Learn about free on-demand learning
Reference our medication guides for helpful information
Make a choice that matters
The best care starts with the best information
The original purpose of the TUG was to test basic mobility skills of frail elderly persons. The test has been used in other populations, including people with arthritis, stroke, and vertigo1.
Measurement of the time in seconds for a person to rise from sitting from a standard arm chair, walk 3 meters, turn, walk back to the chair, and sit down. The person wears regular footwear and customary walking aid.
The TUG was adapted by Podsiadlo and Richardson2 from the Get Up and Go Test of Mathias et al3.
The original Get Up and Go Test3 used an ordinal scoring system based on the observer's assessment of a person's risk of falling. The TUG2 uses time in seconds.
Number of Items in Scale: Not applicable.
Subscales: Not applicable.
The TUG was developed in frail elderly adults 60-90 years of age referred to a geriatric hospital, and it targets community dwelling frail elders.
The TUG has been used in other conditions, including arthritis, stroke, and vertigo1. The TUG was not useful with cognitively impaired frail elderly people because 35.5% of the group were physically unable to perform the test4.
WHO ICF Components: Changing and maintaining body position (d410-d429) and walking (d450).
Method: Easily administered physical performance test with single tester.
Training: None required.
Time to Administer / Complete: 15 minutes or less.
Arm chair with a seat height of approximately 46 centimeters and arm height of 65 centimeters; 3 meter walkway; stopwatch or wrist watch with a second hand.
Availability / Cost: Readily available and inexpensive.
TUG PDF or Video provided by Center for Disease Control and Prevention (CDC).
Responses: Scale. Not applicable. Score range. Not applicable.
Interpretation of Scores
A cut-off score of ≥ 13.5 seconds was shown to predict falls in community-dwelling frail elders, but this score is not verified in other studies4. Scores of ≥ 30 seconds correspond with functional dependence in people with pathology2. Standardized cut-off scores to predict risk of falling have not yet been established.
Method of Scoring: Time in seconds.
Time to Score: Same as time of test.
Training to Score: None required.
Training to Interpret: None required.
No formal normal values are available. Healthy adults over 79 years old took 7-10 seconds2; frail elderly subjects took 10-240 seconds to perform, with 45 out of 57 subjects performing the test in less than 40 seconds2; all healthy community dwelling subjects 65-84 years of age performed the test in < 20 seconds without an assistive5. In a meta-analysis of 21 studies reporting TUG times in healthy older adults, the mean times progressively increased with age with 8.1 seconds (95% confidence interval = 7.1-9.0) among 60 to 69 year olds, 9.2 seconds (95% CI = 8.2-10.2) among 70-79 year olds, and 11.3 seconds (95% CI = 10.0-12.7) among 80-99 year olds6.
Inter-rater reliability is high with a same day, three-rater intra-class correlation coefficient (ICC) of 0.992. The ICC was 0.97 in another study of inter-rater reliability among 3 physiotherapists7. The inter-rater reliability was an ICC of 0.99 for a physical therapist, physician, and patient attendant on consecutive visits, and the consecutive intra-rater reliability was an ICC of 0.99. In another study of intra-session, test-retest reliability, the ICC (model 2, 1) was 0.978. Intra-rater reliability over longer periods (up to 132 days) is not as high with a reliability of 0.749. Test-retest (2-7 days) standard error of measurement has been measured as 1 second7.
Criterion. Moderate to high correlations have been observed with scores on Berg Balance Scale2, 10, gait speed2, 11, 12, stair climbing13, and the Barthel Index of Activities of Daily Living Scale2, 7. TUG scores of greater than 10 seconds were predictive of near-falls in older adults with hip osteoarthritis (Odds ratio 3.1, 95% confidence intervals 1.0-9.9)14
Construct. Validity has been determined by examining differences in scores for patients who were independent and dependent in basic transfers. All subjects who completed the TUG in <20 seconds were independent in transfers. Subjects requiring ≥ 30 seconds were dependent2. The TUG scores for community-dwelling subjects (65–95 years old) with a history of falling were slower than for people with no history of falling14, 15,. Using logistic regression, sensitivity of the TUG to predict falls using a cut-off score of ≥ 13.5 seconds was 0.80 with a specificity of 1.004. TUG scores correlate with mobility and strength complaints15, 16. The TUG is capable of discriminating people at risk of falling from healthy elderly subjects and young control subjects17. In elderly Mexican-American women, those with the best and worst performance on the TUG were more likely to fall than those with moderate performance18.
Responsiveness/sensitivity to change
TUG scores changed following a quadriceps and hamstrings strengthening program for patients with rheumatoid arthritis compared with subjects who received no strengthening19. A small effect size (ES = 0.33, SRM = 0.35) was reported for persons with knee osteoarthritis who had received physical therapy (French)21. The timed test would be more sensitive to change than ordinal measures. Minimum clinically important differences (MCID) or clinically meaningful change of the TUG was established by comparing baseline and 9 weeks scores of 65 people with osteoarthritis who were undergoing physical therapy. A reduction in time greater than or equal to 0.8, 1.4 and 1.2 seconds on the TUG was determined to be the MCID22.
Performance on the TUG is related to multiple factors. A history of arthritis increases the risk of falling as measured by balance tests such as the TUG19. As a test of balance, the TUG may be most useful for patients with rheumatoid arthritis in functional class IV7. Scores differed based on type of footwear worn. They were longest with dress shoes and shortest with walking shoes23. Cognitively impaired subjects took longer to perform the TUG than unimpaired subjects4. Chair type (standard arm chair, armless chair and easy chair) does not affect speeds 1. There is a tendency for TUG times to increase with age8. Including a cognitive or manual task concurrent with the TUG increased the times. Sensitivity for predicting falls was 0.80 and specificity was 0.934. The use of an assistive device increased the TUG times4. A cane increased the time the least, followed by a rolling walker and then a pick up walker5. Female candidates for hip or knee arthroplasty took 2.2 times longer than healthy controls; male candidates took 1.9 times as long24. The TUG measures limited aspects of balance (rising, walking, turning, and sitting).
For questions or comments, contact ARHP@rheumatology.org.
ARHP Research Committee - content last reviewed 2015.