(264) Time to Reach an Adequately Powered Sample Size to Detect Risk of Major Congenital Malformations Using a Prospective Observational Pregnancy Registry and An Administrative Claims Database

Publication Number: 938

Background: The FDA often requires both a prospective observational pregnancy registry (OPR) and a complementary approach, such as an administrative claims database (ACD), to provide clinically relevant safety data to inform healthcare providers on treating patients who are pregnant or anticipating pregnancy about the safety of drugs lacking adequate safety evidence. However, it may not always be timely or cost-effective to conduct both.

Objectives: To assess the time to reach an adequately powered sample size to identify the risk of major congenital malformations (MCM) in infants of women exposed to a medication using an ACD and an OPR.

Methods: We used HealthCore Integrated Research Database (HIRD) data to assess the time to reach an adequately powered sample size to assess the prevalence of MCMs in infants of women exposed to one of four multiple sclerosis or migraine drugs during pregnancy and the associated risk ratio (RR), assuming 1:2 matching. We used HIRD uptake information to calculate the estimated years required to reach the target sample size in a US ACD. As only a proportion of women who take a drug will enroll in an OPR, we assumed a range of recruitment rates (5, 10, 30, and 60 participants/year) to calculate the years required to reach the target sample size to detect the desired RR of 2.0 in an OPR.

Results: To detect an RR of 2.0, 520 women/arm should be included for an adequately powered study sample. Through utilizing HIRD, the average uptake rate for dimethyl fumarate, a multiple sclerosis drug with the lowest uptake of all drugs examined, was 36 pregnancies/year. Using an ACD, it would take 14.3 years to detect a RR of 2.0. It would take 52.0 and 17.3 years to reach adequate power to estimate a RR of 2.0, if 10 (27.8% of women treated) and 30 (83.3% of women treated) women, respectively, were recruited per year in an OPR.

Conclusions: Assessing safety using either an ACD or an OPR will take more than a decade. Actionable safety data from ACDs may be available years before OPR data collection is complete, at which point the clinical landscape may have changed. Incorporating multiple countries into an OPR, enrolling multiple databases in an ACD, raising awareness of registries, and encouraging single arm surveillance programs with a case series approach may facilitate faster enrollment.