AI and federated learning in healthcare : predicting cardiovascular diseases with AI
Kiljander, Tommi (2025)
2025
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2025101926178
https://urn.fi/URN:NBN:fi:amk-2025101926178
Tiivistelmä
The thesis researched the opportunities and challenges of artificial intelligence (AI) and federated learning (FL) in healthcare. One of the biggest challenges in healthcare is obtaining data securely for AI development. The thesis research shows one way to create an AI model that predicts the risk of cardiovascular disease using FL. The development task was also to create a way to transfer a patient's ECG measurement data without compromising the patient's privacy. Research and development work was carried out using the design science method.
The development task of thesis was to create a program that utilizes the principal's platform to transfer a patient's ECG measurement in a secure and authorized manner to a database where the data can be utilized for AI development. A Docker container was created for the program, which communicates with the systems of different organizations using the REST API to obtain the patient's contact information for consent requests, as well as the ECG data itself for sending to the database.
The results showed that FL was a suitable way to develop AI models securely if data cannot be sent outside the device or organization. However, the performance of the model is still limited by the amount of data and health technology regulations. The development task also confirmed that Docker containers and REST API are together good, secure and flexible way to transfer data and communicate between different systems.
The development task of thesis was to create a program that utilizes the principal's platform to transfer a patient's ECG measurement in a secure and authorized manner to a database where the data can be utilized for AI development. A Docker container was created for the program, which communicates with the systems of different organizations using the REST API to obtain the patient's contact information for consent requests, as well as the ECG data itself for sending to the database.
The results showed that FL was a suitable way to develop AI models securely if data cannot be sent outside the device or organization. However, the performance of the model is still limited by the amount of data and health technology regulations. The development task also confirmed that Docker containers and REST API are together good, secure and flexible way to transfer data and communicate between different systems.