When it comes to medical documentation and prescription medication, we are yet to have an efficient and truly secure system. The need for this has been even more prominent due to the COVID-19 pandemic and increased rate of malicious activities on the patient's health record. MEDBUDDY will be able to solve this issue by providing the users with an effective and efficient way of storing and sharing medical information and supplying them with a finer way to notify them of prescribed medication.

What it does

  • Book medical appointments, recall all the complicated names of medicines and spend ages filling in medical forms during your most unwell condition. What if there was a way that allows patients to complete all the preparation for a healthcare visit in just a second. That is MEDBUDDY - the companion for everything medical-related on your phone.
  • Currently, we are lacking in the existing system to help patients in notifying them when they need to take medication effectively. However, MEDBUDDY will be able to solve this problem by implementing a notification system that will notify the user through both their in-built notification system and usage of external calendars like Google Calendar.
  • For a very long time, society has been looking for a way to aid them in diagnosing them by using AI. MEDBUDDY app will be able to collect medical records which will help it create a model to predict users' medical conditions.

How we built it

  • MEDBUDDY is a health-tracking application that is available in both web-app and **mobile app **platforms.
  • MEDBUDDY will accompany and support you with all the complicated steps of storing and sharing medical records with just a quick QR scan.
  • The system is built with Flutter for the mobile app and JavaScript for the web app.
  • The backend for the system was built with FastAPI and MongoDB for the database
  • The system will be using deep learning to do health predictions for patients’ health status and potential diseases based on their current health records. We are using FastAI and a database from the cardiovascular-disease-dataset of Kaggle [update from 2019] Cardiovascular Disease.
  • The model was trained on Google Collab (ipynotebook) using tabular_learner in Tabular model in fastAI to analyse the dataset and train the model.

Challenges we ran into

  • One of the biggest challenges is the time constraints and existence of too many features and the lack of resources to help implement those features.
  • Training on the model on google collab was quite challenging because of the limited GPU and RAM provided by Google.

Accomplishments that we're proud of

  • Setting up the doctor-patient device connection. This application has the basis on which the system can be further improved and can have a good potential to be used in real-life applications. This app sufficiently addresses one of the crucial challenges we are facing in the medical domain.
  • Training a model to predict cardiovascular disease. Being able to use AI in the domain of medical and health science is a topic society has been looking out for.

What we learned

  • Most of us are newcomers, we learned different frameworks and coding languages to implement our website and app.
  • We learned to demonstrate and pitch out ideas effectively to the people by utilising the idea of creating a demo and a visually appealing video.
  • We learned to work in a team environment where most of us are inexperienced. We learned to utilise people's specialities effectively under severe time constraints to provide the best possible outcome.

What's next for MEDBUDDY

  • Extend the availability of AI models into a wider range of health predictions with a diversity of diseases.
  • Broaden the network of application’s API that allows worldwide connection among doctors and patients. This can include utilising medical institutes outside Australia and to make it more appealing to patients around the world, including the usage of different languages and their ways of recording data.
  • Allowing this app to be widely accessible not only helps in the efficiency of medical information storage and sharing but also distributes the data to all the users to reduce the effect of hacking.
  • Provide helpful medical data study for research purposes. There exists an issue with collecting medical data from different educational institutes for their study purposes. This is because most data is collected by private organisations. However, this app will consistently collect user data (and with their permission) and can be given to different studies for them to use. This makes the data collection to be simple and fast.

Built With

+ 63 more
Share this project: