Background: The average postponement of the outcome (gain in time to event) has been proposed as a measure to convey the effect of preventive medications. Among its advantages over number needed to treat and relative risk reduction is a better intuitive understanding among lay persons.
Objectives: To develop a novel approach for modeling outcome postponement achieved within a trial's duration, based on published trial data and to present a formalized meta-analysis of modeled outcome postponement for all-cause mortality in statin trials.
Methods: The outcome postponement was modeled on the basis of the hazard ratio or relative risk, the mortality rate in the placebo group and the trial's duration. Outcome postponement was subjected to a meta-analysis. We also estimated the average outcome postponement as the area between Kaplan-Meier curves. Statin trials were identified through a systematic review.
Results: The median modeled outcome postponement was 10.0 days (interquartile range, 2.9-19.5 days). Meta-analysis of 16 trials provided a summary estimate of outcome postponement for all-cause mortality of 12.6 days, with a 95% postponement interval (PI) of 7.1-18.0. For primary, secondary, and mixed prevention trials, respectively, outcome postponements were 10.2 days (PI, 4.0-16.3), 17.4 days (PI, 6.0-28.8), and 8.5 days (PI, 1.9-15.0).
Conclusions: The modeled outcome postponement is amenable to meta-analysis and may be a useful approach for presenting the benefits of preventive interventions. Statin treatment results in a small increase of average survival within the duration of a trial.
Systematic review registration: The systematic review was registered in PROSPERO [CRD42016037507] .