Inspiration
Long hold times, queues, and overall inefficient ways to get appointments or quick diagnosis.
What it does
HospiAI helps patients and users get medical assistance from the comfort of their home without spending hours on holds with a receptionist. HospiAI offers users a quick and seamless way to get in contact and book appointments with relevant doctors that will be able to help the user according to their symptoms. HospAI also offers users an informational diagnoser that recognizes the user's symptoms and tries its best to diagnose and provide information about the diagnosis and any precautions the user might want to take.
How we built it
With the help of openAI, we are able to classify a user's symptoms into a fixed set of medical specialties which we cross-check with our database of doctors to find the best doctors for the patient. Using openAI, we are also able to diagnose a number of diseases that best fit the symptoms entered and offer any information about a disease and the precautions that might be taken to prevent it.
Challenges we ran into
We found it extremely difficult to securely store the list of doctors and their information and keep the openAI key hidden with just a frontend. Our solution was to implement a backend node.js server to make requests to openAI and a mongoDB database to hold the list of doctors and their information.
Accomplishments that we're proud of
We were able to successfully implement and integrate openAI into HospiAI to classify symptoms into the medical areas of specialties that deal with them as well as diagnose a number of diseases based on the symptoms, output an informational passage on the disease, and offer any precautions to take.
What we learned
We learned just how powerful the openAI is even in its Beta stage. We are excited about what openAI will be able to offer in the near future and how it will shape our communities and the world.
What's next for HospiAI
We plan on adding support for EMTs on emergency calls to be able to relay information about the patients directly to the most suitable doctor that is available based on the emergency situation. We also plan to complement that feature by offering preliminary diagnoses based on the symptoms described by the EMT and the data provided by various readings and measurements of the patient.
Log in or sign up for Devpost to join the conversation.