Decision making support system for conciliation, prediction and mitigation of judicialization using data science in the health area

• Grant number: 20/13762-0
• Support Opportunities: Research Grants - Innovative Research in Small Business - PIPE
• Start date: May 01, 2023
• End date: January 31, 2024
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: Mauro Fernando Jeckel
• Grantee: Mauro Fernando Jeckel
• Company: Datamind Inteligência Artificial S/A
• CNAE: Desenvolvimento de programas de computador sob encomenda
• City: São Paulo
• Associated researchers: DOMINGOS MARCIO RODRIGUES NAPOLITANO ; Fabio Carvalho Mota ; Flavia Palavani da Silva ; Miguel Victor Campanha Manfredi

Abstract

Judicialization in the health area is a recurring issue that directly affects several stakeholders (from litigants to public and private health service providers). There are several problems faced, among them the lack of public sector budget to comply with judicial decisions. In fact, the company SPUMENEWS (SPUME) uses technology to transform data into knowledge, acting in the prediction of events and the relationship between statistical data and human sciences, as well as in the real-time monitoring of various subjects, with the generation of insights, curation and classification content. Within its portfolio, two products were developed that assist the legal department of the most varied types of organizations. They are the Legal Index and the Reading of Legal Texts, both the basis for developing the proposed solution for Phase 1 of Program 'PIPE/FAPESP'. In this context, the objective of this applied research project is to develop an MVP (Minimum Viable Product) of Decision Support System, composed of a series of integrated intelligent tools, based on Data Science technologies, such as Big Data, Natural Language Processing, Machine Learning and Artificial Intelligence, through a more efficient approach to mitigate and minimize judicializations in the health area, stimulating conciliation between the parties as early as possible, in addition to providing strategic information for health management policy makers. From the perspective of three interrelated areas: a) Information Technology, especially in the Machine Learning area; b) Law, which involves knowledge related to the judicial process and its specific vocabulary and from Jurimetry, and c) Health Sciences and its management, which involves technical aspects of the demands imposed on the Health Systems; three scientific challenges were proposed for the evolution of the state of the art, among other technical and market challenges to be overcome, in a niche market that was not identified by the major players. The methodology used will be of an applied nature, involving qualitative, quantitative and experimental approaches, based on bibliographic and documentary analysis, using field research, survey and Delphi research, as well as supported by computational experiments with the techniques to be applied for the development of solution. The innovative research was divided into two stages: 1) Conducting Experiments; and 2) Construction of the tool and tests, totaling 9 months of the project. As an expected result, in addition to the delivery of the Decision Support System itself, four deliverables are proposed: 1) Legal dictionary for health, through a broad knowledge base to identify legal behaviors, potential causes, analyze risks, understand what are the processes won or not, their effects, motivating aspects, in addition to enabling the mitigation of failures; 2) support tool in tactical / operational decision making to help in the classification of these risks and for analysis of processes for analysis and trends in procedural results; 3) support tool in strategic decision making, to support the promotion of public and private health management policies, aiming to mitigate judicialization in the health area, through the identification of standards using aggregated data, based on the development of a model that can classify whether a particular document may or may not become a lawsuit. (AU)