Dödligt virus från Kina orsakar panik (SARS-CoV-2, COVID-19)

Professor Ferguson and the Imperial College team, in a highly influential paper of 16 March 2020, argued that without any policy changes “we predict approximately 510,000 deaths in Great Britain” (p.7). They also suggested that the peak of mortality would take place after approximately 3 months

We note initially that dire predictions of mortality in Sweden were made in the earlier stages of the pandemic. For example, using a model applied to Sweden based on the Imperial one, on 15 April 2020, Gardner and colleagues stated that the model predictions indicated that “at least 96,000 deaths would occur by 1 July without mitigation. Current policies reduce this number by approximately 15%” (p.4).

As noted above, the peak level in EW is twice that of Sweden, but the number of deaths falls
away more sharply in EW. The trajectory of the decline after the peak is shown more clearly in Figure 2, which also llustrates the very similar nature of the rises to the peak. This plots the number of deaths, with, in each country, the data indexed so that the peak level (on 8 April in each case) is equal to 100.

We calculate the cumulative value of the index in each country from 8 April, the peak value, onwards. There are some initial fluctuations in both countries, but by 29 April the cumulative value from 8 April is equalised in the two countries. From 30 April until 7 August, the cumulative value of the index is distinctly higher in Sweden than it is in EW. In terms of index points, the difference is 13.16. In the EW data, an index value of 1 is equivalent to 1349 deaths, the number which occurred in EW on 8 April. We therefore estimate that the number of lives saved in EW by the lockdown is 1349 multiplied by 13.16, which is 17,751.

Compared to the dramatic predictions about death levels if lockdown were not imposed, this number is very low. However, many of the standard models used in epidemiology do not take account of the fact that, faced by a pandemic, people alter their behaviour of their own accord. They do not wait for the decrees of the authorities.

World history is replete with examples. The Black Death in the 14th century seems to have been the most deadly virus in recorded history. As much as 50 per cent of the population of Britain may have died. People adopted many different strategies to try and avoid the disease, some successful, others not. Giving lockdown a legal basis merely reinforces the behavioural adjustments which many people make when faced with a pandemic.

We note, finally, that a larger proportion of deaths took place in care homes in Sweden than in England and Wales. It could therefore be argued that even our estimate of 17,700 is perhaps on the high side.

While considerable uncertainties remain with the epidemiology, there is no doubt that the positive feedback nature of the Covid-19 epidemic will make controlling the outbreak present in May 2020 a very difficult task, especially if the main regulating tool selected is the co-ordinated behaviour of 67 million people. The likely economic effect is analysed of the years-long process to move out of lockdown sufficiently slowly so as not to cause excessive strain on the health services. In the Extended Base Case, annual GDP will fall by 23.5% in 2020 and is not likely to recover to pre-lockdown levels until 2024. Applying the J-value-derived Bristol Curve of population-average life-expectancy versus GDP suggests that the strategy of restriction followed by gradual relaxation is likely to result in a net cost, in terms of average human lives lost, that will be comparable with the UK's sacrifice over the six years of World War Two. A policy of lockdown followed by gradual relaxation is likely to do much more harm to the nation's health than good.

Professor Ferguson and the Imperial College team, in a highly influential paper of 16 March 2020, argued that without any policy changes “we predict approximately 510,000 deaths in Great Britain” (p.7). They also suggested that the peak of mortality would take place after approximately 3 months

We note initially that dire predictions of mortality in Sweden were made in the earlier stages of the pandemic. For example, using a model applied to Sweden based on the Imperial one, on 15 April 2020, Gardner and colleagues stated that the model predictions indicated that “at least 96,000 deaths would occur by 1 July without mitigation. Current policies reduce this number by approximately 15%” (p.4).

As noted above, the peak level in EW is twice that of Sweden, but the number of deaths falls
away more sharply in EW. The trajectory of the decline after the peak is shown more clearly in Figure 2, which also llustrates the very similar nature of the rises to the peak. This plots the number of deaths, with, in each country, the data indexed so that the peak level (on 8 April in each case) is equal to 100.

We calculate the cumulative value of the index in each country from 8 April, the peak value, onwards. There are some initial fluctuations in both countries, but by 29 April the cumulative value from 8 April is equalised in the two countries. From 30 April until 7 August, the cumulative value of the index is distinctly higher in Sweden than it is in EW. In terms of index points, the difference is 13.16. In the EW data, an index value of 1 is equivalent to 1349 deaths, the number which occurred in EW on 8 April. We therefore estimate that the number of lives saved in EW by the lockdown is 1349 multiplied by 13.16, which is 17,751.

Compared to the dramatic predictions about death levels if lockdown were not imposed, this number is very low. However, many of the standard models used in epidemiology do not take account of the fact that, faced by a pandemic, people alter their behaviour of their own accord. They do not wait for the decrees of the authorities.

World history is replete with examples. The Black Death in the 14th century seems to have been the most deadly virus in recorded history. As much as 50 per cent of the population of Britain may have died. People adopted many different strategies to try and avoid the disease, some successful, others not. Giving lockdown a legal basis merely reinforces the behavioural adjustments which many people make when faced with a pandemic.

We note, finally, that a larger proportion of deaths took place in care homes in Sweden than in England and Wales. It could therefore be argued that even our estimate of 17,700 is perhaps on the high side.

While considerable uncertainties remain with the epidemiology, there is no doubt that the positive feedback nature of the Covid-19 epidemic will make controlling the outbreak present in May 2020 a very difficult task, especially if the main regulating tool selected is the co-ordinated behaviour of 67 million people. The likely economic effect is analysed of the years-long process to move out of lockdown sufficiently slowly so as not to cause excessive strain on the health services. In the Extended Base Case, annual GDP will fall by 23.5% in 2020 and is not likely to recover to pre-lockdown levels until 2024. Applying the J-value-derived Bristol Curve of population-average life-expectancy versus GDP suggests that the strategy of restriction followed by gradual relaxation is likely to result in a net cost, in terms of average human lives lost, that will be comparable with the UK's sacrifice over the six years of World War Two. A policy of lockdown followed by gradual relaxation is likely to do much more harm to the nation's health than good.