Risk prediction for postpartum depression based on random forest.

Area of Science:

• Machine learning applications in healthcare
• Reproductive health and maternal mental wellness

Background:

• Postpartum depression (PPD) affects a significant portion of new mothers.
• Early identification of risk factors is crucial for effective intervention.

Purpose of the Study:

• To evaluate the efficacy of the random forest algorithm in identifying risk factors and predicting postpartum depression.
• To assess the predictive performance of the developed model.

Main Methods:

• A cohort of 406 participants was recruited from a tertiary hospital.
• Data on demographic, psychosocial, biological, and obstetric factors were collected.
• A random forest model was trained and validated using a 3:1 data split.

Main Results:

• The incidence of postpartum depression was 36.9%.
• The random forest model achieved 80.10% accuracy, 61.40% sensitivity, 89.10% specificity, and an AUC of 0.833.
• Key predictors included antenatal depression, economic/work worries, and specific biochemical markers.

Conclusions:

• The random forest algorithm demonstrates significant potential for PPD risk prediction.
• The model effectively identifies critical influential factors from complex datasets.
• This approach supports targeted and timely interventions for postpartum depression.