The question of testing keeps coming up with the government under pressure to explain why Germany can manage 500,000 tests a week while we have just belatedly reached 10,000 per day. Matt Hancock admitted at yesterday's press conference that we entered the pandemic "unprepared". I am not usually in the position of defending this government but I have some sympathy with Hancock because he was simply following the plan established by his predecessors (probably several of them).
I was interested in the existing planning documents published by Dr North yesterday and later, a mathematician/epidemiologist named Adam Kucharski tweeted a link to the government's Scientific pandemic Influenza Modelling group summary paper, which I ploughed through as best I could last night. This is quite a recent document, last updated in November 2018. It was quite a revelation and shows the government never intended to do any mass testing of the population at all!
I thought this would be quite a nice blog post only to find this morning that both The Telegraph and Dr North have covered it already. What they don't seem to have done is read the document itself, which is actually remarkably prescient. It forecasts we would see the new virus "in most major UK centres of population within 1 to 2 weeks of the earliest importations" with "the peak of the UK epidemic following about 50 days later".
The first UK case was reported on January 31st and the peak is expected on April 12th, about 70 days later, so not far out. Much of the paper is based on the work of Professor Neil Ferguson at Imperial College's Faculty of Medicine, School of Public Health.
What it does say on page 13 is this:
"Contact tracing (including serological and virological testing of contacts) of the first few hundreds of cases in the UK, community surveys and individual outbreak analysis will be essential for the accurate determination of disease parameters, most importantly generation time and the proportion of cases showing clinical symptoms."
And again on page 35:
"Data for real time modelling in a pandemic will come from two sources, aggregate data during the majority of the UK epidemic and individual data mainly from the first few hundred cases placed on an individual case database, the ‘FF100’ database. The data is split between basic data required to analyse and forecast numbers of cases and deaths, and an extended data set, which would also allow forecasts of the demand for secondary care and absence in both the NHS and elsewhere, as well as a more detailed analysis of development of the UK epidemic."
The intention of the virological (swab) testing was simply to determine accurately the disease parameters. For this, the first few hundred cases were all the guinea pigs they needed. It was never intended to help protect the public by suppressing the spread of the virus - which was assumed to be a type of influenza akin to that in 1918's Spanish Flu outbreak. This is why the government ceased community testing for coronavirus on March 12th.
The planning implications on page 7 includes this statement:
"Encourage construction of realistic and detailed local plans for containment in the source country. (This is different to attempting to contain the virus once it is widespread which has little chance of success, see section 3.3)."
And sure enough, section 3.3 (c) the paper makes clear that
"Because of the probable multiple importations of pandemic flu, and the concentration of the population in cities, attempts at containment (similar to those explained in section 3.1b above) by antiviral prophylaxis and practical social distance measures are almost certain to fail (Ferguson et al. 2006, Nguyen-VanTam et al. 2004)."
In other words, they had no faith that social distance measures would work, and this was probably true - but only because they proposed just two main measures: restricting mass gatherings and school closures: (see 3.6 (a)):
"In addition to the medical countermeasures of vaccination, antivirals and antibiotics, various social distance measures might be used to reduce interpersonal contacts, reduce transmission and hence the progress and extent of the epidemic. Two such measures are restrictions on mass gatherings, and school closures of various kinds - individual classes, local, regional, national, preemptive, scheduled or reactive (Cauchemez et al. 2009)."
There is a clear acceptance in the paper that once the virus has spread in the population there was very little the authorities could or would do. What changed is the way other countries adopted far more draconian lock down measures with people advised or ordered to stay at home in what amounts to a shutting down of national economies. This was never considered by the British government as a way of suppressing a pandemic.
China achieved success, if indeed it was successful - there is some scepticism about their mortality numbers - by taking steps that only a police state could imagine. The New York Times explains that on February 6th Chinese authorities "resorted to increasingly extreme measures in Wuhan on Thursday to try to halt the spread of the deadly coronavirus, ordering house-to-house searches [and temperature checks], rounding up the sick and warehousing them in enormous quarantine centers."
It worked in China but I am not convinced it will work as well in Europe or the USA.
There is also the problem of what happens when restrictions are lifted. The pandemic is likely to return and the paper anticipates that in section 4: Potential subsequent waves. We are almost certain to see at least one and perhaps more waves which will require new lockdown measures until a vaccine is developed. The model assumes this would take 4-6 months and another 8-12 months to produce a "large amount" of vaccine.
I think the government has been forced to switch to suppression simply because that's what other countries were doing. Ireland has a tracing operation, as does Germany and South Korea. Unless you're prepared to trace every contact, mass testing doesn't help anyway.
Finally, on the number of tests being done, the NHS guidance for the Real-Time Polymerase Chain Reaction (RT-PCR) shows why it is hard to speed up. On page 21 (Appendix 5) the protocol explains that the samples of RNA (single strands of nucleotides) undergo a 10 minute cycle followed by 45 cycles, each lasting 3 minutes 45 seconds. This means it takes three hours to process 96 samples (the usual capacity of a PCR machine).
And this does not included the preparation of each sample or data logging of each person. It is not a quick process.
We did enter the pandemic unprepared to do mass testing and probably other countries were similarly unprepared, but they were quicker to to gear up to handle it and to use it as a means of suppressing the virus. It remains to be seen if it is going to be successful. We won't know that for quite a while.
The government really planned, in these pre-vaccine stages, to use the pandemic itself to develop what is referred to as "background immunity" in the paper. That may have been the right approach but Covid-19 is proving to be both more infectious and more lethal than the planners assumed.
