Researchers at Mass General Brigham have developed a machine learning model that can help predict postpartum depression (PPD) using routine clinical and demographic data available at delivery. Their findings, published in the American Journal of Psychiatry, suggest the model could help identify at-risk individuals earlier and improve mental health support after childbirth.
“Postpartum depression is a common but often overlooked challenge,” said lead author Dr. Mark Clapp of Massachusetts General Hospital. “Our goal was to identify higher-risk patients early, so we can offer support sooner.”
The model was developed using health records from over 29,000 patients who delivered at Mass General Brigham hospitals between 2017 and 2022. It was trained on half of the data and validated on the rest. The model effectively ruled out PPD in 90% of low-risk patients and accurately predicted PPD in nearly 30% of those flagged as high-risk — more than double the general population risk estimate.
Importantly, the model performed consistently across racial, ethnic, and age groups and was designed to work even for patients without prior psychiatric diagnoses. Adding prenatal scores from the Edinburgh Postnatal Depression Scale further enhanced the model’s accuracy.
The team is now prospectively testing the model and working with clinicians and patients to explore how it could be implemented in real-world care settings. The goal is to enable earlier, personalized mental health interventions for new parents during a critical time of adjustment and vulnerability.
Related topics:
