« How Could the United States Have a "Health Disadvantage"? | Main | Massachusetts Prevention and Wellness Trust Fund: Cutting Health Care Costs through Prevention »



Feed You can follow this conversation by subscribing to the comment feed for this post.

Scientific methods for testing health improvement hypotheses at the population level are not well developed. Thus, the current evidence base relies heavily on often poorly controlled, retrospective, cohort, comparative designs. Many refer to U.S. States as potential laboratories of experimentation providing sufficient homogeneity and sample size to make more rigorous studies possible. Many population-level interventions now devolve to States. Medicaid is one example. With balanced randomization (or quasi-randomization), population-level interventions could be tested with 25 assigned as intervention States and 25 as controls. Large within-State populations would provide relatively precise estimates of effect so an overall sample size of 50 might provide more useful information than currently available. Feasible? Worth a try?

Thanks for this post. I'm also intrigued by the policy recommendation that implies that "stimulating a national dialogue" will naturally bubble up into policy action, the hopeful assumption of many similar reports (including 2012's IOM report on obesity prevention).

The comments to this entry are closed.

Recent Posts