Use this link to cite:
https://hdl.handle.net/2183/49482 Aplicación web para a simulación de perfís clínicos psiquiátricos mediante LLMs como ferramenta de formación de profesionais sanitarios
Loading...
Identifiers
Publication date
Authors
Lires Saborido, Andrés
Advisors
Other responsabilities
Universidade da Coruña. Facultade de Informática
Journal Title
Bibliographic citation
Type of academic work
Academic degree
Abstract
[Resumo] A formación práctica en Medicina, especialmente no ámbito da Psiquiatría, adoita atopar limitacións á hora de dispoñer de pacientes reais cos que o alumnado poida adestrar as súas habilidades comunicativas e de entrevista clínica. Este Traballo Fin de Grao (TFG) presenta una plataforma web de simulación clínica asistida por Intelixencia Artificial (IA) que permite ao estudantado interactuar con pacientes virtuais a través dun chat en tempo real. O sistema, desenvolvido en colaboración coa área de Psiquiatría da Facultade de Medicina da Universidade de Santiago de Compostela (USC), emprega un Large Language Model (LLM) para xerar o comportamento de perfís clínicos específicos. Ademais, utiliza un enfoque de LLM-as-a-Judge para avaliar automaticamente o desempeño do usuario ao finalizar a consulta, emitindo rúbricas detalladas e xustificadas a partir da transcrición. O obxectivo é proporcionar un contorno seguro no que o alumnado poida gañar destreza antes da práctica real, á vez que se dota ao profesorado dun panel de xestión avanzado para crear casos e supervisar o progreso. A proposta achega unha ferramenta escalable que complementa a docencia tradicional, automatizando a avaliación formativa e mellorando o adestramento do razoamento clínico.
[Abstract] Practical training in Medicine, especially in the field of Psychiatry, often encounters limitations regarding the availability of real patients for students to practice their communication and clinical interviewing skills. This Traballo Fin de Grao (TFG) presents a web-based clinical simulation platform assisted by Artificial Intelligence (IA) that allows students to interact with virtual patients through a real-time chat. The system, developed in collaboration with the Department of Psychiatry at the Faculty of Medicine of the Universidade de Santiago de Compostela (USC), employs a Large Language Model (LLM) to generate the behavior of specific clinical profiles. Furthermore, it utilizes an LLM-as-a-Judge approach to automatically evaluate user performance at the end of the consultation, generating detailed and justified rubrics based on the transcript. The objective is to provide a safe environment where students can gain proficiency prior to real clinical practice, while simultaneously equipping the teaching staff with an advanced management dashboard to create cases and monitor progress. The proposal offers a scalable tool that complements traditional teaching, automating formative assessment and enhancing clinical reasoning training.
[Abstract] Practical training in Medicine, especially in the field of Psychiatry, often encounters limitations regarding the availability of real patients for students to practice their communication and clinical interviewing skills. This Traballo Fin de Grao (TFG) presents a web-based clinical simulation platform assisted by Artificial Intelligence (IA) that allows students to interact with virtual patients through a real-time chat. The system, developed in collaboration with the Department of Psychiatry at the Faculty of Medicine of the Universidade de Santiago de Compostela (USC), employs a Large Language Model (LLM) to generate the behavior of specific clinical profiles. Furthermore, it utilizes an LLM-as-a-Judge approach to automatically evaluate user performance at the end of the consultation, generating detailed and justified rubrics based on the transcript. The objective is to provide a safe environment where students can gain proficiency prior to real clinical practice, while simultaneously equipping the teaching staff with an advanced management dashboard to create cases and monitor progress. The proposal offers a scalable tool that complements traditional teaching, automating formative assessment and enhancing clinical reasoning training.
Description
Editor version
Rights
Todos los derechos reservados. Prohíbese a reprodución, transformación, distribución e comunicación pública da obra por terceiros. En cambio, permítese a súa visualización e a descarga dunha copia privada para o uso persoal.






