Evaluation on different non-pharmaceutical interventions during COVID-19 pandemic: An analysis of 139 countries

Abstract

Since the first case of coronavirus disease 2019 (COVID-19) was confirmed, it has been spreading rapidly in various countries of the world. Various non-pharmaceutical interventions (NPIs) were reported with possible impact to curb both local and global control of COVID19. A recent modeling study evaluated the impact of different NPIs on incidence of COVID-19 in the UK.1Goscé L. Phillips P.A. Spinola P. et al.Modelling SARS-COV2 spread in London: approaches to lift the lockdown.J Infect. 2020; (S0163-4453(20)30315-7)Abstract Full Text Full Text PDF PubMed Scopus (33) Google Scholar The authors found that NPIs such as universal testing, contact tracing and use of facemasks were associated with less burden of COVID-19. Nevertheless, their potential effectiveness to mitigate the COVID-19 pandemic is yet to be testified in real-life settings and in different countries. We noticed an opportunity to examine the impact of these NPIs by a recently released index. The University of Oxford has developed an Oxford COVID-19 Government Response Tracker (OxCGRT) to offer a systematic way to follow the stringency of government responses to the pandemic across countries and time.2University of Oxford. Variation in Government responses to COVID-19. Available at:https://www.bsg.ox.ac.uk/research/publications/variation-government-responses-covid-19. Accessed on 17 May ,2020Google Scholar The Stringency Index consists of school closure (C1); workplace closure (C2); public event cancelation (C3); restrictions on gathering size (C4); public transport closure (C5); staying at home requirements (C6); restrictions on internal movement (C7); restrictions on international travel (C8); and public information campaigns (H1). We examined if the stringency of these containment measures could potentially reduce the number of confirmed infections. We extracted the rate of increase in cumulative incidence for each country between 15 April to 30 April, 2020 from the COVID-19 data repository of the Johns Hopkins center for Systems Science and Engineering.3The 2019 Novel CoronavirusCOVID-19 (2019-nCoV) data repository by Johns Hopkins centre for systems science and engineering (CSSE). Available at:https://systems.jhu.edu/research/public-health/ncov/. Accessed on 17 May2020.Google Scholar We computed the average of all stringency indices for each nation on or before 31 March 2020. A 14-day window period was applied between the closing date of the stringency index and the starting date of incidence change by making reference to recent literature,4Lauer S.A. Grantz K.H. Bi Q. et al.The incubation period of coronavirus disease 2019 (COVID-19) from publicly reported confirmed cases: estimation and application.Ann Intern Med. 2020; 172: 577-582Crossref PubMed Scopus (3248) Google Scholar and was determined by a panel of epidemiologists, physicians and public health practitioners. A linear regression model was constructed to examine the association between average stringency index and increase in incidence of COVID-19 cases as the outcome variable. We controlled for the Gross Domestic Product (GDP) [5The Economist Intelligence UnitWorld Bank and central intelligence agency world factbook.2020https://www.cia.gov/library/publications/the-world-factbook/geos/we.htmlGoogle Scholar] and the population density of each country as potential confounders.6Countries by density by population2020. World population review. Available at:https://worldpopulationreview.com/countries/countries-by-density/. Accessed on 17 May 2020.Google Scholar The distribution of the Government response stringency index in various countries shows its increase over time (01 March, 2020 to 31, March 2020) [7University of Oxford. Coronavirus government response tracker. Available at:https://www.bsg.ox.ac.uk/research/research-projects/coronavirus-government-response-tracker. Accessed on 17 May2020.Google Scholar], probably due to the increase in incidence in this period. In multivariaable regression analysis of data in 139 countries (Table 1), a higher stringency index was significantly associated with lower incidence increase between 15 April to 30 April, 2020 (β coefficient −0.03, 95% C.I. −0.05 to −0.01, p = 0.014). Three indicators also showed an inverse association with incidence increase, namely “School closing” (β coefficient −0.53, 95% C.I. −1.00 to −0.06, p = 0.027), “Workplace Closing” (β coefficient −0.56, 95% C.I. −1.06 to −0.06, p = 0.028) and “Public Information campaign with public officials urging caution about COVID-19” (β coefficient −0.71, 95% C.I. −1.31 to −0.12, p = 0.028). There are no interactions or multicollinearity detected among the predictors.Table 1The association between stringency index and the incidence increase in COVID-19 (N = 139).Rate of increase in cumulative incidence of COVID-19 between 15 April to 30 April, 2020β coefficients95% CIpComposite Stringency Index−0.03−0.05−0.010.014(C1). School closing−0.53−1.00−0.060.027(C2). Workplace closing−0.56−1.06−0.060.028(H1). Public information campaign (public officials urging caution about COVID-19)−0.71−1.31−0.120.028COVID-19: Coronavirus disease 2019. Open table in a new tab COVID-19: Coronavirus disease 2019. The findings of this study showed that more stringent containment and control measures could potentially lead to better COVID-19 pandemic control. In particular, closure of schools and workplace was found to be influential in mitigation of the disease. Stopping schools and workplace attendance involves a substantial number of students and employees, and this could represent a significant containment measure that exerted material effects on the disease incidence. In addition, public information campaign urging caution against COVID-19 was reported to be effective. This highlights the importance of communication among various stakeholders in the community during a pandemic [8Markel H. Lipman H.B. Navarro J.A. et al.Nonpharmaceutical interventions implemented by US cities during the 1918–1919 influenza pandemic.JAMA. 2007; 298: 644-654Crossref PubMed Scopus (459) Google Scholar]. One limitation of the study included the absence of control for some cofounders like personal hygienic measures, testing capability [9Wong M. Teoh J. Huang J. et al.Strengthening early testing and surveillance of COVID-19 to enhance identification of asymptomatic patients.J Infect. 2020; https://doi.org/10.1016/j.jinf.2020.05.048Abstract Full Text Full Text PDF Scopus (7) Google Scholar] and the government's public health resources [10Wong M. Teoh J. Huang J. et al.The potential impact of vulnerability and coping capacity on the pandemic control of COVID-19.J Infect. 2020; https://doi.org/10.1016/j.jinf.2020.05.060Abstract Full Text Full Text PDF Scopus (21) Google Scholar]. Also, our results represent preliminary findings that should be further examined by large-scale confirmatory studies. We recommend future evaluations to explore the effectiveness of these containment strategies in relation to different global health capacities in different countries. We declared no conflict of interests.