Five Steps to a CER Project Portfolio

In a recent article in the New England Journal of Medicine, Harold C. Sox, MD, chair of the Institute of Medicine (IOM) committee to set national priorities for comparative-effectiveness research (CER), and colleagues¹ proposed a process through which the government could begin to fund research to improve medical decision making. The authors acknowledge that the great challenge, in developing a slate of CER projects, will be to maintain focus in working through the 100 priority items identified by the IOM. The authors recommend a five-step process. First, determine funding strategy. The traditional biomedical-research model would be to invite proposals on any of the 100 high-priority topics and award grants to the strongest proposals, but as the authors point out, this would skew toward the research interests of the individual investigators, and not necessarily toward the specific priorities of IOM. The authors believe that the CER program and its stakeholder advisory board—not the research community—should decide on a coordinated portfolio of research on priority topics, infrastructure enhancement, and studies of translation and adoption. Second, perform further winnowing. The authors propose that the CER program develop a portfolio chosen from the top 25 IOM topics by applying the prioritization criteria of the IOM and the Federal Coordinating Council. The program should then determine what type of evidence should be obtained by uncovering evidence gaps (through systematic literature reviews) and performance gaps (using data from insurance claims or electronic health records). Third, focus on evidence gaps. To keep the program within budget, research methods should be selected by the program to fill evidence gaps for specific topics. Citing wide variation in costs for CER (whether the study method is clinical trial, observational study, or qualitative research), the authors believe that the CER program should match the research methods to the research question. The key challenge will be to envision what type of evidence will be required to convince key decision makers to adopt the findings in practice. Fourth, create a balanced, high-impact portfolio. The authors recommend a research-topic portfolio that balances the distribution of topics, outcomes, and target populations. Several complementary studies of each high-priority research question that will ensure a body of evidence strong enough to act upon might be necessary. Fifth, evaluate results and promote transparency. Evaluating progress and reporting to the public should be regularly undertaken. To this end, the authors recommend large-scale, ongoing observational research to measure CER’s effects on clinical practices and patient outcomes.