Estimation of reduced walking speed using simple measurements of physical and psychophysiological function in community-dwelling elderly people: a cross-sectional and longitudinal study.

BACKGROUND

Early detection of reduced mobility function is important in elderly people. Usual walking speed is useful to assess mobility function, but is often not feasible in a community setting.

AIMS: This study aimed to explore a simple surrogate indicator of usual walking speed in elderly people.

METHODS

The participants were 516 community-dwelling elderly people. As a baseline survey, the usual walking speed and candidates of surrogate indicators including physical function and psychophysiological function were measured. After 2 years, the occurrence of mobility limitation was assessed.

RESULTS

In cross-sectional analysis, a linear regression model with maximum step length, age, and sex presented the most favourable adjusted R2 of 0.426 for estimating usual walking speed. Maximum step length (MSL) also showed good predictive accuracy for usual walking speed < 0.8 m/s {area under the curve [AUC] 0.908 [95% confidence interval (CI) 0.811, 1.000]} and < 1.0 m/s [AUC 0.883 (95% CI) 0.832, 0.933)] in receiver-operating characteristic (ROC) analysis. In longitudinal analysis, the predictive accuracy of MSL for mobility limitation [AUC 0.813 (95% CI 0.752, 0.874)] was similar to that of usual walking speed [AUC 0.808 (95% CI 0.747, 0.869)] in ROC analysis.

CONCLUSIONS AND DISCUSSION: The results of this study suggest that MSL may serve as a simple surrogate indicator of UWS in elderly people.