OBJECTIVES: A clinical pathway is defined as the succession of medical events that jointly aim at improving the patients’ condition. It provides efficient care by standardizing processes, reducing the variability of practices and improving the outcomes. Hospital Information Systems contain valuable data about these pathways.  However, they are hardly tapped due to their volume and complexity. The objective of the study is to factually describe the clinical pathways of patients implanted with a CRT defibrillator (CRT-D) on an 8 years timeline from the French PMSI database. The PMSI annually includes the records of all hospital stays in the country. This description will bring new knowledge about the patients care to find out the correlations with their condition. METHODS: From the 2008 database in public hospitals, we extracted all hospital stays with an implantation of a CRT-D in France, leading to 1,602 patients. The patient’s anonymous identifiers were used to find all their hospital stays during the 2 previous years (2006-2008) and the 5 following years (2008-2013), being a total of 16,931 stays. These data were analyzed using Process Mining methods to find similarities in the patient’s sequences of stays. We also developed a specific visualization tool to illustrate the results. RESULTS: We quantified the hospitalization risks on a long-term follow-up: hospitalization for heart failure was observed 3 months prior to implantation in 51% of patients, and 8 months after implantation in 50%. 28% of patients were readmitted for the device’s replacement after 2 years and 2 months, and 20% died at the hospital within 5 years. These risks were also assessed depending on the patient’s comorbidities. CONCLUSIONS: This study shows that Process Mining methods are relevant to analyze clinical pathways at a national scale from big sized database. This descriptive approach is the prerequisite for predictive analytics.