key: cord-0795534-693q80xo
authors: Al Wahaibi, Adil; Al Manji, Abdullah; Al Maani, Amal; Al Rawahi, Bader; Al Harthy, Khalid; Alyaquobi, Fatma; Al-Jardani, Amina; Petersen, Eskild; Al-Abri, Seif
title: COVID-19 epidemic monitoring after non-pharmaceutical interventions: the use of time-varying reproduction number in a country with a large migrant population
date: 2020-08-20
journal: Int J Infect Dis
DOI: 10.1016/j.ijid.2020.08.039
sha: 8eb95f54d9263206d70acff0e633556310c2cb1e
doc_id: 795534
cord_uid: 693q80xo

BACKGROUND: COVID-19’s emergence carries with it many uncertainties and challenges, one of which is epidemic management strategies. Oman has implemented non-pharmaceutical interventions (NPIs) to mitigate the impact of COVID-19. However, responses to NPIs may be different across different populations in a country with a large number of migrants like Oman. This study investigates the different responses to NPIs assessing the use of time-varying reproduction number (R(t)) to monitor it. METHODS: Polymerase chain reaction (PCR) laboratory-confirmed COVID-19 data from Oman was used from February 24th to June 3rd, 2020 and included demographic and epidemiological information. Data were arranged into pairs of infector–infectee, and two main libraries of R software were used to estimate reproductive number (R(t)). R(t) was calculated for both Omanis and non-Omanis. FINDINGS: A total of 13,538 cases were included, 44·9% of which were Omanis. Among all, we identified 2769 infector–infectee pairs to calculate R(t). There was a sharp drop in R(t) from 3·7, (95% confidence interval [CI] 2·8-4·6) mid-March to 1·4 (95% CI 1·2–1·7) in late March in response to NPIs. Then R(t) decreased to 1·2 (95% CI 1·1–1·3) late April when it rose, corresponding to the easing up of NPIs. Comparing the two groups, the response to major public health controls was more evident in Omanis in reducing the R(t) to 1·09 (95% CI 0·84–1·3) at the end of March. INTERPRETATION: Use of real-time estimation of R(t) has allowed us to follow the effect of NPIs. The migrant population responds differently than the Omani population.

The emergence of a new infectious disease carries with it many challenges and uncertainties regarding its natural history, clinical course, transmissibility, and methods of control. For these uncertainties, the leading resort for control when there is no treatment or vaccine is to apply well-known strict public health measures to mitigate and prevent its spread. 1, 2 These measures were implemented by the Chinese government after the reporting of 27 cases of corona virus disease-2019 (COVID- 19) in Wuhan, in China's Hubei province, on the 30 th December 2019. Measures, among others, included the isolation of confirmed and suspected cases and restriction of movement inside the country. 3 During two months after the appearance of COVID-19, and despite the measures that were taken, international travel spread cases to 26 countries around the world, 4 after which the World Health Organization (WHO) announced the COVID-19 to be a worldwide pandemic, 5 with daily cases reaching to 6,080,963 by the 1 st of June. 6 With the increasing number of cases and the introduction of multiple NPIs by most countries, there is a need to assess and monitor the transmissibility of the disease as a measurement of the effectiveness of control measures. The use of the effective reproduction number, Rt, defined as the average number of secondary cases from a partially susceptible population per infectious cases, 7 is a universally used indicator for a control measure's effectiveness. 8 Oman is a country of 4.60 million, and has a heterogeneous population with migrants constituting 41% of the population. 9 The effect of the control and the dynamics of transmission of COVID-19 was expected to be different between these two populations, Omanis and non-Omanis. Similar situations J o u r n a l P r e -p r o o f that led to increasing numbers of cases have been documented in many neighbouring countries of the Gulf Cooperation Council, GCC, 10,11 and also in Singapore. 12 The government of Oman responded to the COVID-19 pandemic much like most countries by implementing multiple NPIs in phases to control the disease beginning in mid-March 2020. Examples include restricting flights from infected countries, to closing schools and commercial activities. (Table   1 ).

In this study, we will analyse the dynamics of COVID-19 infection transmissibility in Oman in the different populations (Omani and non-Omani) and the effects of the introduction of the nonpharmaceutical measures on disease transmissibility. The type of surveillance, whether active (proactive case finding) or passive (regular reporting by health care institutions), were also included in the data when only the passive surveillance cases were included.

The basic reproductive number before mitigation starts are called R0. The reproduction number after mitigation starts, Rt, is a measure of the transmissibility of the infection and defined as the average number of new infections one case can produce. 7 An Rt of more than one means that the infection is J o u r n a l P r e -p r o o f spreading with more cases generated, whereas, an Rt of less than 1, means that the spread of infection is decreasing. Theoretically, we need information about the generation time-defined as the time period between the infection of the index and the next case. However, this information is usually difficult to ascertain, therefore, information regarding the serial interval (defined as the interval between disease onset in the index and the next case) distribution in the data is used instead.

Using R software, we utilised two main libraries to estimate the Rt, Epicontact library 14 and EpiEstim library. 15 The main function of the Epicontact library 14 is to arrange the data and help estimating the distribution parameters of the serial interval in our data. The estimate-R function in the EpiEstim library 15 assumes a gamma distribution of the SI and models the transmission of the infection using a Poisson likelihood to calculate the instantaneous reproduction number.

We arranged our line list into two parts, as required by Epicontact library, 14 the main line list data and the contact data. The daily list of all new PCR-positive cases contains all the demographic and exposure data for each case, whereas, the contact data contains information about the transmission of infection between each identifier number. As case-by-case transmission data is available for many cases of the line list, we used the Epicontact R library to find out the serial interval of our data and its distribution for the entire population, and classified this data by nationality.

We used the serial interval distribution calculated by the Epicontact library to estimate the Rt for the entire population, classified by nationality, for the period between February 24 th until June 3 rd , 2020.

We used the estimate_R function in EpiEstim library to calculate the Rt given the distribution of serial interval (obtained by the Epicontacts library) and incidence time series. This was done through a sliding window of 7 days using the Poisson transmission model. 16, 17 The comparison between different Rt between Omanis and non-Omanis was performed after extracting Rt values from each estimate_R object and plotting the two-time series against each other.

To investigate the behaviour of the transmission in different nationalities, the epidemic curve trend was investigated according to cluster/sporadic types in Omani vs. non-Omanis groups using geom_smooth function in a ggplot2 library. 18

All data cleaning and statistical analysis was done using R software version 3.6.3. 19 We used two R packages to estimate the time-varying reproduction number, Epicontacts 14 and EpiEstim. 15 Plots of incidence were produced using the library incidence in R software. 20 The comparison between the Rt in the different groups was done using simple line plot.

As of June 3 rd , 2020, a total of 13,538 PCR laboratory-confirmed COVID-19 cases were included. Only passive cases presented to the health institutions were included, the active surveillance cases were removed from the dataset, which totalled 1,974. Of these cases, 44·9% were Omanis and the majority of these were in the Muscat governorate (71·1%), where Oman's capital city, Muscat, is located. Among all cases, 2,769 infector-infectee pairs data have been identified and included in the contact dataset.

The median serial interval was estimated to be 6, IQR (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) The daily epidemic curve by date of onset for the passive laboratory-confirmed COVID-19 cases classified by nationality: Omanis and non-Omanis is shown in (Figure 1 ). The first cases registered in Oman were in two women returning from a visit to Iran on February 24 th . There was a doubling of cases from the middle of March to the end of May noticed in both nationality groups. Analysis of the trend of Rt in the different nationality groups is presented in Figure 3 with the corresponding events as per Table 1 . Generally, there is a different behaviour of the Rt trend between the two groups. The response to the major public health control phase, phase 2, was more evident in Omanis in reducing the Rt to 1·09 (95% CI 0·84-1·3) at the end of March. Nevertheless, closure of Muscat and Mutrah (old market area) has a marked reduction in Rt for the non-Omani group reaching 0·9 (95% CI 0·8-1·1) by mid-April. whereas the number of sporadic cases not linked to a known cluster increase (Figure 4 ).

With the increasing numbers of COVID-19 cases in Oman, our study showed the feasibility of using the time-varying Rt to assess and explain transmissibility dynamics and epidemic progression. We J o u r n a l P r e -p r o o f showed that there is a marked reduction in the reproduction number of COVID-19 infections in Oman in response to the major public health control introduced by the government. However, this reduction was not strongly evident in the non-Omani group compared to the Omanis. In fact, the closure of Muscat (specifically the old market area) drastically decreased Rt. The estimated median serial interval in our study was 6, IQR (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) . Previous studies have estimated this parameter to be 4.6 days 21 (95% CrI: 3.5, 5.9) and 4.9 days (95%CI: 3.6−6.2) 22 The daily number of new cases is known to be influenced by testing capacity. However, the use of timevarying Rt in assessing the transmissibility dynamics and epidemic progression was a crucial tool to know how the mitigation measures influenced Rt. Nevertheless, a study from China demonstrated that changes in reporting rates substantially affect estimates of Rt. 23 Thus, the initial Rt is influenced by testing frequency, contact tracing, and reporting of mild cases outside hospital. Because of these reasons, Rt will fluctuate from country to country.

The development of the pandemic in Oman has followed the trend seen in other countries that could not curb the outbreak before the spread of COVID-19 in the community. Oman introduced most of the same mitigation measures implemented in most countries, stopped international travel by closing the points of entries (airports, seaports and ground crossing), closing schools and shops, sending public and private sector workers home, introducing the use of masks in public spaces, and restricted movement to districts with particular high numbers such as the old markets in Muscat. A modelling study from China showed how the comprehensive package of mitigation interventions China implemented in Hubei province caused the Rt to rapidly decline below one in areas where NPIs were implemented. 24 Another study based on data from Wuhan, China, found that the mitigation measures reduced the median number of infections by more than 92% (IQR 66-97) and 24% (13-90) at the peak and at the post peak of the outbreak. 25 On the basis of exponential curve prediction, and the assumption that the duration of infection ranges from 15 to 20 days, a study from Italy estimated the R0 to be from 2·8 to 3·3. 26 This number is similar J o u r n a l P r e -p r o o f to that reported for the initial phase of the infection outbreak in Wuhan, 24 and slightly higher than 2·2 as reported by Li and colleagues in a more recent report. 27 Such numbers were similar to those found in our study.

Throughout the outbreak, the number of cases has been higher in the migrant population compared to Many of the disease clusters in the Omani population are present because Omanis tend to live in extended families (5-12 per households), and thereby increase the possible number of contacts for each primary case. A study from the UK found that a 74% reduction in the average daily number of contacts observed per participant (from 10·8 to 2·8) would be sufficient to reduce Rt from 2·6 prior to lockdown to 0·62 (95% CI 0·37-0·89) after the lockdown, based on all types of contacts. 28 Another study from China also found that transmission of COVID-19 was prominent in family clusters. 29 The rise in Rt by the end of March was due to a considerable number (roughly 8,000) of Omani students returning from abroad even though they were placed in institutional or home quarantine. One study looked at the effect of travel restrictions in China and found that the travel quarantine of Wuhan delayed the overall epidemic progression by only 3 to 5 days in mainland China but had a more marked effect on the international scale where case importations were reduced by nearly 80% until mid-February. 30 The travel restrictions in Oman were effective, except for food supply trucks, which continued over the border with the United Arab Emirates, and between the locked down governorates.

The situation in Oman is complicated by having two major and socio-economically different population groups, Omani nationals and migrants. So far, countries with this population structure have been unable to control the outbreak as efficiently as countries with homogeneous populations, with Taiwan, South Korea and Hong Kong as prime examples. 31, 32 In Singapore, the outbreak expanded rapidly once the J o u r n a l P r e -p r o o f infection spread into the dormitories for migrant workers 12 and by June 16 th , 731 cases per 100,000 population were confirmed. In Qatar, another country with a large migrant population has 3,080 tested positive cases per 100,000 population since the start of the outbreak 10 while Saudi Arabia has reported 445 cases per 100,000 population. 11

Easing mitigation efforts increased the number of cases, and it can be argued that the opening of the small businesses and shops happened too early; however, the overall situation was complex because many small shop owners and self-employed had had no income for months. Nevertheless, with this transmissibility and with the way in which Oman handles communicable diseases, the crude mortality rate from COVID-19 in Oman is still comparably low at 0·4%.

Isolation and contact tracing reduce the time during which cases are infectious in the community, thereby reducing the Rt. The overall impact of isolation and contact tracing, however, is uncertain and highly dependent on the number of asymptomatic cases. 33 As case numbers rise, the burden of contact tracing and quarantining increases. A modelling study from the US indicates that in such a situation, emphasis should be on physical distancing and contact tracing whereas isolation should be prioritised to persons with a high risk of transmission to others. 34 An analysis of the four major clusters in South Korea estimated the Rt at 1·5 (95% CI: 1·4-1·6). The intrinsic growth rate was estimated at 0·6 (95% CI: 0·6-0·7), and the scaling of growth parameter was estimated at 0·8 (95% CI: 0·7-0·8), indicating sub-exponential growth dynamics of COVID-19. The results indicate an early sustained transmission of COVID-19 in South Korea and support the implementation of social distancing measures to rapidly control the outbreak. 35 One model found that the effects of physical distancing strategies vary across age categories; the reduction in incidence is highest among school children and older individuals and is lowest among working-age adults. Children are at a similar risk of infection to the general population, but less likely J o u r n a l P r e -p r o o f to have severe symptoms; hence they should be considered in analyses of transmission and control. 36 It is a limitation of this study that we did not stratify the Rt into different age groups. 37 Another limitation of this study is that the testing capacity increased as the pandemic progressed. Using daily time series of COVID-19 incidence, epidemic curves of reported cases may not always reflect the true epidemic growth rate due to changes in testing rates which could be influenced by limited diagnostic testing capacity during the early epidemic phase. 38 A third limitation of the study is the increasing number of sporadic cases by the end of the study period indicating the lagging in the identification and hence classification of the source of infection. This is likely due to the overwhelmed fatigued public health workforce in the country. There is, therefore, a call for increasing and training this workforce to be able to cope with current and future epidemics. In the short term, the introduction of advanced technologies (such as artificial intelligence and location tracking systems) will help public health professionals in this battle against COVID-19.

The use of real-time estimation of the Rt has allowed us to follow the effect of the mitigation strategies adopted by the government. Our analysis shows that the migrant population behaves differently than the nationals and that the COVID-19 infection is spreading more rapidly in this population mainly because of their special living conditions. . The trend of incidence per 100,000 population (with 95% CI) of daily cases in Omanis and non-Omanis. After June 1 st , the number of cases that could not be linked to known clusters increased rapidly. Cluster of cases: more than two cases that can be linked to a common index case. Sporadic cases: cases that cannot be linked to existing clusters. Allowing 50% of governmental employees to return to work (May 31st) Opening of Muscat Governorate (May 29th)

Effects of non-pharmaceutical interventions on COVID-19 cases, deaths, and demand for hospital services in the UK: a modelling study

Effect of non-pharmaceutical interventions for containing the COVID-19 outbreak in China. Infectious Diseases (except HIV/AIDS)

The effect of human mobility and control measures on the COVID-19 epidemic in China

Impact of international travel and border control measures on the global spread of the novel 2019 coronavirus outbreak

WHO Coronavirus Disease

Complexity of the basic reproduction number (R0)

Covid-19 Outbreak Progression in Italian Regions: Approaching the Peak by March 29th

Health care Workers Section

epicontacts: Handling, Visualisation and Analysis of Epidemiological Contacts

EpiEstim: a package to estimate time varying reproduction numbers from epidemic curves

A New Framework and Software to Estimate Time-Varying Reproduction Numbers During Epidemics

Improved inference of time-varying reproduction numbers during infectious disease outbreaks

Elegant Graphics for Data Analysis

R: A Language and Environment for Statistical Computing

incidence: Compute, Handle, Plot and Model Incidence of Dated Events

Serial interval of novel coronavirus (COVID-19) infections

Estimating the serial interval of the novel coronavirus disease (COVID-19): A statistical analysis using the public data in Hong Kong from

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

First-wave COVID-19 transmissibility and severity in China outside Hubei after control measures, and second-wave scenario planning: a modelling impact assessment

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

COVID-19 and Italy: what next?

Early Transmission Dynamics in Wuhan, China, of Novel Coronavirus-Infected Pneumonia

Quantifying the impact of physical distance measures on the transmission of COVID-19 in the UK

Household secondary attack rate of COVID-19 and associated determinants in Guangzhou, China: a retrospective cohort study

The effect of travel restrictions on the spread of the 2019 novel coronavirus (COVID-19) outbreak

How South Korea Responded to the Covid-19 Outbreak in Daegu

Response to COVID-19 in Taiwan: Big Data Analytics, New Technology, and Proactive Testing

Effectiveness of isolation, testing, contact tracing, and physical distancing on reducing transmission of SARS-CoV-2 in different settings: a mathematical modelling study

Comparative Impact of Individual Quarantine vs. Active Monitoring of Contacts for the Mitigation of COVID-19: a modelling study

Transmission potential and severity of COVID-19 in South Korea

Epidemiology and transmission of COVID-19 in 391 cases and 1286 of their close contacts in Shenzhen, China: a retrospective cohort study

Age-dependent effects in the transmission and control of COVID-19 epidemics

Changes in testing rates could mask the novel coronavirus disease (COVID-19) growth rate

The authors would like to thank Lesley Carson for her editorial assistance in finalizing the manuscript.

The authors declare that they have no conflict of interest

The study was funded by the Ministry of Health, Directorate General for Disease Control and Surveillance as part of an operational research.

The study was approved by the Directorate General for Disease Surveillance and Control, and there was no need for patients' consent as the study was anonymous and used the data produced for public health purposes.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.J o u r n a l P r e -p r o o f

Al Jardani review & editing the manuscript,Eskild Petersen wrote the discussion and contributed to the overall manuscript, Seif Al-Abri supervised the study and participated in all stages of the manuscript.