ViTeM - Intelligent Maternal Care for Ethnic Minorities

Inspiration

Our solution supports women and new mothers from the Tay ethnic minority in remote areas of Dong Nai city, who struggle to understand Vietnamese medical documents due to language barriers and complex medical terminology. As a result, critical instructions about medications, postnatal care, and danger signs often become inaccessible when they are needed most. This gap has serious consequences: maternal mortality rates in ethnic minority regions are 3.5 times higher than in lowland areas. To address this challenge, we built ViTeM, a tool that translates medical documents into clear, actionable guidance in both Vietnamese and Tay language and connects users with doctors when urgent support is needed.

What it does

ViTeM is an AI-powered Crisis-to-Action Translator that turns complex medical documents into clear, actionable steps:

For mothers: They simply take a photo of any medical document: discharge instructions, prescriptions, or check-up slips and AI instantly generates a plain-language summary and a step-by-step action checklist. They can listen to audio instructions in the Tay language and share the results with family members.
For community health workers & midwives (YELLOW cases): They receive case files from the AI for review, confirm or edit medication schedules and appointment details, and add local context before sending the final checklist to the mother.
For doctors (RED cases): When AI detects danger signs like neonatal jaundice, high fever, or bleeding, it immediately escalates the case. AI provides only temporary first-aid guidance, creates a structured case report, and sends it to the doctor for final clinical approval. The doctor then reviews, approves or edits the checklist, and sends the verified version to the mother.

Decision Impact

With ViTeM, Tay mothers are no longer confused by medical documents. No more panic. No more guessing. They finally understand exactly what to do and who to call if it's an emergency. Overall, this means fewer missed information about diagnoses, faster interventions, and ultimately, a reduction in maternal and infant mortality in ethnic minority communities.

How we built it

The frontend of the application was constructed using popular web frameworks like React and Vite, paired with a clean, soft pink UI/UX design optimized for users with limited technological literacy. For the backend database and real-time data storage, we utilized Supabase, while deploying the overall architecture on free hosting platforms like Vercel (front end) and Render (back end). To power the psychological counseling and basic medical advisory feature, we integrated Google's large language model API.

Challenges we ran into

During the development process, our biggest hurdle was gathering accurate local data and designing a user interface that could overcome the strict language and cultural barriers of ethnic minority groups. Because medical literacy varies significantly among the target population in remote districts of Dong Nai, translating standard medical workflows into highly simplified, icon-driven visual modules was a tedious process. We had to undergo multiple data restructuring cycles to ensure our underlying database could logically organize maternal health metrics while remaining respectful of local cultural habits. Stripping away complex technical and medical jargon to create an accessible user interface required continuous adjustments based on direct feedback.

Accomplishments that we're proud of

Through this initiative, we aim to gain valuable experience for future project development while directly empowering ethnic minority communities with better access to information and healthcare support. Ultimately, our vision is to advance this technology to assist a wider range of vulnerable groups and scale our social impact globally.

What we learned

This project provided a vast amount of knowledge that extended far beyond regular classroom learning. On a technical level, we gained practical experience in UI/UX design, technical problem-solving, logical database management, and building functional web systems. More importantly, the fieldwork taught us that implementing technology in real life requires deep empathy and social responsibility. We learned how to analyze actual public health problems, interact with community stakeholders, and understand the necessity of cultural inclusivity when deploying technology for vulnerable populations.

What's next for ViTeM - Intelligent Maternal Care for Ethnic Minorities

Moving forward, we plan to add a smart chatbot that answers pregnancy questions in real time. We'll also build a "Mother's Diary" that learns from user behavior, tracking health updates, remembering ultrasound images, and creating a complete pregnancy journal. A "Flashback Memories" feature will compile baby's milestones into beautiful, shareable moments. These additions transform ViTeM from a translator into a lifelong companion for Tày mothers, one that grows with them, remembers their journey, and provides trusted guidance at every step.