Cluster versus Individual Randomization in Prehospital, Emergency and Critical Care Trials: Protocol for a Meta-Epidemiological Study

Cluster-randomized trials, despite their benefits, carry risks of bias including selection bias, potentially due to challenges in allocation concealment, which may exaggerate treatment effects compared to individually randomized trials. Systematic differences in effect estimates between cluster-randomized and individually randomized trials have previously been documented; however, their specific impact in acute care settings, including prehospital, emergency, and critical care, where selection bias risks may be amplified due to the time-sensitive nature of trial inclusion, remains unclear. This protocol outlines a meta-epidemiological study aimed at assessing these systematic differences in effect estimates between cluster-randomized and individually randomized trials within acute care settings, quantifying potential deviations indicative of systematic bias. A systematic review of meta-analyses published since January 2015, including both trial types, will be conducted. Eligible meta-analyses must address interventions for acutely ill patients in prehospital emergency care, emergency departments, or critical care units, excluding perioperative, endovascular, and exclusively neonatal interventions. Interventions included must occur before or during hospital admission. Mortality at longest reported follow-up will be the primary outcome, while secondary outcomes include days-alive or length-of-stay measures. Data will be extracted independently and analyzed using a multi-level hierarchical regression model to assess the magnitude and consistency of effect-estimate differences between trial designs. Results will be visualized using forest and Bland-Altman plots. No individual trial risk-of-bias assessments or grading of evidence quality will be performed. Findings from this meta-epidemiological study will contribute to improved methodological understanding and interpretation of cluster-randomized trials in acute medical research, potentially influencing future meta-analytic practices.