Building a Machine Learning Model to Predict Postpartum Depression from Electronic Health Records in a Tertiary Care Setting
Aufbau eines Machine-Learning-Modells zur Vorhersage von postpartaler Depression aus elektronischen Gesundheitsakten in einem tertiären Versorgungskontext
Journal of Clinical Medicine · 20.09.2025 · 14 (18) · 6644
DOI: 10.3390/jcm14186644 ↗Auf einen Blick
Bei 12.284 Frauen sagte ein Random-Forest-Modell postpartale Depression mit moderater Güte voraus. Zusätzliche sozioökonomische oder nationale Survey-Daten verbesserten die Leistung nicht.
Fragestellung
Lässt sich postpartale Depression aus elektronischen Gesundheitsakten in einem tertiären Versorgungskontext mit Machine Learning vorhersagen?
Methode
Analyse von EHR-Daten von 12.284 Frauen; Random-Forest-Klassifikationsmodell. Zielvariable über Screening-Tools und Diagnosecodes definiert.
Ergebnisse
AUC 0,733 ± 0,008. Bei Hispanics AUC 0,713 ± 0,040; bei der nationalen Survey-Datenbasis 0,699 ± 0,019 vs. 0,735 ± 0,010 bei White patients (p < 0,01).
Einschränkungen
Ein Teil der Daten war fragmentiert; Frauen mit fehlenden Daten oder ohne pränatale/postpartale Visiten wurden ausgeschlossen. Geringe Überlappung der Variablen in den Zusatzdaten; kleinere Stichprobe bei Hispanics.
Schlussfolgerung der Autor:innen
Das Modell zeigte eine moderate Vorhersageleistung. Vor klinischer Implementierung sind weitere Validierung und prospektive Tests erforderlich.
Kernaussagen
- Random-Forest-Modell mit AUC 0,733 ± 0,008
- Zusätzliche sozioökonomische oder Survey-Daten verbesserten die Leistung nicht
- Leistung bei Hispanics teilweise niedriger, teils statistisch signifikant
Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.
Kurz gefasst
Fachartikel in der Zeitschrift Journal of Clinical Medicine, online erschienen am 20.09.2025. Verfasst von Zhitu Ma und 4 weiteren Autor:innen. Beteiligte Einrichtung: Clinical and Translational Science Institute, Wake Forest University School of Medicine, Winston-Salem, NC 27101, USA. Die Arbeit wurde bisher 9-mal zitiert (Crossref) und nennt 46 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NCI NIH HHS, NCATS NIH HHS, National Center for Advancing Translational Sciences.
Abstract (Originalsprache)
Background
Postpartum depression is a common mental health condition that can occur up to one year after childbirth. Recent studies have increasingly used machine learning techniques to predict its occurrence; however, few have comprehensively explored the use of electronic health record data, particularly in tertiary care settings where such data can be fragmented.
Methods
We analyzed electronic health record data from 12,284 women who delivered at The Birth Center at Atrium Health Wake Forest Baptist Medical Center, excluding those with missing data or no prenatal or postpartum visits. To define the target variable, we examined different combinations of depression screening tools (Edinburgh Postnatal Depression Scale and Patient Health Questionnaire-9), along with diagnosis codes specific to postpartum depression. We then trained a random forest classification model to predict postpartum depression.
Results
The model achieved an area under the receiver operating characteristic curve of 0.733 ± 0.008, which is comparable to previous studies. Adding socioeconomic features from census tract data did not improve predictive performance, underscoring the importance of individual-level data. Incorporating national survey data, such as the Pregnancy Risk Assessment Monitoring System, also did not improve performance due to limited overlap in data features. Interestingly, model performance was slightly lower among Hispanic patients (area under the curve = 0.713 ± 0.040), although this difference was not statistically significant (p = 0.17), likely due to the small sample size. A similar, but statistically significant trend was observed in the larger national survey dataset (area under the curve = 0.699 ± 0.019 for Hispanic patients versus 0.735 ± 0.010 for White patients, p < 0.01).
Conclusions
While our model demonstrates moderate predictive capability, further validation and prospective testing are needed before clinical implementation. This work also identified an optimal approach for digital phenotyping postpartum depression in electronic health record data and highlighted key gaps in data quality and completeness. These findings emphasize the importance of robust data when developing predictive models for real-world clinical use.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Clinical and Translational Science Institute, Wake Forest University School of Medicine, Winston-Salem, NC 27101, USA.
- Department of Obstetrics and Gynecology, Section of Maternal-Fetal Medicine, Wake Forest University School of Medicine, Winston-Salem, NC 27101, USA.
- Center for Artificial Intelligence Research, Wake Forest University School of Medicine, Winston-Salem, NC 27101, USA.
ORCID-Profile: Metin Nafi Gurcan
Förderung
- NCI NIH HHS
- NCATS NIH HHS
- National Center for Advancing Translational Sciences
Zitieren
APA
Ma, Z., Horvath, M., Stamilio, D. M., Sekyere, K., & Gurcan, M. N. (2025). Building a Machine Learning Model to Predict Postpartum Depression from Electronic Health Records in a Tertiary Care Setting. Journal of Clinical Medicine, 14(18), 6644. https://doi.org/10.3390/jcm14186644
Vancouver
Ma Z, Horvath M, Stamilio DM, Sekyere K, Gurcan MN. Building a Machine Learning Model to Predict Postpartum Depression from Electronic Health Records in a Tertiary Care Setting. Journal of Clinical Medicine. 2025;14(18):6644. doi:10.3390/jcm14186644
BibTeX
@article{Ma2025,
author = {Ma, Zhitu and Horvath, Michael and Stamilio, David Michael and Sekyere, Kobby and Gurcan, Metin Nafi},
title = {Building a Machine Learning Model to Predict Postpartum Depression from Electronic Health Records in a Tertiary Care Setting},
journal = {Journal of Clinical Medicine},
year = {2025},
volume = {14},
number = {18},
pages = {6644},
doi = {10.3390/jcm14186644}
}
EUVIMED