ABSTRACT
Purpose@#The purpose of this study was to validate the Edmonson psychiatric fall risk assessment tool (EPFRAT) for psychiatric inpatients. @*Methods@#Data from retrospective study were collected from 670 adult inpatients in two departments of mental health medicine of a tertiary general hospital by reviewing their electronic medical records. There were 41 patients who experienced falls and 629 patients who did not experience falls during the period from January to December 2019. Data were analyzed by sensitivity, specificity, positive predictive value, negative predictive value, and a receiver-operating characteristic curve (ROC) for validity assessment using the IBM SPSS/WIN 26.0 program. @*Results@#Factors affecting falls were the participant’s age, guardian’s residence, high-risk determination at the time of admission, and comorbidity. At the 85 points where the point of sum of the sensitivity and specificity was largest, the sensitivity, specificity, positive predictive value, and negative predictive value of EPFRAT were 92.7%, 79.7%, 22.9%, and 99.4%, respectively. The area under the ROC to assess the overall validity of the tool was .92 (95% CI 0.89~0.94). @*Conclusion@#The EPFRAT was proved to be valid and reasonable for predicting falls in psychiatric inpatients. Based on the results of this study, it could be used for the assessment of high-risk patients for falls in psychiatric units.