Virus strategy must change to protect 90 per cent majority

In over 15 years of writing opinion pieces in newspapers, very few have had such a positive response as this AFR column today proposing that the Prime Minister pivot his current virus response paradigm (including almost 1,000 immediate likes on LinkedIn). It is now enclosed for your benefit below:

While Australia has outperformed most of its global peers in terms of the growth rate of new infections and fatalities, the national virus containment strategy needs to adapt fast.

What we do not need is a national lock-down for six months, or any suggestions of such. Many businesses are confused by the current narrative, which implies that they need to go into hibernation for half a year. On the information available to us, that is totally wrong.

The prime minister has done an excellent job of quickly closing the borders and conservatively managing community expectations. But a six-month hibernation paradigm will dramatically accelerate Australia’s move towards outright depression and mass unemployment, which is where we are heading.

This will have vastly greater economic and human costs than the demonstrably superior strategy of a short-term lockdown followed by a modified “new normal” that only quarantines the minority of vulnerable Australians over the age of 70 that represent less than 10 per cent of our population.

In shaping the optimal policy response, we need to reflect hard on the current data. The first thing we know is that the death rates are significantly overestimated for several reasons: we are only testing a tiny share of the total population; we are not testing the numerous asymptomatic and mild cases; those that are tested are biased towards being seriously ill; and, finally, we are only testing for whether an individual currently has the virus as opposed to whether they have been previously infected.

There are two different tests we can carry out. The first, called a PCR test, evaluates whether the virus is currently in our bodies. A second, antibody test, assesses whether we have had a past infection that make us immune to it. We are doing the first, not second, test currently, which means we are missing all the recovered individuals and hence materially underestimating the population-wide infection rates. The consequence of this is that we are overestimating fatality rates because of the massively underestimated denominator pertaining to infections.

In the one country with a large outbreak, expansive PCR testing, and a decent history of fatalities, we know that for anyone under the age of 50 the death rate is comparable to the flu at circa 0.1 per cent. And this is, again, an upwardly biased estimate of the true fatality rate. Conversely, for anyone over the age of 70 or 80, the South Korean data tells indicates that fatality rates are much more serious at 6.5 per cent (70-79) and 16.2 per cent (above 80), respectively.

We also know that within the next few months we are likely to get access to mass-market anti-viral drugs, such as hydroxychloroquine and remdesivir, that nascent empirical data suggests could further reduce fatality rates. These drugs are already being used on an experimental basis in hospitals throughout the world. And speaking to scientists who are currently working on coronavirus vaccines, they are emphatic that effective vaccines will be available within a year.

What this tells us that we do not need to shutdown, or hibernate, the entire economy for three or six months to combat the virus. Yes, we need to permanently change human behaviours in the immediate term by quickly embracing a hard lockdown for, say, one month. This is not because the virus is a problem for most Australians. It is about radically reducing the transmission rate, or so-called “R0”, from two to three people per infection to ideally less than one person per infection, which would in theory kill-off the pathogen.

While it is unlikely we will achieve this aim, putting downward pressure on the transmission rate slows the number of infections (ie, flattens the curve) and disperses pressure on our health system over time to avoid overwhelming it.

Interestingly, our real-time COVID-19 tracking systems demonstrate that Australia is outperforming its North Atlantic peers both in terms of the growth of new infections, and, more importantly, in respect of the increase in fatalities.

Once human behaviours permanently change through the much-needed shock of a hard lockdown (eg, we learn not to congregate in high densities, shake hands, and so on), most of Australia needs to quickly get back to work via a “new normal” that ruthlessly mitigates transmission.

The new normal must involve extremely aggressive testing for both the virus and anti-bodies to identify who is immune. One of the problems with existing testing for the presence of the virus has been delays in obtaining results. The US has, however, addressed this challenge through a new 15 minute test.

If Australia can get to Germany’s current pace of circa 500,000 tests per week, we could regularly evaluate most of labour force every three months. This should get easier over time as immunity rates increase, reducing the need for re-testing.

The new normal should also encompass an ongoing lock-down, or quarantine, of anyone over the age of, say, 70, who should be strongly encouraged to avoid human contact until we secure tractable anti-viral drugs and/or a cheap, mass-market vaccines.

The alternative to this approach is an economic depression that could result in one-quarter of working age Australians losing their livelihoods. This will have much greater adverse economic, health and welfare consequences than a short lockdown followed by a new normal that protects vulnerable persons on a targeted basis until the virus is eliminated.

My team has constructed real-time forecasting models for the peak in new virus cases that condition on the intensity of the containment strategy each nation adopts. On the basis of the currently available information, the peak in America's new infection numbers should be realised between April 9 and April 18 assuming it is 50 per cent as effective at containment as South Korea and 75 per cent as effective as Italy. In a more gloomy scenario where the US is only half as effective as Italy, the peak new infection numbers materialise around the end of April.

Applying the same assumptions to Australia, we find that the peak in new infection numbers is likely to be between early to mid April. If we do a really poor job at containment, this is deferred to around April 20.

It is important that Prime Minister Scott Morrison and Treasurer Josh Frydenberg quickly embrace the short containment strategy proposed here to minimise the permanent damage that is being done to the economy. Numerous small businesses have advised me they are shutting down for half a year because they have been told that is what they should do. This is madness, and throws 90 per cent of the population under a bus to protect a small minority.