MODELO DE INTELIGÊNCIA ARTIFICIAL PARA IDENTIFICAR MEDICAMENTOS EM PROCESSOS DE JUDICIALIZAÇÃO NO TRIBUNAL DE JUSTIÇA DO ESTADO DE MATO GROSSO DO SUL

Abstract

The healthcare judicialization in Brazil has posed significant challenges to the Unified Health System (SUS), especially with respect to the provision of medicines. At the Court of Justice of the State of Mato Grosso do Sul (TJMS), the growing number of lawsuits related to demands for medicines impacts public policies and requires innovative technological solutions. With the consolidation of the electronic case files, there is a wealth of information to be explored in procedural texts which, once extracted and organized, becomes a powerful tool for management and governance. This study aims to apply an artificial intelligence (AI) model capable of automatically identifying medicines names mentioned in the initial petitions of judicial records, using text-mining and machine-learning techniques focused on case files documents written in Portuguese. The model was trained with transformer-based algorithms, particularly BioBERTpt. After generating the data, an interactive computational platform called MED-SUS-MS was implemented to visualize data related to the healthcare judicialization at TJMS, in order to support decision-making in public policies for medicines provision in the State of Mato Grosso do Sul. The methodology involved constructing an anonymized dataset, applying Named Entity Recognition (NER) techniques, and evaluating the model’s performance using metrics such as precision, recall, and F1-score. The results demonstrate the technical feasibility and institutional relevance of the proposed model, enabling greater transparency, agility, and evidence-based formulation of public health policies. The model is aligned with the guidelines and resolutions of the National Council of Justice (CNJ) and TJMS, promoting ethical and secure governance within the Judiciary and directly contributing to Sustainable Development Goals (SDGs) 3 – Good Health and Well-being and 16 – Peace, Justice and Strong Institutions. This research reinforces the potential of artificial intelligence as a strategic tool for the healthcare de-judicialization by transforming judicial data into qualified inputs for public management and governance. Keywords: artificial intelligence, named entity recognition, medicines mining, healthcare judicialization and unified health system.