Heterogeneous Outcome Selection and Incomplete Prespecification of Outcomes in Systematic Reviews: A Case Study on Pressure Injury.

Abstract

To understand how reviewers select and prespecify outcomes for systematic reviews (SRs), the authors report on the outcomes used in SRs of pressure injury (PI) intervention and treatment and evaluate their completeness of prespecification. The authors searched four electronic databases for SRs involving PI prevention and/or treatments. Inclusion criteria were SRs and meta-analyses evaluating interventions for preventing or treating PI. Studies without systematic search or risk-of-bias assessment, conference proceedings, and articles not in Chinese or English were excluded. Two reviewers extracted and categorized the outcomes in domains, assessing outcome prespecification using a five-element framework. Data items included study characteristics, target population, type of interventions, and outcome variables. This review included 95 SRs that reported a total of 432 instances of 24 different outcome domains. An average of four outcome domains were reported per SR. The most frequently reported domains were PI healing, PI occurrence, and PI status. Of the 62 SRs that prespecified primary outcomes, 40 (64.52%) reported more than one primary outcome. Only 24 of the 432 instances (5.56%) were completely specified. Among the 24 outcome domains, 12 (50.00%) were listed as primary outcomes at least once. Primary outcomes were more completely specified than nonprimary outcomes. Systematic reviews of PI prevention and/or treatment report diverse, incompletely prespecified outcomes, highlighting the need for a core outcome set to standardize key clinical outcomes.