A review of core outcome sets (COS) developed for different settings finds there is a subset of outcomes relevant for both research and routine care

Abstract

Objectives: To compare the outcomes selected for the same condition in core outcome sets (COSs) for research with those in COS for the routine care setting.

Methods: A sample of COS was created from the most frequent five health areas within previous systematic reviews of COS for research and COS for routine care. Outcomes were extracted and categorized using an outcome taxonomy. Frequency of outcome domains included within COS were analyzed in subgroups according to research or care setting, patient involvement in COS development and health area. Matched sets of COS were created, where at least one research COS and one routine care COS exist for the same health condition, to identify the outcomes that were recommended for both settings. A similar process was used for a subset of paired COS matched in scope for both intervention and population as well as health condition.

Results: The sample of COS comprised: 246 COS for research only, 76 COS for routine care only and 55 COS for both research and routine care. Across the 18 sets matched by health condition the median number (range) of outcomes included in both research COS and routine care COS was 6 (3-15), with differences noted across health areas. For the 11 paired COS matched by scope and health condition, the corresponding figures were 2 (2-8). Across all settings, COS that did not include patients as participants were less likely to include life impact outcomes.

Conclusion: Within a given health condition, a small number of core outcomes were found to be relevant for both research and care, offering a meaningful starting point for linking research and real-world evaluation.

Keywords: COSs; Clinical trials; Health outcomes; Methods; Outcome selection; Routine care evaluation.