Why eliminating Covid-19 is critical: Since more waves of infection, hospitalisations and deaths are inevitable for other countries until a vaccine arrives

Posted on by

Prof Rod Jackson (Epidemiologist, Auckland University)

There are really only three ways that a population can beat Covid-19: (i) eliminate the virus and keep infected people out (the approach NZ is taking); (ii) allow a sufficient proportion of the population to get infected and become immune to re-infection so there are no longer enough non-immune people to allow the virus to spread (this is called herd immunity and is estimated to be at least half of any population for Covid-19); or (iii) develop an effective vaccine, which is the ideal way of achieving herd immunity. Given a vaccine might still be a long way off, NZ’s elimination strategy (if successful) is likely to avoid high numbers of deaths and the need for long periods of restrictions to suppress or mitigate the spread of Covid-19.

One of the key pieces of missing information on Covid-19 that will determine the likelihood and implications of each of these strategies is the proportion of people who are currently infected or have been infected and have recovered. These are the potentially immune people. However, it has been extremely difficult to determine because an unknown proportion of people have minimal or no obvious symptoms and so are not detected. Also, there are still concerns about the quality of the antibody tests to determine who has already been infected. Further, the number of people who are infected is a rapidly moving target, so multiple population surveys would be required.

An alternative way to estimate the total number of infected people at any point in time is to use the Infection Fatality Ratio (IFR). This is simply the number of deaths that occur among, say 100 infected people. So, if the IFR was 1 in 100, then there would be one death among 100 infected people.

But calculating the IFR for Covid-19 is very difficult because it requires every infected person, whether they have symptoms or not, to be found. A number of researchers have tried to make this calculation using information from a range of sources, including the quarantined Diamond Princess cruise ship, which provided in essence a ‘captive population’ (and where the IFR was 1.8% on 18 April, albeit an older population).

While estimates of the IFR vary widely, one value can help us predict what is likely to happen in the next year or so, if countries don’t instigate an ongoing strategy of intensive case finding, contact tracing, quarantine, and physical distancing.

If the true IFR value is around 1 in 100 (rounding up from Iceland where 11% of the whole population has been tested and the IFR was 0.5% on 17 April), that means that multiplying the current number of deaths in a country by 100 will give you an estimate of the number of people who had been infected. This is however the IFR of several weeks ago, because it typically takes around 2-4 weeks for infected people to die from the time they were infected. This calculation works only if the people who have died are typical of all people who die from Covid-19, so a country would probably need at least 100 deaths for the calculation to be meaningful. If this IFR is correct, consult the Worldometers Covid-19 website and compare the total confirmed cases per 1 million people with the total deaths per 1 million people for some of the worst effected countries to date. There have been approximately 400 deaths per million people in Spain, which if the IFR is 1 in 100, means there would have been approximately 40,000 infected people per million people (or 4% of the population) several weeks ago, yet the reported cases are under 4,000 per million people, a tenth of the likely number of true cases. This 4% is very far from the 50% plus of the population needed to achieve herd immunity in Spain.

Similarly, in the US the 100 deaths per million would translate into 10,000 infected people per million (or about 1% of the population), but only about 2,000 cases per million have been reported. Again, nearly all of the US population is likely to still be susceptible to further outbreaks.

These calculations have several profound implications.

  1. Herd immunity is likely to be years away in these countries with major epidemics unless we develop an effective vaccine or antiviral or are willing to endure multiple further catastrophic waves of infections, hospitalisations and deaths.
  2. The only reason some severely infected countries appear to be over a recent peak in the epidemic curve is because of the major restrictions they have put in place.
  3. Ongoing vigorous case finding, intensive contact tracing and compulsory quarantine, along with enforced physical distancing (various levels of lock-down) will become the norm in most countries until a vaccine is developed or the virus can be eliminated.

Even if the estimated IFR is overestimated by a factor of 5 and the true ratio is 1 in 500, these implications remain valid, although there would be fewer new waves of infection and herd immunity would be achieved earlier but still at huge further cost to life. If the IFR is underestimated by a factor of 5 and the true ratio is 1 in 20, the implications will be considerably worse than predicted here. Moreover, in the absence of any evidence that infection guarantees subsequent immunity, any scenario is worse than it looks (see Kissler et al).

Fortunately, New Zealand has chosen an elimination strategy which (if successful) means that the country will not need long periods of disease suppression or mitigation measures. That is, after elimination is achieved, NZ will be able to open up its economy to some level of “new normal” – though, with ongoing tight border controls and quarantine, this will still not allow international tourism to be re-established. Then once a vaccine or new anti-viral treatments are available – these border control measures could be reduced.

There are also signs that Australia is also considering elimination – with the Australian Prime Minister recently articulating “suppression/elimination” as the strategy. If Australia does achieve elimination – then the travel and trade between Australia and NZ could be further opened up. Another encouraging sign for the elimination approach in island jurisdictions is that Taiwan has successfully reached zero cases on some days of the last week (see Worldometers data). It also used rigorous border control with quarantine from early on in the Covid-19 pandemic.

This entry was posted in Uncategorized and tagged , , , by Nick Wilson. Bookmark the permalink.

15 thoughts on “Why eliminating Covid-19 is critical: Since more waves of infection, hospitalisations and deaths are inevitable for other countries until a vaccine arrives

  1. Taiwan “also used rigorous border control with quarantine from early on in the Covid-19 pandemic”. NZ did not do this not until two weeks into lockdown. Inexcusable

    • So where exactly would you propose we put the approx 40,000 arrivals coming into NZ from the 16th of March onward , the vast majority were NZ citizens returning home. Perhaps evacuated the whole of Whanganui (population 39k) and turned it into a colony ?
      If your going to say that “well we should be keeping legal citizens out then” that gets very dystopian very quickly.

  2. In reading this, I would imagine that another two weeks would be the correct decision.
    But we will see, I believe that Jacinda is the weak link and will cave under the pressure of a nation.
    This article also makes me think we have tackled this wrong, we had the opportunity and ability to close borders much earlier and failed to do so.
    Mandatory quarantine for 40 or 50 thousand people would have been by far and away less expensive than the current spending Jacinda has endorsed.
    She said it could not have been done but should have staggered it like we are presently directed to avoid queues.
    Too late now but with everyone waxing lyrical about her, in her own words some serious questions need to be answered.

  3. Don’t you need to age adjust the IFR across countries to control for the mix of age groups. I have not seen the age mix of the cruise ship but I assume it’s pretty old. Also it seems there would be a lot of error to apply IFR across different living densities ( ie seems more contagious in places like big cities and cruise ships?)

    • You raise good points especially about density ie Italy and New York being hardest hit illustrates that. I think IFR wouldnt be changed by that much as disease isnt any more or less deadly if you are less densely populated but it would make a big difference to R0 which as you mention is a lot higher in dense populations. This indeed is a point that the author does not seem to take into account

    • Hi – I have based the IFR estimate largely on high income countries and made the assumption that they have similar age structures. You are correct that it varies a lot by age structure but I felt that trying to age-standardise was probably not worthwhile given the uncertainty in the estimates

  4. In today’s Herald Derek Cheng states: “Back in New Zealand, Auckland University epidemiologist Professor Rod Jackson estimates there are 500 undetected but infectious cases of Covid-19 in New Zealand, which he says is reason enough to extend level 4 because physical distancing was the best way to manage them”. But there is no such comment in Rod’s blog. In fact, if the IFR is 1 in 100, and there have been 12 COVID-19 deaths in NZ, we would expect 1200 cases, and over 1400 have been found. Moreover, even if 50% of infections are asymptomatic, and given the wide availability of testing, it seems unlikely that none of these cases have turned up.

    • Here is the 500 missing cases argument.

      Are there still 500 undetected infectious cases of Covid-19 in New Zealand today?

      Rod Jackson
      Professor of epidemiology
      University of Auckland

      As stated in my blog, based on current expert reviews, it seems likely that in western countries, there will ultimately be about one death caused by Covid-19 for every 100 people infected. Using this estimate of the Infection Fatality Ratio (i.e. 1 in 100), most western countries with similar or higher testing rates than New Zealand, have only detected at most 1 out of 3 cases. This can be easily calculated by multiplying the number of reported deaths by 100 and comparing the result with the reported number of confirmed and probable cases. If deaths are under-reported, which is believed to be common, the true proportion of undetected cases will be even greater. In addition, as death usually occurs 2-3 weeks after infection, the calculation estimates the number of cases 2-3 weeks before the reported results, which will also increase the proportion of undetected cases.

      Although there are fortunately too few deaths in New Zealand yet to use this method to estimate the number of undetected cases (at least 100 deaths are probably needed to make the calculation meaningful) and although we have begun to extend our testing out to individuals who have not been selected because they have symptoms, there is, nonetheless, no reason to believe that we are doing a better job at detecting cases than the best of other countries.

      Even if we are only missing about 1 out of 2 cases, this would mean that in addition to the 582 confirmed active cases today (total minus recovered cases), there would be at least another 582 currently undetected but infectious cases in New Zealand. Some active cases may no longer be infectious, so maybe there are 500 undetected but infectious cases.

      Trying to find 500 (about 1 in every 10,000 New Zealanders) currently undetected cases, will be like searching for a needle in a haystack, particularly if many are asymptomatic. Unfortunately, each of them could infect at least 2 other people unless we maintain effective physical distancing.

      These estimates are based on a snapshot and don’t account for the likelihood of rapidly diminishing case numbers, but they do assume we are doing better than any other country at case finding. Please critique these calculations and post a comment.

      Whatever this number is, if it is greater than zero, we need to continue to scale up testing, case finding and rapid contact tracing along with supervised isolation, to ensure elimination before we revert to level 3 or lower. Then we will be able to return quickly to ‘normal’ albeit still inside a New Zealand bubble.

      • Hello Rod, today Dr Bloomfield and PM said in the press conference that a patient who tested positive in Northland yesterday was infected in Auckland in March before the lockdown. They were being almost jovial discussing this case.
        But isnt that an incredibly concerning point that someone can be infectious for what seems like possibly a month maybe more? As they would of had to of done a PCR test so therefore this person had potential to spread the disease so many weeks later.
        Thank you

  5. “But we will see, I believe that Jacinda is the weak link and will cave under the pressure of a nation.”

    We’ll see about this! She’s shown no sign of caving to date. I for one am betting she won’t.

  6. If Pharmac’s QALYs (quality adjusted life years) are priced at about $40,000 and an “elimination of COVID-19 in NZ” QALY is priced at say, $1,000,000, is elimination still a good strategy? Or is there some reason we can’t compare QALYs in relation to drug purchase with QALYs in relation to virus suppression?

    • Hi Andrew – the price per QALY makes sense for informing some policy decisions, but not others. As The costs and benefits of NZ’s short sharp strategy of elimination, would have to be compared with a very long term suppression only strategy. The economics of this are beyond me.

  7. Hi, am not health epidemiologist or health economist but… I’m glad we haven’t out loud used QALYs on this one. The only health economics principle that has come up for me is generally healthy population comes first. Easier to have good looking economy when you have a populace who are healthy, well, happy as opposed to, ill, grieving and fearful.

    • Hi Annie. Sorry for the belated reply, I hadn’t looked at the blog for a while. I completely agree with you. Also, after a long debate with some colleagues about cost-benefit analyses of different covid-19 strategies, I don’t believe this approach is appropriate for addressing a worldwide catastrophe. One of my colleagues suggested that it’s kind of like being on ship and coming across a boatload of refugees who are about to sink and calling on the ship’s health economist to do a cost-benefit analysis before decided whether to save them or not.

Leave a Reply

Your email address will not be published. Required fields are marked *

* *