Abstract
BACKGROUND: We aimed to evaluate the predictive value of the 'Timed Up and Go' test (TUGT) for identifying fall risk in community-dwelling elderly. METHODS: From Aug 2016 to Feb 2017, cluster sampling was conducted among residents aged over 60 from 15 communities in Songjiang district, Shanghai. Face-to-face questionnaire interviews and TUGT measures were conducted to collect data. RESULTS: 6,014 participants were enrolled, with an average age 72.7±7.0 years. 637 (10.6%) elderly people had a fall experience in the past year. TUGT for the non-fall group, one-fall group and recurrent-fall group was 9.02±4.39, 10.00±5.26 and 10.78±4.51 seconds respectively (P<0.001). ROC analysis showed that the TUGT cut-off point for the elderly was 12.5 seconds and AUC was 0.573 for any-fall group and 0.613 for recurrent-fall group respectively. After adjusting for age and gender, the predictive value was not high for any-fall group (AUC=0.614) and recurrent-fall group (AUC=0.648). The TUGT cut-off point for the elderly aged below 65, 65-74, 75-84 and 85 and above was 13.52s, 12.51s, 12.51s, 12.00s, respectively. After adjustment of the confounding factors, the OR values for the risk of fall for the elderly men and women who completed TUGT longer than the cut-off point was 2.404 and 2.075 times higher than those who completed TUGT shorter than the cut-off point, respectively. CONCLUSION: TUGT with the cut-off score of 12.5s has limited capability in predicting fall risk in community-dwelling elderly.