Factors associated with postpartum depression symptoms following antepartum hospitalization.

Background: Hospitalized antepartum patients are at increased risk for postpartum depression (PPD). Developing approaches to identify those at highest risk for PPD would enable timely and targeted intervention. Objective: We aimed to identify factors associated with the development of PPD symptoms in hospitalized antepartum patients and to assess their predictive utility. Study design: This retrospective cohort study included pregnant individuals hospitalized in a regional referral center due to medical or obstetric complications between 2012 and 2025. Data were extracted from the electronic health record, including demographics, medical and obstetric history, hospitalization characteristics, and postpartum Edinburgh Postnatal Depression Scale (EPDS) scores collected within 8 weeks of delivery. Primary outcome was EPDS score ≥10, indicating PPD symptoms. We performed bivariate analyses and multivariable logistic regression modeling. Receiver operating characteristic (ROC) curve analyses assessed model discriminatory ability. Results: Among 4161 included hospitalized antepartum patients, 1291 (31%) reported PPD symptoms within 8 weeks of delivery. Multivariable logistic regression modeling showed that certain demographic and clinical factors were associated with PPD symptoms: single or other marital status, pregestational diabetes, multiple chronic medical conditions, prior anxiety or depression, admission gestational age <28 weeks, preterm prelabor rupture of membranes, placenta accreta spectrum, and pharmacologic treatment for anxiety or depression during antepartum hospitalization. The strongest predictors for the presence of PPD symptoms were prior history of anxiety or depression (adjusted odds ratio [aOR], 2.04; 95% confidence interval [CI], 1.75-2.38) and pharmacologic treatment for anxiety or depression during antepartum hospitalization (aOR, 2.33; 95% CI, 1.86-2.90). Discriminatory ability of the multivariable model was fair, with area under the receiver operating characteristic curve (AUC) of 0.67 (95% CI, 0.65-0.69), positive predictive value of 58%, and negative predictive value of 72%. Conclusions: One in three individuals who experienced antepartum hospitalization at our referral center went on to report PPD symptoms. While certain demographic and clinical factors were associated with PPD symptoms, their overall ability to accurately identify those at increased risk was limited, reducing their utility in guiding targeted interventions. These findings support that postpartum mental health services are broadly needed for individuals who experience antepartum hospitalization, highlighting the need for universal mental health service provision for this high-risk population.