Title
Increasing accessibility and reproducibility of process mining research in healthcare
Abstract
The goal of this thesis was to develop a set of best practices for process mining research in healthcare. Process mining in healthcare has been studied by several researchers already, but these studies suffer from disappointing results, which may be partly to blame to the low accessibility and reproducibility of the process mining research methodologies. Therefore, in this thesis we aimed to synthesize the most common and best research steps, or best practices, of previous process mining research by means of a multiple case study and expert interviews into one process mining methodology for healthcare which would show increased accessibility, reproducibility and results that are better, or at least comparable, to present process mining research. Pattern language was identified as the most suitable method to report the best practices and as a result, 22 best practices patterns that cover a variety of process mining research steps have been developed. Subsequently, 11 patterns were selected as recommendations for use in healthcare, as they showed the most promising results and highest accessibility. In addition, we have introduced the concepts of virtual machines and screencasts to the field of process mining as means to increase accessibility and reproducibility of the best practices pattern language. The results of the best practices and recommendations have been validated by experts at two Dutch healthcare institutions and showed that the process mining best practices pattern language is indeed able to produce process models which are of similar or superior quality compared to process models created
with conventional methodologies.
Online Access
Get the PDF from the TU/e library