The measurement properties of the Lean-and-Release test in people with incomplete spinal cord injury or disease

Abstract

Objective: To evaluate test-retest reliability, agreement, and convergent validity of the Lean-and-Release test for the assessment of reactive stepping among individuals with incomplete spinal cord injury or disease (iSCI/D).

Design: Multi-center cross-sectional multiple test design.

Setting: SCI/D rehabilitation hospital and biomechanics laboratory.

Participants: Individuals with motor incomplete SCI/D (iSCI/D).

Interventions: None.

Outcome measures: Twenty-six participants attended two sessions to complete the Lean-and-Release test and a battery of clinical tests. Behavioral (i.e. one-step, multi-step, loss of balance) and temporal (i.e. timing of foot off, foot contact, swing of reactive step) parameters were measured. Test-retest reliability was determined with intraclass correlation coefficients, and agreement was evaluated with Bland-Altman plots. Convergent validity was assessed through correlations with clinical tests.

Results: The behavioral responses were reliable for the Lean-and-Release test (ICC = 0.76), but foot contact was the only reliable temporal parameter using data from a single site (ICC = 0.79). All variables showed agreement according to the Bland-Altman plots. The behavioral responses correlated with scores of lower extremity strength (0.54, P<0.01) and balance confidence (0.55, P < 0.01). Swing time of reactive stepping correlated with step time (0.73, P < 0.01) and cadence (-0.73 P < 0.01) of over ground walking.

Conclusions: The behavioral response of the Lean-and-Release test is a reliable and valid measure for people with iSCI/D. Our findings support the use of the behavioral responses to evaluate reactive stepping for research and clinical purposes.

Trial registration: ClinicalTrials.gov identifier: NCT02960178.

Keywords: Postural balance; Spinal cord injuries; Validation study.

Figures

Figure 1

Lean-and-Release set up at site 1. Figure originally printed in Chan et al. 2019. At site 2, there were only three force plates, with one wide force plate (60 cm) in front of participants during the initial standing position, and the cable behind the participants was 307 cm long.

Figure 2

Comparison of foot off and foot contact values between site 1 and site 2. Foot off and foot contact values at site 1 (n = 12) and site 2 (n = 5).

Figure 3

Bland–Altman plots of the agreement between the two testing sessions for each Lean-and-Release variable. Limits of agreement are set to the 95% confidence interval.