Розвідувальний аналіз даних для пошуку ознак невідповідності втручань діагнозу в медичній практиці

Abstract

The study addresses the issue of potential abuses in Ukraine’s healthcare system related to the artificial inflation of medical service volumes to increase funding from the National Health Service of Ukraine. The focus is on detecting inconsistencies between diagnoses and medical interventions using exploratory data analysis and large language models. The results indicate that approximately 90% of cases require expert review. The proposed approach demonstrates the potential of large language models for the automated detection of suspicious entries in medical data.