A modelling approach to estimate the transmissibility of SARS-CoV-2 during periods of high, low, and zero case incidence

  1. Nick Golding
  2. David J Price
  3. Gerard Ryan
  4. Jodie McVernon
  5. James M McCaw
  6. Freya M Shearer

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    This paper is interesting, timely and important because it presents a way to understand the transmission potential of a virus even when there are very few local cases. This has a high public health communication and preparedness value. The paper is clearly written, and the results fit with the known epidemiology of the various outbreaks that occurred in Australia in 2020. The paper is likely to be of broad interest within and outside the field of epidemiological modelling.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 and Reviewer #2 agreed to share their name with the authors.)

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Against a backdrop of widespread global transmission, a number of countries have successfully brought large outbreaks of COVID-19 under control and maintained near-elimination status. A key element of epidemic response is the tracking of disease transmissibility in near real-time. During major outbreaks, the effective reproduction number can be estimated from a time-series of case, hospitalisation or death counts. In low or zero incidence settings, knowing the potential for the virus to spread is a response priority. Absence of case data means that this potential cannot be estimated directly. We present a semi-mechanistic modelling framework that draws on time-series of both behavioural data and case data (when disease activity is present) to estimate the transmissibility of SARS-CoV-2 from periods of high to low – or zero – case incidence, with a coherent transition in interpretation across the changing epidemiological situations. Of note, during periods of epidemic activity, our analysis recovers the effective reproduction number, while during periods of low – or zero – case incidence, it provides an estimate of transmission risk. This enables tracking and planning of progress towards the control of large outbreaks, maintenance of virus suppression, and monitoring the risk posed by re-introduction of the virus. We demonstrate the value of our methods by reporting on their use throughout 2020 in Australia, where they have become a central component of the national COVID-19 response.

Article activity feed

  1. Version published to 10.7554/elife.78089 on eLife
  2. eLife

    Author Response

    Reviewer #2 (Public Review):

    In regions that implement an elimination strategy prolonged periods of no local transmission mean that there is no data available to estimate Reff using the currently available methods. Transmission rates from travellers to community members, and between community members, are different when border restrictions occur, as is frequently the case when implementing an elimination strategy. When cases are low and importation risk is high, a reasonable estimation method must acknowledge this transmission heterogeneity, for example, as shown in equations 5-8 and 10-11 of this paper.

    The calculation of transmission potential adds significant data requirements (summarized in Figure 1), such that some regions where the methodology would be valuable may lack the data to estimate the macro- and micro-distancing parameters. In the paper, such parameters are estimated from weekly surveys performed by market research groups and the University of Melbourne. In contrast, using existing methods in regions where local spread does occur, Reff can be calculated and generate reasonable insight with relatively little data. Due to the additional data requirements, the calculation of transmission potential is less accessible than some current approaches to calculate Reff in regions with local spread.

    We agree with these comments about the need for behavioural data. We believe this point is made clearly in our existing discussion text, copied below:

    Despite its demonstrated impact, there are limitations to our approach. Firstly, it relies on data from frequent, population-wide surveys. In Australia, these data are collected for government and made available to our analysis team by a market research company which has access to an established “panel” of individuals who have agreed to take part in surveys of public opinion. Researchers and governments in many other countries have used such companies for rapid data collection to support pandemic response [23, 25]. However, these survey platforms are not readily available in all settings.

    We also believe it is clear throughout the manuscript that transmission potential provides complementary information to Reff, and unlike Reff can be calculated in the absence of transmission.

    The authors describe "macro-distancing": the rate of non-household contacts; and "micro-distancing": the transmission probability per non-household contact. This terminology "micro-distancing" gives the false impression that transmission probability depends solely on distance. In the paper, transmission probability is estimated from survey responses to the question 'are you staying 1.5m away from people who are not members of your household?'. This data is limited to estimate the transmission probability and overlooks the impact of mask use, improved ventilation, and meeting outdoors (all non-distance-based approaches). The paper mentions that self-reported hand hygiene could be used to estimate micro-distancing. COVID-19 spreads through airborne transmission, but the paper gives no mention of ventilation or mask-wearing.

    We agree with these important points and have adjusted the terminology for micro-distancing behaviour to improve clarity. We now refer to it as “precautionary micro-behaviour” since adherence to the 1.5 metre rule is used as a proxy/indicator for change over time in all behaviours that influence transmission (other than those reducing the number of contacts). This includes behaviours such as mask-wearing, preference for outdoor gatherings, hand hygiene etc .

    In addition to changing the terminology for this metric throughout the manuscript, we have added the following explanation to the “Model” section of the manuscript (lines 100-105):

    The modelling framework uses adherence to the 1.5 metre rule as a proxy for all behaviours (other than those reducing the number of contacts) that may influence transmission, and so is intended to capture the use of masks, preference for outdoor gatherings, and hand hygiene, among other factors. The 1.5 metre rule was a suitable proxy because it was consistent public health advice throughout the analysis period and time-series data were available to track adherence to this metric over time.

    Some of the writing lacks precision around the descriptions of Reff. Notably, Reff is not a rate because it does not have units 'per time'. There is a lack of clarity that Reff is infections generated over an individual's entire infectious period. Other metrics of outbreak growth are rates, for example, an exponential growth rate parameter. This lack of clarity in the writing does not impact the methodology.

    Thank you for pointing out this lack of clarity, we have removed references to Reff as a ‘rate’ throughout. We have added to our initial definition of Reff (lines 29-32) that the infections cover the entire infectious period:

    A key element of epidemic response is the close monitoring of the speed of disease spread, via estimation of the effective reproduction number (Reff) — the average number of new infections caused by an infected individual over their entire infectious period, in the presence of public health interventions and where no assumption of 100% susceptibility is made.

    In the paper, model parameters are estimated from multiple independent data sources using carefully derived inference models that include complex considerations such as right-censoring of reported cases. While data availability may be a limitation to calculating the transmission potential, the modelling and statistics in the paper are rigorous, and calculation of the transmission potential fills a gap by allowing regions that implement elimination strategies to estimate a quantity similar to Reff.

    We thank the reviewer for their positive feedback.

  3. eLife

    Evaluation Summary:

    This paper is interesting, timely and important because it presents a way to understand the transmission potential of a virus even when there are very few local cases. This has a high public health communication and preparedness value. The paper is clearly written, and the results fit with the known epidemiology of the various outbreaks that occurred in Australia in 2020. The paper is likely to be of broad interest within and outside the field of epidemiological modelling.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 and Reviewer #2 agreed to share their name with the authors.)

  4. eLife

    Reviewer #1 (Public Review):

    This paper presents a model for estimating the transmission potential of SARS-CoV-2 that can be applied during periods of low or zero case incidence, as well as during periods of sustained high incidence. This novel approach complements and generalises approaches to inferring the time-varying reproduction number from time series of new daily cases by incorporating additional data streams (Google mobility data and social survey data) that are independent of epidemic dynamics and testing. The paper is likely to be of high interest within the field of epidemiological modelling and more broadly.

    The authors have successfully developed and deployed a robust methodological framework that uses a range of data sources to estimate the potential for the spread of SARS-CoV-2. The results can be and were used in real-time to inform situational awareness and policy response, particularly in countries where incidence was low or zero for periods of time. The results are also informative in retrospectively understanding the epidemiological characteristics of different outbreaks and evaluating the effect of interventions. The method could be used by other groups for situational awareness in other countries - the epidemiological data and mobility data the model uses are relatively standardised although some work would be needed to adapt/implement the survey part of the study. The method could potentially also be adapted for other pathogens that spread via close contact.

    The main limitations of the model lie in the types of data that are routinely available and whether these will continue to be available in a comparable form in the future. These limitations are discussed in more detail in the paper. Application of the model in other jurisdictions would of course need careful re-calibration as there are likely to be differences in e.g. testing rates, idiosyncrasies of mobility data, and lack of or different survey data. And use of model results for policy advice would rely on careful and appropriate communication to policymakers of model results, the associated uncertainty and model limitations.

  5. eLife

    Reviewer #2 (Public Review):

    In regions that implement an elimination strategy prolonged periods of no local transmission mean that there is no data available to estimate Reff using the currently available methods. Transmission rates from travellers to community members, and between community members, are different when border restrictions occur, as is frequently the case when implementing an elimination strategy. When cases are low and importation risk is high, a reasonable estimation method must acknowledge this transmission heterogeneity, for example, as shown in equations 5-8 and 10-11 of this paper.

    The calculation of transmission potential adds significant data requirements (summarized in Figure 1), such that some regions where the methodology would be valuable may lack the data to estimate the macro- and micro-distancing parameters. In the paper, such parameters are estimated from weekly surveys performed by market research groups and the University of Melbourne. In contrast, using existing methods in regions where local spread does occur, Reff can be calculated and generate reasonable insight with relatively little data. Due to the additional data requirements, the calculation of transmission potential is less accessible than some current approaches to calculate Reff in regions with local spread.

    The authors describe "macro-distancing": the rate of non-household contacts; and "micro-distancing": the transmission probability per non-household contact. This terminology "micro-distancing" gives the false impression that transmission probability depends solely on distance. In the paper, transmission probability is estimated from survey responses to the question 'are you staying 1.5m away from people who are not members of your household?'. This data is limited to estimate the transmission probability and overlooks the impact of mask use, improved ventilation, and meeting outdoors (all non-distance-based approaches). The paper mentions that self-reported hand hygiene could be used to estimate micro-distancing. COVID-19 spreads through airborne transmission, but the paper gives no mention of ventilation or mask-wearing.

    Some of the writing lacks precision around the descriptions of Reff. Notably, Reff is not a rate because it does not have units 'per time'. There is a lack of clarity that Reff is infections generated over an individual's entire infectious period. Other metrics of outbreak growth are rates, for example, an exponential growth rate parameter. This lack of clarity in the writing does not impact the methodology.

    In the paper, model parameters are estimated from multiple independent data sources using carefully derived inference models that include complex considerations such as right-censoring of reported cases. While data availability may be a limitation to calculating the transmission potential, the modelling and statistics in the paper are rigorous, and calculation of the transmission potential fills a gap by allowing regions that implement elimination strategies to estimate a quantity similar to Reff.

  6. ScreenIT

    SciScore for 10.1101/2021.11.28.21264509: (What is this?)

    Please note, not all rigor criteria are appropriate for all manuscripts.

    Table 1: Rigor

    NIH rigor criteria are not applicable to paper type.

    Table 2: Resources

    No key resources detected.

    Results from OddPub: Thank you for sharing your code and data.

    Results from LimitationRecognizer: We detected the following sentences addressing limitations in the study:
    Despite its demonstrated impact, there are limitations to our approach. Firstly, it relies on data from frequent, population-wide surveys. In Australia, these data are collected for government and made available to our analysis team by a market research company which has access to an established “panel” of individuals who have agreed to take part in surveys of public opinion. Researchers and governments in many other countries have used such companies for rapid data collection to support pandemic response [23, 25]. However, these survey platforms are not readily available in all settings. Further, the sampling strategy did not allow for surveying individuals without internet access, low literacy or limited English language skills, or communication or cognitive difficulties. Further, individuals under 18 years of age were not represented in our surveys. Nor were these survey results available for the pre-pandemic period, limiting our ability to estimate what a true behavioural baseline would be for the Australian population. While the patterns of TP, Reff and C2 observed over time in Australia are consistent with “in field” epidemiological assessments, and while the methods have demonstrated impact in supporting decision making, a direct quantification of the validity of the TP is not straightforward. For example, whether self-reported adherence to the 1.5 m rule is a reliable covariate for change in the per contact probability of transmission over time is difficult to assess....

    Results from TrialIdentifier: No clinical trial numbers were referenced.

    Results from Barzooka: We did not find any issues relating to the usage of bar graphs.

    Results from JetFighter: We did not find any issues relating to colormaps.

    Results from rtransparent:
    • Thank you for including a conflict of interest statement. Authors are encouraged to include this statement when submitting to a journal.
    • Thank you for including a funding statement. Authors are encouraged to include this statement when submitting to a journal.
    • No protocol registration statement was detected.

    Results from scite Reference Check: We found no unreliable references.

    About SciScore

    SciScore is an automated tool that is designed to assist expert reviewers by finding and presenting formulaic information scattered throughout a paper in a standard, easy to digest format. SciScore checks for the presence and correctness of RRIDs (research resource identifiers), and for rigor criteria such as sex and investigator blinding. For details on the theoretical underpinning of rigor criteria and the tools shown here, including references cited, please follow this link.

  7. Version published to 10.1101/2021.11.28.21264509v1 on medRxiv