International Journal of Epidemiology, Volume 33, Number 3, pp. 618-619
IJE vol.33 no.3 © International Epidemiological Association 2004; all rights reserved.
Book Review |
Health Inequalities: Lifecourse Approaches. Davey Smith G (ed.). Bristol: The Policy Press, 2003, pp. 592, £25.00 (PB) ISBN: 1-86134-322-1, £55.00 (HB) ISBN: 1-86134-323-X.
By my count, George Davey Smith (GDS)—the Chair of Clinical Epidemiology in the Department of Social Medicine and Head of the Division of Epidemiology at the University of Bristol—has published well over 400 scientific publications, books, and chapters, around 25–30 per year since receiving his epidemiological training. Thus, there are few epidemiologists who are not aware of his work, although with such a prolific corpus of publications, it is likely that each of our views of GDS is a bit like the blind men and the elephant.
In reprinting 39 papers in a single location, Health Inequalities: Lifecourse Approaches allows us to see some common threads in this body of work. The volume consists of these papers, co-authored with 42 others, commentary by GDS, some wonderful photographs and illustrations, and a short essay occasioned by a traffic accident in Nicaragua. The common threads, as the title suggests, are health inequalities and the life course. To accomplish this, GDS provides us with papers that consider patterns of mortality, cardiovascular diseases, cancers of many sites, respiratory disease, diabetes and insulin resistance, infant mortality, and other outcomes. This is no simple catalogue of risk factors for these outcomes, but a thoughtful, and very often pointed, consideration of these outcomes via a look at health and health inequality trends, measurement of inequalities in health, area-based and geographical analyses, measurement of deprivation, race/ethnicity and culture, the role of psychosocial factors, methodology, voting patterns, historical paradigms for understanding health and development, early origins of disease, and, of, course, the life course.
Most of these papers that have been reprinted in this volume have already appeared in peer-reviewed publications, so it is not useful to consider them individually, searching for a flaw here and there. Instead, it is more valuable to consider whether these papers hang together in a coherent fashion. Certainly, there is a consistent focus on understanding both the broad pattern as well as the specific aetiology of inequalities in health. GDS brings the lens of specificity—historical, social, and pathobiological—to his analyses while concluding at the same time that whatever a society has to offer with respect to health damaging and health enhancing circumstances and exposure will be social patterned and produce health inequalities. This perspective, which admittedly needs more specificity, provides a powerful link between specific patterns of exposure, and their determinants, and the population trends in potential outcomes. While making such connections between social processes that determine exposure and population health trends is very attractive, the papers provide for the most part only an outline of how such connections might be manifest. Still, it is an important perspective that is only beginning to be applied, and GDS deserves considerable credit for promoting such a view.
Another source of coherence is in the application of the life course perspective. As we view these papers covering the period since 1988 we indeed see this perspective evolving. What is particularly impressive is the application of the life course perspective in a way that involves historical processes, specific aetiological pathways, consideration of the diverse pathobiology of common, and not so common, outcomes, and an increasing development of specific life course hypotheses of ‘what’ and ‘when’ might be responsible for specific outcomes.
In the beginning of this volume, GDS provides us with a 36-page introduction that in many respects summarizes much of what he values in the life course perspective on health inequalities. Providing a bit of history on the development of his work in this area as well as an overall perspective, it might be read both before and after reading the other chapters in the book. It seems reasonable then to consider some of the assertions and conclusions that are drawn. Quite rightly, GDS points out that one of the consequences of the observations of Forsdahl, Barker, Kuh and Ben-Shlomo, GDS and others has been an increasing realization that an understanding of life course effects would require ‘... detailed data for periods covering the entire lifecourse ... [in order to] ... to identify the contribution of exposures acting at particular time periods.’ This, of course is no easy task, and we see the grasp for data on the ‘entire life course’ in GDS' use of cohorts that begin with initial data collection as far back as the 1930s, permitting 5–6 decades of observation. Of course, these data are never quite up to the challenges offered by contemporary knowledge, theory, or measurement, but one is struck by the clever and sensitive use of these historical cohorts in the papers in this collection. Still, one wonders to what extent the time-specific patterns of association between childhood and adult measures and various outcomes, and evidence for preponderance of effects from one time period or another, to some extent reflect inadequacies in the measurement of life circumstances and social position at various time points.
As GDS points out, the life course approach made considerable contributions by emphasizing that the ‘clustering of advantage and disadvantage across the life course was key.’ However, combining the quote from the previous paragraph with this last one, indicates a bit of a tension. While the life course approach as promoted by GDS and others places considerable emphasis on estimating independent effects of one time of exposure or another—a critical issue being the calibration of the impact of time-specific exposures—the notions of clustering make such a task difficult if not impossible. Thus the same problems of measurement error and residual confounding that GDS has so importantly pointed out in other areas apply to life course approaches as well. One imagines that the same volume written 10 years from now will, most likely with contributions from GDS, have come up with both new sources of data and analytical techniques that inform this tension. Alternatively, one could imagine an epidemiology of life course trajectories that investigates clustering, both cross-sectional and longitudinal, as an aetiological factor.
Twenty per cent of the introduction is devoted to the issue of general or specific explanations of health inequalities, with GDS contrasting approaches that consider health inequalities as representing the outcome of a general process versus the need for specificity with a case-by-case analysis required. This, of course, is not a new issue and one is certainly aware of the use by some of the specificity argument to deny the causal effects of smoking—how could smoking possibly be related to so many diseases with specific and diverse pathologies? GDS, however, takes a different tack by focusing on several questions—have inequalities always existed, are they the same in contemporary populations, and are there exceptions? As you might expect from his focus on historical and life course contingent processes, the answer to him is ‘no, no, and yes.’ Without denying the fact that health inequalities are not fixed in time in magnitude or direction, but depend on the specific social and economic distribution of their casual risk factors, in this reviewer's opinion there is less to the argument than meets the eye. Indeed, GDS's own comments support my view. On the one hand he argues for extreme specificity, and on the other hand states that there is ‘a background in which most environmental exposures [broadly considered by him] that are known to be—or in a common-sense way appear to be—noxious, are more concentrated on people in adverse social circumstances. ... In this sense there will be a reasonably general association between adverse social position and health, but one that works through social processes leading to the exploited, excluded and oppressed receiving the worst of what life has to offer. ... (pp. xxvi–xxvii).’ Thus, there can be both specific and general processes operating in the production of health inequalities.
In the introduction, GDS takes those to task who argue for a role of psychosocial factors in inequalities in health. Leaving aside the issue of what exactly are psychosocial factors, he points out that he started with a strong belief in considering such factors, but now rejects the view that such factors have a direct [emphasis his] aetiological role in disease, not just inequalities in disease. Some of his arguments are based on issues of confounding, poor measurement, and reporting bias in studies that use ‘soft’ outcomes. These are hardly issues that are absent in some of the studies of life course determinants or, indeed, many other epidemiological studies. In addition, he argues that ‘much of the literature in the field’ argues that psychosocial factors are viewed as direct aetiological agents. While there are certainly some who make such arguments, as one who has worked in this area for a quarter of a century, it is an overstatement to say that ‘much’ of the work argues for direct psychosocial effects, with many researchers presenting models that link psychosocial factors to many of the same structural conditions and experiences that the life course perspective examines. In addition, there are certainly well-designed studies of ‘hard’ outcomes with good measurement of psychosocial factors that indicate their impact on health outcomes. Given the oft-reported association between socioeconomic status and living conditions and psychosocial factors, it is certainly reasonable not to reject that they are part of the process by which social and economic divides become health divides. Indeed, the life course perspective may be very useful here in elucidating the various ways in which time- and stage-dependent conditions generate psychosocial disadvantage. Whether interventions focused on job control or depression, for example, can reduce health inequalities becomes as much an issue subject to test based on empirical findings and feasibility as does an intervention based on a life course perspective.
Davey Smith does not shy away from controversy, and the papers included in this collection represent his rejection of ‘true-believer’ answers to what are solid scientific issues. His work shows sophistication in choice of topic and analysis, and his contributions to the evolving life course perspective on health inequalities are very significant. I highly recommend this book as a window onto this evolving area of research, seen though the work of a remarkable and productive scientist. GDS does not shy away from the tough questions, and there are many connected with the life course approach to health inequalities. My only regret after finishing this book and reminding myself of some of the impressive studies that it reprints, is that it reports a journey of some 14 years by a remarkable and stimulating epidemiologist and I would not want to wait another 14 years to see the next report of his scientific peregrinations.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||