Life with hazard ratios
The article explains what hazard ratios are, how they are commonly misinterpreted in medical studies, and offers guidance on adjusting expectations when encountering them in research findings.
Background
- Hazard ratios (HRs) are a statistical measure widely used in medical and epidemiological studies to compare the risk of an event (e.g., death, disease progression) between two groups (e.g., treatment vs. placebo). An HR of 0.5 means the treated group has half the risk at any given time, while an HR of 2.0 doubles it.
- The technical complaint: HRs assume "proportional hazards"—that the relative risk stays constant over time. In practice, this assumption often fails (e.g., a treatment works for a while, then stops). When it does, a single HR becomes misleading or meaningless.
- The behavioral complaint: Even when a HR is accurate, people routinely misinterpret it as a direct statement about their personal probability of survival, ignoring the crucial element of time. A drug might reduce your risk at any moment by 30% but only extend median survival by a few months.
- This matters because hazard ratios are everywhere in clinical trial reporting, drug advertising, and medical decision-making. Understanding what they actually mean—and what they don't—is a practical skill for anyone trying to make sense of health claims.