Retrospective Validation Of a Patient-Initiated Preconception Screener Against Obstetric Comorbidity Indices To Assess Pregnancy Complications

Overview

This study conducted a retrospective validation of PreMA, a patient-initiated preconception screener implemented within electronic health records, assessing its performance in predicting severe maternal morbidity (SMM) relative to established obstetric comorbidity indices (Bateman and Leonard). The validation encompassed both single-center and multi-center cohorts drawn from University of California health systems, incorporating preconception diagnostic data linked to delivery outcomes.

Methods and approach

A retrospective observational design analyzed de-identified EHR data from live-birth deliveries with documented preconception assessment data. PreMA scores and established comorbidity indices were derived from preconception diagnoses and standardized to z-scores for comparability. Logistic and Poisson regression models estimated associations between standardized scores and SMM and non-transfusion SMM (NT-SMM) outcomes. Analyses included stratification by race, ethnicity, and neighborhood deprivation indices. Item-level analyses using adjusted Poisson regression examined associations between individual PreMA questionnaire domains and maternal morbidity outcomes.

Key Findings

Across both cohorts, elevated standardized PreMA, Bateman, and Leonard scores demonstrated consistent significant associations with increased SMM and NT-SMM risk, with relative risk estimates in the 1.2-1.4 range per standard deviation (adjusted p < 0.001). The magnitude of associations remained similar across the three indices and both cohorts. Stratified analyses by race, ethnicity, and socioeconomic status maintained statistical significance. Item-level examination revealed heterogeneity across PreMA domains, with cardiovascular domains demonstrating the strongest adjusted associations with maternal morbidity outcomes.

Implications

The EHR-derived PreMA score exhibits robust and generalizable associations with severe maternal morbidity outcomes that are comparable in magnitude to established clinician-facing comorbidity indices. This equivalence supports PreMA's validity as a preconception risk assessment instrument and suggests potential utility in clinical settings where patient-initiated screening tools could enhance identification of at-risk pregnancies. The consistency of findings across institutional contexts and demographic subgroups indicates transportability of the PreMA approach across diverse healthcare settings and populations. The differential strength of associations across PreMA domains, particularly the prominence of cardiovascular factors, provides specificity regarding which preconception health dimensions warrant clinical attention in pregnancy risk stratification. Implementation of patient-centered preconception screening through EHR systems may offer a scalable alternative or complement to clinician-administered comorbidity assessments, potentially improving accessibility and early identification of pregnancies at increased risk for severe maternal morbidity. Further investigation into the mechanisms underlying domain-specific associations and the clinical utility of PreMA-guided interventions in reducing maternal morbidity would strengthen evidence for its integration into obstetric care pathways.

Scope and limitations

This summary is based on the study abstract and available metadata. It does not include a full analysis of the complete paper, supplementary materials, or underlying datasets unless explicitly stated. Findings should be interpreted in the context of the original publication.