Accelerometer-based gait characteristics and their discrimination of gait independence in inpatients with subacute...

Gait analysis using inertial measurement devices can identify multifaceted gait disorders after a stroke. Although the usefulness of gait assessment using inertial measurement devices has been reported, its accuracy in discriminating gait independence in patients hospitalized for subacute stroke has not yet been validated.

Can trunk acceleration indices discriminate between dependent and independent walking in patients with subacute stroke?

Thirty-five patients with subacute stroke (mean ± standard deviation, 75.5 ± 9.8 years, 19 males), who were able to understand instructions, had a premorbid modified Rankin scale <3, and were able to walk 16 m straight ahead under supervision were included. The stride regularity, harmonic ratio, and normalized root mean square of trunk accelerations were measured in three directions (mediolateral, vertical, and anterioposterior) during comfortable walking. The Functional Ambulation Categories were used as the dependent variable to classify the patients into two groups (dependent and independent walking groups), and each trunk acceleration index was used as the independent variable to calculate the area under the curve using receiver operating characteristic curves.

Twelve patients were in the dependent group and 23 were in the independent group. The normalized root mean square in both the mediolateral and vertical directions were excellent discriminators of walking independence, with an area under the curve greater than 0.8. The cutoff values (sensitivity/specificity) were 2.20 m²/s² (0.783/0.833) and 2.82 m²/s² (0.739/0.833), respectively.

The magnitude of vertical and lateral acceleration during gait in patients with subacute stroke, has excellent accuracy in discriminating between dependent and independent gaits. The results of this study will be useful for inexperienced clinicians working with stroke patients presenting with gait disturbances to accurately determine gait independence based on objective data.