|
Abstract:
|
In a long-term clinical trial of new drug, patients may experience efficacy and safety endpoints over study period, which are usually evaluated individually. Of key interest to regulatory agencies, patients, and clinicians is whether the drug has a positive risk-benefit profile. A good evaluation of risk benefit must account for patients with more than one event, patient's exposure time, and differences in severity levels and background rates among the events. In this paper, a global event intensity index fulfilled the above requirements is proposed to evaluate the risk-benefit of a drug. A repeated measures Poisson regression model is used to analyze the weighted total number of events of a patient over the study period. Weights are pre-determined based on severity levels. An example from a clinical trial of raloxifene is used to illustrate the method.
|
