1. A machine learning model has been developed to stratify the risk for postpartum depression at the time of hospital discharge.
2. Researchers incorporated prenatal depression screening scores and clinical data into the model.
3. The model demonstrated good discrimination and calibration in an external validation cohort.
4. The study population consisted of individuals who had no prior diagnosis of mood or psychotic disorders or antidepressant prescriptions 1 year prior to delivery.
5. The model's predictor variables included information known prior to delivery discharge, such as prenatal Edinburgh Postnatal Depression Scale scores and sociodemographic factors.
6. The model achieved an area under the receiver operating characteristic curve of 0.721 and a Brier score of 0.087 in the external validation cohort.
7. Overall, 9.2% of individuals in the study met criteria for postpartum depression within 6 months postpartum.
8. The model performed comparably across subgroups, suggesting potential for equitable application in diverse populations.

Source: The American Journal of Psychiatry