Population representative gait speed and its determinants

Although gait speed is an important measure to obtain from older adults, much is uncertain about the measure. Therefore, using a large population-based sample, I sought to: (1) gain an indication of "normal" gait speed, (2) compare speed measured over 8 and 20 feet, (3) determine the relationship of several independent variables with gait speed.

The National Health and Nutrition Examination Survey (NHANES) 2001-2002 provided the following data for individuals (> or = 50 years, n = 1923): gait speed over 8 and 20 feet, age, stature, adiposity, and isokinetic knee extension force. Estimates of normal gait speed were calculated for gender and age strata. Speeds over 8 and 20 feet were compared using a paired t-test and a Pearson correlation. The relationship between the independent variables and gait speed were examined using Pearson correlations and forward multiple regression analysis.

Mean gait speeds were 3.26 ft/sec and 3.29 ft/sec over 8 and 20 feet, respectively. Speeds over the distances differed significantly (p = .002) but not meaningfully; they were correlated strongly (r = .933, p < .001). Age, stature, knee extension force, and waist circumference correlated significantly (p < .004) with gait speed over both distances. Age, gender, knee extension force, waist circumference, and stature were shown by multiple regression to predict walking speed over both 8 feet (R = .459, p < .001) and 20 feet (R = .506, p < .001).

The estimates of normal gait speed provided are slower than many of those previously published but may be more representative of the population. As gait speed measured over 8 and 20 feet are comparable, using the shorter distance appears justifiable. Individuals walk faster who are younger, male, taller, stronger, and have smaller waists.