CO182 Developing Methods to Validate Outcome Measures Captured Via Wearable Activity Trackers

Abstract

While the novel use of wearable activity trackers (producing objective physical activity evidence) as study outcomes is promising, their psychometric properties must be established. These outcomes may not be normally-distributed, and varying tracker wear time may affect the accuracy of raw outcomes, particularly if wear time varies systematically. Here we develop methods for classical test theory validation on endpoints measured via wearable activity trackers, focusing on construct validity and test-retest reliability (TRTR).