Health me if you can!

Home page of the application.
Heart rate record page.
Transferring entered data to the cloud and performing analysis. Finally, notifying Dr if needed.

Inspiration

The inspiration came from both challenges, Capgemini and FabLab. Our application is kind of an integration of both challenges, just without the hardware part.

What it does

Our solution aims at spreading awareness in helping prevent diseases or health problems before they occur. It mainly focuses on two health problems; diabetes, and misuse of medication. The first is one of the leading causes of death in the world, and the latter causes organ damage, and helps bacteria to mutate into anti-biotic resistant bacteria.

It does that by giving a user a personalized interface; when signing up, it asks the user about their health status; if they have diabetes and which type. It asks about all the medication the user is taking and the frequency, and whether or not the user has a fitness tracking device. It then personalizes the application pages with regards to the input data.

If the user has diabetes, it asks then to input the usual time they take their measurements in. It then proceeds to remind users to log their readings. After logging, data is transferred to the cloud using technologies such as Big Data Table, and computing engines where data analysis and machine learning are used to detect anomalies in the data. If anomalies are found, the Dr is notified, and in some cases the user and their emergency contacts are also notified.

This is useful when there is a sudden drop or raise in blood sugar which might lead to a coma, it's also useful in the case where a wrong dosage of insulin is prescribed. This analysis would help detect it. The logging is also of great use to the physician as patients usually take their measurements, check that at that certain time the measurement is okay, and then discard it. Logging gives the doctor a history of data and helps correct diagnosis.

All patients are linked and displayed in the profile of the Dr.

The second main part of our application addresses medications. This is suitable for all age groups as there are multiple channels of entering the data. What happens is that the user can take a picture of the medication name, or manually type it in. They also type in the dosage. The backend of the application then checks if there are any contradictions with other medications currently being administered. If there is, it performs a time scheduling algorithm in where it schedules each dosage such that no drugs interact together. In the future, our platform should support one time entries of pain killers, and could send out notifications warning the user of future liver damage if their trend continues.

In the case of antibiotic treatment; a majority of the people just take the pills until they get better and stop the course as soon as that happens. That's one of the reasons behind antibiotic resistant bacteria. Therefore, for the duration of the required dosage, our app sends out notifications reminding user to take their medicine, and the side effects that might happen if they do not continue treatment.

Finally, if the user owns a fitness tracker, heart rate data are sent to our app, and again uploaded to the cloud where machine learning and data analysis are performed. This helps in detecting signs and patterns of sudden heart attacks or strokes. It warns the user to go to the nearest hospital, and sends notifications to both nearest hospital, and emergency contacts.

There's also an advertisement video that motivates people to use our app.

How we built it

-The structure and wireframes of the application were built on Swift and proto.io. -Python was used to understand and plot data from a diabetes dataset. -We used NodeJS to build a website for our application, and Adobe Spark to create an advertising video. -Finally, we read about the cloud platforms we would need to store and run our data on.