Analysis of Long-Term Adverse Experience Data Using the Weibull Model

Abstract

For most drugs studied by pharmaceutical companies, the comparative period of a clinical trial is relatively short. To accumulate safety experience, especially for drugs meant for chronic administration, trials are continued into “open extensions” in which all patients receive the experimental compound for several years. The information collected on adverse experiences (AEs) is usually tabulated as an overall crude rate.

Survival analysis is more appropriate for the analysis of long-term data and data with variable follow-up time, and provides richer information than a crude rate. The use of the Weibull model for the analysis of adverse experience data is presented and compared to nonparametric survival analytic techniques. Examples of parametric and nonparametric techniques are provided. Use of survival techniques to analyze AE data provides more accurate estimates of cumulative rates than the crude rate and provides a means to test whether or not these rates are increasing, constant, or decreasing with time.