Individual-level exposures and risk of COVID-19: an umbrella review

As of April 2023, over 650 million people have been confirmed to have contracted COVID-19. The risk of transmission of COVID-19 might be influenced by various factors, including immunity from vaccination or previous infection, the use of personal protective equipment (PPE), indoors vs. outdoors environments, the degree of ventilation in indoor environments, and distance from the source of the infection. This review aims to assess the credibility of research conducted on individual risk factor epidemiology and prevention of COVID-19. A search was conducted on PubMed to identify systematic reviews and meta-analyses of studies examining the association between individual-level exposures and COVID-19. The study estimated the summary effect size and its 95% CI using both fixed-effect and random-effects models, and the 95% prediction interval was used to evaluate uncertainty. The study also examined heterogeneity among the studies and the presence of biases. We identified 141 meta-analyses examining 84 types of exposures or treatments. These were categorized into eight macro-groups: demographic factors, behavioural factors, dietary nutrient intake, comorbidities, drugs, PPE, contact types, and vaccination. Interventional studies comprised 5.7% of the total, contact tracing studies were 9.8%, and the remaining studies (84.5%) were other types of observational comparative studies. In these preliminary findings, the heterogeneity index (I2) ranged from large to very large, and there was evidence of small study effect and excess significance bias in many meta-analyses. Preliminary findings suggest that, apart from vaccines, for which abundant randomized evidence exists, the evidence base for COVID-19 prevention consists largely of studies of low or uncertain epidemiological credibility. KEY MESSAGES: • Overall epidemiological credibility of COVID-19 association studies might be low. • Rigorous methodological practices are recommended when conducting epidemiological studies.