key: cord-0721710-drrxuzdt
authors: Valcarcel, B.; Avilez, J. L.; Torres-Roman, J. S.; Poterico, J. A.; Bazalar-Palacios, J.; La Vecchia, C.
title: The effect of public health policies in the transmission of COVID-19 for South American countries
date: 2020-08-12
journal: nan
DOI: 10.1101/2020.08.09.20149286
sha: 4ff017ecf2cbbf11c992cc4a9a29933a1a9eab2c
doc_id: 721710
cord_uid: drrxuzdt

Objectives: The analysis of transmission dynamics is crucial to determine whether mitigation or suppression measures reduce the spread of Coronavirus disease 2019 (COVID-19). This study sought to estimate the basic (R0) and time-dependent (Rt) reproduction number of COVID-19 and contrast the public health measures for ten South American countries. Methods: Data was obtained from the European Centre for Disease Prevention and Control. Country-specific R0 estimates during the first two weeks of the outbreak and Rt estimates after 90 days were estimated. Results: Countries used a combination of isolation, social distancing, quarantine, and community-wide containment measures to contain the spread of COVID-19 at different points in time. R0 ranged from 1.52 (95% confidence interval: 1.13-1.99) in Venezuela, to 3.83 (3.04-4.75) in Chile, whereas Rt, after 90 days, ranged from 0.71 (95% credible interval: 0.39-1.05) in Uruguay to 1.20 (1.19-1.20) in Brazil. Different R0 and Rt values may be related to the testing capacity of each country. Conclusion: R0 in the early phase of the outbreak varied across the South American countries. The adopted public health measures in the initial period of the pandemic appear to have reduced Rt over time in each country.

Coronavirus disease 2019 (COVID-19) is an emerging respiratory infectious disease caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) (1). COVID-19 was first detected in December 2019 in Wuhan, China (2) and have caused serious public health concern in over 213 countries (3) . Several public health measures-social isolation, temporary closure of borders, temporary closure of academic institutions and public places, and quarantine-have been used to reduce the impact of the COVID-19 outbreak. South American (SA) countries have fragile public health systems, thus, COVID-19 has a detrimental impact on their population health (4).

Since the first reported COVID-19 patient on March 26 in Brazil (5) , few SA modeling studies are available in the COVID-19 literature (6) . Given the rapid spread of SARS-CoV-2 in this region, understanding transmissibility is key to guide the implementation of priority prevention and control measures. For this purpose, the basic reproduction number (R0) represents a parameter to determine the transmission of a disease. R0 estimates the number of secondary cases arising from exposure to an infected person in the absence of epidemic containment measures and depends on several biological and sociocultural factors (7) . Another metric, the timedependent reproduction number (Rt) is useful to monitor the transmissibility of SARS-CoV-2 over time and assessing whether current control efforts are adequate. Rt estimates the expected number of secondary infections from an infected individual at time t (8, 9) . A variable that appears in models used to estimate R0 and Rt is the serial interval, which measures the time elapsed between symptomatic cases in a chain of transmission. Determining the probability . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted August 12, 2020. . https://doi.org/10.1101/2020.08.09.20149286 doi: medRxiv preprint distribution of the serial interval of SARS-CoV-2 and then estimating R0 and Rt is crucial in assessing the rate at which the COVID-19 epidemic expands. Knowing the person-to-person transmission rate identifies whether mitigation and suppression measures are effective and when to adopt more or less stringent measures (10) (11) (12) .

Given the importance of assessing the public health interventions to monitor their effectiveness, we estimated both reproduction numbers (R0 and Rt) to identify the impact of the initial public health interventions in South America.

We used the data of the European Centre for Disease Prevention and Control (13) . The database is publicly available and contains the worldwide geographic distribution of COVID-19. The database is updated daily and provide new cases and deaths by country and date of notification.

We extracted the data from ten South America countries: Argentina, Bolivia, Brazil, Chile, Colombia, Ecuador, Paraguay, Peru, Uruguay, and Venezuela. Containment and mitigation decrees issued by each country against COVID-19 were found in government's official webpages. (Appendix 1).

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted August 12, 2020. . https://doi.org/10.1101/2020.08.09.20149286 doi: medRxiv preprint

The following binary variables were recorded for each country for qualitative analysis: Isolation, separation of confirmed cases with COVID-19 in a healthcare facility or their home; Quarantine, social restriction and home containment of persons with suspected or known contact with a patient with COVID-19, or individuals with a travel history to Europe or Asia; Social distancing, group of measures related to the prevention of mass gathering, closure of academic institutions, and cancellation of social and public events; Community-wide containment, mandatory isolation of every citizen of the country in their home, with only permission to acquire life supplies (i.e., food or water) in restricted hours of the day (14) .

Basic reproductive number (R0) and time-dependent reproductive number (Rt) R0 is used to determine the early-outbreak transmission dynamics of a specific infectious pathogen. This parameter estimates the expected number of newly infected people that arise from effective contact with an ill person (15) . We estimated R0 using data for the two weeks after the first laboratory-confirmed case of SARS-CoV-2 in each country. The time frame was selected based on the maximum period to symptoms onset and because the period of exponential growth rate must be used to compute R0 to avoid its underestimation. We specified the prior serial interval distribution using a Gamma distribution with a mean serial time of 3.96 days and a standard deviation of 4.75 days, based on a contact tracing study conducted by the Center for Disease Control and Prevention (CDC) (16) . R0 was estimated with the maximum likelihood method described by White and Pagano (17) .

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted August 12, 2020. . https://doi.org/10.1101/2020.08.09.20149286 doi: medRxiv preprint

The Rt accounts for the expected number of secondary cases that a primary case would produce at time t. We estimated Rt from the time series of daily incidence of SARS-CoV-2 in each country, using the novel methods described by Thompson et al. and specifying the same Gamma distribution specification as we did for R0 (18). We chose a 5-day moving window to elucidate the time trend exhibited by Rt. A previous study in China used a 10-day moving window to report Rt (19) ; however, they specified a mean serial time of 7.5 days, a misspecification of the mean by a factor of two with respect to the CDC contact tracing study. We, thus, chose a finer window to account for the faster spreading dynamics we specify in our models and estimate the number after 90 days for each country.

The statistical analysis was conducted in the software R version 3.6.2. First, we describe the public health measures taken by the SA countries and calculated the cumulative number of cases according to the two periods. Thereafter, the package "R0" was used to compute the basic reproductive number. We employed a Gamma distributed serial interval with mean serial interval (standard deviation) of 3.96 (4.75) days. Estimation of R0 was carried out via the maximum likelihood method, and 95% confidence intervals (95% CI) were computed for each country's R0 value. We used the package "EpiEstim" to compute the time-varying reproductive number; the same serial interval distribution was specified. Country-wise time series for Rt, with 95% credible intervals (95% CrI), were plotted. The code is freely available at https://github.com/jlavileze/covid_sa. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 12, 2020. . https://doi.org/10.1101/2020.08.09.20149286 doi: medRxiv preprint Table 1 gives the total number of cases registered to April 14, and public health measures adopted by SA governments against COVID-19. All the countries adopted isolation, quarantine, and social distancing measures. However, the period between the first laboratory-confirmed case and the implementation of public health measures was different in each country. Seven countries, Argentina, Bolivia, Colombia, Paraguay, Peru, Uruguay, and Venezuela issued their mitigations in ≤7 days. Uruguay enacted the public health measures on the same day of the first case report, while Brazil issued them after 19 days.

Likewise, the time to implement a community-wide containment varied among SA countries.

Two countries, Bolivia and Venezuela, issued this mitigation in ≤7 days after the first COVID-19 case; Paraguay and Peru between 7-14 days; and Argentina, Chile, Colombia, and Ecuador >14 days (Table 1) . Moreover, Colombia (12 days) and Argentina (10 days) had the longest period between the first enacted measure and the community-wide containment. Chile made a stepwise selective community containment, starting with one province at a time. Brazil and Uruguay are the only countries that have not adopted community containment measures as of April 14, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 12, 2020. . https://doi.org/10.1101/2020.08.09.20149286 doi: medRxiv preprint countries with mitigation measures between 5-7 days, the R0 ranged from 1.60 to 2.95, while for the countries with ≤3 days, it ranged from 1.52 to 1.74. All SA countries decreased their Rt over time (Figure 1) . However, all countries experienced peaks throughout the study period. All nations had Rt at the last assessment ranging from 0.99 to 1.13. Although Bolivia and Uruguay had point estimates for Rt below 1, their 95% credible intervals are compatible with a still growing pandemic (Figure 1 and Table 2 ). is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 12, 2020. Italy, France, Germany, and Spain ranged from 3.10 to 6.56 in an overall 20-day period.

The reproduction number -R0 or Rt -is a measure that depends on the socio-cultural dynamics of a specific population (15) . Human behaviors play a critical role in the transmission of SARS-CoV-2, as person-to-person contact exposes a susceptible person through respiratory droplets from an infected individual (23) . Therefore, mitigation measures, such as social distancing or case isolation, are necessary to stop the spread of SARS-CoV-2. Ideal policies should be country-specific, and the overall objective is to reduce R0 (mitigation) or to reach an R0 <1 (suppression). Regardless of the aim, a combination of measures are the best strategy (12).

Three countries, Brazil, Chile, and Uruguay opted for a mitigation strategy. Ferguson et al.

suggest that the best mitigation strategy is the combination of social distancing of high-risk groups (elders and patients at risk of severe disease), case isolation, and quarantine (12).

Uruguay implemented these three measures. Brazil and Chile added a stepwise selective population containment, which allows intermittent circulation of SARS-CoV-2, congruent with a mitigation purpose. The remaining SA countries opted for suppression actions. For example, Peru issued a decree to increase the period of the community-wide containment intervention and . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 12, 2020. . https://doi.org/10.1101/2020.08.09.20149286 doi: medRxiv preprint to fine citizens if they left their home after a specified hour of the day (24) . As a result, all nations managed a reduction of Rt.

Here we highlight several explanations for the change between R0 and Rt. First, the implementation of public health measures by some countries within the first week of the outbreak could have lowered R0. These laws aimed to reduce human contact, which, in turn, reduced the spread of the virus in the community. In contrast, the three countries with the highest R0 in the early phase issued their first mitigation laws ≥14 days, which gave SARS-CoV-2 a higher chance of transmissibility among the population. Second, the reproduction number is susceptible to the ability of each country to detect COVID-19 cases (25) . A lack of testing impairs the identification of cases within a community and provides limited data to estimate the reproduction number, generating unclear information to analyze the effect of mitigation or suppression interventions in a community. In addition, after notification of the first case, a number of cases tends to emerge in the subsequent days, thus inflating R0. Despite the international donations of test kits to identify and isolate cases (26, 27) , the fragile and fragmented structure of healthcare systems in South America deters a prompt diagnostic of true cases.

Although we used national reports of COVID-19 cases, under reporting is likely, and difficult to quantify. True case identification correlates with the testing capacity of both symptomatic and asymptomatic patients, which is difficult in any country and likely be impaired in SA countries.

Second, we did not calculate the case-fatality ratio. During the current course of the pandemic in . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 12, 2020. . https://doi.org/10.1101/2020.08.09.20149286 doi: medRxiv preprint the SA countries, the estimation of this metric is biased for underreporting of cases and time lag between the notification of cases and deaths (28, 29). Third, there are issues arising from the model specification for serial interval. For instance, we assumed the same serial interval distribution for all countries at all points in time, even though these should be space and time dependent, as serial interval distributions vary throughout an epidemic (25) . Also, the CDC serial interval model only considers positive serial times, and hence censors all serial interval observations in which a secondary cases manifests symptom before a primary case. Per their findings, about 12.6% of secondary cases exhibit clinical symptoms before a primary case (i.e. an infected presenting symptoms before an infector); given how sensitive estimates of R0 are to the serial interval distribution, our results should be updated as better model specifications of this variable are elucidated.

Our findings reflected a positive impact of the mitigation and suppression measures in SA nations to reduce the spread of SARS-CoV-2. Despite the fragile health system of most of these countries, the use of a combination of isolation, quarantine, social distancing, and communitywide containment has avoided a skyrocketing number of cases. The difference of R0 in the early phase of the outbreak is probably due to a combination between a shortage of testing and the public health measures adopted by each country. However, the peaks of Rt during the study period suggest that Latin America is still far from containing the spread of COVID.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 12, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 12, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 12, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 12, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted August 12, 2020. . https://doi.org/10.1101/2020.08.09.20149286 doi: medRxiv preprint a Chile and Brazil developed a selective community-wide containment with different dates for specific provinces b Venezuela started the mitigation measures (isolation, quarantine, social distancing, and community-wide containment) at the same date . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted August 12, 2020. . https://doi.org/10.1101/2020.08.09.20149286 doi: medRxiv preprint

The effect of control strategies to reduce social mixing on outcomes of the COVID-19 epidemic in Wuhan, China: a modelling study

A novel coronavirus outbreak of global health concern

Incidence and mortality of lung cancer: global trends and association with socioeconomic status

COVID-19 in Latin America: The implications of the first confirmed case in Brazil

The reproductive number R0 of COVID-19 in Peru: An opportunity for effective changes

Complexity of the basic reproduction number (R0)

Different epidemic curves for severe acute respiratory syndrome reveal similar impacts of control measures

The Effective Reproduction Number as a Prelude to Statistical Estimation of Time-Dependent Epidemic Trends

Evaluation of containment and mitigation strategies for an influenza A pandemic in China

The Efficacy of Contact Tracing for the college/medicine/sph/ide/gida-fellowships/Imperial-College-COVID19-NPI-modelling

European Centre for Disease Prevention and Control. COVID-19 cases worldwide

Isolation, quarantine, social distancing and community containment: pivotal role for old-style public health measures in the novel coronavirus (2019-nCoV) outbreak

Complexity of the Basic Reproduction Number (R(0)). Emerging infectious diseases

Serial Interval of COVID-19 among Publicly Reported Confirmed Cases

A likelihood-based method for real-time estimation of the serial interval and reproductive number of an epidemic

Preliminary estimation of the basic reproduction number of novel coronavirus (2019-nCoV) in China, from 2019 to 2020: A data-driven analysis in the early phase of the outbreak

Monitoring Transmissibility and Mortality of COVID-19 in

IJID : official publication of the International Society for Infectious Diseases

COVID-19 infection: Origin, transmission, and characteristics of human coronaviruses

Legislative Decree No. 1458 to sanction non-compliance with the provisions issued during the national health emergency and other regulations issued to protect the life and health of the population due to the spread of COVID-19

Unraveling R0: Considerations for Public Health Applications

Latin America seeks Chinese medical aid Table 2. The Basic and Time-Varying Reproductive Numbers in South American Countries

R0, basic reproduction number

time-dependent reproduction number; 95% CI, 95% confidence interval; 95% CrI