EDMON - A backend server for an infection detection system monitoring individuals with type 1 diabetes

Abstract

There are a growing number of adults with diabetes worldwide. Within 2045 it is expected to become over 600 million individuals. Since there are no known cures for diabetes, self-monitoring and self-recording are often used to manage the condition. Having tools such as mobile applications allow individuals to do this. The world and society face a significant health threat from communicable diseases, which has resulted in a growth in the detection algorithms of infectious diseases.

This thesis proposes a back-end server with the functionality to implement disease surveillance algorithms on data from monitoring individuals with type 1 diabetes. It has a focus on standardization, security, and privacy. It also offers the opportunity for users to record themselves in a video with each medical recording. The design is devised with a modular approach to provide future scientists and researchers the possibility to extend the functionality.

Experiments and tests are conducted to see that the system is satisfactory and handles enough traffic for the task at hand. The solution handles 100 concurrent clients sending 10 000 requests, with around 800 requests per second. All testing is done on real user data, with calculations to simulate the EDMON infection detection systems performance. The server spends around 11 minutes running the algorithm on almost 3 million medical records, which is sufficient since this algorithm is meant to run once per hour.