Community ambulation in patients with chronic stroke: how is it related to gait speed?

OBJECTIVE To explore the strength of the association between gait speed and community ambulation and whether this association is significantly distorted by other variables. DESIGN Cross-sectional study conducted 3 years after stroke. SUBJECTS A total of 102 patients after first-ever stroke following inpatient rehabilitation who are now living in the community. METHODS Community ambulation was determined by a self-administered questionnaire with 4 categories. Gait speed was assessed by the 5-m walking test. Possible confounding factors included in the analyses were: age, hemisphere, living alone, history of falls, use of assistive walking devices, executive function (Trail Making Test), depression (Center for Epidemiologic Studies-Depression scale), fatigue (Fatigue Severity Scale), motor function (Motricity Index), standing balance (Berg Balance Scale) and walking endurance (SF36). RESULTS Twenty-six percent of the patients were non-community walkers or limited community walkers. The optimal cut-off point for community ambulation was 0.66 m/sec, with an area under the curve of 0.85. Although gait speed was significantly related to community ambulation, this association was confounded by balance, motor function, endurance and the use of an assistive walking device. These factors reduced the regression coefficient of gait speed by more than 15%. CONCLUSION Gait speed is an important factor related to community walking; however, ability to walk in the community is determined by several underlying factors, e.g. balance, motor function, endurance and assistive walking device.