Quantifying gait outside the laboratory with wearable sensors: understanding and leveraging the gap between lab-based assessments and 24/7 monitoring

Wearables enable continuous monitoring of gait in real-world environments over an extended period of time (e.g., week). Previous findings show that in-lab and real-world measures differ; however, it is not fully clear why. This presentation will describe these gaps and illustrate how metrics based on wearables may outperform the conventional gait assessment. We illustrate this in studies among people with Parkinson’s disease (PD), people with mild cognitive impairment, people with multiple sclerosis, and elderly fallers. For example, among patients with PD who have not yet fallen, we find that conventional, in-lab measures of gait, balance and PD severity do not successfully predict future falls, whereas measures based on wearables do. We also quantify the gaps between in-lab usual-walking, in-lab dual-task walking, and daily-living measures of gait. We discuss the reasons for these gaps and why 24/7 monitoring apparently offers a complementary approach to the more conventional, one-time assessment of gait.