TIMED UP AND GO: SECONDS OR ACCELERATION TO MEASURE MOBILITY IN HEALTHY OLDER PEOPLE?

Abstract: This study investigates the convergent validity between acceleration features and the time needed to complete three variations of the Timed Up and Go (TUG) test. The sample consisted of 73 active and healthy older people who were non-fallers and community-dwelling (São Carlos-SP, Brazil). Participants used a triaxial accelerometer at the waist while performing three variations of the TUG. Entropy, amplitude, and frequency features of the signal were extracted from all accelerometry data. The Pearson correlation coefficient was used to investigate convergent validity. The average age of the sample was 70.16 years (SD=6.79), with a maximum age of 88 years and a minimum of 60 years. Women represented 56.6% (n=41) of participants. There were weak correlations between the acceleration data and the time needed to complete the simple TUG and its variations. Moderate to strong correlations were observed between variables extracted from the accelerometer signal, and strong correlations were found between the TUG variations seconds. This result demonstrates that although both can be used to measure mobility, they do not measure the same constructs. Combining TUG with an accelerometer and choosing only one variation of the TUG may provide a good combination.