Application of Decision Trees for Identifying Risk Factors of In-Hospital Mortality in Infective Endocarditis

Introduction: Infective endocarditis is a severe disease with a hospital mortality rate close to 50%, requiring timely interdisciplinary management. In Colombia, there is limited evidence regarding clinical outcomes and factors associated with mortality in patients with infective endocarditis, complicating evidence-based clinical decision-making at the local level.
Methodology: This will be an observational analytical retrospective cohort study using secondary data from the institutional registry of infective endocarditis (ENDOMED). An exploratory analysis will be conducted using decision trees, a statistical technique that can detect complex patterns and interactions among multiple variables. Variables included will encompass initial clinical features, microbiological findings, and treatments received by patients. The analysis will identify variable combinations associated with increased hospital mortality.

To identify specific combinations of clinical, microbiological, and therapeutic characteristics associated with hospital mortality in patients diagnosed with infective endocarditis at Hospital Universitario Mayor Méderi between 2011 and 2025.

Clinical, microbiological, and therapeutic patterns are expected to be identified that allow for more accurate classification of in-hospital mortality in patients with infective endocarditis. These findings will contribute to the early identification of high-risk patients, facilitating timely clinical interventions that may improve
outcomes and reduce mortality in this population.