key: cord-0885339-1wn8b33j
authors: Marimuthu, S.; Joy, Melvin; Malavika, B.; Nadaraj, Ambily; Asirvatham, Edwin Sam; Jeyaseelan, L.
title: Modelling of reproduction number for COVID-19 in India and high incidence states
date: 2020-06-30
journal: Clin Epidemiol Glob Health
DOI: 10.1016/j.cegh.2020.06.012
sha: 75b8a52c99981f3b4dbe3fd351533b79e92e8a13
doc_id: 885339
cord_uid: 1wn8b33j

BACKGROUND: Since the onset of the COVID-19 in China, forecasting and projections of the epidemic based on epidemiological models have been in the centre stage. Researchers have used various models to predict the maximum extent of the number of cases and the time of peak. This yielded varying numbers. This paper aims to estimate the effective reproduction number (R) for COVID-19 over time using incident number of cases that are reported by the government. METHODS: Exponential Growth method to estimate reproduction rate R(0), and Time dependent method to calculate the effective reproduction number (dynamic) were used. “R0” package in R software was used to estimate these statistics. RESULTS: The basic reproduction number (R(0)) for India was estimated at 1.379 (95% CI: 1.375, 1.384). This was 1.450 (1.441, 1.460) for Maharashtra, 1.444 (1.430, 1.460) for Gujarat, 1.297 (1.284, 1.310) for Delhi and 1.405 (1.389, 1.421) for Tamil Nadu. In India, the R at the first week from March 2–8, 2020 was 3.2. It remained around 2 units for three weeks, from March 9–29, 2020. After March 2020, it started declining and reached around 1.3 in the following week suggesting a stabilisation of the transmissibility rate. CONCLUSION: The study estimated a baseline R(0) of 1.379 for India. It also showed that the R was getting stabilised from first week of April (R = 1.1), despite the increase in March. This suggested that in due course there will be a reversal of epidemic. However, these analyses should be revised periodically.

Since the onset of the COVID-19 in China, forecasting and projections of the epidemic based on epidemiological models have been in the centre stage for planning and implementations of strategies to mitigate the epidemic. There has been a great dependence of these due to the unprecedented nature and uncertainties of the disease. Researchers have used various models to predict the maximum extent of the number of cases and the time of peak.

However, the rate of new cases over time, referred to as incidence, serves as a proxy for risk but does not contribute as a metric for the epidemic progression or prediction of its evolution. 1, 2 There are other measures that could serve as proxy for disease progression. One measure of viral spread is the R 0 , which refers to the average number of secondary infections caused by a primary case and is commonly used to characterize the transmissibility potential of a disease in a completely susceptible population. 2 At the beginning of an epidemic, from a couple of days to 1-2 weeks, "Basic reproduction number" (R 0 ) can be estimated. During an ongoing epidemic for few weeks to months, "Time dependent" reproduction number which is otherwise called as effective reproduction number (R) can be estimated. The effective reproductive number (R) can be used to characterize transmissibility once a certain proportion of the population has been infected and is resistant or immune. 3 Serial interval time is the time between the start of symptoms in the primary patient (infector) and onset of symptoms in the patient receiving that infection from the infector (the infectee). 5 The serial interval is observable, while the generation interval usually is not. Therefore, researchers use serial interval time to calculate the reproduction number. Considering the importance, the paper aims to estimate the effective reproduction number (R) of COVID-19

for India and selected high incidence states over time using incident number of cases that are reported by the government.

The The basic reproductive number is affected by several factors such as the rate of contacts in the host population, the probability of infection being transmitted during contact, and the duration of infectiousness. A population will rarely be totally susceptible to an infection in the real world. Some contacts will be immune, for example developing immunity due to prior infection or as a result of previous immunisation. Therefore, not all contacts will become infected and the average number of secondary cases per infectious case will be lower. This is measured by the effective reproductive rate (R). Essentially, R 0 must be >1 for an epidemic to occur in a susceptible population. If R>1, the number of cases will increase, such as at the start of an epidemic. Where R=1, the disease is endemic, and where R<1 there will be a decline in the number of cases. To successfully eliminate a disease from a population, R needs to be less than 1.

There are various methods to estimate the basic reproduction number and effective reproduction number, such as attack rate (AR), exponential growth (EG), maximum likelihood method (ML), sequential Bayesian method (SB) and the Time dependent method (TD). We used exponential and time dependent method to estimate basic and effective reproduction number respectively.

Walling & Lipsitch explained the relationship between basic reproduction number and exponential growth rate (r), which is defined by per capita change in number of new cases per unit of time. 6 This is defined as,

Where, is basic reproduction number, is exponential growth rate and is generation interval distribution which is defined by the time lag between infection in a primary case and a secondary case but it cannot be observed directly, so it is often substituted with the serial interval distribution that measures time between symptoms onsets. The is derived from the above equation and Provided − exists.

= ,

Where, − is moment generating function of serial interval time T and it is finite for all ∈ − , ; a is a positive constant. 7

Time dependant calculation of reproduction number (R) was proposed by Wallinga and Teunis in 2004. 3 In this method, the effective reproduction number is computed by arithmetic mean over the effective reproductive number of the case for all those cases who show the first symptoms of illness on day t.

Where, is number of cases reported in the time unit and the is !, is probability that the case " infected by the case # and is computed by,

The confidence interval of can be obtained by simulation. The data for India and other states were taken from the crowd sourced database available on https://www.covid19india.org, which is given in the Appendix 1. Additional details about the derivation of these equations are provided in Appendix 2. The reproduction number was estimated using "R0" package in R software version 3.6.2. 5 We assume that the serial interval distribution follows gamma distribution with mean (sd) was 4 (2) which was given by Hwang 

The basic reproduction number (R 0 ) for India was estimated at 1.379 (95% CI: 1.297 as compared to other states. The diagrammatic representation of basic reproduction number, R 0 is presented in Figure 1 .

The average effective reproduction number R over 7 days with 95% CI, for India and high incidence states is presented in Table 1 . The state specific details are presented diagrammatically in Figure 2 

In our study, we estimated the basic and time-dependent reproductive number for COVID-19 in India which is critical for developing and assessing the interventions. Though empirical data driven estimates are ideal, this study has attempted to estimate parameters using daily incidence of COVID-19 cases. It is estimated that at effective reproduction rate of 2.5, 90%

the ongoing pandemic can be controlled if 80% of the contacts can be traced and quarantined or isolated effectively. If the effective reproduction number could be brought down to less than 1.5, higher level of control can be achieved with lower level of contact tracing and; if it reaches 3.5 or more then the trajectory of the epidemic will be very rapid. 9 While estimating reproduction number, it is assumed that the number of secondary infections produced by a single case has no variations. However, several factors such as super-spreading events, changing disease control strategies, their effectiveness and other social interventions result in drastic changes in the transmission pattern in countries and within countries within a short time period. As a result, the effective reproductive number is also constantly modified during the epidemic. We estimated the weekly reproduction number, considering the reported median incubation period of 5 to 6 days. 10 The weekly estimation could be useful to monitor and assess the impact of the virus; develop timely and appropriate strategies; and assess the effectiveness of control measures in India and different states.

Everyday cases support the fact that state wise analyses are essential, as the trend of R changes over time which are different between states. Therefore we were able to study the trend in R in the high burden states. The state specific analyses suggested that it is critical to have state specific mitigation strategies as the transmission dynamics vary between states.

Stratification of data would be able to provide a better estimate on which control measures and duration would work best and provide the best benefit for specific states. Some reasons that might explain the observed differences include climatic factors (e.g., temperature variation and humidity), demographic and biosecurity factors (e.g., presence of filtered farms According to the study, almost all the states indicated a significant change in R during

March 15-31, 2020. In India, R was higher during this time period that could be due to a known cluster of around 5000 cases and many of them travelled from the neighbouring countries. It affected the situation in many other states as well. In Tamil Nadu, another cluster was identified in a city vegetable market that served as an epicentre, which became a source for many infections in the state. Due to this, the R in Tamil Nadu increased significantly in the first week of May 2020.

Regarding the serial time interval, the data are restricted to online reports of confirmed cases and therefore may be biased towards more severe cases in areas with a high-functioning healthcare and public health infrastructure. The rapid isolation of such cases may have prevented longer serial intervals, potentially shifting our estimate downwards compared to serial intervals that might be observed in an uncontrolled epidemic. 11 Second, the identity of each infector and the timing of symptom onset were presumably based on individual recollection of past events. If recall accuracy is impeded by time or trauma, cases may be more likely to attribute infection to recent encounters (short serial intervals) over past encounters (longer serial intervals). Therefore there may be under estimation in the serial time interval estimation. However, in order to adjust for bias we need to do a rapid empirical study to validate our estimation. 7 As limitations, at the state level in our study, it is always difficult to observe the initial cases in any new outbreak or epidemic like COVID-19. This will likely to result in overestimation of initial reproduction numbers due to under reporting or delay in reporting of cases. To address this challenge, we have decided not to consider reported numbers of cases for the first 4 weeks starting from January 30, 2020, when summarizing these data. The estimated CIs using the time-dependent method could be wider because of few observed cases at times.

The study estimated a baseline R 0 of 1.379 for India. It indicated an increasing trend in the R, when there was an eventuality in the states. It also showed that the R was getting stabilised from first week of April and remained at the same level of 1.1 which implies that intensive interventions that include aggressive tracing and testing coupled with appropriate clinical management of the infected are essential to control the transmissibility of the disease.

None of the authors have conflicts of interest to report.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Not required.

Estimation of Time-Dependent Reproduction Numbers for Porcine Reproductive and Respiratory Syndrome across Different Regions and Production Systems of the US

The estimation of the basic reproduction number for infectious diseases

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

Measuring Progress on the Control of Porcine Reproductive and Respiratory Syndrome (PRRS) at a Regional Level: The Minnesota N212 Regional Control Project (Rcp) as a Working Example

The R0 package: a toolbox to estimate reproduction numbers for epidemic outbreaks

How generation intervals shape the relationship between growth rates and reproductive numbers

Moment Generating Function, Expectation and Variance of Ubiquitous Distributions with Applications in Decision Sciences: A Review

Basic and effective reproduction numbers of COVID-19 cases in South Korea excluding Sincheonji cases

Feasibility of controlling COVID-19 outbreaks by isolation of cases and contacts

The Incubation Period of Coronavirus Disease 2019 (COVID-19) From Publicly Reported Confirmed Cases: Estimation and Application

Serial Interval of COVID-19 among Publicly Reported Confirmed Cases. Emerg Infect Dis

Since it can be difficult to distinguish between an actual conflict of interest and a perceived conflict of interest, the Journal requires authors to disclose all and any potential conflicts of interest.

The authors whose names are listed immediately below certify that they have NO affiliations with or involvement in any organization or entity with any financial interest (such as honoraria; educational grants; participation in testimony or patent-licensing arrangements), or non-financial interest (such as personal or professional relationships, affiliations, knowledge or beliefs) in the subject matter or materials discussed in this manuscript.Author names: ________________________________________________________________________________________ ______________________________________________________________________________________________________ ______________________________________________________________________________________________________ ______________________________________________________________________________________________________ Section II The authors whose names are listed immediately below report the following details of affiliation or involvement in an organization or entity with a financial or non-financial interest in the subject matter or materials discussed in this manuscript. Please specify the nature of the conflict on a separate sheet of paper if the space below is inadequate.