Evidence-appraisal glossary
Proportional hazards assumption
The core requirement of a Cox regression: the hazard ratio between groups stays constant over the whole follow-up period. In plain terms, one group's relative risk versus the other does not grow or shrink as time passes. If it does, a single reported hazard ratio can mislead.
Also called: PH assumption, proportional hazards, constant hazard ratio assumption.
What it is
A Cox proportional hazards model summarizes a treatment or exposure effect as one number, the hazard ratio. That single number is only trustworthy if the effect holds steady across the entire follow-up. The proportional hazards assumption states exactly this: the ratio of the two groups' hazards is constant over time, even though the underlying risk itself may rise or fall.
How to use it when reading a study
Ask whether the authors checked the assumption. Standard checks include plotting scaled Schoenfeld residuals against time (they should scatter flat around zero, with no trend) and a formal Schoenfeld test, where a p-value below 0.05 signals a violation. Log-log survival curves that cross or converge are another warning sign.
If the assumption fails, a lone hazard ratio is a time-averaged blur. Effects that build up slowly, wane, or reverse (crossing survival curves) get flattened. Look for time-stratified estimates, restricted mean survival time, or landmark analyses instead. A study that never mentions the check leaves this open.
This is a plain-language methodology definition for reading research. It is general education, not medical advice.