Protecting Trial Integrity Through Monitoring and Analysis of Medication Adherence

To better inform patient and physician decisions, it is vital to clearly describe a medication's efficacy and safety, acknowledging that drug exposure may vary due to suboptimal patient adherence to medication schedules. ICH E9(R1) sets out the expectation that randomized controlled trials prespecify an estimand aligned with such research objectives. This involves specifying a well-defined population and anticipation of intercurrent events.Lack of medication adherence, including treatment interruptions and treatment discontinuation, are intercurrent events that complicate the interpretation of all efficacy, safety, and pharmacokinetic data, as well as the interpretation of corresponding estimands and estimated treatment effects. Under a treatment policy strategy, the treatment effect averages over the consequences of adherence-related events occurring under each randomized treatment condition; non-adherence is therefore part of the effect being estimated rather than explicitly acknowledged. However, transportability of resulting effectiveness will still depend on adherence patterns. Equally, our derived understanding of underlying treatment efficacy and its potential under alternative assignment mechanisms depends on adherence.Therefore, adopting a standardized framework for identifying, reporting, interpreting, and limiting lack of medication adherence is essential in explanatory clinical trials. This may need precise measurement methods to avoid biases inherent in traditional methods like pill counts and self-reports, and reliably recognize cases of excessive non-adherence affecting the drug's biological effects. Unanticipated medication non-adherence must be treated as a key risk area within any risk-based quality management (RBQM) framework. Integrating adherence-related risk assessments into quality-by-design processes, monitoring plans, and data review strategies enables early identification and mitigation of adherence issues that could compromise the validity of trial findings.Electronically compiling real-time dosing data surpasses traditional methods in reliability and comprehensiveness and improves the credibility of trial designs. Especially with adaptive components requiring critical decisions from small samples, addressing adherence in interim analyses can prevent premature trial termination. It facilitates pre-specified post-trial analyses of adherence to better understand failures and guides future development. Adherent patients require smaller sample sizes through reduced variance and increased effect size, assuming they represent the target population.