|| Checking for direct PDF access through Ovid
To compare the extent to which gait speed measured in the clinic setting differs from that measured in the community.Participants completed the 10-m walk test at a self-selected speed in a clinic setting. Following this they completed a 300-m community-based walking circuit that covered a variety of environmental conditions. Gait velocity was sampled at different points in the circuit. The same circuit and sampling points were used for all participants. Clinic gait velocity was compared to gait velocity measured on five occasions during the community-based circuit.Physiotherapy clinic and local shopping mall.Twenty-eight chronic stroke patients who regularly accessed the community divided into two groups based on their gait velocity in the clinic.Walking velocity.Spearman rank correlation coefficient indicated that there was a strong correlation between the total time taken to walk the 300-m course and the clinic-based gait velocity (r = −0.88, P < 0.0001). A linear mixed model with repeated measures analysis revealed significant interaction between community measures for group A versus group B (F4,26 = 4.49, P = 0.0068) and significant differences across community conditions (F4,26 = 7.12, P = 0.0005).The clinic-based 10-m walk test is able to predict walking velocity in a community setting in chronic stroke patients who score 0.8 m/s or faster. However, for those who score less than 0.8 m/s in the clinic test, gait velocity in the community may be overestimated.