A unit-free ratio of the odds of onset of a given outcome in any given unit of time for individuals with a given risk, divided by the odds of onset over the same period for individuals without such a risk. Put another way, it is the ratio of the instantaneous probability of the outcome event (e.g., Alzheimer’s disease) in the exposed group compared with that in the non-exposed group, and as such can be considered as a type of a relative risk. Hazard ratios have been used to describe the outcomes of therapeutic trials. Here, the question is to what extent a particular treatment can reduce the duration of an illness. The ratio, however, does not always accurately capture the degree by which an illness has been shortened. Consequently, when such information is required, then time-based parameters derived from the time-to-event curve (e.g., the ratio of the median times of the placebo and drug groups) are recommended to describe the magnitude of the benefit to the patient. One way conveying the difference between hazard-based and time-based measures is to compare them to the odds of winning a race versus the margin of victory: the hazard ratio is the odds of a patient improving faster under treatment, but it does not contain information about how much faster this process may take place.
See Attributable fraction, Epidemiological studies, Epidemiology, Odds ratio, Rate ratio, Risk factors