USE OF CHATGPT-4 AS A SUPPORT TOOL FOR QUALITATIVE ANALYSIS OF MUNICIPAL HEALTH PLANS
DOI:
https://doi.org/10.16891/2317-434X.v14.e1.a2026.id3661Keywords:
Artificial Intelligence, Health Planning, Public HealthAbstract
The study evaluated the application of the natural language model ChatGPT-4 in the qualitative analysis of Municipal Health Plans (PMS), aiming to assess its feasibility as a support tool in the evaluation process. Five PMS from municipalities in Santa Catarina were analyzed, using the Qualitative Analysis Guide for PMS as the reference framework. A persona was created, scopes were defined, and prompts with detailed instructions were developed for each question. The model was instructed to rely exclusively on the information present in the documents, with referenced justifications. The theoretical framework is situated within the context of the Fourth Industrial Revolution, considering the potential of artificial intelligence (AI) in technical and analytical processes. The results showed a 52% valid match between the analyses generated by the AI and those conducted by human evaluators, with higher agreement in affirmative responses. Discrepancies mainly occurred due to out-of-context interpretations. Despite its limitations, ChatGPT-4 was able be used alongside human analyses, serving as a promising tool, although it highlights the need for improved evaluation criteria and refinement of the prompts used.