Objective evaluation of physical activity pattern using smart devices

Physical activity session frequency and distribution over time may play a significant role on survival after major cardiovascular events. However, the existing amount-based metrics do not account for these properties, thus the physical activity pattern is not fully evaluated. The aim of this work is to introduce a metric which accounts for the difference between the actual and uniform distribution of physical activity, thus its value depends on physical activity aggregation over time. The practical application is demonstrated on a step data from 40 participants, half of them diagnosed with chronic cardiovascular disease (CVD). The metric is capable of discriminating among different daily patterns, including going to and from work, walking in a park and being active the entire day. Moreover, the results demonstrate the tendency of CVD patients being associated with higher aggregation values, suggesting that CVD patients spend more time in a sedentary behaviour compared to healthy participants. By combining the aggregation with the intensity metric, such common weekly patterns as inactivity, regular activity and “weekend warrior” can be captured. The metric is expected to have clinical relevance since it may provide additional information on the relationship between physical activity pattern and health outcomes.