Physical Activity Behaviour Before, During and After COVID-19 Restrictions: A Longitudinal Smartphone Tracking Study of 5395 UK Adults

Background: The COVID-19 Pandemic led to the implementation of worldwide restrictive measures to reduce social contact and viral spread. These measures have been reported to have a negative effect on physical activity (PA). Studies of PA during the pandemic have primarily used self -reported data. Only one academic study using tracked data this did not report on demographics. Objective: The study aimed to explore patterns of tracked activity before, during and immediately after Lockdown in the UK and examine differences in sociodemographic characteristics and prior levels of PA. Methods: Tracked longitudinal weekly minutes of physical activity were captured using the BetterPoints smartphone app between January and June 2020. Data was plotted by week, demographics and activity levels at baseline. Non-parametric tests of difference were used to assess mean and median weekly minutes of activity at significant points, before, during and as lockdown was eased. Changes over time by demographics (age, gender, Index of Multiple Deprivation, baseline activity levels) were examined using generalised estimating equations (GEEs). Results: There were 5395 users with mean age of 41 (SD 12), 61% were female. At baseline, 26% of users were inactive, 23% fairly active and 51% active. There was a relatively even spread across deprivation deciles (31% in the least deprived deciles and 23% in the most.). We found significant changes in PA from the week before the first case of COVID-19 was announced (baseline), to the week that social distancing restrictions were relaxed (Friedman test: χ2(2) = 2331, p < 0.001.) By the first full week of lockdown, the median change in PA 57 minutes less than baseline. This represents a 37% reduction in weekly minutes of PA. Overall, 63% of people decreased their level of activity between baseline and the first week of COVID-19 restrictions. Younger people showed more PA before lockdown but the least PA after lockdown. In contrast, the over-65s appeared to remain more active throughout and increased their activity levels as soon as lockdown was eased. Levels of physical activity levels among those classed as active at baseline showed a dramatic drop compared with those considered to be fairly active or inactive. Socioeconomic group and gender did not appear to be associated with changes in PA. Conclusions: Our tracked physical activity data suggests a significant drop in PA during the UK’s COVID-19 lockdown Significant differences by age group and prior PA levels suggests that Government response to COVID-19 needs to be sensitive to these individual differences and react accordingly. Specifically, considering the impact on younger age groups and those that were fairly active not meeting recommended PA levels prior to lockdown