Design Driven Development of a Web-Enabled System for Data Mining in Arthroplasty Registry
MetadataShow full item record
- Master theses 
This research was inspired by the work at the Norwegian Arthroplasty Registry, which serves as a national resource for understanding the longevity of implanted prostheses, analyzing risks, and patient outcomes in general. At this moment, they have no online system that would help and enable several user groups to take advantage of the data for clinical, research, and informative purposes. This thesis has contributed with a high-fidelity prototype of a desktop application named LeddPOR. The system is dedicated to three user groups: patients, physicians, and researchers. The project was completed in collaboration with three other master students, comprising a back-end and front-end development team. Knut T. Hufthamer and Sølve Ånneland, who provided valuable data mining tasks to be incorporated in the prototype, and Arle Farsund Solheim created visualizations that allow interactive data exploration. The project followed the User-Centered Design approach, as a method to produce a prototype that would be appreciated by real users. The Design Science Research methodology allowed five iterations, within which prototypes from low- to high fidelity have taken form. The final, fully interactive prototype is intended for physicians, researchers, and patients. There are two dedicated parts; one for hip, and the other for knee. Under those, a number of data mining tasks could be performed at the convenience of the expert user. The sessions can be saved and reviewed according to users' preferences and needs. The patient part of the system is offering mainly information, but also some resources such as formerly developed applications supporting post-operative care. During this development, we have defined two patient personas, acknowledging their different needs. On the expert side, two personas were created, one for physicians and one for researchers. Usability testing was conducted with both expert and novice users, which suggested a high success rate. The final System Usability Score (SUS) of 95 points, as well as feedback from evaluation, indicate a potential to develop a product that could be valuable for several user groups.