Project Details
Title:
Mumtal Care (Healthcare)
Details:
Postpartum depression (PPD) affects between 3.9% and 25% of new mothers in Ghana, with a 2024 multicenter study of 502 postpartum mothers across five Kumasi hospitals reporting a pooled prevalence of 25%. Despite this burden, Ghana lacks a standardized, AI-powered screening tool designed specifically for its frontline maternal health workforce. Ghana's Community Health Nurses (CHNs) and Community Health Officers (CHOs) — who conduct scheduled postnatal visits at 24-48 hours, day 3, days 7-14, and six weeks after delivery — currently have no evidence-based decision support tool for identifying mothers at risk of PPD during these visits. Ghana's first-ever Maternal Mental Health Policy, launched in April 2026, signals urgent government commitment to addressing this gap, yet no clinical AI tool exists to support its implementation at the frontline. Left undetected, PPD disrupts mother-infant bonding, negatively impacts breastfeeding, and increases risk of child malnutrition — making early identification a critical public health priority.
Mumtal Care aims to: (1) Provide Ghana's CHNs and CHOs with a fast, evidence-based AI tool to identify mothers at risk of PPD during routine postnatal visits, without requiring specialist training or additional infrastructure; (2) Ground risk prediction in real Ghanaian clinical research, calibrated against the GDHS 2022-2023 national dataset; (3) Implement a safety-first screening flow that immediately flags mothers who express thoughts of self-harm, overriding statistical inference with a direct referral action; (4) Ensure accessibility in low-connectivity rural settings through an offline-capable version using ONNX Runtime Web, enabling the same trained AI model to run entirely client-side with zero internet connection; (5) Support Ghana's 2026 Maternal Mental Health Policy by providing the frontline screening infrastructure the policy currently lacks.
Mumtal Care is built exclusively for the Ghanaian maternal health context. The tool targets Ghana's Community Health Nurses and Officers — the specific frontline workforce conducting postnatal visits through Ghana's CHPS network of over 15,000 workers. Every risk factor in the model is validated by published Ghanaian clinical studies from Korle-Bu Teaching Hospital, the Northern Region, and Kumasi. The synthetic training dataset is calibrated against the Ghana Demographic and Health Survey 2022-2023, achieving within 0.6 percentage points of national statistics for marital status, teenage maternal age, and pregnancy loss prevalence. The tool directly supports Ghana's first Maternal Mental Health Policy (April 2026). Its offline capability addresses the real connectivity constraints of rural CHPS compounds. Immediate next steps include piloting in Ashanti Region CHPS compounds and partnering with Korle-Bu and Komfo Anokye Teaching Hospitals for real clinical data validation