Process mining for clinical pathway: Literature review and future directions

Clinical pathway (CP) is a tool to improve service quality and efficiency of the medical institutions. However, most of the clinical pathways designed with the traditional methods are static and non-adaptive. Recently, the process mining techniques are receiving increasing attentions. It can not only help the clinical pathways designers to discover the sequence of activities, but also provide execution information for analyzing variances and correcting design errors. In this article, first, a literature review is presented. In the analysis of 37 studies from the period 2004-2013, three research aspects (process discovery for clinical pathways design, variants analysis and control, continuous evaluation and improvement) are explored, and the weaknesses of the methods are analyzed. It is found that the mining algorithms developed are not efficient enough to deal with the unstructured processes, the models obtained cannot give a good explanation of the variants, and the lack of systemic thinking makes the improvement process of CP very tedious. Based on the analysis, finally, four key trends are identified: (1) further analysis of the variants, (2) integrated process management, (3) customization and (4) self-learning improvement of the clinical pathway.