A Method for Identifying Adverse Events Related to New Drug Treatment

Abstract

This article describes a method for distinguishing between drug-related and non-drug-related adverse events. The method is based on the principles of pharmacology and epidemiology and uses statistical strategies. Data analysis is divided into two categories: between-treatment and within-treatment comparisons. Between-treatment comparisons include data sets from both placebo and test drug groups, while within-treatment comparisons include data from the test drug only. Within-treatment criteria for identifying groups of adverse events include gradient of dose-response relationship, gradient of blood level response, magnitude of incidence of adverse events, investigator assessment of severity and relationship to treatment, and consistency and strength of “presence.” The method is illustrated with data from clinical trials of temafloxacin, a new fluroquinolone antimicrobial drug