(254) Development of a Pregnancy Cohort in Commercial Insurance Claims Data: Evaluation of Inpatient Versus Apparent Outpatient Deliveries

Background: In studies using insurance claims data, pregnancies are often identified by an observed pregnancy outcome (e.g. live or still birth). Some studies have stricter inclusion criteria for pregnancies with only outpatient delivery claims, with the rationale that outpatient claims may be less reliable. It is not known how reliable these outpatient deliveries are nor how different eligibility criteria may impact generalizability.

Objectives: Within a commercial insurance claims database, we examined the strength of evidence confirming the validity of deliveries by apparent delivery setting and described the characteristics of deliveries by apparent setting.

Methods: In a US commercial insurance database (2006-2021), we identified potential pregnancies by the presence of delivery claims from a provider and/or facility. We classified deliveries as inpatient (claim date during inpatient admission) and apparent outpatient (no inpatient admission on claim date). We identified possible confirmatory evidence for each pregnancy including: (1) the presence of both provider and facility delivery codes; (2) the presence of both diagnosis and procedure delivery codes from any source; (3) labor and delivery revenue codes; (4) gestational age diagnosis codes; (5) prenatal care codes; (6) linkage to an infant claim; and (7) infant birthdate within 1 month of the delivery date. We quantified the proportion of deliveries with confirmatory evidence by delivery setting. In the subset of deliveries with ≥1 piece of confirmatory evidence, we compared patient characteristics by apparent delivery setting.

Results: Among 4,084,474 delivery episodes, 96.4% were classified as inpatient and 3.6% apparent outpatient. 99.9% of inpatient and 94.0% of apparent outpatient deliveries had ≥1 piece of confirmatory evidence. Overall, 99.7% (n=4,072,697) had ≥1 piece of evidence. The median number of distinct pieces of confirmatory evidence for inpatient deliveries was 5 (IQR 4-6) and for apparent outpatient deliveries, 2 (IQR 2-4). Prenatal care codes were the most common type of confirmatory evidence (99.0% inpatient, 85.7% apparent outpatient). Gestational age codes were the least common evidence for inpatient deliveries (38.7%) and the presence of both provider and facility delivery codes was the least common for apparent outpatient deliveries (12.9%). Confirmed deliveries classified as inpatient occurred among patients who were older (mean, 30.4 vs 28.4 years) and more clinically complex (i.e., having more pregnancy complications, chronic diseases and prescription medications) than apparent outpatient deliveries.

Conclusions: The vast majority of deliveries had confirmatory evidence. Patient characteristics differed by apparent delivery setting.