Differential Effects of the COVID-19 Pandemic on Physical Activity Involvements and Exercise Habits in People With and Without Chronic Diseases: A Systematic Review and Meta-analysis.

All 36 studies are observational. They describe what happened to activity levels during the pandemic; they cannot show that lockdowns caused the changes. Most of the evidence comes from questionnaires people filled in themselves, and only 15 of the 30 studies using them reported whether that questionnaire was reliable or valid. People misremember how active they were, and the authors note that questionnaire estimates may not match what devices record. Study quality was generally weak: 1 of 36 studies was rated at low risk of bias and 4 at high risk. Common problems were no justification for the sample size, no information about people who did not respond, and, across all 13 retrospective studies, no account of dropouts or of other factors that could explain the results. Only one included study came from a country without lockdown, so the comparison between living with and without lockdown rests on that single study. Some outcomes were measured in only one study, so they could not be pooled, and the range of chronic diseases studied prevented pooling results for each disease separately. The authors report that the variables they pooled showed substantial heterogeneity (I² greater than 60%), which they attribute to differences in sampling methods, how questionnaires were delivered, how long they took to complete and socioeconomic status. The step-count pooling was at I²=100%, meaning those four studies had essentially nothing in common. Twenty-three of the 36 studies were cross-sectional, measuring activity at one point in time. Activity levels shifted over the course of lockdowns, so a single measurement may not represent the whole period.

Do not tell patients the pandemic made everyone less active. Somewhere between a fifth and 62% of surveyed people reported doing less, but as many as 49% reported doing more, and most studies relied on questionnaires that were never checked for reliability. The review's closing suggestion that tele-exercise may help maintain activity is not something any included study tested.