TreeScan to Evaluate the Safety of New Drugs in Pediatric Populations

Children are not small adults; yet we heavily rely on information collected in randomized and non-randomized studies in adult populations to inform treatment strategies for pediatric patients. Historically, children have been protected from research for ethical reasons. However, this well- intentioned protection from the risks and burdens of trials has paradoxically left children more susceptible to the potential risks of drugs used in everyday practice, frequently off-label and without high-quality evidence of efficacy, effectiveness, or safety. The importance of pharmacoepidemiologic drug safety monitoring using routinely collected healthcare data has been increasingly recognized by the FDA, EMA, and other regulators, and forms the backbone for a major component of the nation’s drug safety surveillance system. However, to date, the focus of these initiatives has been on drug safety in adult populations; there is currently no systematic surveillance system targeting drug safety in pediatric populations. To address this evidence gap – which puts pediatric patients at increased risk – we propose to develop and test the performance of TreeScan based approaches for the systematic and simultaneous evaluation of multiple potential adverse outcomes in pediatric populations, and to implement these methods for prospective sequential surveillance of new pediatric medications. TreeScan methods use a hierarchical tree comprising thousands of outcomes and account for multiple testing of correlated hypotheses while systematically screening for potential adverse effects. The methodology has never been used to evaluate drug safety in pediatric patients, where there are unique challenges compared to adult populations. Thus, the methods require further development and refinement. We will test the approach based on real-world examples of established drugs with relatively well characterized safety profiles (Aim 1) as well as plasmode simulations that explore a broad range of plausible clinical contexts in pediatrics (Aim 2). The performance tested approach will then be implemented to prospectively monitor the safety of recently approved drugs and new drugs that will be approved during the early years of the grant (Aim 3). Potential signals will be further evaluated in full-scale epidemiological studies. This project will provide a critically needed tool that could lead to early detection of unsuspected adverse effects of drugs in pediatric populations if they exist and provide reassurance if no safety issues with large effects are detected. The proposed studies are expected to have an immediate and important public health impact by providing comprehensive assessments of the safety profiles for new medications in pediatric populations.