SHM Converge Scientific Abstract ePoster Gallery-External Validation of a Prediction Model for Fall-Related Injuries and Common Predictors Across Multiple Hospitals: A Retrospective Observational Study

External Validation of a Prediction Model for Fall-Related Injuries and Common Predictors Across Multiple Hospitals: A Retrospective Observational Study

Back to course

Pdf Summary

This retrospective observational study externally validated the Saga Fall-Related Injury Risk Model (SFIRM) and searched for common predictors of fall-related injuries across multiple hospitals.<br /><br />The study included consecutive adult inpatients from eight hospitals in Japan between April 2018 and March 2021. From 162,177 total inpatients, 144,777 adults were eligible. After exclusions and random sampling, 2,376 patients were analyzed: 2,341 without fall-related injury and 35 with fall-related injury. Among the sampled patients, the median age was 70 years, 54% were male, the median hospital stay was 9 days, 2.1% experienced falls, and 1.5% had fall-related injuries. Patients with injury were older, with a median age of 76 years, and had longer hospital stays (median 26 days).<br /><br />The main objective was to assess the performance of SFIRM in external settings and identify predictors consistently associated with fall-related injuries. The model’s discrimination was weaker in external validation than in the original setting. Facility-specific AUC values ranged widely from 0.300 to 0.719, indicating variable performance across hospitals.<br /><br />Across eight multivariable analyses, five factors repeatedly appeared as predictors of fall-related injuries: age, male sex, history of falls, diabetes, and bedriddenness rank. These factors were suggested as useful components of a simpler, more broadly applicable prediction model.<br /><br />The authors concluded that SFIRM had lower discrimination in diverse external hospital settings, but the consistent predictors identified across hospitals could help create a practical model to improve fall-related injury prevention and patient safety.

Asset Subtitle

Naoko E. Katsuki

Meta Tag

Author List Chihiro Saito, Eiji Nakatani, Hidetoshi Aihara, Hitomi Shimada, Kaori Amari, Kazuya Kurogi, Kiyoshi Shikino, Maiko Ono, Mariko Yoshimura, Masaki Tago, Midori Tokushima, Motoshi Fujiwara, Naoko E. Katsuki, Risa Hirata, Shizuka Yaita, Shun Yamashita, T

Category Research

Concept SFIRM Prediction Model

Concept Fall-Related Injury

Concept External Validation

Concept Predictor

Concept Retrospective observational study

Distinguished Non-Finalist

Presenter Organization Saga University Hospital

Presenting Author Naoko E. Katsuki

Track Patient Safety

Keywords

fall-related injury

risk model

external validation

hospital inpatients

predictors

age

male sex

history of falls

diabetes

patient safety

SFIRM Prediction Model

Fall-Related Injury

External Validation

Predictor

Retrospective observational study