Mobinurse (Team - GithubNhiAata)

Mobinurse uses AI to auto-generate prescriptions by hearing in on doctor-patient conversations and provides a world-class platform for doctors and patients to book, manage and attend appointments.

Comment

Inspiration

The pandemic not just temporarily affected our lives, but also altered our lifestyle permanently. This calls for a modern healthcare platform that provides No Contact Healthcare and also leverages modern technologies to make this a seamless experience.

There is a need for a platform that not only makes the experience seamless for the patients but also for the doctors. Current solutions available in the market make the experience worthwhile for the patients but not enough for the doctors - which is why doctors are not making the shift to online appointments.

Hence, we came up with Mobinurse - which leverages AI to auto-generate prescriptions and manage appointments - reducing a lot of mundane tasks that doctors do every day.

What it does

MobiNurse is an All-In-One healthcare platform powered by AI on the backend. It offers no contact healthcare where users can book online appointments, and talk with their doctors using our in-house video meeting system. At the same time, we also offer something very unique in nature -

What do doctors hate the most? WRITING PRESCRIPTIONS

What do patients hate the most? READING INCOMPREHENSIBLE PRESCRIPTIONS

Our AI Bot writes the prescription for you, by hearing in on your patient-doctor meetings and don’t worry - we never save your data

  • Patients can book appointments
  • Doctors can register, login and view appointments
  • Patients and doctors can join a video meet
  • Doctors can record the call to generate a draft prescription and approve it to send to the patients

How we built it

Our application is made of three main portions - Here's a description of each along with the tech stack

  • Frontend (HTML, CSS, JS, React, Redux, UI/UX prototyping on AdobeXD)
  • Backend (Mongoose, MongoDB, Firebase, Express, Node, Sockets.io)
  • AI (Keras, Tensorflow)

After the ideation phase, we first designed the prototype to get a clear idea of what our application should look like. Proceeding, we designed the frontend for each user screen using React and Redux and hosted it on Netlify. We then built a video conferencing system using socket.io and peer.js. We then designed the database schema and wrote the backend APIs to render the data to the frontend. Then we hosted the backend layer on Heroku. Along with this, we also developed our ML model that performs ** Named Entity Recognition using a BiDirectional LSTM** and trained it on Artificially Generated Data, since our application is unique and such datasets do not exist. We compiled all these layers and brought everything together.

Challenges we ran into

We ran into multiple challenges and the biggest one of those was to host everything on Heroku. We faced problems with actually connecting the backend with our python model because we hadn't worked on such a complex backend architecture before this hackathon.

Accomplishments that we're proud of

We're proud of being able to complete such an enormous project in less than 48 hours and finally being able to host Node, Tensorflow, and peer.js servers altogether.

What we learned

We learned a lot of things along the way.

  • Team Collaboration
  • Perseverance (We spent more than 4 hours hosting the servers)
  • We learned how to use sockets, and develop a video conferencing system all from scratch

What's next for Mobinurse (Team - GithubNhiAata)

We plan to keep improvising on our platform and hopefully make the software mature enough to be able to launch it as a startup that has the potential to revolutionize how patient-doctor interactions take place. In the future, we also plan to extend the software to more local and foreign languages using a multilingual model on the backend.

Our Innovation

Current solutions available in the market targeted at the niche make the experience worthwhile for the patients but not enough for the doctors - which is why doctors are not making the shift to online appointments.

Our innovation is two fold -

  • AI Bot to auto generate prescriptions trained on artificially generated data of over 10,000+ prescription files.
  • A platform that incentivizes the shift for both stakeholders - the doctors and the patients.
Share this project:

Updates