Home Healthkit

Student Developers:

Megan Biernat, Gabriel Fournier, Anna Lamoureux, Martin Nesbitt, Andrew Reed, Joshua Russett, Michael Turnbach, John Vonelli, Nicholas Zambelli

About the Project

For patients who attend annual check ups, health data is recorded and an opinion of general health is given based on trends over time. This method results in data having long periods of time between new entries, making trends difficult to track. Our sensor platform allows a patient to run tests at home and upload the results to a server, allowing for an increase in testing frequency. The server is accessible anywhere the internet is available, allowing patients to run tests in a doctor’s office, at home, or while traveling. Using the Data Visualization App, medical professionals can view patient data over time in order to track each patient’s overall health. Monitoring health in such a way increases testing flexibility and allows dangerous trends to be corrected before serious illnesses take effect.

The healthkit was developed as a collaboration between Merck Co and Moravian College to provide students an opportunity to work directly with data and new languages and technologies. If you would like to check out the sponsors of this project, funding and project guidance is provided by Merck Co and the developing team of students are from Moravian College. The Github repository for this project is located here.

Architecture

For data collection, we used C++ libraries to read raw data from an e-Health Platform Sensor Kit, connected to a Raspberry Pi, to allow the data input app, built in Python and Kivy, to interpret and display the data. The data server provides a RESTful API, implemented in Flask, that stores the data in a MySQL database. A second Kivy app gets data from the database through the API and allows users to see data over time.

Conclusions

Advantages

Home sensors allow patients to provide doctors with more frequent data.
Other sensors such as Fitbit could be integrated to this system.
The increase in data could lead to additional research opportunities.

Potential Drawbacks

The sensor kit is not medically certified.
Sensor kits are expensive and the sensors are not portable.
While the current architecture works on a very small scale, if it were to be more widely used, the architecture improvements and added security would significantly increase cost.

Sensor Interface

The e-Health Library provided by Cooking-Hacks is written in C++. We used the Ctypes Python library to utilize the e-Health Library in Python. We wrote a sensor library to allow users to interact with the sensors in a Pythonic way. The sensor controls library contains high-level functions that abstract away the details of interacting with the sensors.

Data Input App

The input app (to the left) allows patients to collect their medical data, display the results, and send the data to the server. With the simple interface, speed of data collection and storage, and future compatibility with iOS and Android, the app encourages individuals to test more often, decreasing the time-interval between data points for physicians.

RESTful API

The REST API was implemented in Python using the Flask framework. The API supports an arbitrary number of patients (users). Currently, the API supports data from blood pressure, electrical resistance, and heart rate recovery tests, but it can easily be expanded to support other forms of data. The API is flexible because it supports optional parameters for most functions.

Data Visualization App

This app demonstrates a potential use case for doctors. The doctor can view a patient’s most recent blood pressure reading and historic data simultaneously. The drop-down menu allows the doctor to cycle through patients.