Banca de QUALIFICAÇÃO: RAPHAEL SILVA FONTES

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
STUDENT : RAPHAEL SILVA FONTES
DATE: 17/12/2024
TIME: 09:00
LOCAL: Laboratório de Inovação Tecnológica em Saúde (LAIS)
TITLE:

Artificial Intelligence applied to the National Audit System (SNA) of the Unified Health System (SUS): a study on the regulation of healthcare and the dispensing of orthoses, prostheses, and special materials (OPME)

KEY WORDS:

National Audit System (SNA), Unified Health System (SUS), Audit, Artificial Intelligence, Regulation, OPME, Technological Architecture

PAGES: 15
BIG AREA: Engenharias
AREA: Engenharia Biomédica
SUMMARY:

Context: Audits in the Unified Health System (SUS) are essential to ensure that public funds allocated to healthcare are used correctly and efficiently. Denasus, the agency responsible for these audits, faces challenges such as the large amount of data to be analyzed and the need to protect patients' personal information. In summary, audits in the SUS are important to ensure the quality of healthcare services, but they need more modern and efficient tools to deal with the complexity of data and guarantee patient privacy. Objective: To develop, extend, and implement an architectural framework of artificial intelligence to optimize SUS audit processes, focusing on the regulation of healthcare services and the management of orthoses, prostheses, and special materials (OPME). Methodology: The methodology began with an exploratory study, through a systematic mapping of the literature to support the research. Next, a case study was conducted to deepen the analysis of the context and collect primary data. Finally, experiments were conducted to test the hypotheses and evaluate the effectiveness of the AI algorithms used. Partial results: Among the 16 works identified in the systematic mapping, the selection and extraction criteria consolidated the critical success factors and key points that served as a starting point for the construction of the architectural framework. The adopted approach allowed the development of a robust and flexible solution, such as the use in automated systems for classifying OPME products in invoices, which found 99% accuracy using support vector machines. In addition, a textual content recovery tool allowed the filtering of 56 thousand articles related to public health risks from the 2 million records, speeding up audits and optimizing the resources invested in investigations. Another study revealed irregularities worth R$577,000 in purchases of expired medicines, highlighting the importance of automated monitoring platforms to detect fraud in real time. Conclusion: This study has brought together a set of evidence of the real effectiveness of applying artificial intelligence in optimizing audit processes in the SUS.

COMMITTEE MEMBERS:
Presidente - 2488270 - RICARDO ALEXSANDRO DE MEDEIROS VALENTIM
Interno - 2885532 - IVANOVITCH MEDEIROS DANTAS DA SILVA
Externa ao Programa - 2562782 - KARILANY DANTAS COUTINHO - UFRNExterno à Instituição - GUILHERME MEDEIROS MACHADO
Externo à Instituição - JAILTON CARLOS DE PAIVA - IFRN
Externo à Instituição - METHANIAS COLAÇO RODRIGUES JÚNIOR - UFS