Abstract
The daily growing amount of health data of various forms and sources, as well as the increasing development of novel analysis methods, hold the potential for personalized healthcare. The Digital Health Information Systems team of the AIT Austrian Institute of Technology is developing an infrastructure that addresses this issue: Through privacy-preserving aggregation and standardization of health data, as well as smart analysis methods to create predictive models, services for customized patient treatment are to be made possible. This was implemented in an initial pilot project in a so-called "smart" registry for patients with chronic heart failure. The aim of this work was to support the development of this smart registry by first carrying out a verification analysis in which the smart registry data was compared with the data from a scientific publication. It was shown that, based on the automatically synchronized data of the smart registry, a large part of the manually collected publication data could already be replicated. This analysis also made it possible to identify errors and deviations in the smart registry data and thus improve the data quality. Furthermore, a basic machine learning model was developed that predicts an extension of a second telemonitoring period for smart registry patients who also participated in the "HerzMobil Tirol" telemonitoring programme. The prediction model achieved mixed performance results. Nevertheless, it was possible to demonstrate a first use case for prediction based on aggregated health data in the smart registry. Further research on the networking of health data combined with the application of modern analytical methods is necessary to advance personalized and data-driven healthcare.
Originalsprache | Englisch |
---|---|
Qualifikation | Master of Science |
Gradverleihende Hochschule |
|
Betreuer/-in / Berater/-in |
|
Datum der Bewilligung | 28 Nov. 2024 |
DOIs | |
Publikationsstatus | Veröffentlicht - Okt. 2024 |
Research Field
- Exploration of Digital Health
Schlagwörter
- Verifikation
- Vorhersagemodell
- Smartes Register
- Datengetriebene Gesundheitsversorgung
- Chronische Herzinsuffizienz