Last week the Government released the modelling it has used to inform its pandemic response. They appeared reticent. In the press conferences they were keen to stress its limitations, its reliance on overseas assumptions and so on, as if to say we are not relying on models to decide what to do. They should not have been. Their modelling is sound. It presents a coherent picture of why we are ‘flattening the curve’ and strong justification for what is being done today. It shows the importance of not letting COVID-19 get out of control. And the fact of the matter is that good disease modelling is one of the few reliable tools we have for working out what to do.
Their modelling aligned closely with ours (I would hope so); we predict 32,000 daily ICU admissions at the peak if we do nothing, they predict 35,000. But there is no detailed comparison of the effectiveness of different social distancing measure or how best to relax them. Nor is there any analysis of the economic costs of the measures.
I was struck by the announcement that governments are expanding ICU capacity three-fold to 7,000 beds. It is worth exploring the implications. Assuming patients admitted to ICU need 8 days of treatment, 25% of patients admitted to ICU will die, and that the case fatality rate for COVID-19 with treatment is about 0.5%, it looks like governments are preparing for a peak of around 44,000 new infections a day, much higher than the number today.
I am encouraged by this. It signals that governments are contemplating a managed increase in the spread of the disease. As I said in the AFR last week, in the absence of a vaccine or a cure, the best policy involves a managed increase in spread so that some degree of herd immunity develops, while seeking to protect those most at risk while it does.
We have modelled the health and economic outcomes of different policies to achieve this. One policy that does not work is to try and prevent every death. Why not? Because to do so, we would need to eliminate infections entirely. That can only be done with aggressive and prolonged social distancing at prodigious economic cost (2% of GDP per month for at least 10 months, at least $400bn). Even then as the WHO pointed out last Friday, when you finally relax the measures, there is a high likelihood of a severe second wave. This policy only makes sense if a cure or vaccine is available for the whole community in 6 months’ time.
We examined the gradual removal of social distancing measures from now. Schools reopen by the end of April. Work constraints are relaxed late-May and most people return to work. Other constraints on community social activities such as sports, cinemas and restaurants resume fully in September. We chose this order because, as the table shows, school closures are the costliest means of managing the pandemic, work constraints the next most costly. Quarantine of households for 21 days when there is a confirmed case continues until the end of the year, so we need to test. Isolation of the at-risk including over 70s continues until November.
In this case 64,000 lives are saved, fewer than those saved from prolonged social distancing but only if there is no second wave. The cost to the economy is 6.4% of GDP ($132bn), one third of the cost of retaining strict social distancing until the end of the year. The average cost per life saved is $2.1m. As the figure shows, at no time do we run out of ICU beds.
The second policy we examined is an accelerated removal of social distancing. Schools open and people return to work immediately. Constraints on community activities are removed at the end of June. Household quarantine and isolation of the at-risk are unchanged.
3,000 fewer lives are saved, largely because (as the chart shows) we are short around 500 ICU beds for one month. But the impact on GDP is much smaller at 4.1% ($85bn). The cost per life saved falls to $1.4m. The cost of saving those additional 3,000 lives is $9m each, more than we are normally willing to pay. With the economic savings we could save 6-9,000 extra lives once the pandemic is over rather than 3,000 today.
High compliance rates with quarantine and isolation are essential, which we will only get with effective government and social support for those affected. Without this, other measures will need to stay in place, the pandemic will last longer, cost more, and fewer lives will be saved.
Both scenarios point to a quick relaxation of school closures and workplace constraints once we know that the new ICU beds will be in place.