A data-driven problem-solving approach to performance improvement in healthcare systems: a case study in Tehran heart center hospital

Purpose The primary goal of this research is to introduce a data-driven Problem-Solving Approach for Performance Improvement in Healthcare Systems (DPAPIH). This approach combines process mining and data mining techniques to enhance operational efficiency by identifying bottlenecks in Coronary Artery Bypass Grafting (CABG) procedures, particularly focusing on variability in Length of Stay (LOS) in the Intensive Care Unit (ICU). The study, implemented at Tehran Heart Center, aims to optimize patient flow, reduce ICU congestion and improve hospital efficiency by predicting and managing the occurrence of postoperative Atrial Fibrillation (AF), a significant cause of prolonged ICU stays. Design/methodology/approach The study introduces a data-driven problem-solving approach that integrates process mining and data mining techniques to improve performance in healthcare systems. Focusing on coronary artery bypass grafting (CABG) at Tehran Heart Center, the approach identifies bottlenecks, particularly variability in ICU length of stay (LOS) and predicts postoperative atrial fibrillation (AF). A mixed-methods approach is employed, combining quantitative process mining analyses with qualitative insights from expert consultations. The CHAID decision tree algorithm, alongside other models, is used to predict AF, enabling preemptive interventions, improving patient flow and optimizing resource allocation to reduce hospital congestion and costs. Findings The study reveals that postoperative Atrial Fibrillation (AF) significantly increases the length of stay (LOS) in the Intensive Care Unit (ICU), creating bottlenecks that delay subsequent surgeries and elevate hospital costs. A predictive model developed using CHAID decision tree algorithms achieved a prediction accuracy of 71.4%, allowing healthcare providers to anticipate AF occurrences. This capability enables proactive measures to reduce ICU congestion, improve patient flow and optimize resource allocation. The findings emphasize the importance of AF management in enhancing operational efficiency and improving patient outcomes in Coronary Artery Bypass Grafting (CABG) procedures. Originality/value This study presents an innovative integration of fuzzy process mining and data mining algorithms to address performance bottlenecks in healthcare systems, specifically within the coronary artery bypass surgery process. By identifying atrial fibrillation as a key factor in length of stay fluctuations and developing a robust predictive model, the research offers a novel, data-driven approach to performance improvement. The implementation at Tehran Heart Center validates the model’s practical applicability, demonstrating significant potential for enhancing patient outcomes, optimizing resource allocation and informing decision-making in healthcare management.