Comparison of RCTs and observational studies: implications for benefit assessment in health care and health policy decision-making in light of effect size, PICO and certainty of evidence

Tags: Oral
Toews I1, Anglemyer A2, Meerpohl JJ3, Bero L4
1Institute for Evidence in Medicine, Medical Center, University of Freiburg, 2Global Health Sciences, University of California, San Francisco, California, 3Institute for Evidence in Medicine, Medical Center, University of Freiburg, Germany, 4Charles Perkins Centre and Faculty of Medicine and Health, School of Pharmacy, The University of Sydney, Sydney

Background: Randomised controlled trials (RCTs) are considered the gold standard for the assessment of treatment effects. While observational studies are prone to bias and confounding by known and unknown prognostic factors they can be conducted more easily and rapidly even with limited resources. Stakeholders in health policy decision-making feel the need to make innovative interventions available more rapidly and more widely even though evidence from RCTs is underway and not yet available. Alternative routes for benefit assessments and consideration of evidence from studies other than RCTs are proposed for decision making in health policy. This has the potential to create a conflict with the principals of evidence-based medicine and might create the burden of increased pressure and uncertainty for all stakeholders in the health care system: patients, health care providers, and political decision makers. Therefore, the National Association of Statutory Health Insurance Funds (GKV-SV) in Germany seeks answers to questions about a) the feasibility of RCTs when interventions are already widely available and b) the value of evidence from observational studies compared to RCTs in benefit assessment.

Objectives: To assess the impact of study design (RCTs versus observational studies) on effect measure and to explore whether there are variables that might explain differences and similarities identified, e.g., observational study designs, risk of bias assessments, pharmaceutical vs non-pharmaceutical studies etc..

Methods: As commissioned by the GKV-SV we are a) updating a Cochrane Overview of Reviews comparing the effect estimates of RCTs and observational studies on the same PICO question [1] .

Results: In the initial review, 15 reviews were included and showed that effect estimates were similar between RCTs and observational studies examining interventions. In the updated review titles and abstracts of 6599 records were screened, full texts of more than 75 studies are now scrutinized for eligibility. Date of search is 13 Dec. 2019. The results will provide an up-to-date overview of the evidence comparing effects sizes under consideration by type of intervention, year of publication and other relevant factors. The results can inform future decisions on systematic review design, i.e. inclusion of observational studies. Also, our results will serve as information for processes in German health policy decision-making. The GKV-SV will consider the results for their positioning in health care decision-making. We will present full results by the time of the conference.

Conclusions: We are looking forward to draw conclusions about effect sizes and discuss the value of different study designs for intervention studies in health care decision-making. Further, we will discuss whether and under what conditions benefit assessments and reimbursement decisions can be based on RCTs and observational studies or both.

Anglemyer et al. Cochrane Database Syst Rev, 2014(4): MR000034.