Impact of Dutch COVID-19 restrictive policy measures on physical activity behavior and identification of correlates of physical activity changes: a cohort study.

ABSTRACT

BACKGROUND: Identification of characteristics of individuals that are related to decreases in physical activity (PA) levels during lockdown is needed to develop targeted-interventions. This study aims to evaluate changes in domain-specific (i.e. leisure time, transportation, occupational, and household) and total PA due to the Dutch COVID-19 lockdown, which started on March 15 2020. Furthermore, we aim to identify demographic, health-related, and psychological correlates of these changes. METHODS: Individuals who participated in the Nijmegen Exercise Study during 2017-2019 were invited to this study, which was conducted between April 16 and May 12 2020. Participant characteristics (i.e. age, sex, body mass index (BMI), marital status, education, household composition, and occupation status), living environment (i.e. housing type and degree of urbanization), psychological characteristics (i.e. resilience, outcome expectations, vitality, and mental health), and medical history were collected via an online questionnaire. Short Questionnaire to Assess Health-enhancing physical activity was used to assess PA behavior before and during lockdown. Wilcoxon signed-rank test was used to compare PA levels, in metabolic equivalent of task (MET)-minutes per week (min/wk), before and during lockdown. Multivariable linear regression analyses were performed to examine correlates of PA changes. RESULTS: 4033 participants (57% male; 59 ± 13 years) were included. PA decreased significantly during lockdown with mean ± SD changes of 393 ± 2735 MET-min/wk for total, 133 ± 785 MET-min/wk for transportation, 137 ± 1469 MET-min/wk for occupation, and 136 ± 1942 MET-min/wk for leisure time PA. Household PA did not change significantly. Unemployment, COVID-19-related occupational changes, higher BMI, and living in an apartment or semi-detached/terraced house were significantly related to larger decreases in total and domain-specific PA. Higher vitality was related to smaller decreases in total and domain-specific PA. Higher age was significantly associated with a larger decrease in leisure time PA. Lower education was associated with smaller decreases in transportation and occupational PA compared to higher education. CONCLUSION: PA levels significantly reduced during lockdown compared to before lockdown. Declines were observed during transportation and occupation, but were not compensated by an increase in leisure time PA. We identified subgroups that were more susceptible to reductions in domain-specific or total PA levels and should therefore be encouraged to increase their PA levels during lockdown.