(222) Identification of Preterm Birth in An Administrative Claims Database Using Maternal Claims, Infant Claims, and Gestational Age

Publication Number: 553

Background: Preterm birth (i.e., birth of an infant before 37 weeks gestation) is an outcome of interest when evaluating medication safety during pregnancy. Preterm birth can be identified in health insurance claims databases using International Classification of Diseases, 10th revision (ICD-10) diagnosis codes, which include obstetric (i.e., O) codes for preterm delivery for use on maternal claims and perinatal period (i.e., P) codes for extreme immaturity/preterm newborn for use on infant claims. Additionally, it is possible to calculate gestational age to identify deliveries that occur prior to 37 weeks gestation, independent of codes for preterm birth.

Objectives: To determine the correspondence between 3 separate methods to identify preterm birth: 1) diagnosis code on a maternal claim; 2) diagnosis code on an infant claim; 3) gestational age < 37 weeks at delivery.

Methods: Using Optum’s Dynamic Assessment of Pregnancies and Infants (DAPI), a claims-based pregnancy database that includes mother-infant linkages, women with pregnancies between October 2015 and August 2021 that were linked to infants were identified. A diagnosis of preterm birth was identified based on the presence of ICD-10 codes O60.1, P07.2, or P07.3 on a maternal or infant claim. Gestational age was calculated using the last menstrual period (LMP) as estimated by ICD-10 Z3A codes denoting weeks of gestation and delivery date.

Results: We identified a total of 622,359 pregnancy episodes with linked infants. Among these pregnancies, 56,372 (9.1%) had a preterm birth diagnosis code. Of these, 47.0% had a preterm code on both maternal and infant claims, 39.0% had a preterm code on an infant claim only, and 14.1% had a preterm code on a maternal claim only. Among pregnancies with a preterm birth code on both maternal and infant claims, gestational age was less than 37 weeks for 98.6% compared to 83.1% of pregnancies with a preterm code on an infant claim only and 47.3% of pregnancies with a preterm code on a maternal claim only.

There were 25,974 (4.2%) pregnancies with gestational age < 37 weeks on delivery date but no preterm birth diagnosis code on either a maternal or infant claim. Among these pregnancies, 90% had a gestational age of at least 36 weeks, indicating that these were possibly not preterm births but rather an inaccurate estimation of LMP.

Conclusions: Pregnancies with a diagnosis of preterm birth on both maternal and infant claims were likely to also have an estimated gestational age < 37 weeks. These results study suggest that future studies should consider using a combination of maternal and infant claims to identify preterm birth.