When systematically comparing procedures and therapies in the setting of a comparative effectiveness review (CER), the evidence base often includes single group studies, those that evaluate a single intervention given to all subjects included in the study design. The utility and limitations of single group studies to resolve clinical questions that are inherently comparative in nature should be described clearly.
The purpose of this paper is to review the use and interpretation of single group studies in primary clinical research and to summarize current practices to using single group studies in CERs conducted by Agency for Healthcare Research and Quality (AHRQ) Evidence-based Practice Centers.
We extracted information about the use of single group studies in all published CERs prepared for the AHRQ Effective Health Care (EHC) Program through January 2012. Summary descriptive statistics across the different reports were tabulated.
Of 33 eligible reviews, 21 included single group studies. Ten used single group studies to report only harms, 2 reported only nonharm effects, while the remaining 9 reported both harms and nonharm effects. Ten of 21 did not provide the rationale for including single group studies. Of the 11 that did, the reasons cited included provision of supplementary long-term effect data on surgical interventions, hypothesis generation, and to address a paucity of existing comparative studies. Of the 12 CERs that did not include single group studies, 3 provided reasons for excluding these types of studies. They included specific concerns about confounding, avoidance of bias, and that sufficient data were expected from comparative studies. The terminology used to describe single group studies and their subtypes is not consistent across CERs.
The reporting of inclusion or exclusion of single group studies in the EHC Program's CERs is suboptimal. Our review of published CERs to date indicates that single group studies are commonly included in CERs, but the rationale for including them is not consistently reported, and the methods relevant to their use are not clearly defined. Clarity and transparency in the rationale for including or excluding single group studies in CERs should be promoted. A working group should be convened to develop guidance on the circumstances under which single group studies should be included or excluded from a CER, to discuss how they should be integrated with evidence from other designs and how they may inform the strength of evidence assessment.