ABSTRACT
Out-of-specification (OOS) and out-of-trend (OOT) results may be observed during substance or drug product release or stability testing, which may significantly affect the results of a statistical analysis and of the disposition of a drug substance or product lot. Therefore, when an OOS or OOT result occurs, it is critical to conduct a thorough investigation to identify the root cause(s) so that the results can be appropriately addressed. Over the past decades, many violations of current good manufacturing practice (cGMP) concerning OOS/OOT resulted in legal proceedings, drug recalls, and disruption of supply chain. Despite heightened regulatory oversight, failure to properly investigate OOS results continues to be a significant part of Food and Drug Administration (FDA) Form 483 findings. Although FDA guidance outlines steps for OOS investigation (FDA 2006), it does not provide guidelines for using statistical methods to assist OOS investigations. Numerous working practices and various degrees of understanding of current expectations in this area have led to many different statistical approaches to investigating OOS results. In this chapter, statistical methods for OOS investigation that represent good science and conform to current regulatory guidelines are presented. In addition, various statistical methods for identifying OOT results are also included in the chapter.
