Resposta e assertividade de inteligências artificiais a prompts direcionados ao manejo da lesão por pressão
Palavras-chave:
Inteligência Artificial Generativa, Modelos de Linguagem de Grande Escala, Equipe de Enfermagem, Raciocínio Clínico, Úlcera por PressãoResumo
Objetivo: Comparar as respostas e a assertividade de diferentes Inteligências Artificiais a partir de um conjunto de prompts para o manejo da lesão por pressão.
Método: Estudo descritivo, comparativo, das respostas de Inteligências Artificiais, realizado entre outubro e novembro de 2025, em dois momentos: 1) Estudo de prompts e 2) Teste para obtenção das respostas. Foram avaliadas a Perplexity AI, Grok AI, Blackbox AI, OpenAI ChatGPT-4.0®, Gemini® e Claude AI, divididas aleatoriamente nos grupos hybrid prompting e zero-shot. As respostas foram organizadas em planilha do Excel e analisadas de forma descritiva.
Resultados: A assertividade geral das inteligências artificiais foi 31,9%. O grupo hybrid prompting alcançou 36,1% e o zero-shot obteve 27,7%. A IA Gemini® demonstrou maior assertividade e o ChatGPT-4.0® a pior.
Conclusão: A Assertividade apresentada indica que a recomendação mais segura é o modelo hybrid prompt, integrando inteligência artificial ao julgamento clínico humano.
Referências
1. Pott FS, Meier MJ, Stocco JGD, Petz FFC, Roehrs H, Ziegelmann PK. Pressure injury prevention measures: overview of systematic reviews. Rev Esc Enferm USP [Internet]. 2023 [cited 2025 Aug 15];57:e20230039. Available from: https://doi.org/10.1590/1980-220X-REEUSP-2023-0039en
2. Fontes TLA, de Oliveira BGRB, De Oliveira MF, da Silva MA, da Silva ARG, Pires BMFB, et al. Development of a checklist to prevent pressure injuries in patients with COVID-19. Rev enferm UFPE on line [Internet]. 2024 [cited 2025 Aug 15];18(1):e257602. Available from: https://doi.org/10.5205/1981-8963.2024.257602
3. Macêdo SM, Bastos LLAG, Oliveira RGC, Lima MCV, Gomes FCF. Selection criteria for primary dressings in the treatment of pressure ulcers in hospitalized patients. Cogitare Enferm [Internet]. 2021 [cited 2025 Aug 25];26:e74400. Available from: http://dx.doi.org/10.5380/ce.v26i0.74400
4. de Araújo CAF, Pereira SRM, de Paula VG, de Oliveira JA, de Andrade KBS, de Oliveira NVD. Evaluation of the knowledge of nursing professionals in the prevention of pressure ulcer in intensive care. Esc Anna Nery [Internet]. 2022 [cited 2025 Aug 25];26:e20210200. Available from: https://doi.org/10.1590/2177-9465-EAN-2021-0200
5. Kotp MH, Ismail HA, Basyouny HAA, Aly MA, Hendy A, Nashwan AJ, et al. Empowering nurse leaders: readiness for AI integration and the perceived benefits of predictive analytics. BMC Nursing [Internet]. 2025 [cited 2025 Aug 25];24(56):1-13. Available from: https://doi.org/10.1186/s12912-024-02653-x
6. Dal Sasso GTM, Lanzoni GMM, Alvarez AG, Barra DCC, Barbosa SFF. Potential contribution of Chatgpt® to learning about septic shock in intensive care. Texto Contexto Enferm [Internet]. 2024 [cited 2025 Aug 25];33:e20230184. Available from: https://doi.org/10.1590/1980-265X-TCE-2023-0184en
7. Vitorino LM, Yoshinari Junior GH, Lopes-Júnior LC. Artificial intelligence in nursing: advancing clinical judgment and decision-making. Rev Bras Enferm [Internet]. 2025 [cited 2025 Nov 19];78(4):e780401. Available from: https://doi.org/10.1590/0034-7167.2025780401
8. Seibert K, Domhoff D, Bruch D, Schulte-Althoff M, Fürstenau D, Biessmann F, et al. Application scenarios for artificial intelligence in nursing care: rapid review. J Med Internet Res [Internet]. 2021 [cited 2025 Aug 22];23(11):e26522. Available from: https://www.jmir.org/2021/11/e26522
9. Sampaio RC, Sabbatini M, Limongi R. Diretrizes para o uso ético e responsável da Inteligência Artificial Generativa: um guia prático para pesquisadores. São Paulo: Editora Intercom [Internet]. 2024 [cited 2025 Aug 20]. 62 p. Available from: https://prpg.unicamp.br/noticias/lancamento-diretrizes-para-o-uso-etico-e-responsavel-da-inteligencia-artificial-generativa-um-guia-pratico-para-pesquisadores/
10. Shi J, Chen H, Shi C, Su C, Yue S, Li W, et al. Development and comparative evaluation of knowledge graph-enhanced large language models for domain-specific question answering in nursing. BMC Nursing [Internet]. 2026 [cited 2026 May 5];25:530. Available from: https://doi.org/10.1186/s12912-026-04647-3
11. Gurler N, Sengul T, Bulbul SH, Akyaz DY, Guler O, Yantac AE, et al. Secure and accessible AI in nursing education: a modular agentic chatbot framework comparing ChatGPT-4o with an open-source LLM for chronic wound care. Nurse Educ Pract [Internet]. 2026 [cited 2026 May 5];93:104789. Available from: https://doi.org/10.1016/j.nepr.2026.104789
12. Bsharat SM, Myrzakhan A, Shen Z. Principled instructions are all you need for questioning LLaMA-1/2, GPT-3.5/4. arXiv [Internet]. 2024 [cited 2025 Sep 19];2:1-26. Available from: https://arxiv.org/html/2312.16171v2
13. Pan American Health Organization (PAHO). AI Prompt design for public health: using generative AI responsibly [Internet]. OPAS; 2025. [cited 2025 Sep 30]. 61 p. Available from: https://iris.paho.org/items/b98d9c0f-6008-4b78-9066-01e1e81c29c5
14. Schulhouff S, Llie M, Balepur N, Kahadze K, Liu A, Si C, et al. The prompt report: a systematic survey of prompt engineering techniques. arXiv [Internet]. 2024 [cited 2025 Sep 30];6:1-80. Available from: https://arxiv.org/abs/2406.06608
15. Vilakati, S. Prompt engineering for accurate statistical reasoning with large language models in medical research. Front Artif Intell [Internet]. 2025 [cited 2025 Nov 18];8:1658316. Available from: https://doi.org/10.3389/frai.2025.1658316
16. Dhuliawala S, Komeili M, Xu J, Raileanu R, Li X, Celikyilmaz, et al. Chain-of-verification reduces hallucination in large language models. arXiv [Internet]. 2023 [cited 2025 Oct 12];2:1-19. Available from: https://arxiv.org/abs/2309.11495
17. Bernardes, RM. Prevenção e manejo da lesão por pressão: segurança do paciente [Internet]. São Paulo: Universidade de São Paulo; 2020 [cited 2025 Sep 15]. Available from: http://eerp.usp.br/feridascronicas/recurso_educacional_lp_1_4.html
18. National Pressure Injury Advisory Panel (NPIAP). Prevention and treatment of pressure ulcers/injuries: the international guideline [Internet]. Washington; 2019 [cited 2025 Aug 15]. 405 p. Available from: https://static1.squarespace.com/static/6479484083027f25a6246fcb/t/6553d3440e18d57a550c4e7e/1699992399539/CPG2019edition-digital-Nov2023version.pdf
19. Potter PA, Perry AG, Stockert PS. Fundamentos de enfermagem. 11th. Rio de Janeiro: Guanabara Koogan; 2024. 1644 p.
20. Zhang C, Zhang S, Wu B, Zou K, Chen H. Efficacy of different types of dressings on pressure injuries: systematic review and network meta-analysis. Nurs Open [Internet]. 2023 [cited 2025 Aug 10];10(9):5857-67. Available from: https://doi.org/10.1002/nop2.1867
21. Holman, M. Using tap water compared with normal saline for cleansing wounds in adults: a literature review of the evidence. J Wound Care [Internet]. 2023 [cited 2025 Oct 10];32(8):507-12. Available from: https://doi.org/10.12968/jowc.2023.32.8.507
22. dos Santos TO, da Silva JVL, Rocha RG, Assad LG, da Costa CCP, Pires BMFB. Papain with urea cream in pressure injuries: a case series study. Rev Enferm Atenção Saúde [Internet]. 2024 [cited 2025 Oct 10];13(1):e202404. Available from: https://seer.uftm.edu.br/revistaeletronica/index.php/enfer/article/view/6950
23. O'Connor S, Peltonen LM, Topaz M, Chen LYA, Michalowski M, Ronquillo C, et al. Prompt engineering when using generative AI in nursing education. Nurse Educ Pract [Internet]. 2024 [cited 2025 Nov 19];74:103825. Available from: https://doi.org/10.1016/j.nepr.2023.103825
24. Wang MH, Jiang X, Zeng P, Li X, Chong KKL, Hou G, et al. Balancing accuracy and user satisfaction: the role of prompt engineering in AI-driven healthcare solutions. Front Artif Intell [Internet]. 2025 [cited 2025 Oct 19];8:1517918. Available from: https://doi.org/10.3389/frai.2025.1517918
25. Wang M, Cui J, Lee SMY, Lin Z, Zeng P, Li X, et al. Applied machine learning in intelligent systems: knowledge graph-enhanced ophthalmic contrastive learning with “clinical profile” prompts. Front Artif Intell [Internet]. 2025 [cited 2025 Oct 19];8:1527010. Available from: https://doi.org/10.3389/frai.2025.1527010
26. Flemyng E, Noel-Storr A, Macura B, Gartlehner G, Thomas J, Meerpohl JJ, et al. Position statement on artificial intelligence (AI) use in evidence synthesis across Cochrane, the Campbell Collaboration, JBI, and the collaboration for environmental evidence 2025. JBI Evid Synth [Intermet]. 2025 [cited 2025 Oct 10];23(11):2162-6. Available from: https://doi.org/10.11124/JBIES-25-00480
27. Krifors A, Beskow T, Jonsson M, Lindner KJ, Calås J, Arwsbo V, et al. Improving medication error classification using a reasoning large language model. JAMIA Open [Internet]. 2026 [cited 2025 May 5];9(1):ooag004. Available from: https://doi.org/10.1093/jamiaopen/ooag004
28. Kim SK, Kim GM, Cha S, Jung M. Verification of the validity and reliability of therapeutic communication responses generated by large language models (LLMs): a comparative study of prompt-engineering strategies. J Korean Acad Psychiatr Ment Health Nurs [Internet]. 2025 [cited 2025 May 5];34(Spec):23-35. Available from: https://doi.org/10.12934/jkpmhn.2025.34.S1.23
29. Klein A, Ayoub N, Juhel C, Schuller R, Armstrong F, Pegalajar-Jurado A. Performance of a gelling fibre dressing in management of wounds in a community setting: a sub-analysis of the VIPES study. J Wound Care [Internet]. 2024 [cited 2025 Nov 19];33(7):464-73. Available from: https://doi.org/10.12968/jowc.2024.0125
30. Kimiafar K, Sarbaz M, Tabatabei SM, Ghaddaripouri K, Mousavi AS, Raei M, et al. Artificial intelligence literacy among healthcare professionals and students: a systematic review. Front Health Inform [Internet]. 2023 [cited 2026 May 5];12:168. Available from: https://www.researchgate.net/publication/375600294_Artificial_Intelligence_Literacy_Among_Healthcare_Professionals_and_Students_A_Systematic_Review
Publicado
Como Citar
Edição
Seção
Licença
Copyright (c) 2026 Letícia Pereira Pires, Maithê de Carvalho e Lemos Goulart, Fernanda Garcia Bezerra Góes, Yonara Cristiane Ribeiro, Luíza Lissonger Costa Guimarães, Leidiane Farias de Sousa

Este trabalho está licenciado sob uma licença Creative Commons Attribution 4.0 International License.
A Cogitare Enfermagem se reserva no direito de efetuar, no artigo publicado, alterações de ordem normativa, ortográfica e gramatical, com vistas a manter o padrão culto da língua, respeitando, porém, o estilo dos autores.
O estudo publicado é de inteira responsabilidade do(s) autor(es), cabendo exclusivamente à Cogitare Enfermagem a avaliação do manuscrito, na qualidade de veículo de publicação científica. Não será permitido acréscimo ou mudança de autoria durante a etapa de avaliação ou após aceite do texto submetido.
A Cogitare Enfermagem não se responsabiliza por eventuais violações à Lei nº 9.610/1998, Lei Brasileira de Direitos Autorais.
A Cogitare Enfermagem permite que o autor detenha o copyright dos artigos aceitos para publicação, sem restrições.
Propriedade Intelectual e Termos de uso
Responsabilidade do Autor:
Quanto à autoria, os autores necessitam especificar, na página de identificação, qual o tipo de contribuição individual que exerceu durante a elaboração do artigo. Conforme os critérios estabelecidos pelo ICMJE disponível em: http://www.icmje.org/recommendations/browse/roles-and-responsibilities/defining-the-role-of-authors-and-contributors.html .
É importante destacar que são considerados quatro critérios mínimos de autoria, e todos aqueles designados como autores devem atendê-los, são eles:
- Contribuições substanciais para a concepção ou desenho do estudo; ou a aquisição, análise ou interpretação de dados do estudo;
- Elaboração e revisão crítica do conteúdo intelectual do estudo;
- Aprovação da versão final do estudo a ser publicado;
- Responsável por todos os aspectos do estudo, assegurando as questões de precisão ou integridade de qualquer parte do estudo.
Não será permitido acréscimo ou mudança de autoria durante a etapa de avaliação ou após aceite do texto submetido.
Os artigos publicados estarão licenciados sob a licença Creative Commons CC BY 4.0 Creative Commons - Attribution 4.0 International - CC BY 4.0 – A atribuição adotada pela Cogitare Enfermagem é permitida:
- Compartilhar - copiar e redistribuir o material em qualquer mídia ou formato;
- Adaptar - remixar, transformar e construir sobre o material para qualquer finalidade, mesmo comercialmente;
- Atribuição - Você deve atribuir o devido crédito, fornecer um link para a licença, e indicar se foram feitas alterações. Você pode fazê-lo de qualquer forma razoável, mas não de uma forma que sugira que o licenciante o apoia ou aprova o seu uso;
- Sem restrições adicionais - Você não pode aplicar termos jurídicos ou medidas de caráter tecnológico que restrinjam legalmente outros de fazerem algo que a licença permita.
Responsabilidade do Site:
A Cogitare Enfermagem encoraja os Autores a arquivar seus manuscritos aceitos, publicando-os em blogs pessoais, repositórios institucionais e mídias sociais acadêmicas, bem como postando-os em suas mídias sociais pessoais, desde que seja incluída a citação completa à versão do website da revista.”





















