The intra-rater reliability of the Balance Performance Monitor when measuring sitting symmetry and weight-shift activity after stroke in a community setting

Objective: To examine the intra-rater reliability of sitting symmetry and weight-shift activity measurements in poststroke adults.
Design: An intra-rater reliability study.
Setting: A community setting.
Subjects: Adult stroke survivors attending stroke support groups within the community of Nottingham (UK).
Main measures: The Balance Performance Monitor used to measure sitting symmetry and weight-shift activity. Intraclass correlation coefficients (ICCs) and their 95% confidence intervals (95% CI) were calculated. The Bland-Altman method for assessing agreement is also presented.
Results: We tested 49 participants (median age 73 years; interquartile range 68-81 years). Between-test reliability for sitting symmetry was high: ICC (1,1) = 0.93 (95% CI 0.87 ≤ ICC5 ≤ 0.96). The mean difference between the measures (d̄) was-0.08 (95% CI-0.48 ≤ d̄ ≤ 0.31); the standard deviation of the differences (SDdiff) was 1.383. The coefficient of repeatability was 2.76; the 95% limits of agreement were -2.850 and 2.682. Between-test reliability for weight-shift activity was also high: ICC (1,1) = 0.86 (95% CI 0.775 ≤; ICC5 ≤; 0.92). Bland-Altman d̄= - 0.08 (95%CI-0.19 ≤ d̄ ≤ 0.35), SDdiff = 0.936. The coefficient of repeatability was 1.87; the 95% limits of agreement were-1.792 and 1.952.
Conclusions: The 95% CI for d for both parameters crossed zero, indicating that between-test bias is unlikely. Sitting symmetry and weight-shift activity measures demonstrated acceptable levels of reliability.