Inspiration
New moms in Nigeria get discharge notes like "Amoxicillin 5ml TDS" and have no one to explain it at 2am. Wrong dosing is dangerous. We wanted a fast way to turn hospital shorthand → plain English + dose times.
What it does
A Colab-based translator for moms. Paste medical text → it detects jargon like "TDS/BID/PRN" and returns "3x daily" + safe dose schedule with timing. Includes "Confirm with your doctor" safety note. Built for USAI Hackathon Dir A.
How we built it
Python script in Google Colab. We built a dictionary + rule logic to match abbreviations to plain English and calculate dose times. No ML model yet - just clean code that works fast on any phone.
Challenges we ran into
- Scoping: Qualified for "all medical cases + emergencies" but shipped mom/baby dosing first due to 48hr limit.
- Safety: Had to add disclaimers so users know this explains, doesn’t prescribe.
- Coverage: Built abbreviation dictionary manually since no open dataset fit Nigerian discharge notes.
Accomplishments that we're proud of
- Shipped working code in 48hrs that makes hospital notes less scary.
- Turned "TDS" → "8am, 4pm, 12am" in 2 seconds for sleep-deprived moms.
- Designed safety-first: always tells user to confirm with doctor.
What we learned
- Simple rule-based code can solve real problems faster than waiting for "perfect AI".
- In healthcare, clear language + safety > complex tech.
- Scoping to one user = moms let us actually finish.
What's next
- Upgrade to ML/NLP so it can handle any medical term, not just dictionary matches.
- Add voice + Pidgin/Hausa for low-literacy users.
- Test with real discharge notes from Port Harcourt hospitals.
Built With
- 3
- dictionary
- google-colab
- pydriod
- rule-based-text-parsing
Log in or sign up for Devpost to join the conversation.