External validation of the clinical obstetric co-morbidity index across a diverse health system

OBJECTIVE: The clinically-modified obstetric co-morbidity index (OB-CMI) is a co-morbidity-based scoring system that has been validated to predict severe maternal morbidity (SMM) in a single tertiary, academic hospital using an internal SMM definition. We aimed to validate the OB-CMI for prediction of SMM as defined by the CDC during delivery admissions across a diverse health system.

STUDY DESIGN: This is a retrospective cohort study evaluating all deliveries in a large health system encompassing academic and community hospitals. Data from 2019-2021 were extracted from the electronic health record (EHR) and validated with chart review. An OB-CMI score was calculated for each patient using established diagnosis codes and EHR data. The primary outcome was non-transfusion SMM (defined by the CDC) during the delivery admission. Patient characteristics were evaluated by hospital, and hospital-specific receiver-operator characteristic (ROC) curves were constructed and compared.

RESULTS: 42,130 deliveries were included with significant differences in all demographic, clinical, and obstetric characteristics across the hospitals including age, BMI, race/ethnicity, insurance type, preterm birth and preeclampsia rates. (Table). Median OB-CMI score and rate of elevated OB-CMI score (≥ 6) were also significantly different. ROC curves for OB-CMI and SMM for each hospital are noted in the Figure with an area-under-the curve (AUC) range from 0.77 – 0.83, and no significant differences across hospitals (p = 0.32).

CONCLUSION: In a large cohort of patients delivering across a diverse hospital system, the clinical OB-CMI score similarly predicted SMM despite differences in demographic and clinical characteristics among the hospitals. This validation of the OB-CMI supports the use of this scoring system in variegated clinical settings, which can inform widescale uptake and clinical integration of OB-CMI scoring to improve obstetric risk stratification.

National Institute of Child Health and Human Development

Adina Rachel Kern-Goldberger, Sindhu Srinivas, Michael Harhay, Lisa D Levine