The reliability and validity of a non-wearable indoor positioning system to assess mobility in older adults: A cross-sectional study

Abstract

Background: Chirp is a privacy-preserving radar sensor developed to continuously monitor older adults' safety and mobility without the need for cameras or wearable devices. Our study purpose was to evaluate the inter-sensor reliability, intrasession test-retest reliability, and concurrent validity of Chirp in a clinical setting. Methods: We recruited 35 community-dwelling older adults (mean age 75.5 (standard deviation: 6.6) years, 86% female). All participants lived alone in an urban city in southwestern Ontario and had access to a smart device with wireless internet. Data were collected with a 4-meter ProtoKinetics ZenoTM Walkway (pressure sensors) with the Chirp sensor (radar positioning) at the end of the walkway. Participants walked during normal and adaptive locomotion experimental conditions (walking-while-talking, obstacle, narrow walking, fast walking). Each of the experimental conditions was conducted twice in a randomized order, with fast walking trials performed last. For intra-session reliability testing, we conducted two blocks of walks within a participant session separated by approximately 30 minutes. Intraclass Correlation Coefficient(A,1) (ICC(A,1)) was used to assess the reliability and validity. Linear regression, adjusted for gender, was used to investigate the association between Chirp and cognition and health-related quality of life scores. Results: The Chirp inter-sensor reliability ICC(A,1)=0.999[95% Confidence Interval [CI]: 0.997 to 0.999] and intrasession test-retest reliability [ICC(A,1) =0.921, 95% CI: 0.725 to 0.969] were excellent across all experimental conditions. Chirp concurrent validity compared to the ProtoKinetics ZenoTM Walkway was excellent across experimental conditions [ICC(A,1)= 0.993, 95% CI: 0.985 to 0.997]. We found a weak association between Chirp and cognition scores using the Montreal Cognitive Assessment across experimental conditions (estimated β-value= 7.79, 95% CI: 2.79 to 12.80) and no association between the Chirp and health-related quality of life using the 12-item Short Form Survey across experimental conditions (estimated β-value=6.12, 95% CI: -7.12 to 19.36). Conclusion: Our results demonstrate that Chirp is a reliable and valid measure to assess gait parameters in clinics among older adults.