Poor-quality studies: several public health measures are effective in controlling the spread of COVID-19

Référence

Talic S, Shah S, Wild H, et al. Effectiveness of public health measures in reducing the incidence of covid-19, SARS-CoV-2 transmission, and covid-19 mortality: systematic review and meta-analysis. BMJ 2021;375:e068302.

Sommaire

These authors searched multiple databases, including the World Health Organization’s preprint database, to identify studies that evaluated the effectiveness of individual public health measures (other than vaccinations) to prevent COVID-19. They did not describe “snowballing” to identify any studies that escaped their initial search and did not include studies that evaluated multiple interventions. The authors independently evaluated studies for inclusion and assessed the methodologic quality of the individual studies. Ultimately, they identified 72 studies, 35 of which reported data on individual interventions (only 1 was a randomized trial). Overall, 3 of the 35 studies were of good methodologic quality, with major confounding as the most common source of serious or critical bias. In 3 low-quality studies (n = 10,345 participants), handwashing was associated with a nonsignificant decrease in COVID-19 infection (relative risk [RR] 0.47; 95% CI 0.19 - 1.12). Six low-quality studies (n = 389,228 participants) evaluated mask wearing. Persons who wore masks were less likely to contract COVID-19 (RR 0.47; 0.29 - 0.75), although these data varied greatly among the studies. The authors also report, based on several large natural experiments (in 200 countries and 15 states), that mask wearing was associated with lower rates of COVID-19 transmission and COVID-19 mortality, although they could not pool the available data. They identified a single low-quality study that reported that disinfecting surfaces was associated with a lower risk of secondary COVID-19 transmission (RR 0.23; 0.07 - 0.84). Five low-quality studies (n = 108,933 participants) reported that physical distancing was associated with a lower incidence of COVID-19 (RR 0.75; 0.59 - 0.95), but these data were also quite variable across the studies. Three low-quality studies each found that physical distancing was associated with less COVID-19 transmission. Two lower-quality studies evaluated school closures: one reported a lower rate of COVID-19 incidence and the other found no effect. Additionally, among 3 natural experiments of school closures, 2 studies found a significant decrease in COVID-19 transmission. The other reported a small increase in COVID-19 among the parents and found that teachers in lower secondary schools were twice as likely to become infected as teachers in upper secondary schools. Because of methodologic heterogeneity, the authors did not pool data for several other interventions. All the studies assessing stay-at-home or isolation measures, quarantine measures, business closures, and lockdown were associated with decreased COVID-19 incidence, transmission, or mortality. Finally, the results from 2 low-quality studies of border closures were inconsistent: increased COVID-19 incidence in one and decreased incidence in the other. Overall, these results suggest that better data are needed to determine which measures, other than vaccination, are truly effective.