Commentary: How does socioeconomic disadvantage during childhood damage health in adulthood? Testing psychosocial pathways

The multi-national comparative study by Power and colleagues 1 provides strong evidence for a significant correlation between low socioeconomic status (SES) and health risk in adulthood. As expected, adult SES was related to obesity and smoking behaviours. However, low childhood SES was also significantly related to adult health risk, even after adjusting for contemporaneous adult SES. The evidence supports a long-term association between childhood SES and adult health that is not simply due to the life course continuity of low SES. To obtain such broadly consistent findings in seven samples from six countries is impressive, more so, because of the many factors that work against this outcome (e.g. historical, cultural, sampling, and measurement differences across studies). The challenge is now to explain the reasons for these disparities, ideally by mapping the developmental pathways leading from childhood socioeconomic adversity to poor adult health. Currently little is known about mediating pathways, mainly because most datasets documenting health inequalities have limited data to explore possible explanations for these inequalities. 2 The key limiting factor has been the lack of longitudinal birth cohort studies that contain detailed life history data from birth to adulthood and direct physiological assessments of health. To date, empirical work has been limited by several factors. First, the majority of research on the psychosocial correlates of physical health has been crosssectional. In contrast, longitudinal methods and within-subject comparisons offer a stronger strategy for inferring developmental influences. Second, many available longitudinal studies have suffered high attrition rates, selectively losing unhealthy participants. Third, reported associations between psychosocial risk factors (e.g. psychiatric symptoms) and health outcomes (e.g. obesity) often neglect to control for confounding factors (e.g. medical illness). Fourth, and more generally, epidemiological studies with good health data tend to have weak psychosocial data and vice versa. Gold standard measurement of psychosocial variables is paramount, as are