key: cord-0991067-nkieqheq
authors: Quach, Ha-Linh; Nguyen, Khanh Cong; Hoang, Ngoc-Anh; Pham, Thai Quang; Tran, Duong Nhu; Le, Mai Thi Quynh; Do, Hung Thai; Vien, Chien Chinh; Phan, Lan Trong; Ngu, Nghia Duy; Tran, Tu Anh; Phung, Dinh Cong; Tran, Quang Dai; Dang, Tan Quang; Dang, Duc-Anh; Vogt, Florian
title: Association of public health interventions and COVID-19 incidence in Vietnam, January to December 2020
date: 2021-07-29
journal: Int J Infect Dis
DOI: 10.1016/j.ijid.2021.07.044
sha: 5ac23b7d876dd4eacdf87ba739c92a96f9c11d84
doc_id: 991067
cord_uid: nkieqheq

BACKGROUND: Vietnam implemented various public health interventions such as contact tracing and testing, mandatory quarantine, and lockdowns in response to COVID-19. However, the effects of these measures on the epidemic remain unclear. METHODS: We described public health interventions in relation to COVID-19 incidence. We used maximum likelihood estimations to assess containment delays (time between symptom onset and start of isolation) and multivariable regression to identify associated factors between interventions and COVID-19 incidence. We calculated effective reproductive numbers (Rt) based on transmission pairs. RESULTS: Interventions were introduced periodically in response to the epidemic. 817 (55.4%) among 1,474 COVID-19 cases were imported. Based on a serial interval of 8.72 (±5.65) days, we estimated Rt to reduce below 1 (lowest at 0.02 (95%CI 0-0.12)) during periods of strict border control and contact tracing, and to increase ahead of new clusters. The main method to detect cases shifted over time from passive notification to active case finding at immigration or in lockdown areas, with containment delays showing significant differences between modes of case detection. CONCLUSION: A combination of early, strict and consistently implemented of interventions is crucial to control COVID-19. Low-middle income countries with limited capacity can contain COVID-19 successfully using non-pharmaceutical interventions.

As of June 2021, more than 182 million COVID-19 cases and nearly 3.9 million deaths related to COVID-19 have been reported globally (1) . Various public health measures were implemented at different stages across the world, with varying success (2) (3) (4) . China successfully contained the outbreak through strict lockdown measures (5, 6) . High-income countries like New Zealand, Australia or Taiwan came close to elimination of community transmission for several months during 2020 through strict border control and extensive contact tracing (7) . A number of low-middle income countries successfully applied a mixture of non-pharmaceutical interventions (NPIs) (6) (7) (8) . Among these was Vietnam, where a large variety of interventions were implemented throughout 2020, including travel restrictions, mandatory testing and quarantine of international entry points, social distancing, and regional lockdowns.

Despite recent advances in vaccine development, NPIs will remain paramount until the very end of the pandemic (11, 12) . It is therefore crucial to identify the most effective public health interventions for different stages of the epidemic (13) . However, evidence about the impacts of different public health measures on the epidemic is scarce, in particular for low-middle income countries.

We assessed the association between NPIs and COVID-19 incidence during different epidemic periods in Vietnam during 2020 to better understand how disease spread and response measures relate to one another and to provide public health guidance for the ongoing epidemic. The fact that Vietnam was affected by COVID-19 early on but never experience widespread community transmission nationwide provided opportune conditions for this research.

Data of new daily confirmed COVID-19 cases were provided by the Vietnam Ministry of Health. All COVID-19 cases recorded between January and December 2020 were included in the analyses.

Cases were defined according to the COVID-19 case definition of Vietnam's Ministry of Health (14, 15) . A laboratory confirmed case was defined as a person receiving a positive SARS-CoV-2 test result by real-time reverse transcriptase-polymerase chain reaction (rt-PCR). A suspected case was defined as any person experiencing fever, cough or shortness of breath, with or without a travel history to COVID-19 affected areas during the 14 days before symptom onset, and/or close contact with a suspected or confirmed COVID-19 case during the 14 days before symptom onset. The source of infection was classified as imported (infection acquired most likely outside of Vietnam) or domestic (infection acquired most likely in Vietnam). The mode of case detection was categorized as: i) self-presentation at health facilities; ii) testing and quarantine at immigration points; iii) contact tracing following exposure with a confirmed case; and iv) enhanced mass testing in lockdown areas. All confirmed cases of COVID-19 were required to isolate at designated hospitals immediately after confirmation for monitoring and/or treatment (if needed). Clinical conditions of diseased cases during isolation were categorized as stable or critical. Cases were defined as free of infection, and thus discharged, if they tested negative for SARS-CoV-2 in the row of three consecutive tests, with a sampling interval of one day between each test.

To describe the dynamics of the COVID-19 epidemic, six time periods were distinguished according to important milestones of interventions and virus transmission in Vietnam during 1 spanned from 23 January to 20 March 2020, when only imported cases and importationrelated domestic cases were reported. During Period 2 (21 to 31 March 2020), the first two community outbreaks were reported and extensive testing and contact tracing strategies were enforced. During Period 3 (1 to 30 April 2020), nationwide social distancing measures was implemented, while the third community outbreak was detected. By Period 4 (1 May to 24 July), after social distancing was lifted, international border control was in the focus due to an influx of international travel. During Period 5 (25 July to 15 September), two outbreaks were detected and controlled. In Period 6, Vietnam reported another two outbreaks due to unauthorized entry and quarantine violation, which in turn led to strengthened international border policies.

We grouped public health interventions into three categories: (1) Travel-related measures; (2) Active case finding measures; and (3) Other NPIs. Travel-related measures included travel restriction, arrival screening and quarantine, and border closure to high-risk countries. Active case finding measures included contact tracing, testing, quarantine of contacts, and confirmation of cases. The remaining NPIs included regional lockdowns, school closure, cancellation of mass gathering and non-essential activities, and personal protective measures (i.e. mask wearing, hand hygiene, etc.).

The effective reproduction number (Rt) was defined as the average number of secondary infections generated by a single case in a population at time t (16, 17) . The serial interval, a key parameter for Rt (18, 19) , was defined as the time interval between symptom onset of transmission pairs, which was calculated by interval from date of symptom onset in the primary case to date of symptom onset in the secondary case (20) (see Supplement 2 for details about how transmission pairs were established). Containment delay of a case was defined as the time interval between the date of symptom onset and the date of isolation. For cases who were not symptomatic at testing, the date of positive test confirmation was used.

Imported cases who were quarantined directly at immigration points were excluded from the containment delay, serial interval, and Rt analysis.

An epidemic curve by date of onset was plotted along with implemented interventions, with a detailed description of the three aforementioned intervention groups during each period.

Epidemiological and socio-demographic characteristics of all confirmed cases were compared across the six epidemic periods using Chi-square and Fisher's exact tests.

Multinomial logistic regression was used to compare case characteristics by mode of detection, and risk ratios (RR) and 95% confidence intervals (95%CI) were calculated.

Logistic regression was used to identify factors associated with clinical conditions, based on odd ratios (OR) with 95%CI. Estimations for containment delays stratified by case characteristics were summarized using means and standard deviation. R package "fitdistrplus" (21) was used for maximum likelihood estimations to fit distributions of containment delays, and to compare by goodness of fit test for AIC. Multivariable linear regression was used to explore associated factors of containment delay using odd ratios (OR) with 95% CI. The threshold for statistical significance was set at a p-value of less than 0.05. We used the R package "EpiEstim" (22) to estimate serial interval distributions, and then used these estimation to calculate median Rt with 95% CI based on the daily number of cases via a seven-day moving average using the method developed by Cori et al (23). Table 2 shows the comparison of case characteristics between different modes of case detection while controlling for sex and age. As compared to cases detected at immigration points, cases were more likely to be symptomatic if detected through active notification by hospitals and in lockdown areas (RR 6.35 (95%CI 3.75-10.79) and 2.21 (95%CI 1.35-3.62), respectively). When stratified by epidemic periods, enhanced testing in lockdown areas detected more cases in most periods than immigration point testing. 

By the end of 2020, 89.96% of all confirmed cases had been discharged from isolation. 35 fatalities due to COVID-19 were recorded in Vietnam during 2020 (case fatality ratio, 0.07%); all were in Period 5. More than 90% of cases were discharged during Period 1, the latter half of Period 4, and in Period 6 ( Figure 1B Table 4 and Figure 2A describe the public health interventions implemented across six epidemic periods of COVID-19 in Vietnam during 2020. Border control and NPIs were imposed before the first case was reported, and were expanded in scope throughout the analysis period. Cancellation of mass gatherings and school closures were implemented in late January and continued periodically as soon as community cluster was detected, when mask wearing and personal hygiene practice were already encouraged. From Period 1 onwards, regional lockdowns were enforced frequently during flare ups of clusters, along with extensive contact tracing, testing, and quarantine for close contacts of suspected and confirmed cases. After local transmission was first reported in Period 2, national-level social distancing recommendations, and international and domestic mobility restrictions were introduced swiftly into Period 3. These measures were subsequently relaxed and restricted mostly to limit case importation in Period 4. Period 5 saw the biggest clusters recorded (Supplement 1), and regional lockdowns were extended to city-wide lockdown, with the utilization of community surveillance teams and universal testing. The number of cases soon declined, and returned to predominantly imported cases in Period 6, when measures were focused on border control. During this period Vietnam reported two instances of imported cases among illegal trespassers and one community cluster due to quarantine violation (Supplement 1). Figure 2 is about here 182 transmission pairs were recorded in the database, resulting in a gamma distributed serial interval with a mean of 8.72 days (SD 5.65) (Supplement Figure 1) . Figure 2B Table 1 ). As sensitivity analyses, we also examined the distribution of serial intervals excluding outlier intervals longer than 21 days (Supplement Data of 713 cases were used to calculate containment delays, defined as time between a case's symptom onset and start of quarantine, while 761 imported cases who were immediately isolated upon arrival to Vietnam were excluded. Overall, the mean of containment delay was 1.62 days (95% CI 1.38-1.86) with a standard deviation of 3.31 days (95% CI 3.15-3.50) (Supplement Table 2 ). For the 201 cases who were symptomatic at testing, longer means of containment delay were observed (5.04 days (95%CI 4.40-5.69)).

There were significant differences between the mean containment delay for cases by mode of Table 3 ). Table 5 shows univariate and multivariable regression analyses of containment delays with case characteristics and epidemic periods after adjusting for age and sex. Containment delay was significantly longer for cases who were symptomatic at testing than for cases who were not (RC 4.79 days (95%CI 4.36-5.22)). Cases detected by contact tracing and enhanced testing in lockdown areas had significantly shorter containment delays than cases who selfpresented at health facilities (RC 2.11 days (95%CI 2.73-1.49) and 2.03 days (95%CI 2.65-1.40), respectively). Cases detected during Period 1 had shorter containment delays than cases detected in all other periods, which was statistically significant compared to Period 5 (RC 1.56 days (95%CI 0.66-2.46).

This study provides evidence of the effectiveness of public health interventions in response to the COVID-19 incidence in Vietnam. While the main modes of case detection shifted from self-presentation to pre-emptive testing and quarantine at immigration points and lockdown areas, there were significant positive impacts on containment delays through active case finding measures compared to passive notification. Reproduction numbers reduced with the swift and strict implementation of NPIs.

Given the close proximity to Chinathe initial COVID-19 epicenter, Vietnam was particularly vulnerable to COVID-19 importation from international travel. However, we saw limited domestic cases even during times of high load of imported cases during the first epidemic period , and again in Period 4 and 6, which was a direct result of testing and quarantine for all inbound passengers upon arrival (27). As evidence grew about the infectivity and prevalence of asymptomatic cases (28) (29) (30) , Vietnam rolled out testing irrespective of symptoms very early into the epidemic. This was first done at immigration points for international passengers, and later for all persons with epidemiological links to confirmed cases. This policy was implemented in early March, and following that we saw the majority of cases in all periods being detected before any symptoms emerged. This also explained longer containment delays for domestic cases than imported cases (31), with shorter delays among cases without symptoms at time of testing.

A high number of cases were detected through enhanced testing in lock down areas. Several countries, including China and Italy, implemented severe lockdowns of large geographical areas during the beginning of the pandemic, and succeed in containing widespread transmission (32) . Meanwhile, Vietnam opted for focused lockdowns of clearly defined areas, where mass testing was done to all persons living in these areas irrespective of symptoms with the help of community surveillance teams, who detected any infectious persons at household level. Smaller-scale lockdowns were preferred due to limited response capacity and also to utilize volunteers from within in the community more efficiently with the aim to increase social responsibility among the population (33) .

Similar to the scientific literature encouraging an adaptive combination of quarantine and other public health measures (32, 34, 35) , Vietnam's extensive tracing and mandatory quarantine of all contacts in combination with social distancing measures helped to contain further spread. Increased active case finding, contact tracing and regional lockdowns had significant impacts on shortening containment delays in our study. When comparing periods of small-scale lockdown (Period 1 to Period 3, Period 6) and wide-scale lockdown (Period 5), quarantine measures focusing on close contacts of suspected and confirmed cases led to a significantly shorter containment delay. This might be explained through the small size of lockdown areas, where contacts with confirmed cases were most apparent, and thus easier and quicker to test and identify.

In our study, the effective reproduction number varied over time, especially under the implementation and subsequent relaxation of NPIs. After the first two unlinked community outbreaks in Period 2 quickly unfolded into Period 3, when regional lockdowns and national social distancing were implemented, we saw a clear reduction of Rt. This strongly suggests that timely implementation of active case finding with prompt quarantine and social distancing, even without general lockdowns, can quickly curtail an ongoing outbreak (36, 37) .

These findings confirmed results from a modelling study in eleven European countries showing how Rt began to drop after school closure and other NPIs, even before complete national lockdown (3). In Vietnam, after social distancing was lifted in Period 4, a peak of Rt was observed, which eventually led to the biggest outbreak in Period 5, which can be explained through the reintroduction of imported cases into high-risk areas (38) , and/or the resumption of previous social mixing patterns as observed in several countries following lockdowns (39, 40) . In Period 5, a wide-scale lockdown of two affected cities (Da Nang and Hai Duong), coupled with extensive testing and quarantine policies (see Supplement 1), resulted in a substantial reduction of Rt after only two weeks, and in the containment of outbreaks two months later. This aligns with previous research on the timeliness and effectiveness of combined public health practices (17, 35, 41) , as applied in Singapore (42, 43) , Taiwan (44) , and Hong Kong (45) .

From period 1 to period 3, our Rt estimation approximates those from our previous research on the first 100 COVID-19 cases in Vietnam (46) , and one study during the same period from

Taiwan (47), where the COVID-19 situation was similar to Vietnam's at that time. However, to our knowledge there is no other long-term observation of Rt beyond 6 months, making the comparison of our estimations difficult. We also observed a higher estimation of serial interval compared to previous reviews (48) (49) (50) . High percentage of asymptomatic at testing reported in Vietnam (Table 1 ) could hinder the detection and accuracy of recalled symptom onset for both infectors and infectees, and lengthen observed serial interval.

Towards the latter half of 2020, Vietnam experienced an increasing risk through unauthorized entry without quarantine upon arrival. It has been suggested that the biggest cluster in July (Period 5) was due to undetected imported cases through unauthorized entry, since the city -Da Nangis important for foreign companies and industries in Vietnam (51, 52) . During this time, several cases of illegal trespassing were detected across the country, some of which were detected in the epicenter of the cluster itself (53, 54) . At the end of the year, another three instances of imported cases without quarantine, one resulting in community transmission, shows that strict testing and quarantine at arrival and border control can only limit but not eliminate the introduction of cases. As shown in our study, we see that even with the exclusion of imported cases by border control, transmission chains could not be disrupted In non-infected cities/provinces, from 19 -31

August, non-essential services including outdoor dining, bar, club, and entertainment venues, were required close. All healthcare facilities were required to limit intake caretakers to one per patient per day, and no caretakers visit in emergency wards. Stricter legal punishment was given to illegal trespassers who were confirmed positive with SARS-CoV-2 and/or infected to other members in community. 

Model 2 represents linear regression analyses on the association between containment delay and cases' symptoms onset at testing, adjusted for sex and age Model 3 represents linear regression analyses on the association between containment delay and the modes of case detection, adjusted for sex and age Model 4 represents linear regression analyses on the association between containment delay and epidemic periods, adjusted for sex and age Model 5 represents linear regression analyses on the association between containment delay and cases' source of infection, cases' symptoms onset at testing, the modes of case detection, and epidemic periods, adjusted for sex and age

World Health Organization. WHO Coronavirus (COVID-19) Dashboard

World Health Organization

The impact of non-pharmaceutical interventions on SARS-CoV-2 transmission across 130 countries and territories

Estimating the effects of non-pharmaceutical interventions on COVID-19 in Europe

The temporal association of introducing and lifting non-pharmaceutical interventions with the timevarying reproduction number (R) of SARS-CoV-2: a modelling study across 131 countries

Association of Public Health Interventions With the Epidemiology of the COVID-19 Outbreak in Wuhan

Current epidemiological and clinical features of COVID-19; a global perspective from

Potential lessons from the Taiwan and New Zealand health responses to the COVID-19 pandemic

COVID-19 response in low-and middle-income countries: Don't overlook the role of mobile phone communication

Community interventions in Low-And Middle-Income Countries to inform COVID-19 control implementation decisions in Kenya: A rapid systematic review

Learning from LMICs: Best practices for leveraging sentinel surveillance systems to track the US COVID-19 pandemic

The impact of non-pharmaceutical interventions on SARS-CoV-2 transmission across 130 countries and territories

The impact of non-pharmaceutical interventions, demographic, social, and climatic factors on the initial growth rate of COVID-19: A cross-country study

Role of Non-pharmaceutical Interventions (NPIs) During COVID-19 Pandemic: A Systematic Literature Review

Decision No 3468/QD-BYT Official Guidance on Surveillance

Decision No. 3351/QD-BYT Official Guidelines for the diagnosis and treatment of COVID-19 cases caused by the new Corona virus strain (SARS-Cov-2)

The effective reproduction number as a prelude to statistical estimation of time-dependent epidemic trends

The reproduction number of COVID-19 and its correlation with public health interventions

Serial interval of COVID-19 among publicly reported confirmed cases

Emerging Infectious Diseases

Serial interval of novel coronavirus (COVID-19) infections

A note on generation times in epidemic models

Asymptomatic patients as a source of COVID-19 infections: A systematic review and meta-analysis

A systematic review of asymptomatic infections with COVID-19

Asymptomatic SARS-CoV-2 Infections among Persons Entering China from

Stringent containment measures without complete city lockdown to achieve low incidence and mortality across two waves of COVID-19 in Hong Kong

A Simple Decision Analysis of a Mandatory Lockdown Response to the COVID-19 Pandemic. Appl Health Econ Health Policy

Reaching further by village health collaborators: The informal health taskforce of Vietnam for COVID-19 responses

Mass screening vs lockdown vs combination of both to control COVID-19: A systematic review

Quarantine alone or in combination with other public health measures to control COVID-19: a rapid review

Cochrane Database of Systematic Reviews

Social distancing strategies for curbing the COVID-19 epidemic

Reduction in effective reproduction number of COVID-19 is higher in countries employing active case detection with prompt isolation

Assessing the impact of non-pharmaceutical interventions on the transmissibility and severity of COVID-19 during the first five months in the Western Pacific Region

COVID-19 effective reproduction number dropped during Spain's nationwide dropdown, then spiked at lower-incidence regions

Retrospective analysis of the Italian exit strategy from COVID-19 lockdown

Effectiveness of nonpharmaceutical interventions on COVID-19

Interrupting transmission of COVID-19: lessons from containment efforts in Singapore

Interventions to mitigate early spread of SARS-CoV-2 in Singapore: a modelling study

Combined interventions to suppress R0 and border quarantine to contain COVID-19 in Taiwan

Impact assessment of non-pharmaceutical interventions against coronavirus disease 2019 and influenza in Hong Kong: an observational study

The First 100 Days of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) Control in Vietnam

Comparison of Estimated Effectiveness of Case-Based and Population-Based Interventions on COVID-19 Containment in Taiwan

Rapid review of available evidence on the serial interval and generation time of COVID-19

Serial interval and incubation period of COVID-19: a systematic review and meta-analysis

Estimates of serial interval for COVID-19: A systematic review and meta-analysis

Alert for COVID-19 second wave: A lesson from Vietnam

Estimating and mitigating the risk of COVID-19 epidemic rebound associated with reopening of international borders in Vietnam: a modelling study

21 Chinese people illegal sneaking into Vietnam found and quarantined after their scattering | Vietnam Times

Vietnam Times -The Vietnam Union of Friendship Organizations

COVID-19 in the world: countries are entering the 2nd Covid-19 wave and lessons for Vietnam

Official Online News Outlet -Ho Chi Minh Department of Health

Preventive behavior of Vietnamese people in response to the COVID-19 pandemic

Perceptions and Attitudes Toward COVID-19-Related National Response Measures of Vietnamese: Implications for Pandemic Prevention and Control

Well-Being of Healthcare Workers and the General Public during the COVID-19 Pandemic in Vietnam

Serial interval of SARS-CoV-2 was shortened over time by nonpharmaceutical interventions

Decision 181/QD-BYT Temporary guidance of surveillance, prevention and control acute pneumonia caused by novel coronavirus (nCoV)

Official guidance on the diagnosis and treatment of acute pneumonia caused by the new Corona virus strain

Combating the COVID-19 Epidemic: Experiences from Vietnam

Việt Nam to announce nationwide pandemic

Emerging COVID-19 success story: Vietnam's commitment to containment -Our World in Data [Internet]. Our World in Data

Lesson learnt from COVID-19 outbreak control experience in Da Nang City

The Business Times. Vietnam PM orders halt to international commercial flights