Abstract: Objective To explore the reliability, validity and predictive value of three fall risk self-assessment tools in hospitalized elderly patients. Methods A total of 420 hospitalized elderly patients who met the inclusion and exclusion criteria in Xiangya Hospital of Central South University from January to June 2023 were selected by convenient sampling method. The Chinese version of Self-Assessment of Fall Risk Questionnaire (CSAFR ), the Chinese version of the i Engaging fall self-assessment tool and the Chinese version of the Self-rated Fall Risk Questionnaire (CFRQ ) were used to evaluate the risk of fall. The intra-group correlation coefficient was used to evaluate the investigator reliability of the three tools. Cronbach’ s α coefficient was used to evaluate the internal consistency of the three tools. Factor analysis was used to evaluate the construct validity of the three tools. With the Chinese version of the Morse Fall Scale (CMFS) as the standard, the correlation coefficient of the measured scores of the three tools with CMFS was analyzed. The receiver operating characteristic (ROC) curve was used to analyze and compare the predictive effects of the three scales on falls. Results The intraclass correlation coefficients of CSAFR, the Chinese version of the i Engaging fall self-assessment tool and CFRQ were 0.862, 0.804 and 0.926, respectively. The Cronbach’s α coefficients were 0.515,0.726 and 0.742, and the validity KMO values were 0.618,0.662,0.831 (P<0.01). Orthogonal rotation transformation was performed, and two common factors were extracted from CSAFR, with a cumulative variance contribution rate of 47.4 %. Seven common factors were extracted from Chinese version of i Engaging fall self-assessment tool, with a cumulative variance contribution rate of 58.2%. Three common factors were extracted from the CFRQ, with a cumulative variance contribution rate of 43.3%. The correlation coefficients of CSAFR, i Engaging fall self-assessment tool and CFRO with CMFS were 0.398, 0.376, and 0.478, respectively. The results of ROC curve analysis showed that the AUC values of three tools were 0.814, 0.689 and 0.821, respectively. When the cut-off points of the three tools were 3.5, 5.5 and 4.5 respectively, the Youden indexes were 0.469, 0.259 and 0.503, respectively. The sensitivity was 64.9%, 88.3% and 81.8% respectively, and the specificity was 82.1%, 37.6% and 68.5%, respectively. Conclusions This study suggests that the three fall risk self-assessment tools have certain reliability and validity, and have certain predictive value for the falls in hospitalized elderly patients. The reliability, validity and predictive value of CFRQ for the elderly are highest, and the evaluation time is not long. It is an effective tool and can provide important reference for the self-assessment of fall risk in the hospitalized elderly patients in China.

Key words: fall, risk self-assessment, aged, reliability and validity, predictive value