Peer+

Our AI gets and summarizes patient symptoms and history, giving doctors a clear head start or acting as frontline provider in underserved communities where care is limited, especially for minorities.

Comment

Inspiration

We were inspired in our 9th grade ethnic studies class, which was a class that we learn about marginalized communities. When we learned about healthcare inequalities, it touched our hearts. So, we decided to create Peer+, to ensure that no more marginalized communities feel the inequity of healthcare inequality ever again.

What it does

It analyzes a person's symptoms for a certain disease, and then detects the type of disease, and a potential treatment plan for the disease. This can save a trip to the doctor, and eventually hundreds of dollars, especially for minorities (e.g. Black, Hispanic) who are significantly more likely to be uninsured than their white counterparts.

How we built it

We used multiple types of coding languages, such as HTML, Javascript and Python for a combination of the front end and backend. It analyzes user symptoms via a chat interface, references a Kaggle disease-symptom dataset, and returns likely conditions with treatment plans—supporting English and Spanish. Designed to improve access for uninsured and minority populations.

Challenges we ran into

The two main challenges that we ran into were the integration of the back end and the front end, as well as properly using Open AI's API keys. The time spent analyzing all of the images, trying to provide annotated notes, regarding each one of them (in our data collection). Overall, this hindered the time we spent on other areas, such as the front end, cleaning up the back end, and making it more aesthetically pleasing.

Accomplishments that we're proud of

We're extremely proud of the fact that our model worked, and that it is able to communicate in multiple languages. We are really proud of connecting the user's input with a Kaggle dataset, while also providing the user with a result that is almost perfectly accurate with almost no flaws.

What we learned

We learnt how to effectively connect the back end to the front end, even though we were restricted on time and could not fulfill it. We also learned the skills of effective time management, and the importance of productivity in the workplace. It is really important to leave extra time for testing. There were many instances where in the last moments of time, errors occurred. Our poor time allocation led to there being no time to fix these errors, and therefore, we had to go forward.

What's next for Peer+

We wish to make our website more aesthetically pleasing, as well as getting it cross-analyzed with real doctors, to make the models more accurate. If given more time, we would additionally try to think of new features to add.

Built With

Share this project:

Updates