One of my teachers like to recall a scenario with regrets where a patient developed diabetes as a result of steroids which he has to take to treat nephrotic syndrome. After finishing pharmacology in 4th year, it became obvious to me that doctors can't keep track of all drugs, rather, they tend to know the active components/classes of drugs. Patients are mostly treated objectively. Artificial intelligence and data science in general can help improve pharmacotherapy and make personalized medicine a reality.

What it does

PharmAssist predicts the likelihood of a patient experiencing a drug side effect.

How we built it

It was built using data from MEDDRA. the data was trained and deployed using MAGE AI. Node JS was used to build REST API that get predictions using model deployed on MAGE AI.

Challenges we ran into

  • Lack of good dataset

Accomplishments that we're proud of

I was happy to deploy my first ML model and use it in less than 30 minutes using MAGE AI

What we learned

  • I learnt how to use MAGE AI
  • I learnt how to build an IBM chatbot

What's next for PharmAssist

  • Search for good dataset
  • Research for a better solution to the problem than what i'm thinking right now
  • Ask for help from Subject Matter Experts
  • Build an awesome solution to improve pharmacotherapy

Built With

Share this project: