Normalizing step-to-step variability to age in children and adolescents with hemiplegia

Children with hemiplegia often demonstrate gait deviations including increased variability and asymmetry. Step-to-step gait variability decreases over childhood and increases in the presence of neurologic dysfunction. Gait variability in children with hemiplegia should therefore be interpreted in reference to age-related norms RESEARCH QUESTION: Does conversion of the enhanced gait variability index (eGVI) to age-normalized z-scores improve interpretation of gait variability in children with hemiplegia?

Ten children (11.2 +/- 4.1 years) with hemiparetic gait due to stroke were recruited for a small prospective pilot intervention study. Participants walked at self-selected speed over an instrumented walkway while barefeet and while wearing shoes. eGVI values from baseline sessions were calculated and converted to age-normalized z-scores (eGVI_z) based on published norms. Differences in gait variability between sides and footwear conditions, and its relationship to walking speed, were examined.

There were no differences in raw eGVI or eGVI_z between paretic and nonparetic sides (eGVI p = 0.31; eGVI_z p = 0.31) or between footwear conditions (eGVI p = 0.62; eGVI_z p = 0.33). Average raw eGVI values were just over two standards deviations above the reference mean of 100 (121.2, 122.1, 120.3 for mean (average of both limbs), nonparetic side and paretic side, respectively), indicating significantly greater step-to-step gait variability than in typical gait. However, when converted to age-normalized z-scores (eGVI_z), variability deviated less from the normative sample, averaging just over one standard deviation above the reference mean (1.2, 1.3, 1.1 for mean, nonparetic side and paretic side, respectively). We also observed a relationship between eGVI_z and walking speed in our sample.

We suggest that eGVI values in children be converted to z-scores or otherwise age-normalized so as not to inflate the degree of variability reported in clinical pediatric populations. Future work with larger samples will offer greater insight into gait variability in various clinical pediatric populations.