An examination of the methodologic issues of assessing evidence to inform policies and programs on the social determinants of health argues that randomized controlled trials (RCTs) are inadequate because randomization in studies of social factors is not always ethical, feasible, desirable or useful.
An emphasis on evidence-based medicine has devalued information gained from sources other than RTCs, even in nonclinical research. In addition to its logistical challenges, randomly assigning participants in studies of social factors is problematic because it can create ethical issues and lead to a lack of generalizability. The health effects of social factors can be difficult to document in a controlled fashion, especially for “upstream” factors such as income and education, which influence health in complex ways over a long period of time. Interaction between upstream and downstream factors (such as physical activity), further complicates measuring their effects. RCT therefore cannot be treated as the "gold standard" for evidence.
The authors recommend working with the upstream nature of most social determinants of health, linking existing to new research along causal pathways and employing diverse methods. Standards for evidence to guide social policies should be more diverse than those for medical interventions, but should be no less rigorous.