Social Support and Motherhood: The Natural History of a Research Project 9781447349471

Citation preview

ANN

OAKLEY S O C IA L SU P PO R T A N D MOTHERHOOD T HE NAT URAL H ISTORY O F A R ES EARCH P ROJE CT I NC L UDES NEW I NTRO DU C T IO N

SOCIAL SUPPORT AND MOTHERHOOD Ann Oakley

First published in Great Britain in 1992 by Blackwell Publishers, this edition published in 2018 by Policy Press North American office: University of Bristol Policy Press 1-9 Old Park Hill c/o The University of Chicago Press Bristol BS2 8BB 1427 East 60th Street UK Chicago, IL 60637, USA t: +44 (0)117 954 5940 t: +1 773 702 7700 e: [email protected] f: +1 773-702-9756 www.policypress.co.uk e:[email protected] www.press.uchicago.edu © Ann Oakley 2018 British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data A catalog record for this book has been requested. ISBN 978-1-4473-4616-6 hardcover ISBN 978-1-4473-4949-5 ePub ISBN 978-1-4473-4950-1 Mobi ISBN 978-1-4473-4947-1 ePdf The right of Ann Oakley to be identified as author of this work has been asserted by her in accordance with the 1988 Copyright, Designs and Patents Act. 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 Policy Press. The statements and opinions contained within this publication are solely those of the author and not of The University of Bristol or Policy Press. The University of Bristol and Policy Press disclaim responsibility for any injury to persons or property resulting from any material published in this publication. Policy Press works to counter discrimination on grounds of gender, race, disability, age and sexuality. Cover design by blu inc, Bristol Front cover: image kindly supplied by istock Printed and bound in Great Britain

Contents New Introduction

v

Acknowledgement xii Note on Numbers

xiv

Introduction to the Original Edition

xvi

1

Social Origins

1

2

‘A Friend a Day Keeps the Doctor Away’: Social Support 24 and Health

3

Sickness in Salonica and Other Stories

50

4

Eve in the Garden of Health Research

86

5

A Bite of the Apple

107

6

Who’s Afraid of the Randomized Controlled Trial?

137

7

‘One of Mummy’s Ladies’

169

8

Four Women

227

9

‘Real’ Results

293

10 Women at Risk

337

11 The Poverty of Research

378

12 Models of Knowing and Understanding

411

Appendix I: Study Guidelines

430

Appendix II: Publications from the Social Support and Pregnancy Outcome Study

458

Notes

461

References 466 Index

499

iii

New Introduction1 Reading one’s old texts is a journey into the past. The landscape was different then: the preoccupations were different, and so was the intellectual/social/political context that shaped these preoccupations. But, at the same time, there is usually a strong strand of continuity. The reasons for writing Social Support and Motherhood and for doing the research the book describes remain familiar. The book is what its subtitle says – the natural history of a research project, using ‘natural history’ to mean the systematic study of an organism in its environment. How and why is research conceived, planned, funded, executed, analysed, written up, published, and generally puzzled over? This ambition, to tell the story of how research is actually done, was one of the driving forces behind the writing of the book. (One of its reviewers, interestingly, suggested that title and subtitle should have been reversed (Pilcher 1997).) Not many researchers then (or now) were interested in revealing the often awkward fates of their well-laid plans: the sampling strategies that had no chance of succeeding; the strangely uncooperative ‘subjects’; the mislaid documents; the messy, inscrutable data. They prefer the sanitized versions, which are also the ones that tend to get published. In holding to this fable of sanitary research we deprive ourselves of so many informative stories. The research described in Social Support and Motherhood was undoubtedly the most difficult (and therefore interesting) project I have ever been involved in during the half century and more I have worked in social research. It was difficult mainly because it defied easy description and containment within the framework of either medical or social science research, which was precisely what deepened its attraction for me. What I was intrigued to find out more about was the interface between the social and the biological: the ways in which how we live affects the behaviour of our bodies. The research and the book could, therefore, be seen as products of the sociology of health and illness and the sociology of embodiment, areas of work that were developing rapidly at the time (see, e.g., Shilling 2001; Williams and Bendelow 1999). Sociologies of v

NEW INTRODUCTION

the body demonstrate the falsity of the mind/body person/environment dualism which is both a central feature of Social Support and Motherhood and a main obstacle to the doing of this kind of research. There was also plenty of evidence, added to since, of the importance to health of social ties to others (see, e.g., Feeney and Collins 2015; Holt-Lunstad et al. 2015). As the early chapters of Social Support and Motherhood show, my interest in how social support influences maternity as a biological process developed out of a previous research project on women’s experiences of first-time motherhood (Oakley 1979). A series of interviews with women during pregnancy and afterwards - and, with some of them, attendance at the birth - convinced me that ‘being researched’ could be a positive process (and also that the practice of interviewing was widely misconceived in the research methods textbooks, but that is another story). This notion of supportive interviewing supplemented what was known about the importance of social relations to health, and seemed to provide a good case for exploring in some detail how such an intervention might benefit mothers and their babies. At the time I developed this tricky idea I was working in a research unit which had been set up specifically to look at issues of perinatal health, the National Perinatal Epidemiology Unit in Oxford. In the late 1970s and 1980s Britain’s poor record in the perinatal health field was causing considerable political and public concern and was attracting multiple suggestions that greater investment in medical antenatal care and/ or incentives for women to use it were the answer. Yet the evidence demonstrated that these were unlikely to solve the problem, because much poor health is rooted in stress and poverty (see Chalmers et al. 1980; Oakley et al. 1982). There was, as well, accumulating evidence that social support influences the health of mothers and babies positively – both the support existing in their own social networks and extra support given as a professional or voluntary intervention. I remember being particularly excited by a study carried out in Guatemala in which ‘lay’ companions during labour apparently reduced the incidence of medical problems by two-thirds (Sosa et al. 1980). As Marshall Klaus, a paediatrician who contributed to this and a further similar study (Klaus et al. 1986), said, if social support were a drug, the pharmaceutical companies would surely make a fortune out of it. So there was a definite mood of excitement in the 1980s among medical and social researchers working in the maternity care field about the promise of social support to deliver what intensified and high-technology medical care would probably not be able to do. We conducted systematic reviews of the evidence (see, e.g., Elbourne et al. vi

SOCIAL SUPPORT AND MOTHERHOOD

1989) and held inspirational national and international meetings (see Oakley 2016). The study recounted in Social Support and Motherhood was one of a number carried out around the same time in the USA, Australia, Britain, South Africa and Latin America, all testing the idea that social support is an effective elixir for reproductive problems. Most of these studies were led by medical investigators and used controlled trials to generate similar comparison groups, a design which is much more common in the medical, than the social, research field. Researchers also tended to select physical (‘embodiment’) outcomes such as the occurrence of preterm labour, use of assisted delivery techniques, and infant weight. Birthweight was a particular focus in debates of the time, since it was implicated in that (sadly) enduring conundrum of social class differences in health and illness: low birthweight was, and is, more common among babies born to working-class parents. These are the reasons why the study that occupies centre stage in Social Support and Motherhood settled on the language of pregnancy ‘outcomes’ and put such a stress on birthweight, a measure that would be available for every baby. The book records the adventures of a social scientist engaging with the medical approach to research: the use of an experimental research design, including random allocation; dependence on easily quantifiable measures of ‘outcome’; statistical tests. It is also testimony to some of the challenges of this kind of research: the ethics of random allocation and the behaviour of ethics committees; the limitations of statistical exercises; the restrictions imposed on researchers by the need to ‘pre-specify’ ideas and hypotheses. Re-reading the book now I re-experience the intellectual and emotional struggles of working through some of these difficulties. I am impressed by my obsession with the matter of birthweight – the enormous lengths I went to in an effort to show that our social support intervention did make babies healthier. It was an idea whose imaginative power clearly would not leave me alone. At the time it was published, Social Support and Motherhood was one of only a handful of detailed accounts by social scientists of their encounters with the technology and epistemology of randomized controlled trials. It championed this research design as the only unbiased way of generating comparison groups which are comparable, not only in terms of characteristics already known to be significant, but with respect to those which may well be, but which have not yet been measured. This approach is important in examining the safety, acceptability and effectiveness of all kinds of interventions in people’s lives – not just health care interventions, but those in other fields such as education, criminal justice, social work and social policy. Three decades

vii

NEW INTRODUCTION

on, there has been a good deal of advance in the acceptability of this idea. I went on to take part in three further randomized controlled trials of social interventions: out-of-home day-care for preschool children (Toroyan et al. 2003); peer-led sex education for teenagers (Stephenson et al. 2008); and social support for disadvantaged mothers (Wiggins et al. 2004). These experiences also yielded generic reflections on experimental methodology (Oakley 2000; Oakley et al. 2003). Such connections between projects illustrate the serendipitous logic of a researcher’s career, a logic which often becomes available only with the benefit of hindsight. Social Support and Motherhood was a child of its times. Thus it is sprinkled with archaic references: ‘The Spastics Society’ (an organization concerned with cerebral palsy which changed its name to ‘Scope’ in 1994); ancient technologies such as ‘direct dialing’ and ‘radio-pagers’ (of great importance then to the provision of social support); outdated financial benefit systems (of equal importance); the originally-named Social Science Research Council which, during the efforts to get the Social Support and Pregnancy Outcome project funded, was unhelpfully turned into the Economic and Social Research Council. What has not become archaic are two features of the research world referred to in the book: the immense difficulty of acquiring funding for research that falls in between the discipline-bound agendas of the research councils; and the fragility of research careers, dependent as they are on the endless pursuit of funding to underwrite the researcher’s next employment contract. Most researchers across Europe do not have secure jobs, and this is a major reason for job dissatisfaction (Boman 2017); in the UK more than half the academic work-force is on precarious contracts, with the highest percentages among researchers and early-career staff (UCU 2016). These issues were only beginning to be taken seriously when the funding battles recorded in Social Support and Motherhood took place. My anger with the prevarications, methodological confusions and downright insults of the Social Science/ Economic and Social Research Council with respect to funding requests remains almost palpable in these pages. The book mentions a follow-up study of the mothers and babies who took part in the Social Support and Pregnancy Outcome study which was ongoing when the book was published, and which we did manage to get funded. The findings of this favoured the conclusion that the social support intervention to which mothers had access in pregnancy was continuing to have beneficial effects on the children’s physical health and their mothers’ general wellbeing a year later (Oakley 1992). We also did manage to persuade the Economic and viii

SOCIAL SUPPORT AND MOTHERHOOD

Research Council to fund (as part of a curiously named ‘personal welfare’ initiative) a further survey when the children were seven years old. This demonstrated persisting differences in the health and development of the children and the physical and psychological health of their mothers (Oakley et al. 1996). The Social Support and Pregnancy Outcome study had two other notable research legacies: we were able to use the data to undertake some quite complicated analyses (with expert statistical help) of the relationships between social class, stress and social support (Oakley et al. 1993, 1994a, 1994b); and the experience of doing the study led to the suggestion that using a similar intervention after childbirth might help to mitigate some of the difficulties experienced by socially disadvantaged mothers (Wiggins et al. 2004). Too much research in both the social and medical fields appears as isolated short-term projects, with little connection to other research and not much attempt to examine temporal effects. Mountains of rich data are collected and then never used again. So in these respects the study described in Social Support and Motherhood should be counted a success. In other respects it, and the other studies of this genre, have had disappointingly little impact. My own experiences of presenting its results to different kinds of audiences quickly told me that many doctors found it easy to dismiss because we had not been able to establish a statistically significant effect of social support on birthweight. Midwives, on the other hand, were likely to greet the findings of more satisfaction, less anxiety, fewer Caesarean sections, and better physical health among babies as underlining the importance and effectiveness of the care midwives routinely supply (or would like to be able to, given sufficient resources). The findings of its sister studies, published after Social Support and Motherhood, were variable; most pointed in the same direction, but failed to achieve satisfyingly ‘hard’ outcomes (see, e.g., Villar et al. 1992; Langer et al. 1996). Systematic reviews of research held in the Cochrane Library (the gold standard for such compendia of health care research) show that the Social Support and Pregnancy Outcome study was one of 17 such projects contributing to the evidence that social support does reduce the chances of pregnant women being admitted to hospital and needing to have a Caesarean delivery (Hodnett et al. 2010); and that, according to the data of 26 studies, social support provided during labour does result in a higher likelihood of spontaneous vaginal birth, a reduced need for pain relief, shorter labours, babies in better immediate postnatal health and more positive feelings about the whole childbirth experience (Bohren et al. 2017). As various commentators have noted, the extent to which social

ix

NEW INTRODUCTION

support in pregnancy can be said to ‘work’ or not is bedevilled by the huge difficulty of comparing different approaches to the provision of support: social workers and health educators with a definite mission to teach women how to mother, as used in some studies, are quite likely to be received differently by their target audiences from research midwives who have been trained to provide a listening ear, as in the our study (see Orr 2004, for a discussion). If Social Support and Motherhood was a difficult book to write, it was not an easy one to review, either. The Times Higher decided that the book had multiple uses in debates about feminist methodology, health care and social provision, but that, since none of these tends to produce ‘entertaining writing’, it is easy to see why the author finds it necessary ‘to turn to fiction every now and again’ (Segal 1993). Kate Hunt in the Sociology of Health and Illness delivered a nice understatement: ‘The project itself is of considerable interest and by no means straightforward’ (Hunt 1994:124). The review in Sociology which appreciated the book’s ‘unusual and multilayered nature’ was by a researcher who had herself weighed only 1600 g at birth, and who therefore appreciated the book’s little excursion into the history of birthweight as a cultural artefact (Groves 1994:634). In the British Journal of Sociology Anne Woollett (1994:323) bemoaned the book’s reliance on ‘gripping’ but selective accounts, and asked for more statistical analysis of factors related to birthweight, but still deemed it ‘a fascinating read’; the Journal of Sociology called it ‘innovative’ and ‘poststructuralist’ in its inclusion of ‘the subjectivity of the author’, and especially liked the passages about recalcitrant funders (‘a warning to others and an encouragement to persist’) (Wearing 1994:340-1). The British Medical Journal’s reviewer observed that the book must have been cathartic to write; she recommended reading the book ‘if you are interested in the relation between social stress and health, if you care for women and babies, and if you want to be challenged and stimulated.’ (Coulter 1993). In other words, both medical and social science audiences were able to take from the book messages of relevance to their own disciplines. At a distance of 30 years since the accounts of motherhood printed in the book were gathered, I remain appalled by the circumstances in which many of the women in the Social Support and Pregnancy Outcome study lived. Bad housing, unemployment, inadequate income, fractured relationships, domestic violence, chronic stress – these were the counter-forces with which our efforts to be supportive had to contend. Perhaps it was the height of academic arrogance to assume that extra social support could do anything to help. But the x

SOCIAL SUPPORT AND MOTHERHOOD

tradition of social support and pregnancy studies that flourished in the 1980s was driven by an important policy motive: to highlight how friendship, listening and caring are undervalued aspects of public welfare services. The story of Simone Churchill, told in Chapter 7 of Social Support and Motherhood, was a particularly paralyzing moment for us as researchers: a mother of three, with a history of low birthweight, living in awful damp, cold council housing with a violent partner and not enough money to buy food, getting a lot of help from our research midwife and giving birth to a healthy baby weighing 3530g. Four months later the baby died of pneumonia, having spent most of his brief life struggling with chest infections caused by his family’s housing conditions. The story was headline news on the BBC and in local papers for a time. Sadly such stories are not rare, and, while they may mark the limitations of well-intentioned social support research, they act as salutary reminders of the ways in which reproduction is a social process and must be treated accordingly by any humane society.

xi

Acknowledgements This book has been a long time in the writing, and the research which it describes also has a long (and continuing) history. Many people have had to listen to stories about both the research and the book - different stories over the years, as the problems each has posed have developed, changed, resolved, or given rise, like the phoenix from the ashes, to new flights of fancy. I am grateful to everyone who has provided a listening ear: to everyone, in short, who has given me social support. This includes my friends and family, and colleagues in the three research units where the study was conceived, gestated and born: the National Perinatal Epidemiol­ogy Unit in Oxford, the Thomas Coram Research Unit and the (now) Social Science Research Unit in London. Without the help and support of Sandra Stone, in particular, the project’s own pregnancy outcome would have been a very different story. The role of Lynda Rajan who worked with me on the study from its early stages, and those of the four research midwives are described in this book; they, too, gave unsparingly of their commit­ment and help. Penrose Robertson worked with flair and imagination on the computing side of the study. More recently, Jackie Lee deserves thanks for helping with the preparation of the manuscript, and Gill Bendelow for help with the bibliography. I owe a fundamental intellectual debt to lain Chalmers for drawing my attention to the importance of the experimental method. None of these people are, of course, responsible for the product of the book’s final text. All of us who have been involved in carrying out the various stages of the research owe a particular debt of gratitude to the organizations which had the vision and faith to fund it: to the (then) Social Science Research Council, the Department of Health, the Iolanthe Trust, the Institute of Education, and the (now) Economic and Social Research Council. Those who did not understand or appreciate what we were trying to do may be unhappy to find their names recorded later in this book; but non-funders, like ‘negative’ research findings, are as much a part of the ‘natural history’ of a research project as funders and positive results. xii

ACKNOWLEDGEMENTS

My greatest thanks must go to the women who took up the challenge of taking part in the research - both in the pilot study, and in a fully fledged ‘trial’ of social support. Without their support, there would have been no study to do, and no book to write. Moreover, their commitment to the study demonstrates the possibility of fully informed experimental studies on women’s health issues which this book argues are essential to breaking down the unhelpful divide between different ways of thinking about research, about knowledge, and about the behaviour of bodies and selves in a social, gendered world. The author and publishers gratefully acknowledge permission to repro­duce the paper ‘Social Support and Pregnancy Outcome’, by Ann Oakley, Lynda Rajan and Adrian Grant, which appeared in the British Journal of Obstetrics and Gynaecology, February 1990, Vol. 97, pp. 155162. Thanks for permission to reproduce material are also due to Iain Chalmers for Table 3.1; to Alison Macfarlane for Figure 1.2; and to Raymond Illsley, David Rush and Margaret Stacey for material in Chapter 4. Some of the material in Chapters 6, 10 and 11 has already been published in the form of journal papers; a list of these can be found in Appendix II.

xiii

Note on Numbers This is a book which challenges the divide between quantitative and qualitative research methods. Because of this it uses both. As certain characteristics of quantitative methods are commonly felt to be less easily accessible than those of their qualitative peers, some comments about the way in which numbers are deployed may be required. I have tried to use these so that they are intelligible to those not trained in the science of deciphering them. In the process, however, I will inevitably have sacrificed the refinements needed to satisfy those whose domain they are. For example, in order to aid readability, most of the tables show percentages rather than percentages and figures, with the total figures on which the table is based being given at either the top or the bottom. I have not attempted to standardize the statistical techniques used - indeed, the sprinkling of different techniques illustrates that very disci­plinary and epistemological divide which is the book’s main theme (and which is discussed in more detail in Chapter 9). I apologize for these sins of omission and commission to those who will notice them. Some other comments are needed both about my general use of numbers and for those whose forte is not the domain of numeracy and statistics. Where data from the Social Support and Pregnancy Outcome study are quoted, the total numbers in different tables and for different variables sometimes differ, since the exact numbers depend on the particular data source (interviews, questionnaires, midwives’ records, hospi­tal case-notes, etc.) drawn on. For the same reason, references to other studies may quote different sample numbers in the text and in particular tables. Research is an untidy exercise in which the same full information is rarely available for every research participant. In interviews and when completing questionnaires, for example, not every participant answers every question researchers ask. In the tables, percent­ages have been rounded up to total 100, and they omit people for whom information is missing for that question or variable (because they did not answer the question, or for some other reason).

xiv

NOTES ON NUMBERS

The main statistic used in the tables in this book is the chi-square. This is a test used to evaluate the probability of an apparent difference between groups found in a piece of empirical research occurring by chance. In other words, it provides a rough measure of the extent to which the difference in question can be believed as ‘real’. The ‘p’value shown in the tables is the probability of that particular difference occurring by chance. If the probability is shown as