← Return to Article Details Evaluation and testing of the proportional hazards assumption in analysis of time-to-event data in subgroup analysis of randomised controlled trials: a meta-epidemiological study