Reliability, smallest real difference and concurrent validity of indices computed from GRF components in gait of stroke patients.

We analysed reliability, smallest real difference (SRD) and concurrent validity of indices computed from the ground reaction force (GRF) vertical and fore-aft components in a sample of 56 patients with hemiparesis secondary to stroke. These parameters have been recommended for the assessment of weight-bearing and propulsion. The sample size was calculated based on guidelines for reliability studies and patient levels of impairment ranged from mild to severe. Reliability was assessed by the intraclass correlation coefficient (ICC), SRD was computed as the 95% confidence interval (CI) of the standard error and concurrent validity was assessed using the Spearman correlation coefficient between each index and gait speed, with the latter being used as the criterion standard. Excellent reliability (ICC>0.90) for all indices was achieved by averaging values of three consecutive gait trials. SRD ranged between 5% and 10% of the sample grand mean for vertical GRF-based indices, and between 20% and 40% for fore-aft GRF-based indices. All indices but one showed concurrent validity with walking speed, with correlation coefficients ranging from 0.423 (p<0.01) to 0.834 (p<0.01). Amongst studied indices, the mean value of vertical GRF and the mean value of the propulsive part of the fore-aft component showed the best performance in terms of ICC, SRD and concurrent validity. These appear to be the most appropriate indices to assess weight-bearing and propulsive ability, respectively, in this group.