Search

The objective of this analysis was to apply symbolic aggregate approximation (SAX) time-series analysis to accelerometer data for activity pattern visualization stratified by self-reported mobility difficulty. A total of 2393 (71.6  ±  7.9 years old) participants wore an accelerometer on the hip (4  +  d; 10  +  h) during the national health and nutrition examination survey (NHANES), a biannual series of health assessments of the US population. One minute epoch data was used to perform SAX, which converted accelerometry time series data into four activity levels. Intelligent icons of normalized activity transition prevalence, a visual representation of time-series data, were examined among those who self-reported mobility difficulty. Mobility difficulty questions assessed various levels difficulty performing activities such as walking a quarter mile, walking up ten steps, stooping/crouching/kneeling, and walking between rooms on the same floor. Daily activity counts were estimated across difficulty level using weighted-linear regression after adjusted for demographics, lifestyle factors, medical conditions, and accelerometer wear time. Those reporting higher mobility difficulty tended to be older, female, less educated, not married, and smokers. Additionally, those with higher mobility difficulty self-reported lower education, lower income, lower moderate-to-vigorous physical activity, and higher history of adverse medical conditions. Using SAX-derived intelligent icons, those with no difficulty showed high variations of transitions across all activity levels. With higher difficulty, the variations in transitions were lower and constricted around low activity level transitions. Among those who reported unable/don’t do mobility-related activities, there were only transitions in the lower tiered activity levels with high prevalence of prolonged low level activity. Adjusted estimates of daily activity were lower as higher difficulty reported occurred but only significant for those reporting unable/don’t do mobility-related activity (p  <  0.01). In summary, this analysis showed apparent differences in stratified activity patterns even when traditional regression analyses on volumetric accelerometer data yielded null results. [ABSTRACT FROM AUTHOR]

Actigraphy has attracted much attention for assessing physical activity in the past decade. Many algorithms have been developed to automate the analysis process, but none has targeted a general model to discover related features for detecting or predicting mobility function, or more specifically, mobility impairment and major mobility disability (MMD). Men (N  =  357) and women (N  =  778) aged 70–89 years wore a tri-axial accelerometer (Actigraph GT3X) on the right hip during free-living conditions for 8.4  ±  3.0 d. One-second epoch data were summarized into 67 features. Several machine learning techniques were used to select features from the free-living condition to predict mobility impairment, defined as 400 m walking speed  <0.80 m s−1. Selected features were also included in a model to predict the first occurrence of MMD—defined as the loss in the ability to walk 400 m. Each method yielded a similar estimate of 400 m walking speed with a root mean square error of ~0.07 m s−1 and an R-squared values ranging from 0.37–0.41. Sensitivity and specificity of identifying slow walkers was approximately 70% and 80% for all methods, respectively. The top five features, which were related to movement pace and amount (activity counts and steps), length in activity engagement (bout length), accumulation patterns of activity, and movement variability significantly improved the prediction of MMD beyond that found with common covariates (age, diseases, anthropometry, etc). This study identified a subset of actigraphy features collected in free-living conditions that are moderately accurate in identifying persons with clinically-assessed mobility impaired and significantly improve the prediction of MMD. These findings suggest that the combination of features as opposed to a specific feature is important to consider when choosing features and/or combinations of features for prediction of mobility phenotypes in older adults. [ABSTRACT FROM AUTHOR]