Can a Checklist Cut Deaths? What This Chad Cluster Trial Found

This week’s Three Clinical Papers Worth Knowing covers one primary study that met the evidence gate: CRIMSON, a 34-centre cluster-randomised trial of the RISQ illness-severity and care-recommendation system in a community acute-malnutrition programme in Chad. Mortality occurred in 35/7,494 RISQ programme entries (0.5%) versus 59/7,171 usual-care entries (0.8%). The prespecified Bayesian cluster analysis estimated an adjusted odds ratio of 0.58 (95% credible interval 0.30–1.06) and a 96.2% posterior probability of lower mortality with RISQ. This briefing focuses on absolute effects, why cluster randomisation was appropriate, and what a Bayesian posterior probability can and cannot support. Companion appraisal and evidence register: https://statmd.substack.com/p/can-a-s... Primary paper: https://www.thelancet.com/journals/la... Trial registration: https://clinicaltrials.gov/study/NCT0... Narration is AI-generated and is not the founder’s voice. This video is for education only and does not provide individual medical advice. Chapters 00:00 The question 01:00 Design and population 02:15 The absolute result 03:15 Bayesian interpretation 04:20 Limits and applicability #ClinicalResearch #GlobalHealth #Paediatrics #MedicalStatistics