Estimation of the basic reproduction number of novel coronavirus (COVID-19) in Bangladesh: 1 A 65-day outbreak data-driven analysis

27 Background: The outbreak of novel coronavirus disease (COVID-19), started from Wuhan, China, at 28 the end of December 2019, hits almost the entire world. In Bangladesh, the first case was officially 29 reported on March 8, 2020. We estimated the basic reproductive number, 𝑅 ! , of COVID-19 for 30 Bangladesh using the first 65-day data of the outbreak. 31 Methods: With time-varying disease reporting rate, epidemic curves were estimated using the 32 exponential growth model utilizing daily COVID-19 diagnosis data in Bangladesh from March 8 to 33 May 11, 2020. We estimated 𝑅 ! using the estimated intrinsic growth rate ( 𝛾 ). Serial intervals (SI) have 34 been used from two well-known coronaviruses’ outbreaks, SARS and MERS; and the early estimate of 35 SI of COVID-19 in Wuhan, China. 36 Results : The COVID-19 epidemic in Bangladesh followed an exponential growth model. We found 37 the 𝑅 ! to be 1.84 [95% CI: 1.82 – 1.86], 1.82 [95% CI: 1.81 – 1.84], and 1.94 [95% CI: 1.92 – 1.96], 38 for MERS, COVID-19, and SARS SI respectively without adjusting reporting rate. With the adjusted 39 reporting rate, 𝑅 ! reduced to 1.63 [95% CI: 1.62 – 1.65], 1.62 [95% CI: 1.61 – 1.64], and 1.71 [95% 40 CI: 1.70 – 1.73] for a five-fold increase. Inverse association between the reporting rate and the basic 41 reproduction number was observed. 42 Conclusion : The 𝑅 ! was found to be 1.87 for existing cases and was reduced to 1.65 for the five-fold 43 increase of the early reporting rate. Findings suggest a continued COVID-19 outbreak in Bangladesh 44 and immediate steps need to be taken to control.


Introduction 51
Atypical pneumonia cases with unknown aetiology were first detected in Wuhan, China, at the end of 52 December 2019 (1). This disease subsequently found to be caused by a virus, and the World Health 53 Organization (WHO) named this virus as severe acute respiratory syndrome coronavirus 2 (SARS-54 COV-2) (2). The disease itself was named as a coronavirus (COVID-19) (2). Evidence suggested that 55 the likelihood of spreading this disease for cross border travelling (3) is very high and hence indicated 56 the global spreading (4). On March 11, 2020, the WHO declared that COVID-19 disease outbreaks a 57 global pandemic (5, 6). As of May 11, 2020 (10.00 GMT+ 6), there have been 4,137,591 confirmed 58 cases found in 187 affected countries, among which 17,07,756 (41%) have been closed. Among the 59 closed cases, 14,24,230 (83%) have recovered from the infection while 2,83,526 (17%) cases were 60 closed as death (7) by COVID-19. The most affected country was the United States of America 61 (13,32,609 cases and 79,607 death) which was followed by Spain (2,24,350 cases and 26,621 death), 62 and United Kingdom (2,24,327 cases and 32,140 death) based on the data up to May 11 2020 (7). In 63 Bangladesh, the first three cases were detected on March 8, and the first death reported on March 18, 64 2020, from COVID-19 disease (8 To the best of our knowledge, there was no existing research quantifying the transmissibility of COVID-81 19 in Bangladesh. In this study, we attempt to estimate the transmissibility of COVID-19 using the basic 82 reproductive number, ! utilizing the publicly available daily reported data in Bangladesh.  To account for the effect of increasing surveillance, and to offset the underreported case diagnosis in 100 the early days; we adopted an adjustment mechanism for the daily number of diagnosed cases by a time-101 5 varying reporting rate, motivated by previous studies (14, 18). We assumed that the reporting rate 102 reached at the maximum level on the day the country started testing 2,000 individuals per day. We 103 adjusted the daily number of diagnosed cases before this point in time using a series of reporting rate 104 increments time-varied by the daily number of tests performed. Let ( ) be the daily diagnosed new 105 cases on the day ; ( ) be the number of people tested, and ( ) be the fold change of reporting rate, 106 Then, the adjusted number of daily diagnosed cases, denoted by " ( ), is obtained as 107 (1) Finally, the adjusted cumulative number of cases were computed ( ) = ∑ " ( ) Alongside 108 analyzing the originally diagnosed number of cases, we considered analyzing multiple other scenarios 109 defined by 0.5, 1, 2, 3, 4, and 5-fold increasing of reporting rate. As it is very difficult to assess the 110 exact reporting rate due to unavailability of required data, we considered analyzing these seven scenario 111 covering a broad spectrum of variation in the reporting rate. Figure 1 provides further detail on each of 112 these scenarios that we have investigated for COVID-19 SI. Other scenarios for MERS and SARS SI 113 was given in the supplemental file Figure S1 and Figure S2, respectively. 114

Reproductive Number 115
Following previous studies (14, 19, 20), exponential epidemic curves have been fitted to the data on the 116 daily diagnosed number of cases, both original and adjusted, using the equation (1)   We employed the Gamma distribution as the distribution of serial interval. We used the serial interval 126 (SI) information reported for COVID-19 cases of Wuhan, China using 425 cases (24). We also have 127 conducted our analyses using the SI information for Middle East Respiratory Syndrome (MERS), and 128 the Severe Acute Respiratory Syndrome (SARS) as these two viruses share the same pathogen as 129 COVID-19. The mean ± SD for Gamma distribution was considered as 7.6 ± 3.4 days for MERS (25), 130 7.5 ± 3.4 days for COVID-19 (24); and 8.4 ± 3.8 days for SARS (26). All the data were analyzed using 131 statistical software R using the package R0 (21). 132

Results 133
The estimated exponential epidemic curves are presented in Figure 1  reporting rate, indicated that the epidemic data satisfactorily followed the exponential growth as we 140 assumed. Figure 2 presents  reporting rate is increased to 5-fold to adjust the case numbers in the early phase of the outbreak. 152

Discussion 153
With the ever-growing number of COVID-19 cases, the world is currently in a pandemic situation (5, Besides these, the basic reproduction number estimated in different provinces of China at different 168 period ranged between 1.5 and 6.49 (33). In Korea, this figure was 3.54 (data using between February 169 18-March 1, 2020) (13), which was higher than that in Bangladesh. In Italy, in the early phase (February 170 25-March 12, 2020) of the outbreak, the estimated ! was 2.43 to 3.10 (34), and in Algeria, it was found 171 to be 2.55 [95% CI: 2.15 -2.94] using the actual incidence data of first 25 days of the outbreak (35). 172 The accuracy of the basic reproduction number relies on the selection of the SI of COVID-19; however, 173 as of yet, there is no consistent evidence regarding this interval. Determining the SI requires sufficient

Ethics approval and consent to participate 197
The ethical approval or individual consent was not applicable 198

Availability of data and materials 199
All data and materials used in this work were publicly available.