Error-Correcting Methodology for Evaluating Compliance to Clinical Guidelines: A Case Study on Rectal Cancer

In a clinical setting, physicians often need to tailor general clinical guidelines to match the specific characteristics of individual patients, resulting in deviations from the guidelines. Process mining provides a range of tools to evaluate how well patient pathways align with clinical guidelines. In this work we implement a novel conformance checking method that employs error-correcting techniques to provide a degree of deviation from the clinical guidelines. The described method allows the set of corrective actions to be defined interactively and the cost of each of them to be adjusted on the basis of clinical knowledge. We test the error-correction algorithm on a real-world cohort of locally advanced rectal cancer patients treated at Fondazione Policlinico Universitario A. Gemelli IRCCS to assess the degree of compliance to the European Society of Medical Oncology guidelines, translated into a computer-interpretable version with Pseudo-Workflow formalism. Using this approach, it is possible to extensively analyze the compliance of clinical treatments, from the control-flow to the data perspective, and to discover the most common types of deviations in our cohort.