JMIR Publications

ABSTRACT

Consumer-grade wearable devices are increasingly utilized in remote clinical trials to track physical activity, yet their underlying algorithms are typically calibrated for neurotypical gait patterns. In their recent protocol, Nunez-Gaunaurd et al. integrate Fitbit Inspire 3 step-monitoring data to evaluate the construct validity of patient-reported physical function scores among adults with Autism Spectrum Disorder (ASD). While this telehealth approach significantly enhances study accessibility, we highlight a critical measurement vulnerability: repetitive motor behaviours characteristic of ASD such as hand-flapping, body rocking, and pacing that generate rhythmic limb accelerations that standard step-counting algorithms may misclassify as ambulatory steps. Because daily step metrics serve as a criterion indicator for validating clinical outcomes, unaddressed movement artifacts risk biasing validity findings. To ensure methodological rigor, we recommend incorporating video-coded validation subsamples to isolate and verify true ambulatory steps against stereotypical movements in neurodivergent cohorts.b