A development of drug expiration prediction model

Abstract

Every pharmaceutical product is obliged to indicate its expiration date on its package. There are several alternative drug expiration prediction models that provide a sound statistical prediction, but to date there has been no single model that stands out. This study presents a drug expiration prediction model using the lower prediction interval with proper proportion of lot variability. The results illustrate that when the proportion of lot variability increased, the prediction interval model with proper proportion of lot variability provides a more appropriate drug expiration period than that provided by the confidence interval model. Hence, with a proper proportion of lot variability, the lower prediction interval model is recommended over the traditional lower confidence interval model.