Validity of common physical function tests performed online for older adults

Abstract

Background: Objectively measured physical function tests are important predictors of institutionalization, disability, and premature mortality. Although physical function was usually assessed in person prior to the COVID-19 pandemic, there is a need to investigate whether physical function tests performed online are valid.

Objective: The purpose of this study was to determine the validity of common physical function tests conducted online compared to in-person testing in older adults.

Methods: Physical functions included gait speed, one leg stance balance, 30-second chair stands, and the 2-minute steps were tested online and in-person using a random order for community dwellers aged 65 years and above. Using an alpha two way mixed model, average intraclass correlation coefficients (ICC) were calculated between the two settings and one sample T-test performed on the difference between the results of each test between the two settings. Finally, Bland-Altman plots were created, and proportional biases tested via linear regressions.

Results: Besides the one-leg stance balance with eyes closed, for which the ICC was 0.47 (0.23-0.74) the average ICC's were excellent ranging from 0.87 to 0.94. No proportional biases were observed based on Bland-Altman graphs.

Conclusion: For older adults living in the community, common physical function tests can be performed online.

Keywords: Bland-Altman; Intraclass correlation coefficient; Older adults; Physical function tests; Validity; Virtual.