Propuesta de automatización para proceso de agendamiento de citas médicas post-alta a partir de epicrisis en el servicio de medicina interna del Hospital Las Higueras.
Loading...
Date
2024
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad de Concepción
Abstract
El aumento exponencial de la producción de datos en el ámbito de la salud ha provocado la subutilización de la información contenida en ellos, una de las razones es debido a su naturaleza no estructurada. Este informe aborda la problemática del proceso manual de agendamiento de citas médicas post-alta, que genera ineficiencias y una carga administrativa solucionable mediante tecnologías y aprovechamiento de datos existentes.
Para resolver esta situación, se propuso un sistema automatizado basado en técnicas de Machine Learning, Data Mining y Procesamiento de Lenguaje Natural, utilizando herramientas y bibliotecas de Python y Google. Este sistema permitió clasificar y priorizar episodios de pacientes para su agendamiento en controles post-alta, transformando la información extraída de las epicrisis en un formato estructurado y utilizable para análisis y toma de decisiones, considerando variables demográficas, diagnósticas y tiempos de espera.
La finalidad del proyecto fue mejorar la eficiencia y efectividad del agendamiento, reduciendo errores en la lectura e ingreso de datos, además de mejorar la coordinación entre el personal hospitalario desde la emisión de la epicrisis hasta la programación de la cita del paciente. Esto resulta en una disminución de los desplazamientos innecesarios del personal médico hacia los mesones de los policlínicos de interconsulta para coordinar controles post-alta, facilitando una transmisión rápida y precisa de la información clave.
La metodología incluyó el desarrollo de un algoritmo para extraer y analizar datos de las epicrisis de la base de datos del Hospital Las Higueras de Talcahuano, así como su integración con aplicaciones de Google para visualización y gestión. Los resultados demostraron que el sistema implementado procesa los datos de manera eficiente y coherente en el entorno hospitalario, proporcionando información sobre los episodios y su flujo para el agendamiento en diferentes policlínicos. Donde la herramienta, desde el punto de vista de un usuario experto, responde a las necesidades de información del proceso de alta y controles del Servicio de Medicina Interna. Esto destaca el potencial de las tecnologías de la información en la gestión médica actual, sirviendo como modelo para innovar y mejorar entornos dinámicos.
The exponential increase in data production in the healthcare sector has led to underutilization of the information contained within them, primarily due to their unstructured nature. This report addresses the issue of manual scheduling processes for post-discharge medical appointments, which create inefficiencies and administrative burdens that can be resolved through technology and leveraging existing data. To address this situation, an automated system was proposed based on Machine Learning, Data Mining, and Natural Language Processing techniques, utilizing Python tools and libraries and Google. This system enabled the classification and prioritization of patient episodes for scheduling post-discharge appointments, transforming information extracted from discharge summaries into a structured and usable format for analysis and decision-making, considering demographic variables, diagnoses, and waiting times. The project aimed to enhance scheduling efficiency and effectiveness by reducing errors in data reading and entry, as well as improving coordination among hospital staff from discharge summary issuance to patient appointment scheduling. This results in fewer unnecessary trips by medical personnel to outpatient clinic desks for post-discharge appointment coordination, facilitating rapid and precise transmission of key information. The methodology involved developing an algorithm to extract and analyze data from discharge summaries at Hospital Las Higueras de Talcahuano's database, integrating it with Google applications for visualization and management. The results demonstrated that the implemented system efficiently and consistently processes data within the hospital environment, providing insights into patient episodes and their flow for scheduling across various outpatient clinics. From the perspective of an expert user, the tool meets the information needs of the discharge and follow-up process in the Internal Medicine Service. This underscores the potential of information technologies in current medical management, serving as a model for innovation and improvement in dynamic environments.
The exponential increase in data production in the healthcare sector has led to underutilization of the information contained within them, primarily due to their unstructured nature. This report addresses the issue of manual scheduling processes for post-discharge medical appointments, which create inefficiencies and administrative burdens that can be resolved through technology and leveraging existing data. To address this situation, an automated system was proposed based on Machine Learning, Data Mining, and Natural Language Processing techniques, utilizing Python tools and libraries and Google. This system enabled the classification and prioritization of patient episodes for scheduling post-discharge appointments, transforming information extracted from discharge summaries into a structured and usable format for analysis and decision-making, considering demographic variables, diagnoses, and waiting times. The project aimed to enhance scheduling efficiency and effectiveness by reducing errors in data reading and entry, as well as improving coordination among hospital staff from discharge summary issuance to patient appointment scheduling. This results in fewer unnecessary trips by medical personnel to outpatient clinic desks for post-discharge appointment coordination, facilitating rapid and precise transmission of key information. The methodology involved developing an algorithm to extract and analyze data from discharge summaries at Hospital Las Higueras de Talcahuano's database, integrating it with Google applications for visualization and management. The results demonstrated that the implemented system efficiently and consistently processes data within the hospital environment, providing insights into patient episodes and their flow for scheduling across various outpatient clinics. From the perspective of an expert user, the tool meets the information needs of the discharge and follow-up process in the Internal Medicine Service. This underscores the potential of information technologies in current medical management, serving as a model for innovation and improvement in dynamic environments.
Description
Tesis presentada para optar al título de Ingeniera Civil Biomédica
Keywords
Alta del paciente, Agendas Programas para computador, Medicina interna