A Latent Class Analysis of Pre-Pregnancy Multimorbidity Patterns in a Delivery Cohort at a Safety-Net Hospital.

Abstract

Background: Multimorbidity affects approximately 1 in 3 adults and is associated with adverse health outcomes. However, there is a paucity of information describing patterns of multimorbidity among the birthing population. The objective of this study was to describe the clustering of pre-pregnancy chronic conditions in the birthing population by age, race and ethnicity, insurance status, and parity using latent class analysis (LCA). Study design: We conducted a retrospective cohort study of deliveries using medical record data between 2015 and 2019. Multimorbidity was defined as having at least two chronic conditions before the start of the index pregnancy, using adapted versions of obstetric comorbidity indices. The final LCA model was selected based on clinical interpretability and statistical fit. We also compared the distribution of sociodemographic factors across classes. Results: Of 6,455 deliveries, 1,870 (29%) deliveries were to patients with multimorbidity. LCA resulted in a 3-class model: Class 1 (45% of individuals with multimorbidity) was characterized by mood/anxiety and substance use disorders; class 2 (39%) was defined by body mass index ≥30 kg/m² and chronic hypertension; and class 3 (16%) was characterized by reproductive conditions and infertility. Individuals who were <25 years or non-Hispanic White were more frequently in class 1; individuals who were ≥35 years or non-Hispanic Black were disproportionately in class 2. Nulliparas and individuals with private insurance were more frequently in class 3. Conclusion: Multimorbidity is prevalent in pregnancy and distinct chronic condition clusters vary across sociodemographic sub-groups, demonstrating the need for integrative approaches to periconceptional care for birthing individuals with multimorbidity.