post by Paul Kelleher
Earlier this week I contrasted a recent quote of Michael Marmot's about the source of the social gradient in health with findings from a 2010 study that conflict with his appraisal. To review, Marmot was quoted as saying that "[Lifestyle factors] accounted for about a third of the gradient." Here Marmot seems to refer to the gradient found in the original Whitehall Study (Whitehall Cohort I), where those at the top of the occupational hierarchy had the best health, while those in the occupations just below the top had the second-best health, and so on down the line. Yet 2010 research by Stringhini et al., which analyzed data from the follow-up Whitehall II study, seemed to show that when health behaviors are assessed more accurately, they can explain the vast majority of the gradient that Marmot found and attributed to other causes.
In a comment on that post, Daniel Goldberg drew my attention to a follow-up article by Stringhini et al. from February 2011. I'll explain in a moment why this second article will likely be very important to the research on the health gradient. But the thing that excited me most was that Marmot himself is listed as a co-author of this second article. This suggested to me that the conversation about the social gradient is unfolding as it should, with the original pioneer joining forces with the new iconoclasts to push the debate forward. That is a noteworthy development.*
From a research perspective, the February 2011 article is exciting because it claims to show that not all health gradients are created the same--that is, not all pronounced social gradients in health are caused by the same factors. As we have seen, Whitehall II appears to show that the gradient in health between individuals of different occupational status is mediated by a gradient in risky healthy behaviors. Those at the bottom of the occupational hierarchy engage in more risky behavior, and this explains 75% of the association between occupational position and all-cause mortality. After first reading Stringhini et al.'s report of these findings, I said that "the question" must now become, Why are risky health behaviors themselves socially patterned?
But the latest Stringhini et al. article may prove me wrong. That article also reports on a second study, the GAZEL study of French national gas and electricity employees from 1989 to around 2005. In GAZEL, there was a pronounced occupational gradient in health, with those in the lowest occupations having the worst health and those at the top having the best. In fact, GAZEL workers in the lowest status occupations actually had a 1.94 times higher risk of death than workers at the top, while in Whitehall II that hazard ratio was 1.62. And yet when the researchers crunched the data to determine the role that health behaviors played in creating the GAZEL gradient, they found that such factors could explain only 19% of it. This doesn't mean that the GAZEL cohort doesn't engage in risky behaviors. In fact they do, and to a higher degree than members of the Whitehall II cohort. It's just that everyone, from the top of the occupational hierarchy to bottom, engages in risky health behaviors to a similar degree in GAZEL. So something else must be causing the pronounced GAZEL gradient in health between the top and the bottom. But what? The authors offer a list of hypotheses that echoes the original list Marmot made when he thought the Whitehall gradient couldn't be explained by the distribution of health behaviors:
In this study, we show that, not surprisingly, health behaviours contribute little to socioeconomic inequalities in mortality when their social patterning is weak. Other factors are likely to play an important role in the SES-mortality association in both cohorts. Material deprivation or financial insecurity, work stress and work environment, psychosocial factors such as job control or social support, or differential access to health care may be other possible mechanisms through which SES influences health [57–70]. Moreover, GAZEL includes both blue and white-collar workers; it is therefore possible that physical occupational hazards and working conditions contribute more to socioeconomic differences in health in this cohort [67,71]. However, restricting the analysis in GAZEL to only the white-collar workers did not substantially change the results.
At this point I simply quote and register agreement with the authors' own driving motivation: " A better understanding of these mechanisms is essential in order to identify targets for intervention aimed at reducing social inequalities in health." Since not all gradients are created in the same way, we cannot assume that they will be eliminated in the same way.
*The follow-up article by Stringhini, et al. does not withdraw the claim that health behaviors account for the majority of the gradient found in Whitehall II. This leaves it curious why Marmot--a co-author on the new article--evidently reiterated to the reporter recently that lifestyle factors can explain only one-third of the gradient. I won't speculate as to what was going on there.