Re: “Reconsidering the Role of Social Disadvantage in Physical and Mental Health: Stressful Life Events, Health Behaviors, Race, and Depression”

Mezuk et al. (1) tested and provided empirical support for the hypothesis that engaging in unhealthy behaviors modifies the impact of social stress on the risk of depression and that the relation between stress, behavioral coping, and depression operates differentially by race. This hypothesis was initially proposed and tested by Jackson et al. (2, 3) as a potential explanation for the ostensible epidemiologic paradox of lower rates of psychological distress among blacks compared with whites, despite greater exposure to social disadvantage and higher rates of physical morbidity. We recently conducted a third and more exhaustive empirical test of the hypothesis and failed to replicate these findings (4). Here, we highlight a number of theoretical and methodological concerns raised by Mezuk et al. (1).

The proposition that coping behaviors might have different implications for physical illness and psychopathology (hereafter referred to as the Jackson hypothesis) is an interesting one that is worthy of investigation. However, as an explanation of the differences in the prevalence of depression in blacks and whites, the theory is problematic. For the Jackson hypothesis to plausibly explain the “paradox,” it must be true that 1) the physiologic impact of unhealthy behaviors on depression is differential by race and/or 2) unhealthy behaviors protect against depression in both blacks and whites, but blacks are more likely to engage in them, especially under conditions of stress. There is no established empirical or theoretical basis for the first supposition, and the authors do not explicate a mechanism by which the physiologic impact of health behaviors might differ by race. The second supposition is inconsistent with substantial empirical evidence of 1) comorbidity between substance disorders, obesity, and depression (5–7) and 2) levels of substance use and dependence among blacks that are lower than or comparable to those among whites (8–11). Of the behaviors being considered, only unhealthy eating is consistently more prevalent among blacks in the epidemiologic literature (12, 13). In our data, however, high body mass index was only protective against depression among blacks who reported no stressful life events (4), which is inconsistent with the theory's prediction of protective effects at high levels of stress.

Mezuk et al. (1) also suggested that their findings may reflect in part the inadequacies of standard measures of social disadvantage in capturing the uniquely chronic and noxious stresses blacks experience. If the measures underestimate stress and disadvantage among blacks, however, the prevalence of both depression and poor health behaviors should be higher in blacks than in whites at a given level of stress. These patterns were not presented in the article by Mezuk et al. (1), and when examined in our analysis, they were not evident (4).

Finally, Mezuk et al. (1) analyzed stressful life events as a continuous measure, assuming a linear relation between stress and depression among both blacks and whites. This assumption can and should be tested in their data; in our data, we observed that, among blacks, a count of stressful life events was not linearly related to depression. However, when modeled (incorrectly) as such, engaging in unhealthy behaviors did appear to be protective against depression at high levels of stress (4).

Theories that generate specific testable hypotheses contribute enormously to the field of epidemiology. In addition, substantive explanations for the black/white paradox in depression are an important complement to the methodological interpretations that are more commonly advanced. Further theoretical and empirical attention to this issue is certainly warranted. However, a more promising theory would both posit specific etiologic mechanisms and account for established population-level empirical patterns. The Jackson hypothesis fails to attend fully to the latter and therefore cannot be regarded as a “comprehensive explanation” for the paradoxical racial patterning in physical and mental health outcomes (1, p. 1243).