(298) Pregnancy Outcomes and Maternal Morbidity Using Administrative Claims Data in the US

Background: Pregnant women are actively excluded from experimental research, i.e., trials, thus there is limited or no safety data for this large and important segment of the US population, where nearly half of pregnancies are unintended and may experience potential inadvertent exposure to pharmacological or biologic products. In recent years, there is increased use of administrative health claims data for product safety evaluation in pregnancy using claims code algorithms. Therefore, appropriateness of the data source for algorithm application is crucial for valid results.

Objectives: To apply published, validated algorithms applicable for pregnancy safety studies in a large representative closed-claims database in the US.

Methods: Women in the US, aged 15-49, with a pregnancy episode between November 2021 and November 2022 and with continuous enrolment for 12-months before and after pregnancy, were identified. Pregnancy episodes were identified using validated diagnosis, procedure and service codes. Outcomes included live, full term, preterm and still births (LB, FTB PTB and SB), ectopic pregnancies (EP), elective abortions (AB), spontaneous abortions (SA) and trophoblastic and other abnormal products of conception (TRO). Severe maternal morbidity (SMM) were identified during the delivery-related hospitalization using the Center for Disease Control and Prevention (CDC) index, and were categorized into hypertension (HTN), non-HTN cardiopulmonary, infectious and other SMM-related events.

Results: We identified 648,830 pregnancy episodes with a median maternal age of 29 years. A total of 71.26% (n=462,361) of pregnancies resulted in LB, with 39% (n=252,206) FTB, 3.21% (n=20,835) PTB, and 0.63% (n=4,073) stillbirths. 29.18% (n=189,320) had unknown status. 12.75% (n=82,704) of pregnancies resulted in SA and 9.79% (n=63,536) in AB, 2.46% (n=15,976) were EP and 0.8% (n=5,317) were TRO.
Overall, 1.4% (n=9,337) of pregnancies showed a SMM events during the delivery-related hospitalization. Among these, 69.9% (n=6,529) were HTN-related, 24.7% (n=2,308) were non-HTN cardiopulmonary related and 19.1% (n=1,783) were infectious-related events.

Conclusions: The study demonstrates the suitability of a large US administrative claims database for identifying pregnancy episodes, pregnancy outcomes and maternal outcomes. Plans to link to electronic health record (EHR) data are ongoing to validate the algorithms, improve their specificity for this data source and apply them to estimate gestational age, exposure during first trimester and associated outcomes. Demonstrating fitness-for-purpose of secondary data sources including the validation of algorithms is an essential step to ensure feasibility and validity of real-world pregnancy studies.