Concurrent and discriminant validity and reliability of an Android App to assess time, velocity and power during sit-to-stand test in community-dwelling older adults.

ABSTRACT

INTRODUCTION: Nowadays, smartphones are equipped with the most sophisticated hardware which provides the opportunity to develop specific smartphone apps to analyze kinetic and kinematic parameters during sit-to-stand test in a clinical setting. The aims were to ascertain whether a new Android video-analysis based-App is comparable to the previously validated Apple-App for measuring time, velocity and power during sit-to-stand test, to determine its reliability and discriminant validity. METHODS: One-hundred sixty-one older adults (61-86 years) were recruited from an elderly social center. Sit-to-stand variables were simultaneously recorded through the Android and Apple-App. Their validity and inter-rater, intra-rater, and test-retest reliability was tested using an intraclass correlation coefficient (ICC2-1). Low gait speed (< 1.0 m/s), low physical performance (Short Physical Performance Battery < 10 points), and sarcopenia (EWGSOP2 guideline) were used to determine discriminant validity which was reported as the area under the curves (AUC) and their effect sizes (Hedges' g) for independent sample t-test. RESULTS: Excellent reproducibility (ICC2-1 > 0.85) and strong agreement (ICC2-1 > 0.90) between operating systems for sit-to-stand variables derived from the App was found. Older adults classified as sarcopenic (11.2%), low physical performance (15.5%), or reduced gait speed (14.3%) showed worse sit-to-stand time, velocity and power with large effect sizes (Hedges' g > 0.8) compared to their respective counterpart. These variables showed the acceptable-to-excellent ability to identify low gait speed, low physical performance, and sarcopenic older adults (AUC-range 0.73-0.82). CONCLUSION: The new Sit-to-Stand App running on the Android operating system is comparable to the previously validated Apple App. Excellent reproducibility and acceptable-to-excellent discriminant validity were found.