Conectar
The aim of this study was to innovatively utilise the BERTopic model for topic modelling in order to comprehensively identify and understand the factors contributing to bed falls.
Retrospective study.
The study collected 241 reports of bed fall accidents recorded by nurses from Peking University Third Hospital Nursing Department from 2014 to 2024. Among them, 102 reports met the inclusion and exclusion criteria.
This study follows the Minimum Information for Medical AI Reporting (MINIMAR). It collected patient bed fall reports from Peking University Third Hospital between 2014 and June 2024, preprocessed the texts, utilised the BERTopic library in Python for topic modelling, and manually aggregated secondary topics by combining visualisation results and professional knowledge.
We utilised cluster bar charts to visually display the distribution of the 22 secondary topics and further consolidated them into five core topics through the use of a topic distribution diagram and a topic similarity matrix diagram. These topics were related to patient factors, ward equipment and surroundings factors, medication risk factors, caregiver factors, and nursing practice factors. The study highlights the environment's specificity in bed falls, especially bedside safety and patient-bed rail interaction.
The innovation of this study lies in the successful utilisation of BERTopic technology to identify topics of risk factors for bed falls through alternative data sources, providing a scientific basis for formulating preventive measures. The findings aim to optimise nursing processes, improve ward environments and enhance educational training, ultimately reducing patient bed falls and enhancing medical safety, nursing quality and patient experience.
This study not only helps nurses identify risk factors for patient bed falls, but also provides important guidance for developing effective prevention strategies.
No patient or public contribution applied.
FreshRSS 