Statistical Methods Used in Interim Monitoring

Abstract

In Chap. 16, the administrative structure was discussed for conducting interim analysis of data quality and outcome data for benefit and potential harm to trial participants. Although statistical approaches for interim analyses may have design implications, we have delayed discussing any details until this chapter because they really focus on monitoring accumulating data. Even if, during the design of the trial, consideration was not given to sequential methods, they could still be used to assist in the data monitoring or the decision-making process. In this chapter, some statistical methods for sequential analysis will be reviewed that are currently available and used for monitoring accumulating data in a clinical trial. These methods help support the evaluation of interim data and whether they are so convincing that the trial should be terminated early for benefit, harm, or futility or whether it should be continued to its planned termination. No single statistical test or monitoring procedure ought to be used as a strict rule for decision-making, but rather as one piece of evidence to be integrated with the totality of evidence [1–6]. Therefore, it is difficult to make a single recommendation about which should be used. However, the following methods, when applied appropriately, can be useful guides in the decision-making process.