Process Mining for Quality Improvement in Malawi

For non-communicable disease (NCD) program management in Malawi, long waiting times are frequently cited as a reason for poor clinic attendance and treatment default. Current quality-improvement (QI) practices, such as root-cause analysis, are often subjective. This study introduces process mining (PM) as an objective, data-driven approach to complement existing QI methods through analysis of patient records and facility registers. We describe the first protocol to implement PM in the Malawian healthcare system. The study employs a sequential mixed-methods design combining two focus group discussions (FGDs) conducted before and after quantitative PM, key-informant interviews, and document review. Qualitative data will be analysed thematically to identify perceived workflow bottlenecks, while quantitative PM will involve data extraction, pre-processing, clustering, mining, and visualisation using discovery algorithms such as the Heuristic or Fuzzy miner. Results will be displayed as Business Process Model and Notation (BPMN) nets to reveal process variations and inefficiencies. This work aims to demonstrate the feasibility and utility of PM for clinical efficiency and quality-improvement initiatives within Malawi's NCD program. With increasing adoption of electronic health-record systems and the national Quality of Care program, PM can provide a scalable tool for evidence-based monitoring of healthcare processes.