Good research practices for comparative effectiveness research: defining, reporting and interpreting nonrandomized...

Health insurers, physicians, and patients worldwide need information on the comparative effectiveness and safety of prescription drugs in routine care. Nonrandomized studies of treatment effects using secondary databases may supplement the evidence based from randomized clinical trials and prospective observational studies. Recognizing the challenges to conducting valid retrospective epidemiologic and health services research studies, a Task Force was formed to develop a guidance document on state of the art approaches to frame research questions and report findings for these studies.

The Task Force was commissioned and a Chair was selected by the International Society for Pharmacoeconomics and Outcomes Research Board of Directors in October 2007. This Report, the first of three reported in this issue of the journal, addressed issues of framing the research question and reporting and interpreting findings.

The Task Force Report proposes four primary characteristics-relevance, specificity, novelty, and feasibility while defining the research question. Recommendations included: the practice of a priori specification of the research question; transparency of prespecified analytical plans, provision of justifications for any subsequent changes in analytical plan, and reporting the results of prespecified plans as well as results from significant modifications, structured abstracts to report findings with scientific neutrality; and reasoned interpretations of findings to help inform policy decisions.

Comparative effectiveness research in the form of nonrandomized studies using secondary databases can be designed with rigorous elements and conducted with sophisticated statistical methods to improve causal inference of treatment effects. Standardized reporting and careful interpretation of results can aid policy and decision-making.