Sepsis is considered a worldwide health and economic burden. In Brazil, sepsis is the major cause of death in Intensive Care Units, as well as, one of the main causes of late hospital mortality. In this thesis, we first provide a 10 years population-level epidemiology report of sepsis in Brazil, using data from the Brazilian Unified Health System. Secondly, we present a research study that supports health care facilities in the evaluation and optimization of their sepsis clinical pathways (CP) using process mining techniques. A CP consists of a well-defined care plan, which includes a clear order and time for the execution of interventions with expected outcomes. During the execution of this study, it became clear for us the lack of existing process mining techniques for the optimization of CPs. Thus, we proposed, implemented and tested a novel process mining technique that supports users to improve their processes and we applied it for CP improvement. Our developed technique (Multi-CAT) identifies and highlights a set of activities and sub-sequences that provide positive or negative outcomes considering multiple simultaneous criteria. We successfully applied our technique in a real sepsis CP, and we acquired more optimization insights that we got in our previous manual analysis. We conclude that Multi-CAT has high potential to help in the optimization of processes with a good performance. In the most complex test scenario, with 58 activities, 50,000 cases and 14,451 variants, Multi-CAT took 2.04 minutes to execute the analysis. Finally, the tool was validated in two different processes, indicating potential capability to be applicable to other business areas.