Clustering Patients' Drug Usage Based on ATC Codes
Name
Charleen Konsa
Abstract
Healthcare data provides an opportunity to study patients’ drug trajectories. This thesis aims to create a workflow that clusters patients based on their drug use using ATC codes. In addition, a user interface is developed that can be used to interactively run the workflow. The workflow consists of 5 parts: filtering, drug trajectories compilation, drug trajectories comparison, drug trajectories clustering, and cluster analysis. As a result of the workflow, patients are divided into clusters, which are given a simple overview. The results can be used for further research to find the reasons for the different drug trajectories. The user interface consists of 4 parts. Its sidebar displays the main user inputs that can influence patient selection, clustering, and analysis. Tabs display the results of clustering and analysis. The results and the used parameters in the user interface can be downloaded as RDS and CSV files.
Graduation Thesis language
Estonian
Graduation Thesis type
Bachelor - Computer Science
Supervisor(s)
Markus Haug
Defence year
2024