skip to main content

Survival analysis hazard ratio — MRCPsych Paper A MCQ

Instant feedback + full explanation. One question, done properly.

ModerateEpidemiology and research methodsSurvival analysis hazard ratioMRCPsych Paper A

A prospective cohort study compares two antidepressant maintenance strategies. The outcome is time from treatment initiation to first depressive relapse. Follow-up duration varies, and some participants leave the study before experiencing relapse. Censoring is considered non-informative, and the proportional hazards assumption is reasonable. Investigators want an adjusted estimate of the relative effect of treatment on relapse over the follow-up period. Which effect measure is most appropriate?

Educational content. Not a substitute for clinical judgement or local policy.

Reveal the answer and explanation

Correct answer: A — Adjusted hazard ratio from Cox proportional hazards regression

The correct answer is A. Time to first relapse is a time-to-event outcome, and participants who leave before relapse contribute censored observations. Cox proportional hazards regression uses the available follow-up time, accommodates censoring and provides an adjusted hazard ratio comparing the instantaneous relapse hazards between groups, assuming proportional hazards. Logistic regression and an end-of-follow-up risk ratio reduce the outcome to relapse versus no relapse and discard information about when relapse occurred; they also handle differing follow-up poorly. A mean difference based only on observed relapse times excludes or mishandles censored participants and may be biased. Pearson correlation is inappropriate because treatment group is categorical and the outcome is censored time-to-event data.

Reference: In J, Lee DK. Survival analysis: Part I—analysis of time-to-event. Korean Journal of Anesthesiology. 2018;71(3):182–191. https://pubmed.ncbi.nlm.nih.gov/29768911/