Composite Primary Outcomes Reported in Studies of Critical Care: A Scoping Review

Objective: Composite primary outcomes (CPO) (incorporating both mortality and non-mortality outcomes) offer several advantages over mortality as an outcome for critical care research. Our objective was to explore and map the literature to report on CPO evaluated in critical care research.

Data sources: PubMed, Embase, Cumulative Index to Nursing and Allied Health Literature, Scopus, and Cochrane Library from January 2000 to January 2024.

Study selection: All studies (both non-randomized controlled trial [RCT] and RCT) conducted in ICUs treating adult patients (18 yr old or older) that described CPOs and their definitions, were included for mapping, reporting, and analyzing CPOs without any restrictions.

Data extraction: Independent double-screening of abstracts/full texts and data extraction was performed using a pilot-tested extraction template. The data collected included characteristics of CPO, definitions, trends, and death handling techniques used while reporting the CPO.

Data synthesis: Seventeen CPOs were extracted from 71 studies, predominantly reported in the setting of pharmaceutical studies (48/71, 67.6%), used RCT methodology (60/71, 84.5%), and were mostly single-center studies (55/71, 77.5%). Ventilator-free days were the most commonly reported CPO (29/71, 40.8%) with marked variability in the definition used and death handling (0 d in 33 studies and -1 d in 7 studies). The most common statistical paradigm used was frequentist (63/71, 88.7%) and the study follow-up time was 90 days with 28 studies using this timeline (28/71, 39.4%). Narrative synthesis highlighted the variability in defining CPO.

Conclusions: CPOs are an emerging set of outcomes increasingly reported in critical care research. There was significant heterogeneity in definitions used, follow-up times, and reporting trends.