MedCase

Inspiration

The inspiration for this project came from previous experience watching family and friends who are medical students and are facing challenges while learning complex material. We aimed to develop a tool that would enable medical students to practice complex patient diagnosis while offering a fun and interactive game-like interface for this purpose.

What it does

Do you know what medical students often state is one of the biggest challenges before seeing patients? It’s the jump from textbook knowledge to confident real-world decision-making. Right now, medical students often learn through textbooks, slideshows, and high-pressure role-plays, but there aren’t too many tools that make learning fun and interactive. That is why we created MedCase. Think of MedCase as a flight simulator for future doctors. It’s a low-stakes interface where medical students can get access to different situations and practice applying their medical knowledge to real scenarios. Users have the option to toggle between practicing their listening skills and reading/comprehension skills. There are two modes: learning and testing modes. Learning mode allows users to retry diagnosing a patient if they get the diagnosis wrong. Testing mode allows the user to proceed if the diagnosis is incorrect on the first attempt. We make medical scenario practice fun and engaging while also providing feedback on how the user can ask better questions to get the correct diagnosis with as much efficiency as possible. Both modes include a timer (5 minutes for easy questions, 7 minutes for medium-level questions, and 10 minutes for hard questions). The level of difficulty is determined by the average number of follow-up questions a user might have before determining the diagnosis correctly.

How we built it

We built this website using two components. The frontend was coded using languages such as HTML, CSS, and JavaScript. The backend was coded using languages such as JavaScript and TypeScript. We also integrated Deepgram’s API and OpenAI’s API into our website.

Challenges we ran into

Some challenges we ran into included when we were trying to implement Deepgram’s API system in our platform. We were having difficulty implementing the API as we had no prior knowledge, and the generative voice would work sometimes, but other times would be delayed. It was also difficult to limit the scope of our project as we had too many ideas and too little time to implement all of them. We definitely had to pick and choose which features to focus our time on. It was all of our first time participating in a hackathon, and it was difficult since we had no clear instructions and had to come up with our own ideas.

Accomplishments that we're proud of

Some accomplishments we are proud of are implementing Deepgram’s API, completing a whole website within 24 hours, as we are all first-time hackathon participants, understanding frontend and backend integration, and collaborating with team members who have different ideas, and being able to implement them into one website.

What we learned

Key major skills we learned were efficient and effective use of AI in coding, learning how to use new programs like Cursor and Deepgram, as well as familiarizing ourselves with different coding languages and generative AI used in building apps and websites.

What's next for MedCase

In the future, we intend to implement a streak tracking/point system, a public leaderboard to encourage friendly competition between players, and a competitive mode where players can compete against each other to rack up the most points. Additionally, a tone-detection system is a feature that will be extremely useful in letting medical students practice proper bedside manner for their future patients.

Built With

• css
• deepgram
• express.js
• html
• javascript
• node.js
• openai
• typescript