Routine vs. Early Postpartum Depression Screening: A Pragmatic Clinical Trial

This is a pragmatic clinical trial comparing early EPDS screening (2 weeks postpartum) with standard care (6 weeks postpartum) in 428 patients without a history of depression and delivering and receiving obstetric care at an MGH-affiliated clinic. Patients will be stratified as high-risk or low-risk for PPD using a validated prediction model and then randomized (1:1) to one of two arms: early screening (intervention group) or routine care (control group). Patients and clinical team members will be blinded to risk assignment. The early EPDS screening group will be sent an electronic EPDS (PROMS) at 2-3 weeks postpartum by the MGH Obstetric Clinic Staff; results will be directed to the patient's primary obstetric provider, mirroring the clinical process for the routine 6-week postpartum EPDS screen. The routine care group will continue to receive standard postpartum care, which includes a visit around 6 to 8 weeks postpartum and administration of the EPDS survey. Clinic staff will review and manage screening results according to local clinic protocols (e.g., arranging virtual or in-person visits, social work referrals, etc.); this includes managing reports of self-harm, which is a component of the scale. The research team will not be responsible for sending or reviewing EPDS results in real-time. The study will be conducted until 214 high- and 214 low-risk patients have been randomized and followed through 6 months postpartum.

This study is a pilot project designed to generate preliminary data for a larger definitive trial. In the retrospective data, 10.1% of patients in the high-risk group and 2.0% of patients in the low-risk group had an EPDS score>12 at 6 weeks postpartum (non-responders were considered not to have a positive screen). Assuming the same rates in this prospective cohort, 107 individuals in each risk grouping would provide 80% power at alpha=0.05 to detect the same effect (one-sided Z-test). The study will enroll 107 individuals in the control arm in each risk grouping.

The study will employ an intent-to-treat approach, analyzing all randomized participants in their assigned groups, with a secondary per-protocol analysis limited to participants who completed their assigned screening protocol.

The primary endpoint, EPDS-positive screen (EPDS >12 vs ≤12), will be compared using chi-square tests within the standard screening group between the high- and low-risk groups. EPDS score at 2 weeks postpartum in the early screening group will be compared between high-risk and low-risk participants using Mann-Whitney U tests for median scores or t-tests for mean scores, if normally distributed. Secondary EPDS score endpoints at 6 weeks postpartum will be analyzed using the same methods. The binary endpoints of PPD incidence and mental health referral/treatment at 6 months (defined as the presence of a mood/depressive disorder diagnosis code, psychiatric medication prescription, or EPDS >12) will be compared between groups using chi-square tests. Feasibility endpoints, including EPDS completion rates at 2 and 6 weeks and positive screen rates (EPDS >10 and >12), will be analyzed using chi-square tests. Time to mental health referral or treatment initiation will be evaluated using Kaplan-Meier survival curves with log-rank tests for between-group comparisons. Model validation will be assessed through AUROC analysis for discrimination between those who develop PPD and those who do not, along with calibration plots comparing predicted versus observed PPD rates.

Using an intention-to-treat analytic plan, the primary analysis will compare rates of positive EPDS screening among the high- and low-risk groups in the control arm; non-responders to the EPDS screen will be considered "screen negative" to mirror the approach used for our sample size calculation. For other analyses, a complete-case approach will be used as the primary comparison method.

Regression will adjust for patient characteristics, such as race (White vs. non-White), language (English vs. non-English), age (< vs ≥35 years), and mode of delivery (vaginal vs. cesarean delivery), to examine effect modification across various subgroups.