2022 will be better: COVID-19 Pandemic Tradeoffs modelling
September, 2021 This report may be cited as: Blakely, T., Wilson, T., Andrabi, H., Thompson, T. (2021). “2021 will be better: COVID-19 Pandemic Tradeoffs Modelling.” Population Interventions Unit, Melbourne School of Population and Global Health. Corresponding author: Tony Blakely, [email protected] Ref: https://populationinterventions.science.unimelb.edu.au/pandemic-trade-offs-september-2021/ (This tool allows you to explore how COVID-19 policy responses (restrictions, vaccination roll-out, and border opening) impact future SARS-CoV-2 infection, hospitalisation and mortality rates, and the future probability of lockdowns.) And quoted below: Ref: https://populationinterventions.science.unimelb.edu.au/posts/pandemic-trade-offs-september-2021/2022-will-be-better_COVID-19-Tradeoffs-modelling-21-SEP.pdf Modelling Summary 2022 will be better than 2021. For us to ‘live with the virus’ will take more effort that what many of us assume, but by effectively using the tools we have now and innovating, we can achieve a well-functioning society in 2022. A commonly held view is that we can ‘open up’ at 80% vaccination coverage of adults, in a scenario we call the Default Scenario. In our modelling this is a loose suppression policy designed to limit hospitalisations to a level that our health care system can handle. We expect travel to increase to the point that, on average, after screening, one vaccinated but infected person unwittingly crossing our borders undetected per day. The health loss of this scenario is arguably tolerable, at about 4000 hospitalisations over the year (range 2300 to 7300) in a state the size of Victoria. But – in our COVID-19 Pandemic Tradeoffs modelling at least – this default scenario requires us to spend more than half the year in lockdown. We have to do better than this. In an Upgraded Scenario that extends 80% vaccination coverage to include children (5+ year olds), and keeps moderate public health and social restrictions in place even when case numbers are low (e.g. some density limits in hospitality), we will be ‘okay’. ‘Okay’ under this Upgraded Scenario actually looks pretty good in health loss terms with a range of between 130 to 1800 hospitalisations from COVID-19 over the year, and 36 to 490 deaths. Not to belittle preventable deaths from infectious disease, these base scenario estimates of mortality are about 5% to 50% of the deaths per year from influenza and pneumonia in Victoria. But the flipside of this contained health loss is the social cost to keep the pandemic under control. Even for the Upgraded Scenario we might expect 14% of time is expected in some form of lockdown, with a wide uncertainty range of 0% to 50% of the year in lockdown. These scenarios only show us we can achieve in 2022 without stretching ourselves too much. In fact, we can do better: 1. Increase vaccine coverage to 90%: Achieving 90% vaccination coverage of both children and adults will slash the hospitalisation and death rates by about 80%, and we will most likely have no time in lockdown at all (so long as we keep moderate public health and social measures in place at all times). 2. Reducing overseas/interstate infected incursions: Reducing the expected number of vaccinated but infected arrivals that get into our community undetected from 1 per day to 1 every five days (equivalent to the current risk from 200 vaccinated quarantine-free arrivals per day from the UK in a State the size of Victoria) achieves the same reductions in health loss and time in lockdown as 90% vaccination. These two improvements are for interventions we understand reasonably well. We also need to innovate to reduce our reliance on lockdowns as the main tool to augment high vaccine coverage. Our modelling suggests that improved air filtration and ventilation of buildings (e.g. school rooms and office buildings), higher rates of mask use even when we are not in lockdown, a third dose of an mRNA vaccine to all those double-dosed with AstraZeneca, deployment of mass rapid antigen testing when we need to dampen transmission without resorting to lockdowns, and technological enhancements to contact tracing (e.g. Bluetooth enable apps that both work and satisfy privacy concerns) can all have important impacts – reducing health loss, and reducing the need for lockdowns even more. It is critical to note that it is not the vaccination coverage alone that determines what opening up and 2022 will be like. Rather, it is the full package of measures – of which vaccination coverage is just one. Public and policy discourse should reflect this reality. To achieve a better way of living in 2022, we also need to watch out for a few things. There is convincing evidence emerging of substantial waning vaccine immunity for both AstraZeneca and Pfizer to the Delta virus. We first need to complete the job of vaccinating the global population. This is important for equity, and also because it reduces the chance of a dangerous new variant emerging. But when we can, we will need to roll out third or booster vaccines to everyone. Especially and first to recipients of AstraZeneca. Assuming and hoping a more infectious, lethal and vaccine resistant variant of the virus does not emerge, we should be optimistic that 2022 will be substantially better than both 2020 and 2021. We have choices as to what mix of measures we use to chart our way to and through next year, including known interventions (vaccines, border controls, suppression policies within country) and innovations we can see coming (ventilation, mass rapid testing). This report covers 432 possible scenarios, each modelled 100 times in an agent-based model to capture as many futures as possible. All results are publicly available at an interactive webtool, COVID-19 Pandemic Tradeoffs (www.pandemictradeoffs.com). Our modelling finds that predictions are sensitive to two important and poorly understood input parameters. First, the proportion of Delta infections that are asymptomatic. If in our modelling we use the estimates used in the Doherty-led report, the situation deteriorates. Second, there is genuine uncertainty about the effectiveness of current vaccines at reducing onwards transmission if a vaccinated person is unlucky enough to become infected. In our model we assume this reduction is 25% on average. If we replace this with the 65% reduction assumed in the Doherty-led modelling, the situation improves dramatically. However, we fear that the 65% reduction assumed in the Doherty-led report – based on evidence accruing since their modelling – is too optimistic. Pulling back, we all need to be cautious about the sensitivity of modelling predictions to inputs we do not yet fully understand. We need to use modelling to plan our opening up, then nimbly alter how we open up as actual data arrives in real-time. Our modelling supports a key finding in the Doherty-led Report that keeping ‘light restrictions’ as a minimum at all times dramatically reduces the need for lockdowns. We concur that, unfortunately, allowing society to go back to near normal settings when case numbers are low often allows transmission to gain hold, and requiring longer lockdowns to bring surges back under control. As 2022 progresses, and we move into 2023, we can probably ease these minimal restrictions as immunity from natural infection creeps up and we revaccinate the whole population with better vaccines that (hopefully) reduce transmission risk more than current versions. Our modelling also extends on the Doherty-led modelling in important and policy-relevant ways. For example, we do not start from a baseline of 30 infected cases, but account for ongoing community transmission as Australia is experiencing now, and how case numbers respond dynamically to restrictions and other measures. The time-window of our work also extends beyond 6 months to the end of 2022, including the first year after opening up. That is our modelling attempts to represent the patterns of infection growth and decline we are likely to experience from now through 2022. Melbourne School of Population and Global Health September 2021 _______________________________________________ Link mailing list [email protected] https://mailman.anu.edu.au/mailman/listinfo/link
