Predictive Validity of Different Walking Measures to Identify the Incident Long-Term Care Needs in Older Adults.

OBJECTIVES: A comfortable walking speed is a suitable measurement of functional status in older adults. In addition to assessing their comfortable walking speed, two complex walking tests were administered to a cohort of older people, assuming that these tests would be a more sensitive predictor of the incident long-term care needs than comfortable walking speed. DESIGN: A prospective observational study was conducted to collect data. SETTING AND PARTICIPANTS: Among the initial 5,563 community-dwelling independent older adults (aged ≥ 65 years), 935 were excluded and the data of 4,628 (mean age, 73.9 ± 5.5 years, 65-94 years; 2,052 men, 2,576 women) older adults were finally analyzed. METHODS: Three walking tasks were administered: comfortable, complicated balance, and Go-stop walking. Complicated balance walking was measured under comfortable walking conditions, with participants having to walk with their hands crossed at the shoulder joint at 90°. For the Go-stop walking test, the time taken to walk 2 meters was measured using a stopwatch. For two years following baseline assessments, participants received monthly follow-ups for incident certification of the need for care under the long-term care insurance (LTCI) system. RESULTS: Low performance in comfortable, complicated balance, and Go-stop walking were 29.8%, 37.7%, and 35.1%, respectively. During the 24-month follow-up period, 246 participants (5.3%) required LTCI certification. The Youden Index was used to determine the cut-points of the incident long-term care needs in the comfortable, complicated balance, and Go-stop walking conditions, which were 1.055 m/s, 0.936 m/s, and 3.205 seconds, respectively. Participants classified as exhibiting low performance included 1,381 (29.8%) under comfortable walking, 1,746 (37.7%) under complicated balance walking, and 1,623 (35.1%) under the Go-stop walking tests. The C-indices of the comfortable, complicated balance, and Go-stop walking tests were 0.72 (95% confidence interval (CI) 0.69-0.76), 0.71 (95% CI 0.67-0.74), and 0.65 (95% CI 0.61-0.69), respectively. Cox proportional hazards regression model revealed significant relationships between the incident long-term care needs and the comfortable (hazard ratio (HR) 2.14, 95% CI 1.62-2.84), complicated balance (1.81, 1.36-2.41), and Go-stop (1.46, 1.12-1.91) walking conditions. CONCLUSIONS AND IMPLICATIONS: The findings suggest that slow walking speed has a considerably greater impact on the incident long-term care needs in older adults. However, the complex walking task did not improve the predictive performance. Comfortable walking speed tests, which can easily be measured to predict the future incident long-term care needs, are effective tools in community health promotion and primary care.