Simulation study of frailty as a baseline confounder: the need to improve reporting intensive care trials
Authors' conclusion
Baseline distributional imbalance in frailty may create clinically significant differences in apparent mortality estimates.
Study context
- Population
- Patients admitted to ICU (database used to simulate trial enrolment of 100 to 1000 patients per arm)
- Participants
- 6,968 database patients
- Comparison
- 1:1 randomisation vs block randomisation in simulated two-arm trials
- Outcome
- Proportion of simulated 'trials' with frailty imbalance (Clinical Frailty Scale) sufficient to generate ≥2% and ≥5% differences in in-hospital mortality
Methodological detailssimulated two-arm trials, 1:1 and block randomisation, 100 simulations
- Simulated two-arm trials
- 1:1 and block randomisation
- 100 simulations
Main finding
As reported in the abstract
Proportion of 'trials' with frailty imbalance generating mortality differences
≥2% mortality difference, 100 patients/arm
36%
95%CI 31–42%
≥2% mortality difference, 400 patients/arm
9%
95%CI 6–13%
≥5% mortality difference, 100 patients/arm
3%
95%CI 1–5%
≥5% mortality difference, all larger sample sizes
0%
95%CI 0–1.3%
Further findings
Ordinal frailty imbalance despite binary non-frail vs frail balance, ≥2% mortality difference
- 100 patients/arm
- 10% (95%CI 7-14%)
Block randomisation
- Only reduced the proportion of 'trials' with frailty-related mortality differences to a minor and variable degree
Original publication
Simulation study of frailty as a baseline confounder: the need to improve reporting intensive care trials. Crit Care. Published 2 June 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →