220 96 4MB
English Pages 290 Year 2012
The Biological Consequences of Socioeconomic Inequalities
The Biological Consequences of Socioeconomic Inequalities
Barbara Wolfe, William Evans, and Teresa E. Seeman, editors
Russell Sage Foundation New York
The Russell Sage Foundation The Russell Sage Foundation, one of the oldest of America’s general purpose foundations, was established in 1907 by Mrs. Margaret Olivia Sage for “the improvement of social and living conditions in the United States.” The Foundation seeks to fulfill this mandate by fostering the development and dissemination of knowledge about the country’s political, social, and economic problems. While the Foundation endeavors to assure the accuracy and objectivity of each book it publishes, the conclusions and interpretations in Russell Sage Foundation publications are those of the authors and not of the Foundation, its Trustees, or its staff. Publication by Russell Sage, therefore, does not imply Foundation endorsement. BOARD OF TRUSTEES Robert E. Denham, Esq., Chair Kenneth D. Brody W. Bowman Cutter III John A. Ferejohn Larry V. Hedges Lawrence F. Katz
Nicholas Lemann Sara S. McLanahan Nancy L. Rosenblum Claude M. Steele Shelley E. Taylor
Richard H. Thaler Eric Wanner Mary C. Waters
Library of Congress Cataloging-in-Publication Data The biological consequences of socioeconomic inequalities / Barbara Wolfe, William Evans, Teresa E. Seeman, editors. pages cm Includes bibliographical references and index. ISBN 978-0-87154-892-4 (pbk. : alk. paper) 1. Poverty—Social aspects. I. Wolfe, Barbara. HC79.P6B5396 2012 362.1´042—dc23 2012032504 Copyright © 2012 by the Russell Sage Foundation. All rights reserved. Printed in the United States of America. 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 written permission of the publisher. Reproduction by the United States Government in whole or in part is permitted for any purpose. The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences-Permanence of Paper for Printed Library Materials. ANSI Z39.48-1992. Text design by Genna Patacsil. RUSSELL SAGE FOUNDATION 112 East 64th Street, New York, New York 10065 10 9 8 7 6 5 4 3 2 1
Contents
List of Tables and Figures
vii
Contributors
xi
Preface Chapter 1
Chapter 2
Chapter 3
Chapter 4
Chapter 5
xiii
T he SES and Health Gradient: A Brief Review of the Literature William Evans, Barbara Wolfe, and Nancy Adler P romise of Biomarkers in Assessing and Predicting Health Arun S. Karlamangla, Tara L. Gruenewald, and Teresa E. Seeman iological Imprints of Social Status: B Socioeconomic Gradients in Biological Markers of Disease Risk Tara L. Gruenewald, Teresa E. Seeman, Arun S. Karlamangla, Elliot Friedman, and William Evans issecting Pathways for Socioeconomic D Gradients in Childhood Asthma Edith Chen, Hannah M.C. Schreier, and Meanne Chan Cardiovascular Consequences of Income Change David H. Rehkopf, William H. Dow, Tara L. Gruenewald, Arun S. Karlamangla, Catarina Kiefe, and Teresa E. Seeman
v
1
38
63
103
126
vi Contents Chapter 6
Cognitive Neuroscience and Disparities in Socioeconomic Status Jamie Hanson and Daniel A. Hackman
Chapter 7
Brain Development and Poverty: A First Look Jamie Hanson, Nicole Hair, Amitabh Chandra, Ed Moss, Jay Bhattacharya, Seth D. Pollak, and Barbara Wolfe
Chapter 8
eversing the Impact of Disparities in R Socioeconomic Status over the Life Course on Cognitive and Brain Aging Michelle C. Carlson, Christopher L. Seplaki, and Teresa E. Seeman
Chapter 9
158 187
215
Conclusions William Evans, Teresa E. Seeman, and Barbara Wolfe
248
Index
263
List of Tables and Figures
Table 1.1 Table 2.1 Table 2.2 Table 5.1 Table 5.2 Table 5.3 Table 5.4 Table 5.5 Table 5.6 Table 7.1 Table 7.2 Table 7.3 Table 7.4 Figure P.1 Figure 1.1 Figure 1.2 Figure 1.3 Figure 1.4
Disparities in Health by Socioeconomic Status Major Physiological Systems and Corresponding Biomarkers Clinical High-Risk Criteria for Commonly Used Biomarkers Characteristics of CARDIA Participants Socioeconomic Indicators, 1992 to 1993 Log Income, 1992 to 2006 Random-Effect, Fixed-Effect, and Long-Difference Models Models Run with Alternative Categorizations of Outcome Variables, Odds Ratios Models Run with Alternative Categorizations of Income Exposure Variables Demographic Summary Attrition by Income Summary Statistics for Brain Regions of Interest Model Estimates for Association Between SES Measures and Brain Regions of Interest Interdisciplinary Schematic The Income-Health Relationship Marginal Effects on Income Dummy Variables, Children, Fair or Poor Health Marginal Effects on Income Dummy Variables, Children, School Absence Ten Days or Longer Marginal Effects on Income Dummy Variables, Children, Limitation on Activity
vii
16 41 51 136 138 144 148 151 152 193 194 194 207 xv 2 5 5 6
viii List of Tables and Figures Figure 1.5 Figure 1.6 Figure 1.7 Figure 1.8 Figure 1.9 Figure 1.10 Figure 1.11 Figure 1.12 Figure 1.13 Figure 1.14 Figure 1.15 Figure 1.16 Figure 1.17 Figure 1.18 Figure 1.19 Figure 1.20 Figure 1.21 Figure 1.22 Figure 2.1 Figure 2.2 Figure 2.3 Figure 3.1 Figure 3.2
Marginal Effects on Income Dummy Variables, Children, Hospital Stay Marginal Effects on Income Dummy Variables, Children, Emergency Room Visit Marginal Effects on Income Dummy Variables, Children, Injury or Poisoning Marginal Effects on Income Dummy Variables, Children, Asthma Marginal Effects on Income Dummy Variables, Adults, Fair or Poor Health Marginal Effects on Income Dummy Variables, Adults, Mental Health Days Marginal Effects on Income Dummy Variables, Adults, Bad Physical Health Days Marginal Effects on Income Dummy Variables, Adults, Current Smoker Marginal Effects on Income Dummy Variables, Adults, Obese Marginal Effects on Income Dummy Variables, Adults, Overweight Marginal Effects on Income Dummy Variables, Adults, No Exercise Marginal Effects on Income Dummy Variables, Adults, Ages Eighteen to Seventy-Four, Limited Fruits and Vegetables Odds Ratio for Income Variables, Adults Marginal Effects of Household Income, Australian Adults, Fair or Poor Health Marginal Effects of Household Income, Australian Adults, Psychological Distress Risk Marginal Effects of Household Income, Australian Adults, Long-Term Health Condition Odds Ratio of General Physical Health Measures, Europe Odds Ratio of Self-Perceived Health, Europe Adjusted Seven-Year All-Cause Mortality Odds All-Cause Mortality Rate in Those Younger than Sixty-Five Years Old All-Cause Mortality Rate in Those Sixty-Five Years Old or Older Conceptual Model of SES and Health Links AUC Cortisol Area as Function of Quintile of SEP
6 7 7 8 10 10 11 11 12 12 13 13 15 17 17 18 18 19 50 52 53 65 70
List of Tables and Figures ix Figure 3.3 Figure 3.4 Figure 3.5 Figure 3.6 Figure 3.7 Figure 3.8 Figure 3.9 Figure 3.10 Figure 3.11 Figure 3.12 Figure 4.1 Figure 4.2 Figure 5.1 Figure 5.2 Figure 5.3 Figure 5.4 Figure 5.5 Figure 5.6 Figure 6.1 Figure 6.2 Figure 7.1 Figure 7.2 Figure 7.3 Figure 7.4 Figure 7.5 Figure 7.6 Figure 7.7 Figure 7.8 Figure 7.9 Figure 7.10
Mean Overnight Norepinephrine Levels Marginal Effects, C-Reactive Protein Marginal Effects, Glycated Hemoglobin Levels Marginal Effects, Total Cholesterol Marginal Effects, High-Density Lipoprotein Marginal Effects, Waist-to-Hip Ratio Marginal Effects, Systolic Blood Pressure Marginal Effects, Diastolic Blood Pressure Marginal Effects, Resting Pulse Regression Coefficients, Allostatic Load SES and Inflammatory Responses Model: SES Effects and Clinical Health Outcomes Metabolic Syndrome Score versus BMI Income Change Between Exams, Men Income Change Between Exams, Women Income Change Between Exams, Men Without Marital Change Income Change Between Exams, Women Without Marital Change First Difference Nonlinear Models, 1992 to 2005 Four Basic Lobes of the Brain Five Neurocognitive Systems of Interest Axial Brain Slice Sagittal Brain Slice Association Between Hippocampal Volume and Family Income Association Between Superior Prefrontal Cortex Volume and Family Income Association Between Ventral Medial Prefrontal Volume and Family Income Association Between Cerebellar Gray Matter Volume and Family Income Association Between Total Cerebellar Volume and Family Income Association Between Occipital Gray Matter Volume and Family Income Ventral Medial Prefrontal Cortex Superior Prefrontal Cortex
73 84 84 85 85 86 86 87 87 88 111 117 137 140 141 142 143 146 160 162 195 196 197 198 199 201 202 203 205 206
Contributors
Barbara Wolfe is professor of economics, population health sciences, and public affairs and faculty affiliate at the Institute for Research on Poverty at the University of Wisconsin–Madison. William Evans is Keough-Hesburgh Professor of Economics at the University of Notre Dame. Teresa E. Seeman is professor of medicine and epidemiology at the University of California, Los Angeles.
Nancy Adler is Lisa and John Pritzker Professor of Psychology in the Departments of Psychiatry and Pediatrics at the University of California, San Francisco. Jay Bhattacharya is associate professor of medicine at the Stanford University School of Medicine. Michelle C. Carlson is associate professor of mental health and associate director of the Center on Aging and Health at Johns Hopkins Bloomberg School of Public Health. Meanne Chan is Ph.D. candidate in psychology at Northwestern University. Amitabh Chandra is professor of public policy at the Harvard Kennedy School of Government.
xi
xii Contributors Edith Chen is professor of psychology at the Institute for Policy Research at Northwestern University. William H. Dow is Henry J. Kaiser Professor of Health Economics at the University of California, Berkeley School of Public Health, and associate director of the Berkeley Population Center. Elliot Friedman is assistant professor of human development and family studies at Purdue University. Tara L. Gruenewald is assistant professor of gerontology in the Davis School of Gerontology at the University of Southern California. Daniel A. Hackman is Ph.D. candidate in psychology at the University of Pennsylvania. Nicole Hair is Ph.D. candidate in economics at the University of Wisconsin–Madison. Jamie Hanson is Ph.D. candidate in psychology at the University of Wisconsin–Madison. Arun S. Karlamangla is professor of medicine in the David Geffen School of Medicine at the University of California, Los Angeles. Catarina Kiefe is professor of quantitative health sciences at the University of Massachusetts Medical School. Ed Moss is director of the Developmental Neuropsychology Service in the Department of Pediatric Psychology of the Children’s Seashore House (Department of Pediatrics, University of Pennsylvania School of Medicine). Seth D. Pollak is College of Letters and Science Distinguished Professor of Psychology at the University of Wisconsin–Madison. David H. Rehkopf is assistant professor of medicine at the Stanford University School of Medicine. Hannah M. C. Schreier is Ph.D. candidate in psychology at the University of British Columbia. Christopher L. Seplaki is assistant professor of community and preventative medicine at the University of Rochester Medical Center.
Preface
In 2001, the Russell Sage Foundation and the Carnegie Corporation of New York began a research program designed to examine the implications of rising economic inequality in the United States. The initial program funded interdisciplinary working groups at a number of universities, representing the fields of economics, sociology, political science, education, public health, and public policy. In the early phases of the program, scholars were asked to consider three broad questions about social inequality. First, are those groups that have increasingly been left behind economically also those that have lost ground in other ways that limit their full participation in society? This activity produced detailed descriptive research documenting changes in social inequality across a wide spectrum of outcomes including the financial well-being of families, the quality and quantity of time parents spend with children, participation in the democratic process, the quality of education from preschool to college, interactions with the criminal justice system, and the quality of health care and health outcomes. A second group of papers examined the consequences of rising inequality within a particular domain and sought to consider whether rising social inequality has impacts that tend to work alongside economic inequality. The third question asked why social and economic disparities exist, why they have changed so dramatically over the past forty years, and whether any policy levers are available to potentially reduce these disparities. After this initial phase of research, the Russell Sage Foundation continued its work on social inequality by turning to in-depth examinations of key institutions in the United States. It is out of this second phase of the foundation’s initiative that the project discussed here emerged. The third question, about why inequality exists and may have grown in scope, may be the most difficult to answer. The question we explore here is one aspect of the question: the nature of the links between socioeconomic status (SES) and health. The literature is vast and the results to date are rather stark.
xiii
xiv Preface Thousands of studies across a variety of disciplines have documented a gradient between greater SES and better health. This pattern holds across all ages and all countries for which such studies have been conducted; for virtually all measures of health such as mortality, morbidity, measures of general health, health habits, and functional limitations; and for a variety of measures of SES such as income, wealth, occupation, and education. Despite this extensive literature, scholars have only begun to scratch the surface on some of the important questions concerning mechanisms, causal relationships, and possible policy responses. Why do we observe a relationship between SES and health? What does higher income or greater wealth do for families that allows them to produce better health outcomes? Does greater income allow one to purchase health insurance, better health care, or better neighborhoods? Does low income or having a low-skill job produce more stress and hence the associated poor health consequences? Alternatively, do other explanations for this relationship exist? For example, might it be that low income is caused by poor health and not the other way around? Without answers to these important questions, it is hard to fashion appropriate policy responses to the question of whether income transfers reduce health disparities. The purpose of this book is to suggest a possible research path that can potentially provide answers to these questions. Specifically, we outline a research program surrounding the biology of disadvantage that attempts to get under the skin and quantify if and how material deprivation affects basic physiological processes. To date, much of the work that has attempted to answer these questions has been produced by scholars working within their field of study. Physicians, epidemiologists, psychologists, and public health officials have attempted to identify the antecedents of the gradient by examining possible links between factors such as stress and body physiology. By contrast, economists and other social scientists have studied the links between SES and a variety of outcomes such as health and how health influences labor market activity and SES. Although one of the primary lessons of economics is the gains from specialization, it was clear to us from the start that for research to begin to formulate answers to the difficult why questions, a truly interdisciplinary group open to the ideas of other disciplines would have to coalesce behind the question. In that spirit, to better understand the set of links between SES and health, we felt that it was imperative to unite the focus of biological studies that characterize stress and the environment to those of social scientists studying income health and wealth; in the process, we have brought together an interdisciplinary team that has truly worked together on joint projects. Our ability to pursue this interdisciplinary question was enhanced by a critical series of meetings in which we could come to better
Preface xv Figure P.1 Interdisciplinary Schematic
Biological processes
Socioeconomic status
Health
Source: Authors’ compilation.
understand the various perspectives and approaches and begin to build relationships and mutual respect that could be the basis of the research as reported in the chapters that follow. The basic set of links we try to understand in this volume and a broad approximation of what different disciplines have examined is captured in figure P.1. Work from a variety of authors has established a clear link between SES and health. Unfortunately, as we outline in chapter 1, assigning a clear causal pathway is difficult. Envisioning how SES alters health is easy, but poor health alters earnings capacity, so the direction of causation can go either way. Likewise, the intervening variable we explore in detail in this book is how SES affects basic biological processes. Poor social standing may influence health indirectly by heightening stress, impairing brain growth, triggering asthma, and slowing cognitive processes. From a research standpoint, that causal pathways can potentially move in both directions complicates the exercise. SES may affect health, but poor health also reduces wealth by requiring out-of-pocket spending and reducing work. Likewise, poor health in one area may directly influence biological processes, which in turn reduces earnings capacity and affects SES status. This book has three main purposes. First, we hope to provide a resource for scholars interested in questions relating to SES and health who may not have expertise in either more socioeconomic dimensions or more biological dimensions. To that end, the first three chapters provide a guide to the current state of the literatures in the respective fields.
xvi Preface The current literature in social sciences surrounding the SES-health gradient is outlined in chapter 1 by Barbara Wolfe, William Evans, and Nancy Adler. This literature is vast, and reviewing the entire field would require hundreds of pages. As a result, the authors focus on one particular relationship—between income and health—as a conical example. This focus is strategic. First, income helps illustrate the basic statistical relation between the two variables that is indicative of other measures of SES. Second, all of the possible basic correlations between SES and health signal both reverse causation and omitted variables bias, and it is quite easy to outline how these problems arise in the case of the income-health relationship. For example, poor health reduces work and hence income, so some of the relationship may simply signal reverse causation. Third, it is arguably easier to outline the pathways through which income influences health compared to other measures of SES such as occupation or education. Fourth, among the various measures of SES, if income causally affects health, it is easy to imagine various policy levers that can be employed to increase the income levels for those most at risk. Policies such as lowering tax rates, raising the minimum wage, increasing Earned Income Tax Credit payments, raising welfare payments, and providing subsidized day care are but a few legislative remedies that can be employed, with varying degrees of success, to directly raise the incomes of citizens. In contrast, it is less obvious how one can alter the assignment of workers to occupations or, unfortunately, how governments can raise the education levels of recipients. Complementing this chapter by social scientists, chapters 2 and 3 provide data from the medical sciences and outline how biomarkers are collected, what they measure, and what future health events they predict; outline some potential pathways through which SES status can potentially affect the physiological operation of the body; and provide some evidence of the SES gradients for these biomarkers. In chapter 2, Arun Karlamangla, Tara Gruenewald, and Teresa Seeman outline the promise of biomarkers in assessing and predicting health. They describe the major physiological systems, how they are measured by biomarkers, and the predictive capacity of particular tests for health endpoints such as disease incidence or death. They discuss the variation in quality of various biomarkers and the limitations associated with using particular values. The chapter ends with a discussion noting that more information can be obtained about the underlying health of an individual by aggregating results from multiple tests. In chapter 3, Tara Gruenewald, Teresa Seeman, Arun Karlamangla, Elliot Friedman, and William Evans outline the biological imprint of deprivation on health by examining the SES gradients in biomarkers of disease
Preface xvii risk. They note that the premise underlying such efforts is that everyday experiences such as behaviors, stress, and cognitive-emotional processes lead to variation in biological functioning and subsequent disease risk. This chapter therefore serves as our transition from the measurement of biomarkers to how they can be used in the analysis of the SES gradient in health. The chapter begins by “getting under the skin” and outlining how variation in SES can potentially impact the physiology of body operation. Particular focus is paid to two regulatory systems: the hypothalamic-pituitary-adrenal (HPA) axis, which regulates many body processes such as energy consumption, and the sympathetic nervous system (SNS), which controls the body’s fight-or-flight response to stress. The authors outline how stress induced by low SES may alter biomarkers associated with these two systems and how these metrics are measured. The authors also outline the SES gradient in metabolic and immune system disorders as measured by biomarkers, and like the previous chapter, the authors outline how combining information about multiple systems provides more information than would any single marker. To close the chapter, the authors provide data on the income gradient for eight of the most frequently used biomarkers (C-reactive protein, glycated hemoglobin, diastolic and systolic blood pressure, total cholesterol and high-density lipoproteins [HDL], hip-to-waist ratio, and resting pulse rate), plus a composite measure of health referred to as allostatic load that is derived from these eight measures. The second purpose of this book is to illustrate how teams from the social and biological sciences can combine the benefits of biomarkers with the methods used by social scientists to isolate causal relationships in nonexperimental settings. These strategies are illustrated in chapters 4 through 8 of this book. This section begins with a chapter by Edith Chen, Hanna M. C. Schreier, and Meanne Chan on the link between SES and asthma. Asthma is the most common chronic condition in childhood, affecting 12.5 percent of children during their lifetimes. Asthma is also steeply correlated with SES. When deciphering why this relationship exists, the authors advocate the importance of understanding the pathophysiology of the disease. Using this approach, the authors illustrate how SES can be linked to the asthma inflammatory process at the cell level. Initially, the authors describe in detail the process that generates the inflammation necessary to produce an asthma attack. Next, they explore at the cell level whether the biological markers necessary for asthma attacks are more prevalent among low-SES families. For example, as the chapter outlines, one of the steps in the biology of asthma attacks is the activation of cells called eosinophils, which bring on edema, muscle constriction, and mucus production in the airways. The authors illustrate that these cells are more
xviii Preface prevalent in low-SES than in higher-SES children. Work by the authors is some of the first to establish a link between SES and disease-specific biological markers in patient populations. The authors then report on a series of studies that illustrate how stress at the individual, family, and neighborhood levels can affect biological markers at the cell level, increasing the incidence of asthma. In chapter 5, using data from a variety of sources, David Rehkopf, William Dow, Tara Gruenewald, Arun Karlamangla, Catarina Kiefe, and Teresa Seeman examine the impact of income changes on metabolic syndrome. A primary concern among public health officials is the rising obesity epidemic, and, as many authors have established, obesity rates are much higher in lower-income groups. A key question is whether low income is a contributing factor to the crisis or whether there is something endemic about certain groups that manifests as both low income and greater rates of obesity. To get closer to identifying whether this represents a causal relationship, the authors exploit longitudinal data from CARDIA—the Coronary Artery Risk Development in Young Adults study—a dataset of roughly 5,100 young adults as they aged over a twenty-year period. The key outcome is a composite measure that identifies whether a respondent had three or more risky levels of biomarkers associated with metabolic syndrome: hip-to-waist ratios, body mass index, triglycerides, HDL, blood pressure, or blood glucose levels. The authors begin the chapter by reproducing the basic cross-sectional results showing higher than average rates of metabolic syndrome among the low-income respondents. The panel nature of the data allows one to hold constant the time-invariant characteristics that would typically contaminate a cross-sectional model and instead examine whether metabolic syndrome changes over time for an individual as his or her income varies. Chapter 4 showed an important role for income in childhood asthma. By contrast, the authors in chapter 5 find little if any role for time-series changes in individual-level income explaining the changes in metabolic syndrome over time. Chapter 6 expands the use of biomarkers in a new and exciting way by considering cognitive neuroscience and SES disparities. The chapter, written by Jamie Hanson and Daniel Hackman, begins by outlining the basics of brain organization. The brain comprises two hemispheres, four lobes, and three substances. The chapter outlines the purposes and function of all these divisions. The authors then review how advances in neuroimaging have greatly facilitated the measurement of the volume of various components of the brain. Next, the authors outline how the brain develops during childhood and describe some ways that material deprivation during these years might manifest itself in brain development and what current evidence suggests about these hypotheses. This chapter thus serves
Preface xix as a primer on the use of a new biomarker as well as a literature review to help future researchers understand the current landscape. In chapter 7, armed with these new tools, Jamie Hanson, Nicole Hair, Amitabh Chandra, Ed Moss, Jay Bhattacharya, Seth Pollak, and Barbara Wolfe use an exciting new longitudinal dataset to provide some preliminary evidence on the impact of poverty on brain development. The data for this project are taken from the National Institutes of Health MRI (magnetic resonance imaging) study of normal brain development. This public access database contains a large sample of MRI brain scans for a sample of children age four to eighteen. The data are longitudinal in nature in that children are scanned every two years for up to seven years. Exploiting the longitudinal nature of the data, the authors find evidence that material deprivation is associated with smaller sizes of the hippocampus, prefrontal cortex, and cerebellum, areas of the brain that have previously been associated with the quality of environmental inputs and stress. Chapter 8, by Michelle Carlson, Christopher Seplaki, and Teresa Seeman, extends the focus from chapter 7 into later life by examining the role of SES on risks for cognitive decline among older adults. The chapter begins by summarizing the literature about how early-life and environmental conditions potentially affect cognitive decline and brain functioning later in life. The authors note the potential role of cognitive reserve in explaining the role of SES in later life brain function. According to this hypothesis, the greater the store of cognitive ability generated in reserve by environment and life experiences, the greater the insult necessary to negatively impact cognitive ability later in life. Despite large disparities in cognitive function among the elderly based on early-life experiences, the authors note that the areas affected most by early-life events, the pre- frontal-limbic circuits, remain plastic and responsive to the environment in late-life development. The chapter moves past the question of whether SES alters health outcomes but considers whether disparities can be altered by tailored interventions. The authors report some exciting positive results from interventions designed to assist low-SES elderly in improving cognitive function through basic interventions such as increased physical activity and greater social engagement. Overall, this volume seeks both to provide necessary background for those who may be less familiar with socioeconomic and biological brain development and function and to illustrate, through selected examples, the ways in which incorporating these biological and neurocognitive processes allows for enhanced interdisciplinary work to elucidate the processes through which SES gets under the skin and leads to the consistent and widespread SES health disparities seen worldwide. The remaining goal of this project is to encourage joint research projects that will bring
xx Preface the theory and methods of economists together with complementary knowledge and methods of biological scientists, along with available data and measures, to greatly improve research and understanding of the tie between income and income inequality and changes in health. Only time will tell whether we have been successful. This work could not have been done without the willingness of the Russell Sage Foundation to support a series of meetings where a mix of social and biological scientists could get to know one another’s work and perspectives. And it is unlikely to have taken place without the foundation’s earlier activities in the area of inequality. The earlier Russell Sage project on inequality was the direct stimulus to this project, and we thank both the president of the foundation, Eric Wanner, and the board for their initiatives in this area. We also thank the early participants in our project, whose encouragement and enthusiasm contributed to the interest of participants in this work. These include David Cutler, Stephanie Robert, and Thomas Boyce.
Chapter 1 The SES and Health Gradient: A Brief Review of the Literature William Evans, Barbara Wolfe, and Nancy Adler
Numerous studies have documented a positive gradient between socioeconomic status (SES) and health—the better off individuals are, the better their health. The positive relationship between good health and higher SES is generally accepted, but until we understand both the nature of the relationship and what explains the link, policy may be ineffective in substantially reducing disparities across groups. The graded association between various indicators of SES and health holds across all ages and for all countries in which it has been studied. The gradient emerges in relation to a range of health indicators, including mortality, morbidity, measures of general health, health habits, and functional limitations. These health indicators are associated with a range of alternative measures of SES, such as income, wealth, occupation, and education. These indicators of SES are in turn related to one another, but each has unique aspects. Each provides different material and social resources. In addition, they differ in terms of their potential role in serving as a cause of health and as an outcome of health status. For example, income may fluctuate as a result of poor health, while simultaneously poor health may be the result of financial constraints. In contrast, education is generally established relatively early in life and is less likely to be subject to changes in health status. Figure 1.1 illustrates the general shape of the relationship between income and health when compared across individuals or groups or countries. The horizontal axis measures income, the vertical axis measures a
1
2 Biological Consequences of Socioeconomic Inequalities Figure 1.1 The Income-Health Relationship Health (H) Hb* H = f(Y) Hb Ha* Ha
Ya
Ya + 100
Yb
Yb + 100
Income (Y)
Source: Authors’ figure.
positive health outcome such as life expectancy, and the curve represents the empirical relationship between the two variables. Although higher income is associated with better health at all levels, the steepest association is at the bottom of the income distribution. As a result, and as shown in the figure, the relative gain in a given health outcome as the result of adding $100 to a person’s income (Ya to Ya + 100 versus Yb to Yb + 100) is greater for those whose incomes are lowest. This graph clearly portrays that the marginal benefit of additional income declines as income rises. Adding an extra $100 to income at Ya improves outcomes Ha to Ha*, but that same $100 increment at Yb improves outcomes only marginally from Hb to Hb*. The income-health gradient portrayed in figure 1.1 is widely interpreted to indicate that income causally influences health. At the same time, poor health can reduce a person’s productivity and hence income and wealth. These two scenarios lead to the question of whether low income leads to poor health or whether poor health leads to low income. Given that both may be true, the more appropriate question is the extent to which income affects health and the extent to which health affects income.
The SES and Health Gradient 3 A third scenario is also possible: a correlation between SES and health may not simply represent the impact of a given aspect of SES on health or the impact of health on SES but also reflect an underlying common determinant of both health and SES. For example, factors such as motivation or genetics could account for the presence of both low income and poor health. To date, these alternatives remain as active hypotheses of what lies behind the income-health gradient. In this chapter, we attempt to set the groundwork for the volume by reviewing the existing evidence on the SES-health relationship. This includes discussions of the basic descriptive models that may enable us to better test the nature of the gradient, two of the more influential streams of empirical literature attempting to understand the gradient, and finally some assessment of which alternative approaches may allow us to make progress in increasing our understanding of the SES gradient in health.
Descriptive Evidence Literally thousands of papers document the SES-health gradient. These studies use different samples, outcomes, measures of SES, and statistical methods and cover very different periods. Rather than try to summarize this vast literature, we present a number of samples and similar models to document the persistence of the SES-health link and its changing nature over time. Although the gradient occurs in relation to health, illness, and mortality at every stage of life, the strength of the gradient varies at different ages. The gaps in health are greatest in mid- to late adulthood, when rates of disease begin to rise and more variation is linked to socioeconomic factors. The gap narrows after age sixty-five, possibly because of differential survival and the buffering effects of safety net programs, including Medicare, that are available starting at age sixty-five. Despite the somewhat weaker gradient in childhood, this period is important to examine for two reasons. First, the SES-health gradient for children is less susceptible to reverse causation concerns because it is less likely that poor health is “causing” low income.1 Second, although the magnitude of SES differences is greater in adulthood, previous work has provided evidence that the origin of the SES-health gradient among adults has its roots in childhood (Case and Paxson 2008; Singh-Manoux et al. 2004). To illustrate the breadth of the income gradient for children, we use data from the 2001 through 2003 National Health Interview Surveys (NHIS), an annual survey designed to measure the health status of the U.S. noninstitutionalized population. From the NHIS, we select a popula-
4 Biological Consequences of Socioeconomic Inequalities tion of school-age children, age six through seventeen, giving us 39,357 observations.2 We focus on seven measures of child health. All of the measures are characterized as dummy variables, in which the variable equals one if the child has the condition and zero otherwise, and all are constructed such that the realization of the outcome is a measure of poor health. These outcomes are whether the child has fair or poor health (on a 5-point scale) as reported by the adult in the house; has missed ten days or more of school in the past year due to injury or illness; has a physical, mental, or emotional condition that limits activity; had a hospital stay in the previous twelve months; had an emergency room visit in the previous twelve months; had an injury or poisoning in the past year; and has ever been diagnosed with asthma. For each outcome, we run a simple probit model controlling for a variety of characteristics.3 The key covariate in these models is a measure of family income, which is reported by an adult within the household. The variable is categorical, and we break it into six broad income categories ( 3, ≤ 4
> 4, ≤ 5
Poverty-to-Income Ratio versus Reference Group (>5) Source: Authors’ calculations based the Third National Health and Nutrition Examination Survey (NHANES III; Centers for Disease Control and Prevention n.d.). Note: Risky level of biomarker outcomes, NHANES III, adults age twenty to seventy-four. Error bars represent 95 percent confidence intervals.
Figure 3.7 Marginal Effects, High-Density Lipoprotein
Marginal Effect from Probit
0.2
Fraction Answering Yes = 0.237
0.1
0.0
≤1 −0.1
> 1, ≤ 2
> 2, ≤ 3
> 3, ≤ 4
> 4, ≤ 5
Poverty-to-Income Ratio versus Reference Group (>5)
Source: Authors’ calculations based on the Third National Health and Nutrition Examination Survey (NHANES III; Centers for Disease Control and Prevention n.d.). Note: Risky level of biomarker outcomes, NHANES III, adults age twenty to seventy-four. Error bars represent 95 percent confidence intervals.
Figure 3.8 Marginal Effects, Waist-to-Hip Ratio
Marginal Effect from Probit
0.25
Fraction Answering Yes = 0.634
0.20 0.15 0.10 0.05 0.00
≤1
> 1, ≤ 2
> 2, ≤ 3
> 3, ≤ 4
> 4, ≤ 5
Poverty-to-Income Ratio versus Reference Group (>5) Source: Authors’ calculations based on the Third National Health and Nutrition Examination Survey (NHANES III; Centers for Disease Control and Prevention n.d.). Note: Risky level of biomarker outcomes, NHANES III, adults age twenty to seventy-four. Error bars represent 95 percent confidence intervals.
Figure 3.9 Marginal Effects, Systolic Blood Pressure
Marginal Effect from Probit
0.10
Fraction Answering Yes = 0.125
0.05
0.00 ≤1
> 1, ≤ 2
> 2, ≤ 3
> 3, ≤ 4
> 4, ≤ 5
Poverty-to-Income Ratio versus Reference Group (>5) −0.05 Source: Authors’ calculations based on the Third National Health and Nutrition Examination Survey (NHANES III; Centers for Disease Control and Prevention n.d.). Note: Risky level of biomarker outcomes, NHANES III, adults age twenty to seventy-four. Error bars represent 95 percent confidence intervals.
Figure 3.10 Marginal Effects, Diastolic Blood Pressure 0.10
Fraction Answering Yes = 0.064
Marginal Effect from Probit
0.08 0.06 0.04 0.02 0.00 −0.02 −0.04
≤1
> 1, ≤ 2
> 2, ≤ 3
> 3, ≤ 4
> 4, ≤ 5
Poverty-to-Income Ratio versus Reference Group (>5) Source: Authors’ calculations based on the Third National Health and Nutrition Examination Survey (NHANES III; Centers for Disease Control and Prevention n.d.). Note: Risky level of biomarker outcomes, NHANES III, adults age twenty to seventy-four. Error bars represent 95 percent confidence intervals.
Figure 3.11 Marginal Effects, Resting Pulse
Marginal Effect from Probit
0.10
Fraction Answering Yes = 0.066
0.05
0.00 ≤1
> 1, ≤ 2
> 2, ≤ 3
> 3, ≤ 4
> 4, ≤ 5
Poverty-to-Income Ratio versus Reference Group (>5) Source: Authors’ calculations based on the Third National Health and Nutrition Examination Survey (NHANES III; Centers for Disease Control and Prevention n.d.). Note: Risky level of biomarker outcomes, NHANES III, adults age twenty to seventy-four. Error bars represent 95 percent confidence intervals.
88 Biological Consequences of Socioeconomic Inequalities Figure 3.12 Regression Coefficients, Allostatic Load
Ordinary Least Squares Coefficient
1.00
Mean Count = 1.62
0.80 0.60 0.40 0.20 0.00 ?1 −0.20
> 1, ≤ 2
> 2, ≤ 3
> 3, ≤ 4
> 4, ≤ 5
Poverty-to-Income Ratio versus Reference Group
Source: Authors’ calculations based on the Third National Health and Nutrition Examination Survey (NHANES III; Centers for Disease Control and Prevention n.d.). Note: Allostatic load regressions, NHANES III, adults age twenty to seventy-four. Error bars represent 95 percent confidence intervals.
A similar regression for an overall allostatic load score constructed by summing high-risk scores on the eight individual biomarkers also demonstrates a pronounced income gradient. These results are graphically illustrated in figure 3.12. The impact of the income-to-poverty ratio on counts of high-risk biomarker levels declines monotonically as income rises, with the four lowest income groups generating coefficients statistically different from zero. The results are qualitatively large. Those in the lowest three income groups have AL scores 0.58, 0.35, and 0.27 points higher, respectively, than those in the highest income group, representing values 36, 25, and 16 percent of the sample mean, respectively. The results in figures 3.4 through 3.12 are by construction rather limited. A much deeper analysis is possible. The data could be analyzed for separate subgroups (such as by age, race, ethnicity), or we could use different measures of SES (education, for example). Such a detailed analysis is beyond the scope of this review. Despite these limitations, the results in the figures are broadly consistent with the outlined literature and provide three key results. First, the income gradient in biomarkers is rather broad
Biological Imprints of Social Status 89 based, encompassing many different measures of biological risk. Second, these measures of health show some heterogeneity, the most noticeable being the lack of any income gradient for diastolic blood pressure. Finally, aggregating biomarkers and generating a cumulative measure of insults appears to be advantageous. The fraction of the population with risky biomarkers is large. Although most risky biomarkers show some income gradient, the results are sometimes of weak precision and the results are not always monotonic. However, when the multiple risks are accumulated, the gradient is more pronounced and is estimated with a great deal more precision.
Do Observed Biomarker Gradients Mediate SES Disparities in Morbidity and Mortality? As this review makes clear, SES disparities in biomarkers levels and physiological functioning are evident across a wide range of physiological systems. These disparities are present early in life and persist across the life course. Although the frequency and persistence of these observations are striking, the relevance of such disparities is borne out only in the ability of SES variations in biomarker level and function to predict SES variations in clinical health outcomes. Although available data sources—which include information on SES characteristics, biomarkers, and clinical outcomes— are sparse, information is slowly accumulating that suggests that social status variations in biomarker level and function may translate into SES variations in health outcomes. Most investigations examining the role of biomarkers in explaining clinical health outcomes have examined the role of individual or clusters of biomarkers in accounting for SES variations in cardiovascular events or mortality. Individual inflammatory biomarkers (IL-6, CRP, fibrinogen) have been found to account for small to moderate proportions (8 percent to 22 percent) of SES gradients in the incidence of cardiovascular disease and events (Marmot et al. 2008; Rosvall et al. 2008) and mortality in older adults (Ramsay et al. 2009). Interestingly, support is less consistent for a mediating role of more traditional cardiovascular risk factors, such as blood pressure and metabolic biomarkers, with some studies finding no or little mediating role (Loucks et al. 2009; Ramsay et al. 2009) and others finding that they account for a small proportion of SES gradients in cardiovascular outcomes (Marmot et al. 2008). Biomarkers have been found to account for a greater share of SES gradients in disease when examined simultaneously in analytic models or when composite indices of biomarker risk are examined. For example, a set of inflammatory biomarkers
90 Biological Consequences of Socioeconomic Inequalities explained 23 percent and another of cardiovascular-metabolic biomarkers explained 27 percent of the variance in coronary heart disease incidence in the Whitehall cohort (Marmot et al. 2008). Including both cardiovascular- metabolic and inflammatory biomarkers together in an analysis indicated that these biomarkers accounted for 42 percent of the social class gradient in heart disease incidence. As noted, similar greater explanatory power of aggregated biologic risk was observed in a study examining mortality in older adults from the MacArthur Studies of Successful Aging (Seeman et al. 2004); a composite allostatic load index of biological risk explained more than one third of the education gradient in late-life mortality, but the mediating role of individual biomarkers was much lower when examined separately in analyses.
Conclusions This overview documents SES gradients in biomarkers of a wide array of physiological regulatory systems in the body, including systems involved in organizing biological responses to stressful experiences and our cognitive and affective responses to our social world. These include neuroendocrine systems, such as the HPA axis and SNS hormones, as well the cardiovascular, metabolic, and immune systems. The sensitivity of biomarkers of each of these systems to SES-related exposures, experiences, and behaviors represents potential routes through which social status may get under the skin to affect health and well-being. Although most research has examined cross-sectional associations between SES and biomarkers, longitudinal research is accumulating that suggests that the imprints of SES adversity can be observed in the level and activity of biomarkers of physiological regulatory systems across the life course. However, longitudinal research designs with comprehensive assessments of SES, biomarkers, and the psychosocial and behavioral pathways through which socioeconomic status may be linked to physiological functioning remain few in number. A related challenge is the need to document that observed SES gradients in biomarkers actually translate into SES variations in disease risk and thus could serve as early warning signals of greater risk for poor health in later life. Despite these limitations, the large number of studies and the wide array of biomarker targets for which SES gradients in biomarker levels and function have been found do suggest that our status in socioeconomic structures is intimately linked with the functioning of the internal regulatory systems in our body. The continued inclusion of biomarker measures in studies of social status and health, hopefully facilitated by advancements in the ease and reliability of biomarker measurement, may prove central to understanding the mechanisms through which SES may be
Biological Imprints of Social Status 91 linked to health as well as aiding in the identification of those most at risk or those in greatest need of intervention. Indeed, using childhood asthma as an example, chapter 4 is designed to illustrate how a more targeted program of research can further delineate the mechanisms through which socioeconomic status may affect health.
References Abraham, Nader G., Eric J. Brunner, Jan W. Eriksson, and R. Paul Robertson. 2007. “Metabolic Syndrome: Psychosocial, Neuroendocrine, and Classical Risk Factors in Type 2 Diabetes.” In Stress Responses in Biology and Medicine: Stress of Life in Molecules, Cells, Organisms, and Psychosocial Communities, edited by Péter Csermely, Tamás Korcsmaros, and Katalin Sulyok. Oxford: Blackwell Publishing. Adam, Emma K., and Meena Kumari. 2009. “Assessing Salivary Cortisol in Large- Scale, Epidemiological Research.” Psychoneuroendocrinology 34(10): 1423–36. Adler, Nancy E., and David H. Rehkopf. 2008. “U.S. Disparities in Health: Descriptions, Causes, and Mechanisms.” Annual Review of Public Health 29(2008): 235– 52. Anderson, Norman B., and Cheryl A. Armstead. 1995. “Toward Understanding the Association of Socioeconomic Status and Health: A New Challenge for the Biopsychosocial Approach.” Psychosomatic Medicine 57(3): 213–25. Barkley, G. S. 2008. “Factors Influencing Health Behaviors in the National Health and Nutritional Examination Survey, III (NHANES III).” Social Work in Health Care 46(4): 57–79. Black, Paul H. 2002. “Stress and the Inflammatory Response: A Review of Neurogenic Inflammation.” Brain Behavior and Immunity 16(6): 622–53. ———. 2003. “The Inflammatory Response Is an Integral Part of the Stress Response: Implications for Atherosclerosis, Insulin Resistance, Type II Diabetes and Metabolic Syndrome X.” Brain Behavior and Immunity 17(5): 350–64. Blane, D., C. L. Hart, G. D. Smith, C. R. Gillis, D. J. Hole, and V. M. Hawthorne. 1996. “Association of Cardiovascular Disease Risk Factors with Socioeconomic Position During Childhood and During Adulthood.” British Medical Journal 313(7070): 1434–38. Brunner, Eric J., Michael G. Marmot, K. Nanchahal, M. J. Shipley, S. A. Stansfeld, M. Juneja, and K. G. M. M. Alberti. 1997. “Social Inequality in Coronary Risk: Central Obesity and the Metabolic Syndrome. Evidence from the Whitehall II Study.” Diabetologia 40(11): 1341–49. Brunner, Eric J., George Davey Smith, Michael G. Marmot, R. Canner, M. Beksinska, and J. O’Brien. 1996. “Childhood Social Circumstances and Psychosocial and Behavioural Factors as Determinants of Plasma Fibrinogen.” Lancet 347(9007): 1008–13.
92 Biological Consequences of Socioeconomic Inequalities Casas, J. P., T. Shah, A. D. Hingorani, J. Danesh, and M. B. Pepys. 2008. “C-Reactive Protein and Coronary Heart Disease: A Critical Review.” Journal of Internal Medicine 264(4): 295–314. Centers for Disease Control and Prevention (CDC). n.d. National Health and Nutrition Examination Survey (NHANES) III. Available at: http://www.cdc.gov/ nchs/NHANES.htm (accessed June 10, 2012). Chen, Edith, David A. Langer, Yvonne E. Raphaelson, and Karen A. Matthews. 2004. “Socioeconomic Status and Health in Adolescents: The Role of Stress Interpretations.” Child Development 75(4): 1039–52. Chichlowska, K. L., K. M. Rose, Anna V. Diez-Roux, S. H. Golden, A. M. McNeill, and G. Heiss. 2008. “Individual and Neighborhood Socioeconomic Status Characteristics and Prevalence of Metabolic Syndrome: The Atherosclerosis Risk in Communities (ARIC) Study.” Psychosomatic Medicine 70(9): 986–92. Clougherty, Jane E., and Laura D. Kubzansky. 2009. “A Framework for Examining Social Stress and Susceptibility to Air Pollution in Respiratory Health.” Environmental Health Perspectives 117(9): 1351–58. Cohen, Sheldon. 1994. “Psychosocial Influences on Immunity and Infectious Disease in Humans.” In Handbook of Human Stress and Immunity, edited by R. Glaser and J. K. Kiecolt-Glaser. San Diego: Academic Press. Cohen, Sheldon, William J. Doyle, and Andrew Baum. 2006. “Socioeconomic Status Is Associated with Stress Hormones.” Psychosomatic Medicine 68(3): 414–20. Cohen, Sheldon, and T. B. Herbert. 1996. “Health Psychology: Psychological Factors and Physical Disease from the Perspective of Human Psychoneuroimmunology.” Annual Review of Psychology 47(1996): 113–42. Cohen, Sheldon, George A. Kaplan, and J. T. Salonen. 1999. “The Role of Psychological Characteristics in the Relation Between Socioeconomic Status and Perceived Health.” Journal of Applied Social Psychology 29(3): 445–68. Cohen, Sheldon, Joseph E. Schwartz, Elisa Epel, Clemens Kirschbaum, Steve Sidney, and Theresa Seeman. 2006. “Socioeconomic Status, Race, and Diurnal Cortisol Decline in the Coronary Artery Risk Development in Young Adults (CARDIA) Study.” Psychosomatic Medicine 68(1): 41–50. Colhoun, H. M., H. Hemingway, and N. R. Poulter. 1998. “Socio-Economic Status and Blood Pressure: An Overview Analysis.” Journal of Human Hypertension 12(2): 91–110. Coppack, S. W. 2001. “Pro-Inflammatory Cytokines and Adipose Tissue.” Proceedings of the Nutrition Society 60(3): 349–56. Crimmins, Eileen M., Jung K. Kim, and Teresa E. Seeman. 2009. “Poverty and Biological Risk: The Earlier ‘Aging’ of the Poor.” The Journals of Gerontology Series A 64(2): 286–92. Dallongeville, Jean, Dominique Cottel, Jean Ferrieres, Dominique Arveiler, Annie Bingham, Jean B. Ruidavets, Bernadette Haas, Pierre Ducimetiere, and Philippe
Biological Imprints of Social Status 93 Amouyel. 2005. “Household Income Is Associated with the Risk of Metabolic Syndrome in a Sex-Specific Manner.” Diabetes Care 28(2): 409–15. Danese, Andrea, Terrie E. Moffitt, HonaLee Harrington, Barry J. Milne, Guilherme Polanczyk, Carmine M. Pariante, Richie Poulton, and Avshalom Caspi. 2009. “Adverse Childhood Experiences and Adult Risk Factors for Age-Related Disease Depression, Inflammation, and Clustering of Metabolic Risk Markers.” Archives of Pediatrics & Adolescent Medicine 163(12): 1135–43. Danesh, John, Rory Collins, Paul Appleby, and Richard Peto. 1998. “Association of Fibrinogen, C-Reactive Protein, Albumin, or Leukocyte Count with Coronary Heart Disease: Meta-Analyses of Prospective Studies.” Journal of the American Medical Association 279(18): 1477–82. Dietrich, Denise F., Christian Schindler, Joel Schwartz, Jean-Claude Barthelemy, Jean-Marie Tschopp, Frederic Roche, Arnold von Eckardstein, Otto Brandli, Philipppe Leuenberger, Diane R. Gold, Jean-Michel Gaspoz, and Ursula Ackermann-Liebrich. 2006. “Heart Rate Variability in an Ageing Population and Its Association with Lifestyle and Cardiovascular Risk Factors: Results of the SAPALDIA Study.” Europace 8(7): 521–29. Dowd, Jennifer B., and Allison E. Aiello. 2009. “Socioeconomic Differentials in Immune Response.” Epidemiology 20(6): 902–98. Dowd, Jennifer B., Allison E. Aiello, and D. E. Alley. 2009. “Socioeconomic Disparities in the Seroprevalence of Cytomegalovirus Infection in the U.S. Population: NHANES III.” Epidemiology and Infection 137(1): 58–65. Dowd, Jennifer B., and Noreen Goldman. 2006. “Do Biomarkers of Stress Mediate the Relation Between Socioeconomic Status And Health?” Journal of Epidemiology and Community Health 60(7): 633–39. Dowd, Jennifer B., Mary N. Haan, Lynn Blythe, Kari Moore, and Allison E. Aiello. 2008. “Socioeconomic Gradients in Immune Response to Latent Infection.” American Journal of Epidemiology 167(1): 112–20. Dowd, Jennifer B., Amanda M. Simanek, and Allison E. Aiello. 2009. “Socio- Economic Status, Cortisol and Allostatic Load: A Review of the Literature.” International Journal of Epidemiology 38(5): 1297–309. Epel, Elissa, Rachel Lapidus, Bruce McEwen, and Kelly Brownell. 2001. “Stress May Add Bite to Appetite in Women: A Laboratory Study of Stress-Induced Cortisol and Eating Behavior.” Psychoneuroendocrinology 26(1): 37–49. Evans, Gary W. 2003. “A Multimethodological Analysis of Cumulative Risk and Allostatic Load Among Rural Children.” Developmental Psychology 39(5): 924–33. Evans, Gary W., and Kimberly English. 2002. “The Environment of Poverty: Multiple Stressor Exposure, Psychophysiological Stress, and Socioemotional Adjustment.” Child Development 73(4): 1238–48. Evans, Gary W., C. Gonnella, L. A. Marcynyszyn, L. Gentile, and N. Salpekar. 2005. “The Role of Chaos in Poverty and Children’s Socioemotional Adjustment.” Psychological Science 16(7): 560–65.
94 Biological Consequences of Socioeconomic Inequalities Evans, Gary W., and Pilyung Kim. 2007. “Childhood Poverty and Health: Cumulative Risk Exposure and Stress Dysregulation.” Psychological Science 18(11): 953– 57. Evans, Gary W., Pilyung Kim, Alber H. Ting, Harris B. Tesher, and Dana Shannis. 2007. “Cumulative Risk, Maternal Responsiveness, and Allostatic Load Among Young Adolescents.” Developmental Psychology 43(2): 341–51. Evans, Gary W., and Michelle A. Schamberg. 2009. “Childhood Poverty, Chronic Stress, and Adult Working Memory.” Proceedings of the National Academy of Sciences 106(16): 6545–49. Evans, Gary W., Elaine Wethington, Meridith Coleman, Margo Worms, and Edward A. Frongillo. 2008. “Income Health Inequalities Among Older Persons: The Mediating Role of Multiple Risk Exposures.” Journal of Aging and Health 20(1): 107–25. Everson, Susan A., Slobhan C. Maty, John W. Lynch, and George A. Kaplan. 2002. “Epidemiologic Evidence for the Relation Between Socioeconomic Status and Depression, Obesity, and Diabetes.” Journal of Psychosomatic Research 53(4): 891– 95. Fernald, Leah C.H., and Nancy E. Adler. 2008. “Blood Pressure and Socioeconomic Status in Low-Income Women in Mexico: A Reverse Gradient?” Journal of Epidemiology and Community Health 62(5): e8. Finkelstein, Daniel M., Laura D. Kubzansky, John Capitman, and Elizabeth Goodman. 2007. “Socioeconomic Differences in Adolescent Stress: The Role of Psychological Resources.” Journal of Adolescent Health 40(2): 127–34. Fries, Eva, Lucia Dettenborn, and Clemens Kirschbaum. 2009. “The Cortisol Awakening Response (CAR): Facts and Future Directions.” International Journal of Psychophysiology 72(1): 67–73. Gallo, Linda C., Laura M. Bogart, Ana-Maria Vranceanu, and Karen A. Matthews. 2005. “Socioeconomic Status, Resources, Psychological Experiences, and Emotional Responses: A Test of the Reserve Capacity Model.” Journal of Personaltiy and Social Psychology 88(2): 386–99. Gallo, Linda C., Laura M. Bogart, Ana-Maria Vranceanu, and L. C. Walt. 2004. “Job Characteristics, Occupational Status, and Ambulatory Cardiovascular Activity in Women.” Annals of Behavioral Medicine 28(1): 62–73. Gallo, Linda C., and Karen A. Matthews. 2003. “Understanding the Association Between Socioeconomic Status and Physical Health: Do Negative Emotions Play a Role?” Psychological Bulletin 129(1): 10–51. Gehi, Anil K., Rachel Lampert, Emir Veledar, Forrester Lee, Jack Goldberg, Linda Jones, Nancy Murrah, Ali Ashraf, and Viola Vaccarino. 2009. “A Twin Study of Metabolic Syndrome and Autonomic Tone.” Journal of Cardiovascular Electrophysiology 20(4): 422–28. Geronimus, Arline T., Margaret Hicken, Danya Keene, and John Bound. 2006.
Biological Imprints of Social Status 95 “‘Weathering’ and Age Patterns of Allostatic Load Scores Among Blacks and Whites in the United States.” American Journal of Public Health 96(5): 826–33. Gimeno, David, Jane E. Ferrie, Marko Elovainio, Laura Pulkki-Raback, Liisa Keltikangas-Jarvinen, Carita Eklund, Mikko Hurme, Terho Lehtimaki, Jukka Marniemi, Jorma S.A. Viikari, Olli T. Raitakari, and Mika Kivimaki. 2008. “When Do Social Inequalities in C-Reactive Protein Start? A Life Course Perspective from Conception to Adulthood in the Cardiovascular Risk in Young Finns Study.” International Journal of Epidemiology 37(2): 290–98. Goodman, Elizabeth, Stephen R. Daniels, and Lawrence M. Dolan. 2007. “Socioeconomic Disparities in Insulin Resistance: Results from the Princeton School District Study.” Psychosomatic Medicine 69(1): 61–67. Goodman, Elizabeth, Bruce S. McEwen, Bin Huang, Lawrence M. Dolan, and Nancy E. Adler. 2005. “Social Inequalities in Biomarkers of Cardiovascular Risk in Adolescence.” Psychosomatic Medicine 67(1): 9–15. Greiser, Karina H., Alexander Kluttig, Barbara Schumann, Cees A. Swenne, Jan A. Kors, Oliver Kuss, Johannes Haerting, Hendrik Schmidt, Joachim Thiery, and Karl Werdan. 2009. “Cardiovascular Diseases, Risk Factors and Short-Term Heart Rate Variability in an Elderly General Population: The CARLA Study 2002–2006.” European Journal of Epidemiology 24(3): 123–42. Gruenewald, Tara L., Sheldon Cohen, Karen A. Matthews, Russell Tracy, and Teresa E. Seeman. 2009. “Association of Socioeconomic Status with inflammation Markers in Black and White Men and Women in the Coronary Artery Risk Development in Young Adults (CARDIA) Study.” Social Science & Medicine 69(3): 451–59. Guize, L., C. Jaffiol, M. Gueniot, J. Bringer, C. Giudicelli, M. Tramoni, F. Thomas, B. Pannier, K. Bean, and B. Jego. 2008. “Diabetes and Socio-Economic Deprivation. A Study in a Large French Population.” Bulletin de l’academie nationale de medecine 192(9): 1707–18. Hansen, Ase Marie, Anne Helene Garde, and Roger Persson. 2008. “Sources of Biological and Methodological Variation in Salivary Cortisol and Their Impact on Measurement Among Healthy Adults: A Review.” Scandinavian Journal of Clinical & Laboratory Investigation 68(6): 448–58. Hanson, Margaret D., and Edith Chen. 2007. “Socioeconomic Status and Health Behaviors in Adolescence: A Review of the Literature.” Journal of Behavioral Medicine 30(3): 263–85. Hemingway, Harry, Martin Shipley, Eric Brunner, Annie Britton, Marek Malik, and Michael G. Marmot. 2005. “Does Autonomic Function Link Social Position to Coronary Risk? The Whitehall II Study.” Circulation 111(23): 3071–77. Hemingway, Harry, M. Shipley, M. J. Mullen, M. Kumari, Eric Brunner, M. Taylor, A. E. Donald, J. E. Deanfield, and Michael G. Marmot. 2003. “Social and Psychosocial Influences on Inflammatory Markers and Vascular Function in Civil Servants (the Whitehall II Study). American Journal of Cardiology 92(8): 984–87.
96 Biological Consequences of Socioeconomic Inequalities Ishizaki, Masao, Pekka Martikainen, Hideaki Nakagawa, and Michael G. Marmot. 2000. “The Relationship Between Employment Grade and Plasma Fibrinogen Level Among Japanese Male Employees: YKKJ Research Group.” Atherosclerosis 151(2): 415–21. Janicki-Deverts, Denise, Sheldon Cohen, Nancy E. Adler, Joseph E. Schwartz, Karen A. Matthews, and T. E. Seeman. 2007. “Socioeconomic Status Is Related to Urinary Catecholamines in the Coronary Artery Risk Development in Young Adults (CARDIA) Study.” Psychosomatic Medicine 69(6): 514–520. Jousilahti, P., V. Salomaa, V. Rasi, E. Vahtera, and T. Palosuo. 2003. “Association of Markers of Systemic Inflammation, C Reactive Protein, Serum Amyloid A, and Fibrinogen, with Socioeconomic Status.” Journal of Epidemiology and Community Health 57(9): 730–33. Kemeny, Margaret E., and Tara L. Gruenewald. 2000. “Affect, Cognition, the Immune System and Health.” In Progress in Brain Research, vol. 122, The Biological Basis for Mind Body Interactions, edited by E. A. Mayer and C. B. Saper. Maryland Heights, Md.: Elsevier. Kendzor, Darla E., Michael S. Businelle, Carlos A. Mazas, Ludmila M. Cofta- Woerpel, Lorraine R. Reitzel, Jennifer I. Vidrine, Yishen S. Li, Tracy J. Costello, Paul M. Cinciripini, Jasjit S. Ahluwalia, and David W. Wetter. 2009. “Pathways Between Socioeconomic Status and Modifiable Risk Factors Among African American Smokers.” Journal of Behavioral Medicine 32(6): 545–57. Kivimaki, Mika, Debbie A. Lawlor, Markus Juonala, G. Davey Smith, Marko Elovainio, Liisa Keltikangas-Jarvinen, Jussi Vahtera, Jorma S. Viikari, and Olli T. Raitakari. 2005. “Lifecourse Socioeconomic Position, C-Reactive Protein, and Carotid Intima-Media Thickness in Young Adults: The Cardiovascular Risk in Young Finns Study.” Arteriosclerosis, Thrombosis, and Vascular Biology 25(10): 2197–202. Kivimaki, Mika, Debbie A. Lawlor, George D. Smith, Liisa Keltikangas-Jarvinen, Marko Elovainio, Jussi Vahtera, Laura Pulkki-Raback, Leena Taittonen, Jorma S. A. Viikari, and Olli T. Raitakari. 2006. “Early Socioeconomic Position and Blood Pressure in Childhood and Adulthood: The Cardiovascular Risk in Young Finns Study.” Hypertension 47(1): 39–44. Kivimaki, Mika, George D. Smith, Marko Elovainio, Laura Pulkki, Liisa Keltikangas-Jarvinen, Leena Talttonen, Olli T. Raitakari, and Jorma S. Viikari. 2006. “Socioeconomic Circumstances in Childhood and Blood Pressure in Adulthood: The Cardiovascular Risk in Young Finns Study.” Annals of Epidemiology 16(10): 737–42. Kivimaki, Mika, George D. Smith, M. Juonala, Jane E. Ferrie, Liisa Keltikangas- Jarvinen, Marko Elovainio, Laura Pulkki-Raback, Jussi Vahtera, M. Leino, Jorma S. Viikari, and Olli T. Raitakari. 2006. “Socioeconomic Position in Childhood and Adult Cardiovascular Risk Factors, Vascular Structure, and Function: Cardiovascular Risk in Young Finns Study.” Heart 92(4): 474–80.
Biological Imprints of Social Status 97 Koenig, Wolgang, Malte Sund, Margit Frohlich, Hans-Gunther Fischer, Hannelore Lowel, Angela Doring, Winston L. Hutchinson, and Mark B. Pepys. 1999. “C- Reactive Protein, a Sensitive Marker of Inflammation, Predicts Future Risk of Coronary Heart Disease in Initially Healthy Middle-Aged Men: Results from the MONICA (Monitoring Trends and Determinants in Cardiovascular Disease) Augsburg Cohort Study, 1984 to 1992.” Circulation 99(2): 237–42. Koster, Annamarie, Hans Bosma, Brenda W. Penninx, Anne B. Newman, Tamara B. Harris, Jacques Thm. van Eijk, Gertrudis I. Kempen, E. M. Simonsick, K. C. Johnson, Ronica N. Rooks, Hilsa N. Ayonayon, Susan M. Rubin, and Stephen B. Kritchevsky. 2006. “Association of Inflammatory Markers with Socioeconomic Status.” The Journals of Gerontology Series A 61(3): 284–90. Kubzansky, Laura D., Ichiro Kawachi, and D. Sparrow. 1999. “Socioeconomic Status, Hostility, and Risk Factor Clustering in the Normative Aging Study: Any Help from the Concept of Allostatic Load?” Annals of Behavioral Medicine 21(4): 330–38. Kumari, Meena, Ellena Badrick, T. Chandola, Nancy E. Adler, Elissa Epel, Teresa E. Seeman, Clemens Kirschbaum, and Michael G. Marmot. 2010. “Measures of Social Position and Cortisol Secretion in an Aging Population: Findings from the Whitehall II Study.” Psychosomatic Medicine 72(1): 27–34. Langenberg, Claudia, Diana Kuh, Michael E. J. Wadsworth, Eric Brunner, and Rebecca Hardy. 2006. “Social Circumstances and Education: Life Course Origins of Social Inequalities in Metabolic Risk in a Prospective National Birth Cohort.” American Journal of Public Health 96(12): 2216–21. Li, Leah, Chris Power, Shona Kelly, Clemens Kirschbaum, and Clyde Hertzman. 2007. “Life-Time Socio-Economic Position and Cortisol Patterns in Mid-Life.” Psychoneuroendocrinology 32(7): 824–33. Lidfeldt, Jonas, Tricia Y. Li, Frank B. Hu, JoAnn E. Manson, and Ichiro Kawachi. 2007. “A Prospective Study of Childhood and Adult Socioeconomic Status and Incidence of Type 2 Diabetes in Women.” American Journal of Epidemiology 165(8): 882–89. Light, Kathleen C., Kimberly A. Brownley, J. Rick Turner, Alan L. Hinderliter, Susan S. Girdler, Andrew Sherwood, and Norman B. Anderson. 1995. “Job Status and High-Effort Coping Influence Work Blood Pressure in Women and Blacks.” Hypertension 25(4): 554–59. Loucks, Eric B., John W. Lynch, Loiuse Pilote, Rebecca Fuhrer, Nisha D. Almeida, Hugues Richard, Golareh Agha, Joanne M. Murabito, and Emelia J. Benjamin. 2009. “Life-Course Socioeconomic Position and Incidence of Coronary Heart Disease.” American Journal of Epidemiology 169(7): 829–36. Loucks, Eric B., Kristjan T. Magnusson, Stephen Cook, David H. Rehkopf, Earl S. Ford, and Lisa F. Berkman. 2007. “Socioeconomic Position and the Metabolic Syndrome in Early, Middle, and Late Life: Evidence from NHANES 1999–2002.” Annals of Epidemiology 17(10): 782–90.
98 Biological Consequences of Socioeconomic Inequalities Loucks, Eric B., David H. Rehkopf, Rebecca C. Thurston, and Ichiro Kawachi. 2007. “Socioeconomic Disparities in Metabolic Syndrome Differ by Gender: Evidence from NHANES III.” Annals of Epidemiology 17(1): 19–26. Loucks, Eric B., Lisa M. Sullivan, Laura J. Hayes, Ralph B. D’Agostino Sr., Martin G. Larson, Ramachandran S. Vasan, Emelia J. Benjamin, and Lisa F. Berkman. 2006. “Association of Educational Level with Inflammatory Markers in the Framingham Offspring Study.” American Journal of Epidemiology 163(7): 622–28. Lubbock, Lindsey A., Anne Goh, Sadia Ali, James Ritchie, and Mary A. Whooley. 2005. “Relation of Low Socioeconomic Status to C-Reactive Protein in Patients with Coronary Heart Disease (from the Heart and Soul Study).” American Journal of Cardiology 96(11): 1506–11. Lynch, J. W., George A. Kaplan, and J. T. Salonen. 1997. “Why Do Poor People Behave Poorly? Variation in Adult Health Behaviours and Psychosocial Characteristics by Stages of the Socioeconomic Lifecourse.” Social Science & Medicine 44(6): 809–19. Marmot, Michael G., Rebecca Fuhrer, Susan L. Ettner, Nadine F. Marks, Larry L. Bumpass, and Carol D. Ryff. 1998. “Contribution of Psychosocial Factors to Socioeconomic Differences in Health.” Milbank Quarterly 76(3): 403–48. Marmot, Michael G., M. J. Shipley, H. Hemingway, J. Head, and Eric J. Brunner. 2008. “Biological and Behavioural Explanations of Social Inequalities in Coronary Heart Disease: The Whitehall II Study.” Diabetologia 51(11): 1980–88. Marmot, Michael G., and Richard G. Wilkinson. 2001. “Psychosocial and Material Pathways in the Relation Between Income and Health: A Response to Lynch et al.” British Medical Journal 322(7296): 1233–36. Matthews, Karen A., Katri Raikkonen, Susan A. Everson, Janine D. Flory, Christine A. Marco, Jane F. Owens, and Catherine E. Lloyd. 2000. “Do the Daily Experiences of Healthy Men and Women Vary According to Occupational Prestige and Work Strain?” Psychosomatic Medicine 62(3): 346–53. Maty, Siobhan C., Susan A. Everson-Rose, Mary N. Haan, Trivellore E. Raghunathan, and George A. Kaplan. 2005. “Education, Income, Occupation, and the 34-Year Incidence (1965–99) of Type 2 Diabetes in the Alameda County Study.” International Journal of Epidemiology 34(6): 1274–81. Maty, Siobhan C., John W. Lynch, Trivellore E. Raghunathan, and George A. Kaplan. 2008. “Childhood Socioeconomic Position, Gender, Adult Body Mass Index, and Incidence of Type 2 Diabetes Mellitus over 34 Years in the Alameda County Study.” American Journal of Public Health 98(8): 1486–94. McEwen, Bruce S. 1998. “Stress, Adaptation, and Disease: Allostasis and Allostatic Load.” Annals of the New York Academy of Sciences 840(1): 33–44. ———. 2002. “Protective and Damaging Effects of Stress Mediators: The Good and Bad Sides of the Response to Stress.” Metabolism 51(6 Suppl 1): 2–4. McLaren, Lindsay. 2007. “Socioeconomic Status and Obesity.” Epidemiologic Reviews 29(1): 29–48.
Biological Imprints of Social Status 99 McNeill, Lorna H., Matthew W. Kreuter, and S. V. Subramanian. 2006. “Social Environment and Physical Activity: A Review of Concepts and Evidence.” Social Science & Medicine 63(4): 1011–22. Mikolajczyk, Rafael T., Walid El Ansari, and Annette E. Maxwell. 2009. “Food Consumption Frequency and Perceived Stress and Depressive Symptoms Among Students in Three European Countries.” Nutrition Journal 8(2009): 31–39. Morello-Frosch, Rachel, and Edmond D. Shenassa. 2006. “The Environmental ‘Riskscape’ and Social Inequality: Implications for Explaining Maternal and Child Health Disparities.” Environmental Health Perspectives 114(8): 1150–53. Owen, Natalie, Terry Poulton, Frank C. Hay, Vidya Mohamed-Ali, and Andrew Steptoe. 2003. “Socioeconomic Status, C-Reactive Protein, Immune Factors, and Responses to Acute Mental Stress.” Brain Behavior and Immunity 17(4): 286–95. Panagiotakos, Demosthenes B., Christos E. Pitsavos, Christina A. Chrysohoou, John Skoumas, Marina Toutouza, Dennis Belegrinos, Pavios K. Toutouzas, and Christodoulos Stefanadis. 2004. “The Association between Educational Status and Risk Factors Related to Cardiovascular Disease in Healthy Individuals: The ATTICA Study.” Annals of Epidemiology 14(3): 188–94. Perel, Pablo, Claudia Langenberg, Jane Ferrie, Kath Moser, Eric Brunner, and Michael G. Marmot. 2006. “Household Wealth and the Metabolic Syndrome in the Whitehall II Study.” Diabetes Care 29(12): 2694–700. Petersen, Karen L., Anna L. Marsland, Janine Flory, Elizabeth Votruba-Drzal, Matthew F. Muldoon, and Stephen B. Manuck. 2008. “Community Socioeconomic Status Is Associated with Circulating Interleukin-6 and C-Reactive Protein.” Psychosomatic Medicine 70(6): 646–52. Phillips, Jennifer E., Anna L. Marsland, Janine D. Flory, Matthew F. Muldoon, Sheldon Cohen, and Stephen B. Manuck. 2009. “Parental Education Is Related to C-Reactive Protein Among Female Middle-Aged Community Volunteers.” Brain Behavior and Immunity 23(5): 677–83. Pincus, Theodore, and Leigh F. Callahan. 1995. “What Explains the Association Between Socioeconomic Status and Health: Primarily Access to Medical Care or Mind-Body Variables?” ADVANCES: Journal of Mind-Body Health 11(1): 4–36. Pollitt, Ricardo A., Jay S. Kaufman, Kathryn M. Rose, Ana V. Diez-Roux, Donglin Zeng, and Gerardo Heiss. 2007. “Early-Life and Adult Socioeconomic Status and Inflammatory Risk Markers in Adulthood.” European Journal of Epidemiology 22(1): 55–66. ———. 2008. “Cumulative Life Course and Adult Socioeconomic Status and Markers of Inflammation in Adulthood.” Journal of Epidemiology and Community Health 62(6): 484–91. Poulton, Richie, Avshalom Caspi, Barry J. Milne, W. Murray Thomson, Alan Taylor, Malcolm R. Sears, and Terrie E. Moffitt. 2002. “Association Between Children’s Experience of Socioeconomic Disadvantage and Adult Health: A Life- Course Study.” Lancet 360(9346): 1640–45.
100 Biological Consequences of Socioeconomic Inequalities Pressman, Sarah D., and Sheldon Cohen. 2005. “Does Positive Affect Influence Health?” Psychological Bulletin 131(6): 925–71. Pruessner, J. C., O. T. Wolf, D. H. Hellhammer, A. Buske-Kirschbaum, K. von Auer, S. Jobst, F. Kaspers, and C. Kirschbaum. 1997. “Free Cortisol Levels After Awakening: A Reliable Biological Marker for the Assessment Of Adrenocortical Activity.” Life Sciences 61(26): 2539–49. Ramsay, S. E., R. W. Morris, P. H. Whincup, O. Papacosta, A. Rumley, L. Lennon, G. Lowe, and S. G. Wannamethee. 2009. “Socioeconomic Inequalities in Coronary Heart Disease Risk in Older Age: Contribution of Established and Novel Coronary Risk Factors.” Journal of Thrombosis and Haemostasis 7(11): 1779–86. Ranjit, Nalini, Elizabeth A. Young, and George A. Kaplan. 2005. “Material Hardship Alters the Diurnal Rhythm of Salivary Cortisol.” International Journal of Epidemiology 34(5): 1138–43. Rathmann, Wolfgang, Burkhard Haastert, Guido Giani, Wolfgang Koenig, Armin Imhof, Christian Herder, Rolf Holle, and Andreas Mielck. 2006. “Is Inflammation a Causal Chain Between Low Socioeconomic Status and Type 2 Diabetes? Results from the KORA Survey 2000.” European Journal of Epidemiology 21(1): 55–60. Rod, N. H., M. Gronbaek, P. Schnohr, E. Prescott, and T. S. Kristensen. 2009. “Perceived Stress as a Risk Factor for Changes in Health Behaviour and Cardiac Risk Profile: A Longitudinal Study.” Journal of Internal Medicine 266(5): 467–75. Rosero-Bixby, Luis, and William H. Dow. 2009. “Surprising SES Gradients in Mortality, Health, and Biomarkers in a Latin American Population of Adults.” The Journals of Gerontology Series B 64(1): 105–17. Rosmond, Roland, and Per Björntorp. 2000. “Occupational Status, Cortisol Secretory Pattern, and Visceral Obesity in Middle-Aged Men.” Obesity Research 8(6): 445–50. Rosvall, Maria, Gunnar Engstrom, Goran Berglund, and Bo Hedblad. 2008. “C- Reactive Protein, Established Risk Factors and Social Inequalities in Cardiovascular Disease: The Significance of Absolute Versus Relative Measures of Disease.” BMC Public Health 8:189–199. Schroeder, Emily B., Lloyd L. Chambless, Duanping P. Liao, Ronald J. Prineas, Gregory W. Evans, Wayne D. Rosamond, and Gerardo Heiss. 2005. “Diabetes, Glucose, Insulin, and Heart Rate Variability: The Atherosclerosis Risk in Communities (ARIC) Study.” Diabetes Care 28(3): 668–74. Schroeder, Emily B., Duanping P. Liao, Lloyd E. Chambless, Ronald J. Prineas, Gregory W. Evans, and Gerardo Heiss. 2003. “Hypertension, Blood Pressure, and Heart Rate Variability: The Atherosclerosis Risk in Communities (ARIC) Study.” Hypertension 42(6): 1106–11. Seeman, Teresa E., Eileen Crimmins, Mei-Hua Huang, Burton Singer, Alexander Bucur, Tara Gruenewald, Lisa F. Berkman, and David B. Reuben. 2004. “Cumu-
Biological Imprints of Social Status 101 lative Biological Risk and Socio-Economic Differences in Mortality: MacArthur Studies of Successful Aging.” Social Science & Medicine 58(10): 1985–97. Seeman, Teresa E., S. Sharon Merkin, Eileen M. Crimmins, Brandon Koretz, Susan Charette, and Arun Karlamangla. 2008. “Education, Income and Ethnic Differences in Cumulative Biological Risk Profiles in a National Sample of U.S. Adults: NHANES III (1988–1994).” Social Science & Medicine 66(1): 72–87. Senese, Laura C., Nisha D. Almeida, Anne K. Fath, Brendan T. Smith, and Eric B. Loucks. 2009. “Associations Between Childhood Socioeconomic Position and Adulthood Obesity.” Epidemiologic Reviews 31(1): 21–51. Sesso, Howard D., Lu Wang, Julie E. Buring, Paul M. Ridker, and J. Michael Gaziano. 2007. “Comparison of Interleukin-6 and C-Reactive Protein for the Risk of Developing Hypertension in Women.” Hypertension 49(2): 304–10. Shrewsbury, Vanessa, and Jane Wardle. 2008. “Socioeconomic Status and Adiposity in Childhood: A Systematic Review of Cross-Sectional Studies, 1990–2005.” Obesity 16(2): 275–84. Singer, Burton, and Carol D. Ryff. 1999. “Hierarchies of Life Histories and Associated Health Risks.” Annals of the New York Academy of Sciences 896(1999): 96–115. Sloan, Richard P., Mei-Hua Huang, Stephen Sidney, Kiang Liu, O. Dale Williams, and Teresa Seeman. 2005. “Socioeconomic Status and Health: Is Parasympathetic Nervous System Activity an Intervening Mechanism?” International Journal of Epidemiology 34(2): 309–15. Smyth, Joshua M., Margit C. Ockenfels, Amy A. Gorin, D. Catley, Laura S. Porter, Clemens Kirschbaum, Dirk H. Hellhammer, and Arthur A. Stone. 1997. “Individual Differences in the Diurnal Cycle of Cortisol.” Psychoneuroendocrinology 22(2): 89–105. Sobal, Jeffery, and Albert J. Stunkard. 1989. “Socioeconomic Status and Obesity: A Review of the Literature.” Psychological Bulletin 105(2): 260–75. Steptoe, Andew, Samantha Dockray, and Jane Wardle. 2009. “Positive Affect and Psychobiological Processes Relevant to Health.” Journal of Personality 77(6): 1747–76. Steptoe, Andrew, Sabine Kunz-Ebrecht, Natalie Owen, Pamela J. Feldman, Ann Rumley, Gordon D. Lowe, and Michael G. Marmot. 2003. “Influence of Socioeconomic Status and Job Control on Plasma Fibrinogen Responses to Acute Mental Stress.” Psychosomatic Medicine 65(1): 137–44. Steptoe, Andrew, Sabine Kunz-Ebrecht, Natalie Owen, Pamela J. Feldman, Gonneke Willemsen, Clemens Kirschbaum, and Michael G. Marmot. 2003. “Socioeconomic Status and Stress-Related Biological Responses over the Working Day.” Psychosomatic Medicine 65(3): 461–70. Stone, Arthur A., Joseph E. Schwartz, Joshua Smyth, Clemens Kirschbaum, Sheldon Cohen, Dirk H. Hellhammer, and Steven Grossman. 2001. “Individual Differences in the Diurnal Cycle of Salivary Free Cortisol: A Replication of Flattened Cycles for Some Individuals.” Psychoneuroendocrinology 26(3): 295–306.
102 Biological Consequences of Socioeconomic Inequalities Su, Jason G., Rachel Morello-Frosch, Bill M. Jesdale, Amy D. Kyle, Bhavna Shamasunder, and Michael Jerrett. 2009. “An Index for Assessing Demographic Inequalities in Cumulative Environmental Hazards with Application to Los Angeles, California.” Environmental Science & Technology 43(20): 7626–34. Taylor, Shelly E., Rena L. Repetti, and Teresa E. Seeman. 1997. “Health Psychology: What Is an Unhealthy Environment and How Does It Get Under the Skin?” Annual Review of Psychology 48(1997): 411–47. Thayer, Jullan F., and Richard D. Lane. 2007. “The Role of Vagal Function in the Risk for Cardiovascular Disease and Mortality.” Biological Psychology 74(2): 224– 42. Thorand, Barbara, Hannelore Lowel, Andrea Schneider, Hubert Kolb, Christa Meisinger, Margit Frohlich, and Wolfgang Koenig. 2003. “C-Reactive Protein as a Predictor for Incident Diabetes Mellitus Among Middle-Aged Men: Results from the MONICA Augsburg Cohort Study, 1984–1998.”Archives of Internal Medicine 163(1): 93–99. Tuomisto, K., P. Jousilahti, J. Sundvall, P. Pajunen, and V. Salomaa. 2006. “C- Reactive Protein, Interleukin-6 and Tumor Necrosis Factor Alpha as Predictors of Incident Coronary and Cardiovascular Events and Total Mortality: A Population-Based, Prospective Study.” Thrombosis and Haemostasis 95(3): 511– 18. Turnbull, Fiona, Bruce Neal, Toshiaru Ninomiya, Charles Algert, Hisatoma Arima, Federica Barzi, Christopher Bulpitt, John Chalmers, Robert Fagard, A. Gleason, Stephane Heritier, N. Li, Vlado Perkovic, Mark Woodward, and Stephen MacMahon. 2008. “Effects of Different Regimens to Lower Blood Pressure on Major Cardiovascular Events in Older and Younger People: Meta-Analysis of Randomised Trials.” British Medical Journal 336(7653): 1121–23. Wamala, Sarah P., Mittleman A. Murray, Myriam Horsten, Margita Eriksson, Karin Schenck-Gustafsson, Anders Hamsten, Angela Silveira, and Kristina Orth- Gomer. 1999. “Socioeconomic Status and Determinants of Hemostatic Function in Healthy Women.” Arteriosclerosis Thrombosis and Vascular Biology 19(3): 485– 92. Weinstein, Maxine, Noreen Goldman, A. Hedley, L. Yu-Hsuan, and Teresa E. Seeman. 2003. “Social Linkages to Biological Markers of Health Among the Elderly.” Journal of Biosocial Science 35(3): 433–53. Williams, David R. 1990. “Socioeconomic Differentials in Health: A Review and Redirection.” Social Psychology Quarterly 53(2): 81–99. Wilson, Thomas W., George A. Kaplan, Jussi Kauhanen, Richard D. Cohen, Melien Wu, Riitta Salonen, and Jukka T. Salonen. 1993. “Association Between Plasma Fibrinogen Concentration and Five Socioeconomic Indices in the Kuopio Ischemic Heart Disease Risk Factor Study.” American Journal of Epidemiology 137(3): 292–300.
Chapter 4 Dissecting Pathways for Socioeconomic Gradients in Childhood Asthma Edith Chen, Hannah M.C. Schreier, and Meanne Chan
The goal of this chapter is to describe a program of research on socioeconomic status (SES) and childhood asthma as a specific, in-depth illustration of an integrated biological and psychosocial approach to establishing the mechanisms underlying SES and health relationships. Beginning with an established clinical phenomenon—that is, the link between low SES and asthma morbidity—we focus on the importance of understanding the basic pathophysiology of a disease to determine which steps in the disease process are plausibly altered by social factors. Researchers will be able to develop a more accurate understanding of why health disparities are so pervasive in our society and what types of interventions may hold the most promise for reducing these disparities. Socioeconomic status has profound effects on physical health outcomes throughout the lifespan (Adler et al. 1994; Braveman et al. 2010; Chen, Matthews, and Boyce 2002). For years, researchers have sought to understand why these relationships exist, but compelling explanations have proven elusive. For example, lack of insurance and access to health care is clearly one reason why low-SES individuals suffer worse health. And yet countries that have universal health-care systems show the same gradient relationship of SES with health as countries that do not, indicating that differential access to care is not the primary explanation for SES disparities (Adler et al. 1993). Similarly, low-SES individuals are known to engage in poorer health behaviors, and yet the impact of SES on health per-
103
104 Biological Consequences of Socioeconomic Inequalities sists after health behaviors are controlled for, suggesting that health behaviors also do not provide a complete explanation for the gradient (Lantz et al. 1998). Hence a need exists for more sophisticated models that convincingly explain how SES has such pervasive effects on health. One approach is to simultaneously consider factors at multiple levels, spanning broad neighborhood and family influences to basic genomic processes within the individual, all in an effort to more thoroughly understand the contributors to health disparities. This type of approach is important because rarely have studies that conducted in-depth assessments of basic, biological processes also probed social contexts comprehensively. Similarly, research on neighborhood-level effects often does not incorporate individual-level factors into models, and vice versa. By considering factors across multiple levels of influence, researchers will be able to better answer the challenging question of why disparities by SES exist. Thus the overall goal is to be able to explain how broad, distal social environment characteristics such as SES get manifest within an individual in a way that affects clinical disease outcomes. In this chapter, we focus on two primary research questions aimed at addressing this overall goal— using childhood asthma as our illustrative health outcome. First, what are the plausible biological mechanisms by which low SES can exert effects on physical health? To make convincing arguments about the social environment affecting disease, we need credible biological explanations for how social factors could plausibly influence disease processes. Second, what are the more proximal social factors that can explain the effects that distal variables such as SES have on individuals? To understand why SES has effects on individual health, we need models that articulate the neighborhood, family, and individual characteristics shaped by SES that have implications for biological processes linked to disease.
An Approach to Conducting Mechanistic Research We have previously articulated an approach to conceptualizing a search for biological mechanisms to explain links between social factors and disease (Miller, Chen, and Cole 2009). First, a robust association needs to be documented between a social factor such as SES and a disease outcome; at that point, mechanisms can be investigated on both the biological and social fronts. Biologically, it can be helpful to understand the basic processes that drive the progression of the specific disease linked to the social factor of
Socioeconomic Gradients in Childhood Asthma 105 interest. This can allow researchers to draw on basic biomedical research regarding the pathophysiology of a disease and to systematically test which steps within the pathophysiological processes leading to disease are patterned by the social factor. In this way, researchers can begin to build a systematic and convincing argument about the causal chain of biological mechanisms that underlie the links between a social factor and a clinical outcome. Complementing the biological approach, one also needs to understand the processes on the social end that operate to bring a social environment variable such as SES to the level of the individual. Thus one needs to move across different levels of social factors, such as neighborhood influences, family factors, individual characteristics, and test whether these are associated with biological disease processes. In this way, one can start broadly on the social end with a construct such as SES and broadly on the clinical end with an outcome like mortality and systematically establish the links in between that bring the social and the clinical health worlds closer together. The ultimate goal is to lay out a step-by-step mechanistic model of the linear progression from broader social environment to physical health outcome. In the remainder of this chapter, we illustrate this approach by describing research on childhood asthma disparities.
Asthma Asthma is the most common chronic illness in childhood. Approximately 9 million children (12.5 percent) in the United States have had asthma during their lifetime (Dey and Bloom 2005). Asthma is the third-ranking cause of hospitalizations among children (Kozak, Owings, and Hall 2004), resulting in close to 200,000 hospitalizations a year (Akinbami 2006). The economic impact of asthma, in terms of the annual estimated cost of asthma care for children, lies around $3.2 billion (Weiss, Sullivan, and Lyttle 2000). Asthma is also one of the leading causes of school absenteeism, resulting in 12.8 million missed school days a year (Akinbami 2006).
SES Disparities in Childhood Asthma Importantly, asthma morbidity is not equally distributed across the population. Children with lower-SES backgrounds are twice as likely to be hospitalized for asthma, and to suffer activity limitations due to asthma, than their counterparts (Miller 2000; Simon et al. 2003). More specifically, poor children with asthma are significantly more likely to have greater asthma symptoms, to have more severe asthma episodes, and to be hospi-
106 Biological Consequences of Socioeconomic Inequalities talized for asthma than more affluent children with asthma (Miller 2000; Simon et al. 2003; Wood et al. 2002). Fewer years of parent education also are associated with greater risk of asthma hospitalizations and emergency department visits in children with asthma (Dales et al. 2002; Maziak et al. 2004). At the neighborhood level, similar relationships exist. Neighborhoods with lower income levels and higher unemployment rates have been found to have higher rates of pediatric asthma hospitalizations (Castro et al. 2001; Claudio, Stingone, and Godbold 2006; Goodman, Stukel, and Chang 1998). These relationships are stronger for predicting outcomes among those already diagnosed with disease (morbidity) than for predicting risk of getting the disease among those who are initially healthy (prevalence; Chen, Matthews, and Boyce 2002). Overall, however, there is a pressing public health need to better understand the reasons why disparities by SES in asthma exist: to identify promising targets for interventions aimed at alleviating asthma disparities in our society.
Why Is Low SES Associated with Greater Asthma Morbidity? Increasing health disparities research is being conducted on general biological markers that can be linked to low SES. This work has implicated hormonal pathways such as cortisol, markers of systemic inflammation such as C-reactive protein, and the accumulation of long-term pathogenic mechanisms such as allostatic load (Cohen, Doyle, and Baum 2006; Cohen, Schwartz, et al. 2006; Evans et al. 2007; Evans 2003; McDade, Hawkley, and Cacioppo 2006; Seeman et al. 2004, 2008). However, these studies have largely been conducted in healthy individuals. Hence, though this work is able to shed some light on general biological mechanisms, it may not be as useful for explaining specific disease outcomes, given that different diseases will have different underlying mechanisms. Hence to understand associations of SES with asthma morbidity, we turn to models of the biology of asthma. Biological Models of Asthma Asthma is a disease involving inflammation of the airways. Certain cytokines (chemical messengers of the immune system) are important for orchestrating cellular events related to airway inflammation (Busse and Lemanske 2001; Chung and Barnes 1999). These cytokines are produced by T helper (Th) cells, often in response to an external stimulus such as allergen exposure. Th cells are now recognized to have multiple phenotypes, but the best characterized are Th-1 and Th-2 cells. Th-1 cells generally coordinate cellular immune responses by de-
Socioeconomic Gradients in Childhood Asthma 107 ploying cytokines such as IL-2 and IFN-γ. In contrast, Th-2 cells coordinate what are called humoral responses—that is, those that involve the production of antibodies. These cells do this by inducing B cells to proliferate, and it is the B cells that then secrete antibodies. Th-2 cells release specific cytokines such as IL-4, IL-5, and IL-13, and these Th-2 cytokines have been implicated in asthma. For example, secretion of IL-4 and IL-13 induces B cells to produce IgE antibodies, which initiates an inflammatory cascade leading to airway constriction and mucus production (Bacharier and Geha 2000). IL-5 also recruits eosinophils to the airways and activates them, leading to a late-phase asthma response including airway inflammation and obstruction (Kamfar, Koshak, and Milaat 1999; Ying et al. 1997). What Biological Pathways Relevant to Asthma Are Patterned by SES? To establish the biological mechanisms linking SES and asthma, we tested whether SES could be linked to the various types of inflammatory processes implicated in asthma. As we alluded to earlier, the literature does link SES to systemic inflammatory markers (Hemingway et al. 2003; McDade, Hawkley, and Cacioppo 2006; Owen et al. 2003; Panagiotakos et al. 2005; Phillips et al. 2009). However, these studies typically use healthy populations and measure risk markers, such as C-reactive protein. As far as we are aware, our studies are among the first to establish associations of SES with disease-specific biological markers in patient populations. In a sample of thirty-seven children, age nine through eighteen, recruited from the community, with a physician-diagnosis of asthma, and a comparison sample of thirty-nine healthy children in the same age range with no chronic illnesses, blood samples were collected, and parents were interviewed about family SES in terms of family savings and home ownership. As mentioned, one of the last steps in the biology of asthma exacerbations is the recruitment and activation of cells called eosinophils, which bring about edema, smooth muscle constriction, and mucus production in the airways, resulting in clinical symptoms of asthma like wheezing, chest tightness, and shortness of breath. We found that children with asthma who had lower-SES backgrounds had significantly greater eosinophil counts than those with higher-SES backgrounds, even after controlling for a variety of medical and demographic characteristics (Chen et al. 2006). In contrast, no associations were found among healthy children. We next investigated the immune processes that foster the production and activation of eosinophils—that is, Th-2 cytokines in this same sample. Because cytokines are released only when immune cells are activated, we set up a laboratory model for activating immune cells in vitro and tested
108 Biological Consequences of Socioeconomic Inequalities whether low SES would be associated with the production of IL-5 (which specifically activates eosinophils) after participants’ mononuclear cells were stimulated with a mitogen cocktail. This approach provides a laboratory approximation of what the immune cells might do if they encountered allergens in real life and also allows researchers greater control by equalizing the dose of exposure (to the mitogen cocktail) across all participants’ cells. We found that lower SES was associated with significantly greater stimulated production of IL-5 in children with asthma (Chen et al. 2006), suggesting that even if the exposure is equivalent, children with asthma and lower-SES backgrounds will exhibit heightened inflammatory responses to a stimulus compared to children with asthma but higher-SES backgrounds. Can SES Effects Be Seen at the Level of the Genome? Thus far, we have highlighted research that focuses on what cells do, such as produce cytokines, and whether this function can be potentially shaped by SES. Another question is what regulates the function of these immune cells and whether this process may also be patterned by SES. To better understand these mechanisms, we focused on the genomic processes that control the production of proteins secreted by immune cells. Proteins such as cytokines are made by immune cells when the genes coding them get switched on by transcription factors. Transcription factors are molecules that serve as conduits between a cell’s environment and the genes that govern its behavior. For example, when a receptor on an immune cell registers the presence of external stimuli (such as bacteria), it initiates a signaling cascade that results in a specific transcription factor migrating to the cell’s nucleus. There it binds to a specific segment of DNA and switches on a gene that codes for a certain protein, such as a cytokine. Dysregulation of a number of transcription factors has been implicated in asthma inflammation (Barnes and Adcock 1998; Busse and Lemanske 2001). Bringing this biological understanding back to the social world, we asked the question of whether we could document associations between SES and the activity of transcription factors that regulate inflammation. Other aspects of social adversity have been linked in recent studies to gene expression profiles (Cole et al. 2007; Lutgendorf et al. 2009; Miller et al. 2008). However, never has this understanding of genomic pathways been used as a way to understand SES and health relationships. We conducted genome-wide transcriptional profiling on the immune cells of a sample of thirty-one children ages nine through eighteen recruited from the community with a physician diagnosis of asthma and either a low-SES or a high-SES family (bottom or top 15 percent of the
Socioeconomic Gradients in Childhood Asthma 109 distribution in terms of both family income and parent education). Bioinformatic analyses suggested that across overexpressed genes, low-SES children with asthma showed a pattern consistent with significantly increased activity of NF-κB and decreased activity of CREB transcriptional signaling compared to high-SES children with asthma (Chen et al. 2009). NF-κB is a transcription factor that mediates induction of pro-inflammatory cytokines, whereas CREB is a transcription factor that relays messages between catecholamines and immune cells via beta 2 adrenergic receptors. This was the first study to provide evidence that even molecular pathways involved in the regulation of inflammation can be shown to be patterned by SES in children with asthma. Note that in all of these studies, we find more robust associations of SES with asthma-relevant biological markers when using resource-based SES measures, such as family income or savings. This suggests that material resources—more so than status or prestige (such as parental occupation)— may play an important role in affecting the physiological processes underlying diseases such as asthma. Overall, these findings demonstrate that if one takes an approach of systematically investigating the steps in the pathophysiology of a disease, one can develop a working model of the specific biological processes that may be susceptible to alterations based on changes in the broader social environment. Here we document this specifically with respect to asthma and show that SES is linked to peripheral markers related to asthma inflammation such as eosinophil counts, cellular processes like the production of Th-2 cytokines that govern inflammation, and genomic pathways that coordinate these cellular responses such as the activation of transcription control pathways that regulate inflammation such as NF-κB. Importantly, all of these relationships are in a direction consistent with the clinical phenomenon of lower-SES children experiencing greater asthma morbidity.
What Social Processes Govern SES Effects on Asthma Inflammation? As mentioned, in developing plausible models linking the social environment and physical health outcomes, it is important to identify the intervening social processes that explain how distal variables such as SES register at the level of the individual. To do this, one needs to understand the various neighborhood, family, and individual levels of social factors linked to SES. We present several examples of such an approach in the context of asthma.
110 Biological Consequences of Socioeconomic Inequalities What Individual Factors Link SES to Asthma Inflammatory Processes? One of the most promising psychosocial explanations at the individual level for the SES and health relationship is the notion that lower-SES individuals experience greater stress in their daily lives, which in turn takes a physiological toll, affecting physical health outcomes (Adler et al. 1994). Low SES has been associated with more frequent exposure to stressful life events (Attar, Guerra, and Tolan 1994; Brady and Matthews 2002; Garbarino, Kostelny, and Dubrow 1991; Hatch and Dohrenwend 2007), and children who live in low-SES environments report greater subjective stress (Goodman et al. 2005). Moreover, stress has been found to form one pathway between SES and health (Cohen, Kaplan, and Salonen 1999; Khang and Kim 2005; Lantz et al. 2005). Asthma itself has been linked to psychological stress (for a review, see Chen and Miller 2007). Clinically, experiencing stressful life events is associated with an increased risk for a subsequent asthma attack in children with asthma (Sandberg et al. 2000). Daily diary studies have shown that daily stress is associated with poorer lung functioning as well as increased reports of asthma symptoms (Smyth et al. 1999). In young children, perceived stress in parents has been linked prospectively to children’s risk of developing wheezing in the first two years of life (Wright et al. 2002). In turn, research has linked experiences of stress to immune responses relevant to asthma. For example, college students with asthma showed a greater immune response to allergen challenge during high-stress periods (final exams) than in low-stress (no exam) periods (Liu et al. 2002). High school students with asthma showed greater IL-5 production after high stress (post-exam) than students who were healthy and experiencing stress (Kang et al. 1997). Similarly, atopic individuals show a decreased Th-1/Th-2 ratio in response to exam stress relative to healthy control participants (Hoglund et al. 2006). However, missing from this earlier research is an integration of stress into the broader social environment. That is, it remains unclear whether stress can provide one explanation for why low SES comes to affect asthma inflammatory pathways. We tested this idea with respect to how children with different SES backgrounds perceive the events in their social world. In previous work, we hypothesized that children who grow up in low-SES environments develop a pattern of thinking about the world as a threatening place that requires constant vigilance based on previous negative life experiences (Chen and Matthews 2003). We hypothesized that these children would develop a tendency toward interpreting even ambiguous events as threatening. That is, when outcomes are ambiguous, such as a teacher asking to speak with you after class, low-SES children will be more
Socioeconomic Gradients in Childhood Asthma 111 Figure 4.1 SES and Inflammatory Responses
SES
β = −.40**
Threat appraisal
βs= .28† − .37*
IL-5 Eosinophils
Source: Authors’ compilation based on Chen et al. (2006). Note: Indirect pathway from SES to asthma inflammatory pathways signficant at p 100 mg/dL). Categories of BMI used are both obese class I and above (BMI >30 kg/m2) and obese class II and above (BMI >35 kg/m2; see table 5.5).
Statistical Models The first type of models we fit were standard cross-sectional regressions controlling for potential confounding factors:
yi = b0 + b1log incomei + b2xi + εi,
(5.1)
where yi is the outcome of interest (BMI or metabolic syndrome score) for individual i, b0 is the intercept, b1 is an estimate of the association between the primary exposure of interest and the dependent variable, b2 is a vector
Cardiovascular Consequences of Income Change 133 of estimated effect of observed control variables xi, and εi is the error term that we are concerned may contain omitted confounders as discussed earlier. Models of the association between BMI and the metabolic syndrome in years 1992, 1995, 2000, and 2005 were examined separately in a series of gender stratified regression models for both BMI and the metabolic syndrome (we only show results from 1992, as associations did not differ substantially across time). The independent variables included in five different models are as follows. Model 1a is log income, age, age squared, married indicator, divorced indicator, number living in the household, study site indicator, and race indicator. Model 1b is identical to model 1a except that it replaces the variable log income with a covariate for education level in 1992. Model 1c is identical to model 1a except that it replaces the variable log income with a covariate for parental education. Model 1d adds to model 1a a covariate for education level in 1992. Model 1e adds to model 1d a covariate for parental education. The next set of models exploit the longitudinal nature of the data to estimate individual-level fixed-effect models using years 1992, 1995, 2000, and 2005 to examine the association between changing BMI or the metabolic syndrome and changing log income.
yit = b1log incomeit + b2xit + b3yeart + μi + νit,
(5.2)
where yit is the outcome of interest (BMI or metabolic syndrome score) for individual i at time t, and b1 is again the main coefficient of interest. The key difference between this model and the equation 5.1 models (aside from now including multiple years of data in the analysis) is that the error term is decomposed into two parts: εit = μi + νit. The former part μi represents time-invariant (fixed) unobserved characteristics of individuals such as genetics and preferences formed early in life. To the extent that these are correlated with socioeconomic indicators such as income, their omission from the vector of observed control variables x may cause bias in b1. For example, higher time discount rates (focused more on present gratification than the future) may lead to both lower income and higher obesity, resulting in overly negative estimates of the obesity effects of income. To address this concern, we estimate equation 5.2 using an individual-level fixed-effects estimator that removes from the error term those unobservables μi that are fixed across survey years (this is equivalent to adding to the model a dummy variable for each person). This fixed-effects model can roughly be interpreted as capturing the effects of changes over time in income on changes over time in cardiovascular risk. Essential to the model is the assumption that time-varying confounders (such as changes in mar-
134 Biological Consequences of Socioeconomic Inequalities ital status) are included in the xit vector, and thus the remaining error component νit is unrelated to income. Although these fixed-effects models will be more robust to omitted variables bias, a drawback is their lower efficiency and hence wider confidence intervals if omitted variables bias was not indeed a problem. To examine this trade-off, we use Hausman tests to compare these fixed- effects models against a null specification in which the μi were instead modeled as random effects. If the models yield statistically similar estimates, then we would prefer the random-effects models to the less efficient fixed-effects approach; if the models yield different estimates, then we reject the random-effects models as biased (and implicitly also reject the cross-sectional models in equation 5.1). For all but one model using the CARDIA data, we rejected (p .05) for the household income and BMI model in women, but for consistency of presentation we report only the fixed-effects specifications. Similarly, for all but one model using the National Longitudinal Survey of Youth (NLSY) data we rejected (p .05) for the household income and BMI model in men, but, again for consistency of presentation, the NLSY tables include only fixed-effects results. We also examined the fixed-effect models with interaction terms between income and indicators for having become married, having become divorced, or having had an income increase rather than decrease since the last wave of observation. We found none of these interaction terms to be statistically significant at p