Designing randomized trials that are efficient and ethical is a challenge. Investigators need to collect enough data to identify meaningful benefits or harms when they exist, but also avoid exposing more individuals than necessary to experimental conditions.

The amount of data that needs to be collected often depends on characteristics of the population and treatment that aren’t well understood at the outset of a study. Many studies are designed to use prior estimates, often from smaller studies or different populations, to inform sample size requirements. If these estimates are incorrect, studies may collect too much data, or not enough to provide the precision needed to meaningfully answer scientific and clinical questions.

Information monitoring allows investigators to use the data being collected to determine when enough data has been collected to answer the primary aim of the study. Rather than using prior estimates to determine sample size requirements, such designs instead use the precision of the estimate to determine when data accrual should stop. It can be viewed as a generalization of a type of study known as an event-driven trial for studies with a time-to-event outcome.

impart allows investigators to plan, monitor, and analyze information-monitored trials. It also can be used to ensure that covariate adjusted analyses maintain appropriate type I error when used in trials with interim analyses for efficacy and futility.