15(53.6)

36(28.1) 21(58.3)

50(34.2) 7(38.9)

49(34.0) 8(40.0)

X

X

Urban

39(73.6) 21(53.8) 23(67.6) 13(61.9) 11(64.7)

94(69.1) 13(46.4)

92(71.9) 15(41.7)

96(65.8) 11(61.1)

95(66.0) 12(60.0)

X

X

Rural

x

x

t

0.053

1.488

–3.65

0.274

–0.038

Midwife

0. 957

0.100

0.000

0.785

0.97

x

x

P-value

14(93.3) 14(100) 12(85.7) 13(92.9) 13(92.9)

56(91.8) 10(100)

13(76.5) 53(98.1)

39(88.6) 27(100)

53(91.4) 13(100)

42(89.4) 24(100)

Urban

8/24/11, 4:25 PM

1.096

0.664

–2.545

1.599

0.002

–1.16

t

0.277

0.509

0.014

0.115

0.999

0.251

P-value

continued on next page

1(6.7) 0 2(14.3) 1(7.1) 1(7.1)

5(8.2) 0

4(23.5) 1(1.9)

5(11.4) 0

5(8.6) 0

5(10.6) 0

Rural

Medical

TABLE 9.17 Logistic Regression of Students’ Intention to Work with Regard to Socio-Economic Characteristics [N (%)] Understanding the Rural Health Ser vice in Cambodia

229

Attitude towards altruism ≥ 4.1 ≤ 4.1 ≥ 4.2 ≤ 4.2 Attitude towards professionalism ≥ 4.6 ≤ 4.6 ≥ 4.7 ≤ 4.7 Attitude towards policy interventions ≥ 5.2 ≤ 5.2 ≥ 5.4 ≤ 5.4 Attitude towards working and living in r ural areas ≥ 3.9 ≤ 3.9 ≥ 4.3 ≤ 4.3

09 Improving Health

Rural

230

X

8/24/11, 4:25 PM

x x

72(86.7) 11(13.3) 59(67.0) 29(33.0)

X

60(76.9) 18(23.1) 71(76.3) 22(23.7)

63(81.8) 14(18.2) 68(72.3) 26(27.7) X X

67(79.8) 17(20.2) 64(73.6) 23(26.4) X X

Urban

Nurse

2.79 x

x

0.006 x

x

0.197

x

x

–1.3

0.164

x

x

1.397

0.967

P-value

–0.04

t

39(46.4) 18(22.5)

25(33.3) 32(36.0)

27(33.3) 30(36.1)

22(27.8) 35(41.2)

x

Urban

45(53.6) 62(77.5)

50(66.7) 57(64.0)

54(66.7) 53(63.9)

57(72.2) 50(58.8)

x x

t

x 3. 896

–0.711

x

–1.111

–1.847 x

Midwife Rural

TABLE 9.17 — cont’d

x 0.000

0.478

x

0.268

0.067 x

x

P-value

29(93.5) 38(92.7)

28(87.5) 38(97.4)

28(93.3) 38(92.7)

29(93.5) 37(92.5)

Urban

2(6.5) 3(7.3)

4(12.5) 1(2.6)

2(6.7) 3(7.3)

2(6.5) 3(7.5)

Rural t

0.517

–1.599

–0.181

–0.101

Medical

0.607

0.177

0.857

0.92

P-value

230 Seng Bundeth, Net Neath, P agaiya Nonglak and Sok Sethea

Understanding the Rural Health Ser vice in Cambodia

231

urban rather than rural areas, but not in the case of the midwifery students. Childhood background was one of the strongest predictors of location preference. Students from rural backgrounds were more likely to prefer rural service. The results were statistically significant for midwives and medical students. This confirms findings from studies in other countries. Lastly, a high level of altruism in midwives translates into a preference for rural service. Altruism, here, means motivations to help poor and other people, to assist people in trouble, to take care of other people who need support, urge the government to help the poor, and see help for criminals rather than punishment.

Respondents’ Attitudes towards Government Policy Nurses and midwives were asked to score possible policy interventions to attract health workers into rural areas on a six-point Likert scale. The mean score of their opinion of each intervention is reported in Table 9.18. This survey module was not presented to medical students. All possible interventions were looked upon favourably, each having an average score higher than five. A notable exception was “being able to choose the rural area”, which obtained the lowest score, perhaps suggesting

TABLE 9.18 Attitude towards Measures to Attract Health Workers to Rural Areas Government Policy

Nurse Midwife mean ( SD) mean (SD)

Medical mean (SD)

I want to work in a place where managers support me It is important to work in a place where MoH most needs health w orkers If decent housing was pro vided with posts in rural areas, I would be happ y to go Career advancement for those who wor k in rural should be faster I want to work close to where I g rew up Being able to choose the r ural area is impor tant I want to work in a place which is w ell organized

5.22 (0.97)

5.38 (0.83)

5.28 (0.93)

5.23 (0.88)

5.28 (0.92)

5.54 (1.01)

5.18 (0.86)

5.18 (0.93)

4.97 (1.07)

5.57 (0.71)

5.65 (0.68)

5.01 (1.19)

5.05 (1.00) 4.53 (1.31) 5.58 (0.84)

5.47 (0.77) 4.90 (1.26) 5.70 (0.71)

4.54 (1.16) 5.11 (1.21) 4.38 (1.52)

09 Improving Health

231

8/24/11, 4:25 PM

232

Seng Bundeth, Net Neath, P agaiya Nonglak and Sok Sethea

that health workers see rural areas as relatively homogeneous, which is not the case in reality. The correlation between the rural-urban background of the student and attitude towards choosing location of placement is 0.7043 for medical students, 0.2541 for nursing students, and 0.4366 for midwifery students Nursing students attached the most importance to a well organized workplace and good career progress as factors that could attract them to rural areas. They also viewed working where the Ministry of Health most needs workers as important. Midwifery students reported very similar results, their first and second most valued interventions being identical to those of nurses. However, they put rural hardship allowances in third place.

Respondents’ Job Preferences

Nursing Students Nursing student preferences regarding rural service are reported in Table 9.19. On the one hand, we describe the results of the DCE analysis, that is, the relative contribution of each attribute to attracting nurses into rural posts and urban posts. On the other hand, we report on the relationship between rural choice and socio-economic and other personal characteristics. All job attributes specified in Table 9.10 were found to be significant: the higher the odds ratio, the higher the contribution of the attribute. Increasing the salary from US$45 to US$120 per month had the most important effect on potential rural job uptake for nurses. Somewhat surprisingly, having a workplace close to one’s hometown had the second largest effect. Increasing salaries from US$45 to US$95 and US$70, respectively, were the third and fourth priority attributes. Other important attributes affecting rural service uptake were: adequate equipment, housing provision, transportation allowance, faster promotion, and four weeks’ training per year (compared with two weeks). These results suggest that both financial and non-financial incentives impact rural choice for nurses. Salary increase, transportation allowance, and one year’s work counting as two years for promotion were included only in rural choice. Similarly factors significantly influencing urban choice included working close to one’s hometown, provision of housing, adequate equipment, and opportunity for training. We found that nursing students who grew up in a rural hometown were more likely to choose a rural post. Respondents who had an average

09 Improving Health

232

8/24/11, 4:25 PM

Understanding the Rural Health Ser vice in Cambodia

233

TABLE 9.19 Conditional Logit Regression Results of Final Year Nursing Students Job characteristics

Odds Ratios (SE)

95% CI

P-value

0.049 (0.034)

0.013, 0.187