Novel methods for pregnancy drug safety surveillance in the FDA Sentinel System

Abstract

Purpose: It is a priority of the US Food and Drug Administration (FDA) to monitor the safety of medications used during pregnancy. Pregnancy exposure registries and cohort studies utilizing electronic health record data are primary sources of information but are limited by small sample sizes and limited outcome assessment. TreeS-can™, a statistical data mining tool, can be applied within the FDA Sentinel System to simultaneously identify multiple potential adverse neonatal and infant outcomes after maternal medication exposure. Methods: We implemented TreeScan using the Sentinel analytic tools in a cohort of linked live birth deliveries and infants nested in the IBM MarketScan ® Research Database. As a case study, we compared first trimester fluoroquinolone use and cephalosporin use. We used the Bernoulli and Poisson TreeScan statistics with compatible propensity score-based study designs for confounding control (matching and stratification) and used multiple propensity score models with various strategies for confounding control to inform best practices. We developed a hierarchical outcome tree including major congenital malformations and outcomes of gestational length and birth weight. Results: A total of 1791 fluoroquinolone-exposed and 8739 cephalosporin-exposed mother-infant pairs were eligible for analysis. Both TreeScan analysis methods resulted in single alerts that were deemed to be due to uncontrolled confounding or otherwise not warranting follow-up. Conclusions: In this implementation of TreeScan using Sentinel analytic tools, we did not observe any new safety signals for fluoroquinolone use in the first trimester.TreeScan, with tailored or high-dimensional propensity scores for confounding control, is a valuable tool in addition to current safety surveillance methods for medications used during pregnancy.

Six Years of the US Food and Drug Administration's Postmarket Active Risk Identification and Analysis System in the Sentinel Initiative: Implications for Real World Evidence Generation

“…To address these shortcomings, Sentinel built a mother‐infant linkage table in ARIA in 2018, a necessary data element for birth outcome studies. It developed the analytic tools for targeted signal evaluation studies in pregnancy using propensity score adjustment in 2019, 15 and expanded its existing signal identification programs to cover pregnant patients and their offspring in 2022 16,17,48 . In 2022, Sentinel added Medicaid data via the Transformed Medicaid Statistical Information System into ARIA, 49 which covers nearly half of all pregnancies in the United States 50 .…”

“…ARIA's capabilities have grown over time as new electronic data sources and new querying programs are added to the pharmacoepidemiologic toolkit. Generating high-quality real-world evidence for drug and therapeutic biologic safety using primarily administrative claims data and noninterventional study Addressable with the addition of EHR data elements into ARIA 35,36 Inability of ARIA tools to perform required analysis 82 Insufficient signal identification tool ARIA has integrated signal identification abilities (Figure 1) [16][17][18] Study…”

“…It developed the analytic tools for targeted signal evaluation studies in pregnancy using propensity score adjustment in 2019, 15 and expanded its existing signal identification programs to cover pregnant patients and their offspring in 2022. 16,17,48 In 2022, Sentinel added Medicaid data via the Transformed Medicaid Statistical Information System into ARIA, 49 which covers nearly half of all pregnancies in the United States. 50 Finally, in the most recently published Prescription Drug User Fee Act commitments, 51 the FDA has agreed to a series of demonstration projects focusing on pregnancy and postmarket evidence generation.…”

“…13 ARIA's querying tools are a set of standardized, parameterizable computer programs that have grown in scope over time to constitute a full pharmacoepidemiology toolbox (Figure 1). 14 In the last 5 years, Sentinel has added the standard ability to enable birth outcome studies following in utero exposure, 15 signal identification analyses, [16][17][18] interrupted time series analysis, 19 and propensity score adjusted analyses using inverse probability treatment weighting and fine stratification, 20 among others. Importantly, large patient populations are only available for querying in the ARIA system if returned analytic datasets contain aggregate or summary level data.…”

Detection of Drug Effect Signals Associated with Adverse Pregnancy Outcomes Using Propensity Score Matching at Scale

“…A similar practice was done in an at-scale study conducted by Sentinel System, one of the US Food and Drug Administration (FDA) efforts in surveillance medical products. [20] We excluded comorbidities specific to the prenatal period, such as gestational diabetes; the obstetric comorbidity index is designed to assess the mortality risk at delivery. Third, we selected the fifteen most common comorbidities before and during the pregnancy.…”