Predictive modeling of postpartum depression risk using electronic health record data

Postpartum depression (PPD) affects approximately one in seven women within the first year following childbirth. Current screening methods, such as the Edinburgh Postnatal Depression Scale (EPDS) and PHQ-9, are primarily reactive and rely on the onset of clinical symptoms. This study aimed to develop machine learning (ML) models to proactively predict PPD risk utilizing multidimensional electronic health record (EHR) data and to identify key predictive risk factors. A retrospective analysis was conducted on a cohort of 7,184 women aged 18 to 45 with a delivery encounter between January 1, 2022, and January 1, 2025. Demographic, clinical, and psychosocial variables were extracted. Statistical analysis identified 42 predictive risk factors and 24 protective factors. Strong predictors included a prior history of EPDS screening (OR: 2.54), bipolar disorder (OR: 1.98), depressive disorders (OR: 1.93), and abnormal weight gain during pregnancy (OR: 1.93). Several ML algorithms were trained and evaluated using the testing set metrics including AUROC, Precision, Recall, and F1-score. To address class imbalance, the Synthetic Minority Over-sampling Technique (SMOTE) was applied. The analytic cohort demonstrated a PPD prevalence of 7.15% (n=514). The Logistic Regression model achieved the highest predictive performance with an AUC of 0.750, outperforming traditional baseline screening tools in sensitivity. Feature importance analysis identified the number of postpartum visit days, the week at delivery, and the presence of a support system as the top model contributors. Advanced computational models can effectively leverage EHR data to identify patients at high risk for PPD prior to symptom onset. Implementing these predictive tools offers a proactive approach that addresses critical gaps in maternal mental healthcare, bypassing diagnostic barriers to improve early intervention strategies.

Name thesis:degree_name

M.S. (Master of Science)

Level thesis:degree_level

Masters

Discipline thesis:degree_discipline

Bioinformatics (UMKC)

Grantor

University of Missouri--Kansas City

Year dc:date.issued

2026

Author dc:creator

• Tyrrell, Margaret Anne

Advisor dc:contributor.advisor

• Dua, Prerna

Language dc:language.iso

en_US

Handle dc:identifier.uri

https://hdl.handle.net/10355/112344

OAI identifier oai:identifier

oai:mospace.umsystem.edu:10355/112344