Alfreda Alfred Uhega
AI & Data Science EATCMamaAlert — AI-Powered Predictive Referral & Triage System for Maternal Care
Problem Statement
Tanzania has cut maternal deaths sharply (556 to 104 per 100,000 births), but preventable deaths still happen because of delays in the referral chain especially the "three delay," where mothers wait too long for proper care even after reaching a facility. Hospitals, community health workers, and the national m-mama transport network currently respond to emergencies instantly but there's no way to flag high-risk pregnancies before labor starts, and no tool to help staff quickly decide who needs urgent care first.
Proposed Solution
Mama Alert adds a lightweight AI layer on top of infrastructure Tanzania already has health records and the national m-mama transport network rather than building something new. It has two working parts: first a risk classification model that scores a mother's ANC data to flag high-risk pregnancies early, and second a triage tool that combines clear clinical rules with anomaly detection to instantly rank arriving patients by urgency. Both give health workers a simple flag and reason.
Innovation
Instead of proposing new infrastructure, Mama Alert is the missing analytics layer on top of two things Tanzania has already built and proven nationally: the DHIS2/ANC data system and the m-mama transport network. It's also fully explainable combining transparent clinical rules with a simple, auditable model, so a nurse can see exactly why a case was flagged rather than trusting an opaque system.
Expected Impact
Earlier identification of high-risk pregnancies, faster triage at arrival (directly targeting the delay shown to cause the most facility level maternal deaths in Tanzanian studies), smarter ambulance dispatch, and over time fewer preventable maternal deaths, while extending the cost savings m-mama has already proven.