What comparative effectiveness research is needed? A framework for using guidelines and systematic reviews to identify evidence gaps and research priorities

Abstract

The authors developed and tested a framework for identifying evidence gaps and prioritizing comparative effectiveness research by using a combination of clinical practice guidelines and systematic reviews. In phase 1 of the project, reported elsewhere, 45 clinical questions on the management of primary open-angle glaucoma were derived from practice guidelines and prioritized by using a 2-round Delphi survey of clinicians. On the basis of the clinicians' responses, 9 questions were classified as high-priority. In phase 2, reported here, systematic reviews that addressed the 45 clinical questions were identified. The reviews were classified as at low, high, or unclear risk of bias, and evidence gaps (in which no systematic review was at low risk of bias) were identified. The following comparative effectiveness research agenda is proposed: Two of the 9 high-priority questions require new primary research (such as a randomized, controlled trial) and 4 require a new systematic review. The utility and limitations of the framework and future adaptations are discussed.

Figure 1

Key steps and stakeholder involvement in a framework that combines practice guidelines and systematic reviews to identify evidence gaps and prioritize comparative effectiveness research. * Patients and consumers were not involved in this project but could be engaged in future adaptations.

Figure 2

Systematic reviews on the management of primary open-angle glaucoma that satisfied each methodological quality criterion. * Not applicable because the reviews did not assess methodological quality or did not perform a meta-analysis, † The denominator was the 32 systematic reviews with ≥1 meta-analysis.

Appendix Figure 1

Summary of evidence search and selection. POAG = primary open-angle glaucoma; SR = systematic review. * One systematic review is associated with 4 articles. A review could address >1 clinical question depending on the breadth of the review topic. Consequently, the number of clinical questions does not match the number of reviews.

Appendix Figure 2

Example of pooling data from treatment groups. IOP = intraocular pressure. Reproduced from Holmstrom and colleagues (39) with permission of LibraPharm Ltd.

Appendix Figure 3

Example of an incorrect formula used to estimate the SD of the percentage of change in IOP from baseline. IOP = intraocular pressure. Reproduced from Zhang and colleagues (61) with permission from the BMJ Publishing Group Ltd.

Appendix Figure 4

Example of transcription and use of an incorrect formula to estimate the SD of the percentage of change in IOP from baseline. IOP = intraocular pressure. Reproduced from Cheng and colleagues (19) with permission. The publisher for this copyrighted material is Mary Ann Liebert, Inc. Publishers.