Evaluate multiple adverse events in crossover design bioequivalence clinical trials

AIM: To establish a statistical model to appropriately evaluate the relationship between clinical adverse event and treatment. METHODS: By splitting the duration of each treatment period into several time intervals (use day as unit), a clinical adverse event was sampled in each time interval as presented or not presented. The number of presented cases was added for each time interval, and time of subjects spent in each time interval was cumulated as person-days. The onset of a clinical adverse event and its duration were represented as repeatedly measured count data. By using the generalized linear mixed model with fixed- and random-effects, the relative rate of clinical adverse events relative to different treatments was modeled by the generalized estimate equation (GEE) technique. RESULTS: Example shows that not only the onset of adverse events, but also its duration and total person-days subjects spent in study would influence the relative rate of clinic adverse events. CONCLUSION: Our proposed approach is a good alternative and supplemental method for evaluating clinical adverse events.