Accessible AI-Driven Health Diagnostics for Marginalized Communities

College student holding her robotic toy at robotics classroom at school

Project Overview

Portable, AI-powered tools for non-invasive cholesterol and dengue diagnostics, improving care for underserved groups in Malaysia.

What problem was the project designed to solve?

Cardiovascular diseases (CVD) and dengue remain significant public health burdens in Malaysia, with disproportionate impact on women, children, migrants, and underserved rural communities. For cholesterol and lipid disorders, invasive blood-based tests are costly, time-consuming, and impractical for daily or frequent use, leaving many individuals unaware of elevated CVD risks until it is too late. For dengue, limited access to rapid and accurate diagnostics in low-resource areas leads to delayed treatment, unnecessary hospital referrals, and poor outbreak control. This integrated project aimed to bridge gaps in early detection and monitoring by developing low-cost, portable, AI-driven diagnostic tools for frequent, accessible, and reliable health monitoring.

What did the project do and who was involved? How were you involved?

The project consists of three complementary innovations:

SmartScan: A patented, non-invasive cholesterol monitoring device using bioelectrical impedance analysis (BIA) and AI modeling to deliver lab-comparable accuracy (~90%) for daily lipid profiling.

DenKit: A portable dengue triage tool combining impedance biomarkers, clinical symptoms, and an AI-powered smartphone app to classify dengue severity in under one minute.

DMF Dengue Biosensor: A digital-microfluidics platform that automates rapid, finger-prick blood testing for dengue using interchangeable chemistries, providing accurate early and confirmatory diagnostics at the point of care.

The initiative was a collaboration among biomedical engineers, clinicians, AI researchers, and community health workers from universities, hospitals, and local organizations. I am the primary lead of the project.

Prof. Fatimah Ibrahim

Universiti Malaya
Biomedical Engineering

What was the outcome?

These tools have collectively expanded access to affordable, accurate health diagnostics: SmartScan enabled proactive daily cholesterol tracking for women and children, supporting early interventions to prevent CVD progression. DenKit allowed community health workers to triage patients quickly, reducing unnecessary hospital admissions and accelerating urgent care for high-risk cases. DMF Dengue Biosensor brought laboratory-grade dengue testing to rural and underserved areas, improving outbreak response and ensuring timely treatment initiation. Together, these innovations empowered marginalized groups—including rural families, migrants, and women in remote areas—with accessible health tools, reducing diagnostic delays and health inequalities.

What challenges did you address and how were they addressed?

Technical Accuracy Across Populations

Challenge: Variability in body composition and health profiles affected device calibration.
Solution: Collected large, diverse datasets to train and validate AI models for consistent accuracy.

Community Adoption and Digital Literacy

Challenge: Low awareness and limited digital literacy among older adults and rural communities.
Solution: Designed intuitive interfaces, offered multilingual guidance, and provided training sessions for community health workers and patients.

Connectivity and Infrastructure Limitations

Challenge: Poor or no internet connectivity in remote regions hindered real-time data sharing.
Solution: Enabled offline operation with secure data synchronization when internet access becomes available.