Process mining is a process-oriented data analytics approach that has been applied in many domains, including healthcare. Healthcare process mining aims to improve health care services through treatment enhancement and disease traj ectory analysis, among others. One of the common challenges in process mining for healthcare is the high variability in the healthcare data. This paper proposes clustering as a method to improve process mining for disease trajectory analysis. The proposed method is applied in a case study coming from a real dataset of Indonesia health insurance, consisting of records of patient treatments and diagnoses of Indonesia health insurance participants. Clustering was done to group patient profiles into clusters based on their similarities. This approach comes from a prior assumption that similar patients might have similar sequent of treatments. This research contributes to two directions: a technical contribution to improve process mining projects in handling high variability data using categorization and a research contribution to promote Indonesia health insurance dataset as a real-world healthcare dataset.

Link File:

https://ieeexplore.ieee.org/abstract/document/10604436