With aging, daily physical activity (PA) becomes less frequent and more fragmented. Accumulation patterns of daily PA—including transitions from active-to-sedentary behaviors—may provide important insights into functional status in older, less active populations. Participants of the Baltimore Longitudinal Study of Aging (n = 680, 50% male, aged 27–94 years) completed a clinical assessment and wore an Actiheart accelerometer. Transitions between active and sedentary states were modeled as a probability (Active-to-Sedentary Transition Probability [ASTP]) defined as the reciprocal of the average PA bout duration. Cross-sectional associations between ASTP and gait speed (m/s), fatigability (rating-of-perceived-exertion [RPE]), 400 m time (seconds), and expanded short physical performance battery score were modeled using linear and logistic regression, adjusted for chronic conditions. Further analyses explored the utility of ASTP over-and-above total daily PA. In continuous models, each 0.10-unit higher ASTP was associated slower gait (β = −0.06 m/s, SE = 0.01), higher fatigability (β = 0.60 RPE, SE = 0.12), slower 400 m time (β = 16.31 s, SE = 2.70), and lower functioning (β = −0.13 expanded short physical performance battery score, SE = 0.03; p < .001). In categorical analyses, those in the highest tertile of ASTP were >2 times more likely to have high fatigability (rating of perceived exertion ≥10), slow 400 m time (>300 seconds) and reduced functional performance (expanded short physical performance battery score < 3.07) than those in the lowest tertile (p < .01). Further analyses demonstrated ASTP provided additional insight into functional outcomes beyond total daily PA. Fragmented daily PA—as measured by ASTP—is strongly linked with measures of health and functional status and may identify those at risk of high fatigability and reduced functional performance over and above traditional PA metrics.

Physical activity benefits health and quality of life, particularly for adults aged 65 years and older.¹ With aging, functional capability declines, physical activity decreases,² and mortality risk increases. Previous research^3,4 shows that volume and/or intensity measures of physical activity are associated with mortality risk, but technological advances (eg, accelerometers) present the possibility to evaluate and test whether detailed patterns of activity may be more informative to health outcomes than traditional measures of total activity, thus providing an earlier marker of future health and longevity.⁵ This is particularly important with regard to older adults, who are one of the most sedentary and rapidly growing segments of the US population.⁶

Digital health technologies (DHTs) enable us to measure human physiology and behavior remotely, objectively and continuously. With the accelerated adoption of DHTs in clinical trials, there is an unmet need to identify statistical approaches to address missing data to ensure that the derived endpoints are valid, accurate, and reliable. It is not obvious how commonly used statistical methods to handle missing data in clinical trials can be directly applied to the complex data collected by DHTs. Meanwhile, current approaches used to address missing data from DHTs are of limited sophistication and focus on the exclusion of data where the quantity of missing data exceeds a given threshold. High-frequency time series data collected by DHTs are often summarized to derive epoch-level data, which are then processed to compute daily summary measures. In this article, we discuss characteristics of missing data collected by DHT, review emerging statistical approaches for addressing missingness in epoch-level data including within-patient imputations across common time periods, functional data analysis, and deep learning methods, as well as imputation approaches and robust modeling appropriate for handling missing data in daily summary measures. We discuss strategies for minimizing missing data by optimizing DHT deployment and by including the patients' perspectives in the study design. We believe that these approaches provide more insight into preventing missing data when deriving digital endpoints. We hope this article can serve as a starting point for further discussion among clinical trial stakeholders.