Ethical and Methodological Features of Cluster Randomized Trials in Critical Care Settings: A Systematic Review

Cluster randomized trials (CRTs) are useful for evaluating system-level interventions in critical care but raise complex ethical and methodological issues that are amplified in this setting. We aimed to describe ethical and methodological features of CRTs in critical care settings.

We performed a systematic literature search in MEDLINE and Embase.

We included primary reports of CRTs in critical care settings published from 2004 to 2022 in English.

Two reviewers independently extracted data from each CRT, resolving any discrepancies by consensus.

Among 102 CRTs, 77 CRTs (78%) reported trial registration and 101 CRTs (99%) reported research ethics committee review; five CRTs (5%) were exempted from review. Common designs were parallel-arm (48, 47%), two-period crossover (23, 23%), stepped-wedge (20, 20%), and multi-period crossover (8, 8%). A justification for using cluster randomization was provided in 47 CRTs (46%). Patients and caregivers were identified as research participants in 99 CRTs (97%). Of these, 74 CRTs (75%) reported on consent, which was most commonly waived consent (36, 49%). Healthcare providers were identified as research participants in 48 CRTs (47%). Of these, 11 CRTs (23%) reported on consent. Most (82, 80%) reported a sample size or power calculation, with 67 CRTs (82%) accounting for the intracluster correlation; 77 CRTs (76%) accounted for the intracluster correlation in the primary analysis. Among 85 CRTs with fewer than 40 clusters, 11 CRTs (13%) reported a small sample correction.

Ethical and methodological shortcomings persist in CRTs. Enhanced reporting standards and adherence to established ethical frameworks, including the Ottawa Statement, are needed to ensure both scientific and ethical integrity in this essential design in critical care research.