Substantial heterogeneity found in reporting mortality in Cochrane systematic reviews and Core Outcome Sets in COMET database

Abstract

ObjectiveTo explore mortality outcome usage in Cochrane systematic reviews and Core Outcome Sets for research. Study Design and SettingCochrane PICO searches identified Cochrane reviews (published January 2015-March 2021) including mortality outcomes. These outcomes were categorized according to terminology used: all-cause mortality, cause-specific mortality, infant mortality, maternal mortality, survival.Mortality outcomes in Core Outcome Sets (published until 2019 on the Core Outcome Measures in Effectiveness Trials (COMET) database) were also extracted and categorized. ResultsIn total, 2454 mortality outcomes were reported in 49% (1978/3999) of Cochrane reviews published January 2015-March 2021: all-cause (37%), infant (23%), maternal (11%), survival (10%), cause-specific (9%). Due to reviews not specifying mortality outcome type or including studies reporting no data, 11% (273/2208) remained uncategorized. Infant mortality and maternal mortality were frequently used together in reviews reporting two mortality outcomes.In total, 226 mortality outcomes were reported in 37% (165/449) of Core Outcome Sets: all-cause (48%), survival (27%), cause-specific (12%), infant (9%), maternal (4%). Mortality measurement timing varied. ConclusionMortality outcome usage varies in Cochrane reviews and Core Outcome Sets. This is problematic for evidence-based decision-making. Greater standardization is necessary for effective utilization of health research.