The first 100 days of SARS-CoV-2 control in Vietnam

Background: One hundred days after SARS-CoV-2 was first reported in Vietnam on January 23rd, 270 cases have been confirmed, with no deaths. We describe the control measures used and their relationship with imported and domestically-acquired case numbers. Methods: Data on the first 270 SARS-CoV-2 infected cases and the timing and nature of control measures were captured by Vietnam's National Steering Committee for COVID-19 response. Apple and Google mobility data provided population movement proxies. Serial intervals were calculated from 33 infector-infectee pairs and used to estimate the proportion of pre-symptomatic transmission events and time-varying reproduction numbers. Results: After the first confirmed case on January 23rd, the Vietnamese Government initiated mass communications measures, contact tracing, mandatory 14-day quarantine, school and university closures, and progressive flight restrictions. A national lockdown was implemented between April 1st and 22nd. Around 200,000 people were quarantined and 266,122 RT-PCR tests conducted. Population mobility decreased progressively before lockdown. 60% (163/270) of cases were imported; 43% (89/208) of resolved infections were asymptomatic. 21 developed severe disease, with no deaths. The serial interval was 3.24 days, and 27.5% (95% confidence interval, 15.7%-40.0%) of transmissions occurred pre-symptomatically. Limited transmission amounted to a maximum reproduction number of 1.15 (95% confidence interval, 0.37-2.36). No community transmission has been detected since April 15th. Conclusions: Vietnam has controlled SARS-CoV-2 spread through the early introduction of communication, contact-tracing, quarantine, and international travel restrictions. The value of these interventions is supported by the high proportion of asymptomatic cases and imported cases, and evidence for substantial pre-symptomatic transmission.

Vietnam has controlled SARS-CoV-2 spread through the early introduction of communication, contacttracing, quarantine, and international travel restrictions. The value of these interventions is supported by the high proportion of asymptomatic cases and imported cases, and evidence for substantial presymptomatic transmission.
. 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 (which was not certified by peer review)

Introduction
The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) emerged in Wuhan city, Hubei Province, China, in late 2019 1 . On January 30 th , the WHO declared the outbreak a 'Public Health Emergency of International Concern', and on March 11 th a global pandemic. By May 1 st 2020, the virus had infected more than 3 million people and killed over 200,000. The numbers of cases and deaths continue to rise in many countries worldwide and the virus now represents the greatest acute infectious threat to humankind since the influenza pandemic of 1918.
SARS-CoV-2 is antigenically different from known human and zoonotic coronaviruses and there is no known pre-existing population immunity 2 . It is highly transmissible through respiratory secretions expelled from an infected person, with a basic reproduction number (R0) estimated between 2 and 3 in the absence of control measures [3][4][5][6] . Many infections are asymptomatic 7 , while others lead to symptoms of coronavirus disease (COVID-19) of varying severity 5 . Analyses of serial intervals suggest that contagiousness starts before the onset of symptoms 8 . The exponential rise in infections has threatened to overwhelm even the world's best developed health systems and cause major loss of life. Methods to control the virus and reduce the impact of COVID-19 have become a global priority.
The preparedness, timing, and nature of the response to SARS-CoV-2 have varied substantially between countries. Many affected countries have resorted to extreme social distancing measures through so-called 'lockdowns', where populations isolate themselves within their homes, reducing all but essential contact with others. As first observed in Hubei Province in China, and subsequently in other countries, these measures slow transmission and reduce disease incidence, but at significant social and economic cost [9][10][11] . However, 'lockdowns' represent a combination of potentially independent interventions (for example, closing schools and universities, suspending public transport, banning public gatherings, closing non-essential businesses), the effects of which in isolation are uncertain. Determining their relative contributions to SARS-CoV-2 control is critical to understanding how they might be safely and incrementally lifted, or partially reinstated. Such information may be acquired from studying the measures employed by countries that have so far controlled the virus.
. 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 May 16, 2020. . https://doi.org/10.1101/2020.05. 12.20099242 doi: medRxiv preprint Vietnam is the 15 th most populous country on earth, with 97.3 million people. It was one of the first countries affected by SARS-CoV-2, recording its first case on January 23 rd 2020. Yet, as of May 1 st , 270 cases have been confirmed, with no deaths 12 . In this report we describe the first 100 days of SARS-CoV-2 control in Vietnam, including the timing and types of interventions and their impacts on imported and domestically-transmitted case numbers.

Methods
Epidemiological and policy data came from Vietnam's National Steering Committee for COVID-19 response. Data from 270 SARS-CoV-2-confirmed cases to May 1 st 2020 described their age, gender, nationality, dates of symptom onset (if any), entry to the country and quarantine (if any), hospital admission and discharge, and the results of all RT-PCR tests. Imported cases were distinguished from those acquired domestically, with information on quarantine at or after entry to the country. Imported cases were denoted G0; among domestically-acquired infections, those acquired directly from G0 cases were denoted as G1, others were denoted G2+.
Intervention data consisted of daily time-series of the numbers in quarantine and RT-PCR tests performed. Diaries of the Ministry of Health and Vietnam's National Steering Committee for COVID-19 response identified key milestones in national SARS-CoV-2 control measures. Apple mobility data 13 and Google community mobility data 14 provided proxies of population movements.
Serial intervals were calculated from infector-infectee pairs and fitted to a normal distribution by maximum likelihood 8 . The estimated distribution parameters were used to estimate the proportion of pre-symptomatic transmissions and three time-varying reproduction numbers 15 : between G0 and G1 (step 1), between G1 and G2+ (step 2), and between G0, G1, and G2+ combined (step 1 and 2 combined).
We used logistic regression to investigate the link between the proportion of asymptomatic infections and age, gender, nationality (Vietnamese versus non-Vietnamese), and imported versus domesticallyacquired infection. We used Gamma regression to investigate the link between the duration of . 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 May 16, 2020. . https://doi.org/10.1101/2020.05.12.20099242 doi: medRxiv preprint hospitalization and the same variables listed above, plus symptomatic versus asymptomatic. To account for potential non-linear effects of age in both models, this variable was binarized according to an age threshold, the value of which was estimated by maximum likelihood in the model fitting. To correct for potential confounding effects between the explanatory variables, we used Type-II likelihood ratio tests 16 . All analyses were done with R 4.0.0 17 using the packages EpiEstim 18 2.2-1 and fitdistrplus 19 1.0-14.

Epidemic description and control measures
The epidemic timeline for Vietnam, including the numbers quarantined and hospitalized, tests performed, cases confirmed, population movements, and the timing and nature of major Government-led control measures are summarized in Figure 1. The control measures are also summarized in Table 1 and in greater detail in Table S1. To date, two waves of transmission have occurred: the first began on January 23 rd and resulted in 16 cases (9 imported, 7 acquired in-country), and the second on March 6 th , leading to 254 cases (154 imported, 100 acquired in-country).
The first confirmed cases of SARS-CoV-2 infection presented in Hanoi and Ho Chi Minh City during the lunar New Year holiday (23-29 th January). Cases were travelers from Wuhan city or their contacts, and were identified by the public health laboratory network using improvised molecular diagnostics, including sequencing, prior to implementation of the WHO-approved assays 20,21 . Amongst the cases were the first confirmed human-to-human transmissions outside of China 22 .
Entry of airline passengers into Vietnam from Wuhan city and elsewhere in China was monitored and progressively limited ( Table 1), and cases and their contacts were quarantined for 14 days in Government facilities to prevent onward transmission. Schools and universities remained closed after the lunar New Year holiday, with staggered re-opening from May 4 th (closures lasted ~3 months). The National Steering Committee for COVID-19 response was established in late January, composed of 24 members from 23 ministries charged with coordinating the epidemic response. A hotline was set-up . 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 May 16, 2020. . https://doi.org/10.1101/2020.05. 12.20099242 doi: medRxiv preprint by the Ministry of Health on January 27 th , a nationwide SMS push notification system was put in place through all mobile phone providers on February 3 rd , and a mobile phone app for contact tracing and symptom reporting was launched on February 8 th .
In early February, following the repatriation of a number of Vietnamese nationals from Wuhan city, a cluster of community transmitted infections was detected in two communes in Vinh Phuc province, . 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 May 16, 2020. . https://doi.org/10.1101/2020.05. 12.20099242 doi: medRxiv preprint After further measures to prevent entry of infected international travelers (Table 1), a nationwide lockdown was enforced on April 1 st , including closure of all shops except gas stations, food stores, and pharmacies; suspension of public transport including all taxis; and mandatory mask wearing in all public spaces. Mobility data show that population movement decreased substantially very soon after the start of the second infection wave in early March, reaching a nadir in early April at the start of the lockdown (Figure 1). Movements increased slowly during the last week of the lockdown and more rapidly once the lockdown was partially lifted on April 23 nd . On April 15 th , the last case of wave two was identified; subsequent cases (n = 2) have since been detected among international travelers quarantined on arrival. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Epidemiological parameters over time
From 33 infector-infectee pairs, the mean serial interval was estimated to be 3.24 days (95% confidence interval (CI), 1.38-5.10 days) with a standard deviation of the distribution of 5.46 days (95% CI, 4.14-6.78 days). An estimated 27.5% (95% CI, 15.7%-40.0%) of the distribution was below zero, suggesting these transmissions occurred prior to the onset of symptoms in the infector. From the (non-quarantined) imported cases (G0) and onward infected cases (G1 and G2+), we calculated the effective reproductive number R by date (Figure 1). Limited transmission amounted to a maximum R of 1.15 (95% CI, 0.37-2.36). R rarely exceeded 1 and a decrease of R is seen as more mitigating measures were implemented from the end of March before the nationwide lockdown. When analyzing R from G0 to G1 (step 1) and from G1 to G2+ (step 2) separately, we found that R was drastically decreased for step 1 simultaneously with suspension of all international travel (March 18 th ), while for step 2 transmission continues with R slightly above 1 despite intense contact tracing and quarantine. Only the nationwide lockdown was able to eventually bring R below 1 (Figure 1).

Discussion
On January 23 rd 2020, Vietnam was one of the first countries to report SARS-CoV-2 infection and the first to report human-to-human transmission outside of China 22 . Yet 100 days later, it confirmed just 270 cases despite extensive testing, with no community transmission since April 15 th . In the three weeks prior to May 1 st , there were only two imported cases and no reported cases elsewhere in the country. The nature, timing, and success of the control measures introduced may have relevance to other countries seeking to control SARS-CoV-2 transmission.
Vietnam has experience in responding to emerging infectious diseases. In the last 20 years, it has confronted outbreaks of SARS 24 , avian and pandemic influenza 25,26 , hand-foot-and-mouth disease 27 , measles 28 , and dengue 29 . Its outbreak responses are coordinated by the Ministry of Health, a . 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 May 16, 2020. The high proportion of cases that developed symptoms after isolation (73.9%) or never developed symptoms (43%) highlights one of the major challenges of controlling SARS-CoV-2 and the strengths of Vietnam's approach. Suspected cases were identified and quarantined based on their epidemiological risk of infection (recent contact with a confirmed case or travel to a COVID-19 affected country), rather than on exhibiting symptoms. Without the implementation of strong control . 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 May 16, 2020. The strength of our report is that it provides a complete national picture of case numbers, their clinical and demographic characteristics, and the testing performed and various interventions made by the Government over time. Further, the use of systematic quarantine measures allowed clear distinction between imported and domestically-acquired cases, thus allowing for estimation of the efficiency of various interventions. The limitations are that the data are descriptive, contain relatively small numbers of confirmed cases, and only include the first 100 days of an epidemic that is likely to continue for many months. It is therefore impossible to conclude definitively which of these control measures have resulted in the current control of SARS-CoV-2 in Vietnam and whether they will continue to work in the future.
There are, however, two distinctive features of Vietnam's response. First, the Government acted quickly, placing restrictions on international flights, closing schools and universities, and instituting exhaustive case-contact tracing from late January, well before these measures were advised by WHO.
Second, they placed the identification, serial testing, and minimum 14-day isolation of all direct contacts of cases, regardless of symptom development, at the heart of the response. Our findings suggest the latter measure was likely to be especially effective given nearly half of those infected did not develop symptoms.
In summary, Vietnam has controlled SARS-CoV-2 spread by acting early, maintaining clear and consistent public communications, introducing meticulous contact-tracing and quarantine, and implementing progressive international travel restrictions. The value of these interventions in controlling the infection is supported by the high proportion of asymptomatic cases and imported cases, and evidence for substantial pre-symptomatic transmission.
. 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 May 16, 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 May 16, 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 May 16, 2020. . https://doi.org/10.1101/2020.05.12.20099242 doi: medRxiv preprint Table 1 Table S1). The colors shown in the phase column indicate the intensity of control measures taken over different periods (white, initial; light yellow, early; light orange, intermediate; orange, preepidemic; brown, epidemic/lockdown; dark orange, post-lockdown), and correspond to those used in Figure 1 and Table S1.

. The timing and nature of major Government-led control measures, including international border control, internal control, and Ministry of Health-led communications (further details provided 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.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 16, 2020. 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 May 16, 2020. .

Figure 1. Timeline of SARS-CoV-2 emergence and response in Vietnam.
The background color reflects the intensity of the interventions taken by the Vietnamese Government to control the COVID-19 epidemic, with darker shades indicating more intense disease control measures. The main events of these periods are described in detail in Table 1 The shaded blue area shows the 95% confidence intervals.
. 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 May 16, 2020. . https://doi.org/10.1101/2020.05.12.20099242 doi: medRxiv preprint 18 . 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.

Figure 2. Demographics of the 270 SARS-CoV-2 positive cases in Vietnam.
Age distribution for the 163 imported cases (left column) and the 107 cases of local transmission (right column), by nationality (top row) and gender (bottom row).
. 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 May 16, 2020. . https://doi.org/10.1101/2020.05.12.20099242 doi: medRxiv preprint . 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 May 16, 2020. . https://doi.org/10.1101/2020.05.12.20099242 doi: medRxiv preprint 21 . 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 May 16, 2020. . https://doi.org/10.1101/2020.05.12.20099242 doi: medRxiv preprint Table S1. The timing and nature of major Government-led control measures, including international border control, internal control, Ministry of Health-led communications, and enhancement of diagnostic capability. The colors shown in the phase column indicate the intensity of control measures taken over different periods (white, initial; light yellow, early; light orange, intermediate; orange, pre-epidemic; brown, epidemic/lockdown; dark orange, post-lockdown), and correspond to the colors used in Table 1 and Figure 1 in the main text.