Construção colaborativa de algoritmo de tomada de decisão para implantação de serviço de farmácia clínica em oncologia

Abstract

Introduction: Cancer is the second leading cause of death worldwide, posing a challenge to public health systems. In Brazil, the SUS (Unified Health System) has prioritized strengthening cancer care, notably through the National Policy for Cancer Prevention and Control (Law No. 14,758/2023), the Cancer Prevention and Control Network (GM/MS Ordinance No. 6,591/2025), and the creation of the Pharmaceutical Assistance Component in Oncology - AF-Onco (Ordinance No. 8,477/2025). In this context, clinical pharmacy in oncology emerges as a strategy to improve care, promoting patient safety, therapeutic adherence, and cost control. However, decision-making regarding its implementation in public hospitals lacks objective and standardized instruments aligned with the specificities of the SUS. Simultaneously, the DATASUS databases remain an opportunity for planning and strategic decisions. General objective: To collaboratively build the architecture of a decision-making algorithm for the implementation of clinical pharmacy services in oncology within the Brazilian Unified Health System (SUS). Method: Applied research, of a methodological nature, structured in three stages: a) theoretical (documentary analysis and scope review); b) theoretical-empirical (regulatory analysis and preliminary algorithm design); c) collaborative empirical (evaluation with oncology specialist pharmacists). Additionally, the project integrated the Technological Initiation Program (PIT/EBSERH nº02/2025), with a scholarship recipient dedicated to the inventory of DATASUS systems, the use of generative AI (DeepSeek®), and the development of the OncoFarmaSUS prototype. Results: The work was organized into scientific articles and technical-technological products. These included: a case study on hospital pharmaceutical management; a scoping review highlighting a gap in algorithms adapted to the Brazilian Unified Health System (SUS); collaborative construction of the algorithm architecture and validation of the layout and content by experts (agreement rate = 100%); an opinion article on public policies in the pharmaceutical area; and validation of the algorithm architecture in simulated scenarios, demonstrating its operability. The resulting technical-technological product was the architecture of an expert system (symbolic AI) composed of 8 attributes (Oncological Care, Pharmacist Team, Protocols, Humanized Follow-up, Pharmacovigilance, Dispensing and Costs, Teaching and Research, Patient-Centered Care) and 36 inference mechanisms, based on the author's professional experience and on Ordinance GM/MS No. 8,477/2025 (AF-Onco). From this, the navigable prototype of the OncoFarmaSUS application was developed, with a software registration request currently being processed by EBSERH/INPI. Originality and relevance: This research proposes an expert system focused on decision-making regarding the implementation of clinical pharmacy in oncology within the Brazilian Unified Health System (SUS), innovating in the following dimensions: technological (interoperability with DATASUS), methodological (integration between documentary research, generative AI, and prototyping), institutional (articulation between professional master's programs, technological initiation programs, and teaching hospitals), and social (improvement of public management, access to healthcare, and safety of cancer patients). Practical contributions and impact: For the teaching hospital, the results offered a readiness diagnostic tool aligned with AF-Onco; for the HU Brasil network, a replicable model; for public management, the innovative use of government data with application in direct patient care; and for training, it allows integration between postgraduate, undergraduate, and practice settings. Attached institutional statements attest to the hospital's interest in future implementation phases. Final considerations: The dissertation fulfilled its proposed objective, progressing from theoretical conception to functional prototyping and software registration. Integration with PIT/EBSERH enhanced the results and the training of human resources. Next steps include the implementation of the inference engine, effective integration with DATASUS, and pilot testing. OncoFarmaSUS embodies the integration between academic research, technological innovation, and public service practice scenarios, contributing to the strengthening of the SUS (Brazilian Unified Health System) and the improvement of oncological pharmaceutical care in Brazil.