Mental health conditions affect nearly 1 in 7 new parents during pregnancy and postpartum. Nearly a quarter of preventable maternal deaths are caused by such conditions . Yet a majority of new parents do not receive the treatment they need, often due to lack of access, stigma, gaps in screening, or a shortage of mental health specialists. Trayt Health partners with state psychiatry Access Programs serving pregnant and postpartum women by giving OB/GYNs, pediatricians, and more access to real-time behavioral health support through Consults.
A new study1 shows how everyday medical record data and a simple model could help spot who’s most at risk for postpartum depression—before they even leave the hospital.
Why Build a Risk-Prediction Model for Postpartum Depression?
Postpartum depression (PPD) is one of the most common complications after childbirth, affecting about 1 in 7 new parents. It’s not just about feeling down—PPD can seriously affect a person’s health and is a leading cause of maternal death, especially from suicide. The tough part? It’s often preventable, but our current system doesn’t catch everyone who’s at risk.
People with fewer socioeconomic and healthcare resources are both more likely to develop PPD and less likely to be screened during the postpartum period.2 That’s where early prediction could help.
How Did the Study Work?
Researchers from Massachusetts General Hospital and Harvard Medical School looked at data from more than 29,000 births across eight hospitals between 2017 and 2022. They split the data into two groups—one to build the model and one to test it. They used only information that would already be available by the time a patient is discharged after giving birth, like:
- Results from prenatal depression screening (EPDS scores)
- Basic demographics (age, education, marital status, etc.)
- Medical history and past prescriptions
- Details about prenatal care and delivery
To keep the model focused on predicting new cases, they excluded anyone with a recent history of mood disorders or antidepressant use.
Using these records, they trained a machine learning model called elastic net regression to predict PPD—defined as a mood disorder diagnosis, an antidepressant prescription, or a high postpartum EPDS score (≥13).
Then they tested how well the model performed, including whether it worked just as well for patients of different races, ages, and hospital settings.
What Did They Find?
The model did a solid job. It could accurately flag people at low risk 95% of the time. And while not every high-risk flag turned into a confirmed case, about 1 in 4 people flagged as high risk did develop PPD—a rate much higher than average. This model would allow care teams to focus limited resources on the people most likely to develop PPD.
Importantly, the model worked equally well across race, ethnicity, age, and whether the hospital was academic or community-based. And, as the figure shows, the model also performed very well in the validation data set, evidence that it could be useful beyond the hospitals and patients used to train the model.
Why Does This Matter?
This kind of tool could help care teams identify people at highest risk for PPD before they leave the hospital—a time when nearly everyone is reachable, unlike the postpartum period where follow-up can be spotty.
Since the model only uses data that’s already collected during hospital stays, it’s easy to implement without adding burden. And by pinpointing those most at risk, it lets health systems focus their limited time and resources on the people who need help most.
Down the line, this could mean earlier treatment, fewer missed diagnoses, and better outcomes for parents and families.
Partner with Trayt Health
Contact us to learn more about how Trayt Health utilizes workflows, data, and insights to improve behavioral health care delivery.
Citations:
- Clapp MA, Castro VM, Verhaak P, et al. Stratifying Risk for Postpartum Depression at Time of Hospital Discharge. Am J Psychiatry. 2025. doi:10.1176/appi.ajp.20240381
- Goyal D, Gay C, Lee KA. How much does low socioeconomic status increase the risk of prenatal and postpartum depressive symptoms in first-time mothers? Womens Health Issues. 2010 Mar-Apr;20(2):96-104. doi: 10.1016/j.whi.2009.11.003. Epub 2010 Feb 4. PMID: 20133153; PMCID: PMC2835803.
