Unadjusted time-varying reproduction number (Rt)

Confidence intervals are provided to reflect the uncertainty inherent in these estimates. Users are reminded that the reproduction number is only an estimate. Due to the recent lockdowns and effective public health controls, some states - if not all - have experienced low incidences of COVID-19 cases and even a long sequence of no new daily cases, e.g. Victoria. Note that wider confidence intervals may indicate potential disease extinction, e.g. Victoria (23 days of no new cases as at 22 November 2020). Information is scarce and the accuracy of the reproduction number estimate may be affected. No method has been implemented in the application to adjust for the estimate uncertainty and estimates are only illustrative and for educational purposes. Thank you to covid19data.com.au for the underlying data set which allows for a breakdown by transmission source.

Rt for Australia

Rt by Australian states/territories

Daily cases by Australian states/territories

Adjusted time-varying reproduction number (Rt)

The unadjusted time-varying reproduction number uses the popular EpiEstim method (Cori et al., 2013) which is very useful for providing real-time estimates of transmissibility based on past incidence data. However, a major limitation exists under this method when there are periods of small incidence and/or signs of epidemic extinction. With small or zero cases, estimates become substantially unreliable. The adjusted time-varying reproduction number adopts a novel method, termed EpiFilter (Parag, 2020), which allows us to handle periods of scarce data. Note that this is only a partial implementation of this method and is not the only way to smooth out reproduction numbers. These adjusted time-varying reproduction numbers should only be used for educational and illustrative purposes.

Note that some states have very minimal cases, for example, NT and ACT. Due to the limited data, estimates may not be reliable.

Monitoring the Victorian COVID-19 accommodation program

Current residents in quarantine (as at 11pm)

Returned overseas travellers
Community members
Frontline workers
Voluntary support person

Daily new overseas arrivals

Confirmed cases by LGA

Regional summary of confirmed cases

Daily new cases for selected local government area (LGA)

Reproduction number for selected local government area (LGA)

Projected confirmed cases

Data Downloads

Case incidence data is supplied below and is extracted from covid19data.com.au and contains reported cases by transmission source and Australian state and territory. This data can be downloaded in different formats.

The adjusted time-varying reproduction number (Rt) is provided below. These estimates have been calculated using Parag (2020) method for dealing with low/zero case incidence. These are only estimates and should not be treated as an official source of information for measuring transmissibility in Australia. This data can be downloaded in different formats.

Overview

This is a developmental R Shiny app designed to provide some tools and analyses to monitor the COVID-19 situation in the state of Victoria in Australia. It is important to note that none of the estimates produced in the application are official. Simulations and output produced from this tool is for educational purposes only and should not be used for decision-making.

Collaborations

This R Shiny application is developed and maintained by Hung Vo (Senior Manager and Independent Consultant):

Hung Vo
coeus_thinks
anevaluator

Data is produced and managed by digital journalist and communications consultant Juliette O’Brien (covid19data.com.au):

Juliette O’Brien
juliette.io

Acknowledgements

I would like to acknowledge the extremely useful work of Kris Parag (MRC Skills Development Fellow; Epidemiologist) from Imperial College London. His work on EpiFilter and EpiSmooth (Parag, 2020) has been used to improve Australian state-based estimates of time-varying reproduction numbers which were highly inflated when there were zero-to-low case incidence.

Kris Parag
Kris Parag
kpzoo

Data Sources

Data from covid19data.com.au has been used to report on daily confirmed cases and to calculate the reproduction number. The data is managed and owned by covid19data.com.au and I am grateful for the amount of effort their team has put into such exhaustive data collection. See covid19data.com.au for more information.

Reproduction Number

The basic reproduction number (\(R_{0}\)) is the fundamental epidemiological estimate for measuring the transmissibility of an epidemmic (Heesterbeek & Dietz, 1996). It is generally seen as the number of infections that is caused by a reference cases in a completely susceptible population. The limtation to the \(R_{0}\) estimate is that it does not reflect the time lapsed in an epidemic (Ng & Wen, 2019).

The time-varying reproduction number is estimated in this app to provide a measure of transmissibility over a time series (\(R_{t}\)). This uses the EpiEstim package (Cori, Ferguson, Fraser, & Cauchemez, 2013) to quantify the transmissibility of the COVID-19 outbreak in Victoria using past incidence data.

The estimation of the \(R_{t}\) accounts for the incidence of imported cases. The mean and standard deviation (in days) for the serial interval of COVID-19, by default, is set as 4.7 days and 2.9 days. This is adopted from Nishiura, Linton, & Akhmetzhanov (2020).

Incidence Projections

The Projections tab allows for simulations of future COVID-19 incidence. These require an assumed reproduction number and distribution of the serial interval. For further details about the forecasting process, take a look at Nouvellet et al. (2018).

Incidence Rates

Crude incidence rates for confirmed COVID-19 cases in the population are calculated for simplicity and are not adjusted for any variable(s). The equation for calculating crude incidence rate is described below:

\(Crude Rate = \frac{Confimed Cases}{Population} \cdot 10000\)

Disclaimer

Note that the Doherty Institute (Price et al., 2020) has produced time-varying reproduction number estimates and these appear to be very similar to the time-varying estimates announced previously by the Chief Medical Officer of Australia. The estimates produced are independent of those estimates and are intended to be illustrative. I have no affiliation with Doherty Institute and official information announced by the Australian Government should be referred to for public health information. Please do not use this app for official public health monitoring and information.

Projected cases are only illustrative. No one can precisely predict when an outbreak occurs and as such, projections generated cannot and should not be used for planning purposes.

References

Cori, A., Ferguson, N. M., Fraser, C., & Cauchemez, S. (2013). A new framework and software to estimate time-varying reproduction numbers during epidemics. American Journal of Epidemiology, 178(9), 1505–1512.

Heesterbeek, J., & Dietz, K. (1996). The concept of ro in epidemic theory. Statistica Neerlandica, 50(1), 89–110.

Ng, T.-C., & Wen, T.-H. (2019). Spatially adjusted time-varying reproductive numbers: Understanding the geographical expansion of urban dengue outbreaks. Scientific Reports, 9(1), 1–12.

Nishiura, H., Linton, N. M., & Akhmetzhanov, A. R. (2020). Serial interval of novel coronavirus (covid-19) infections. International Journal of Infectious Diseases.

Nouvellet, P., Cori, A., Garske, T., Blake, I. M., Dorigatti, I., Hinsley, W., … others. (2018). A simple approach to measure transmissibility and forecast incidence. Epidemics, 22, 29–35.

Parag, K. V. (2020). Improved estimation of time-varying reproduction numbers at low case incidence and between epidemic waves. medRxiv.

Price, D., Shearer, F. M., Meehan, M. T., McBryde, E., Moss, R., Golding, N., … McCaw, J. M. (2020). Early analysis of the australian covid-19 epidemic. https://www.doherty.edu.au/uploads/content_doc/COVID_19_early_epidemic_analysis_Doherty.pdf.

About

Change History

9 December 2020

  • Added a Data Downloads tab.

5 December 2020

  • Added disclaimer collapsible boxes.
  • Made tab default to Adjusted Reproduction Number.
  • Added the extraction date for the adjusted reproduction number.

4 December 2020

  • Partially implemented Parag (2020) method on smoothing reproduction numbers.
  • Acknowledgements added.
  • Updated disclaimers and added a cron scheduled task.

28 November 2020

  • Added in the time-varying reproduction number for Australia overall.
  • Added key epidemiological events for the Australian Rt.
  • Minor changes to About page.

27 November 2020

  • Reorganised the menu bar tabs.

22 November 2020

  • Created loop for EpiEstim reproduction number to be estimated for each state.
  • Expanded bar chart to provide daily new cases by state.

18 November 2020

  • Fixed up the reproduction number charts. Issue was NA values were appearing in the raw data.

3 November 2020

  • Reproduction number troubleshooting. Modified API link.
  • Updated the Glyph icons in the menu.
  • Updated the Introduction page.
  • Changed the data source for the LGA map.

26 July 2020

  • Added daily new cases for LGA.

28 June 2020

  • Minor fix to data scraper for LGA-level COVID-19 cases.
  • Minor fix to data scraper from covid19.com.au.

7 June 2020

  • Added crude incidence rate based on currently active cases.
  • Added leaflet map for confirmed COVID-19 cases for local government areas (LGA).

29 May 2020

  • Fixed scraper to find position number of “VIC”.

20 May 2020

  • Added brief information about projections.
  • Renamed the first tab.
  • Added projections tab.

11 May 2020

  • Minor changes to content and plot.

9 May 2020

  • Modified caption for the \(R_{t}\) plot.
  • Minor changes to the bibliography.
  • Added rangeslider for the \(R_{t}\) plot.
  • Added more information to the sidebar for reproduction tab.

6 May 2020

  • Fixed crontab schedule.
  • Fixed non-loading Shiny app issue by coding the new Interstate transmission source.

3 May 2020

  • Added more information to Introduction tab.
  • Added rangeslider for the Victorian daily COVID-19 cases.
  • Time-varying reproduction number now accounts for imported cases.
  • Added estimation parameters for EpiEstim::estimate_R() in the time-varying reproduction number.
  • Added plotly for estimated time-varying reproduction number in Victoria.

2 May 2020

  • Extract/scraper for covid19data.com.au.
  • Add import of John Hopkins University COVID-19 data.
  • Changed glyph icon for the effective reproduction number tab.
  • Added a new tab for monitoring the effective reproduction number \(R_{eff}\).

References

Cori, A., Ferguson, N. M., Fraser, C., & Cauchemez, S. (2013). A new framework and software to estimate time-varying reproduction numbers during epidemics. American Journal of Epidemiology, 178(9), 1505–1512.

Heesterbeek, J., & Dietz, K. (1996). The concept of ro in epidemic theory. Statistica Neerlandica, 50(1), 89–110.

Jenness, S. M., Goodreau, S. M., & Morris, M. (2018). EpiModel: An r package for mathematical modeling of infectious disease over networks. Journal of Statistical Software, 84.

Ng, T.-C., & Wen, T.-H. (2019). Spatially adjusted time-varying reproductive numbers: Understanding the geographical expansion of urban dengue outbreaks. Scientific Reports, 9(1), 1–12.

Nishiura, H., Linton, N. M., & Akhmetzhanov, A. R. (2020). Serial interval of novel coronavirus (covid-19) infections. International Journal of Infectious Diseases.

Nouvellet, P., Cori, A., Garske, T., Blake, I. M., Dorigatti, I., Hinsley, W., … others. (2018). A simple approach to measure transmissibility and forecast incidence. Epidemics, 22, 29–35.

Parag, K. V. (2020). Improved estimation of time-varying reproduction numbers at low case incidence and between epidemic waves. medRxiv.

Price, D., Shearer, F. M., Meehan, M. T., McBryde, E., Moss, R., Golding, N., … McCaw, J. M. (2020). Early analysis of the australian covid-19 epidemic. https://www.doherty.edu.au/uploads/content_doc/COVID_19_early_epidemic_analysis_Doherty.pdf.

Swerdlow, D. L., & Finelli, L. (2020). Preparation for possible sustained transmission of 2019 novel coronavirus: Lessons from previous epidemics. Jama, 323(12), 1129–1130.