EU Referendum


Coronavirus: ripples in a pond


13/05/2020




I have been trying to visualise images that might best explain how the statistics in this epidemic are being wrongly presented giving a completely false impression of what has been going on and what is likely to happen.

The starting point, familiar to most readers of this blog, is that the single epidemic curve which is offered to us is a statistical artefact. It does not exist in real life, but represents a composite of the records of multiple outbreaks. As such, I have asserted, it has next to no epidemiological value.

As to a visualisation of what has been happening, the best I can offer for the moment is the picture of ripples on the still water of a pond. If you can imagine the stone on the pond as the original outbreak (assuming only one), then the ripples radiating out from the point of disturbance are new outbreaks, with the height of each ripple (amplitude) representing the intensity (number of cases).

If we now take the single epidemic curve on offer, what is happening is that the compilers are summating, over a period of time, the amplitudes of all the ripples in the pond, to produce single values for each time point. Thus, we end up with a single, tsunami-like wave which no more represents reality than does the rhythmic crashing of tidal waves on our shores.

Now stretching the analogy (possibly to breaking point), if we take the epidemic in its early stages and try to represent it, we could perhaps imagine a relatively short sequence of quite high ripples, each representing a large community outbreak – as might occur in workers travelling to and from a large, city centre office complex – totalling thousands of cases.

Summate the amplitudes at their collective highest and that might give you the peak we have experienced on the national epidemic curve. But then comes the lockdown. The large ripples are no more, replaced by many more (hundreds even) of tiny ripples, each representing either household outbreaks of 2-3 persons, or care homes, which might be anywhere up to 20-30.

While there may be many more ripples (outbreaks), the total of the amplitudes may be several orders of magnitude lower than the value immediately before the lockdown. This is the famous "turning of the tide" scenario, where we are being encouraged to see a decline in the epidemic. Johnson has squashed his "sombrero".

But what we could actually have done is swap size for quantity. The national curve would then conceal a situation where we have vastly fewer large outbreaks, but a massive expansion in the number of small ones, with more emerging as soon as the earlier ones peter out.

The epidemiological significance of this is huge: each of the outbreaks constitutes a potential (or actual) reservoir of infection. Although damped down by the lockdown, when restrictions are lifted, they may seed communities with new infection, causing multiple new outbreaks.

At this point the ripples on the pond analogy has almost served its usefulness, but not quite. The new ripples, despite being many in number, may each be quite small – the outbreaks only affecting the communities in which they are located.

As likely as not, aggregate value of the amplitudes may not (at least, not immediately) match the pre-lockdown values. As the national curve stays flat, an incautious politician might even believe that the relaxations are safe and can be continued.

The trouble is that the infection will be in the community. And here we can switch from ripples in a pond to a smouldering fire. At any time, the fire could burst into life. But the time-lag could be considerable and, by the time we see the flames, we could have a major problem on our hands.

Such a scenario, therefore, lends weight to yesterday's report by David King's Independent Sage Group. Despite being infested with prestige, and its message blurred by lefty nostrums, it asserts that there is "an urgent need to rebuild an integrated public health infrastructure of the form required to deliver optimal protection of the people of Britain and Northern Ireland".

This ties in with one of the main recommendations of the group, which suggests that: "The government should take all necessary measures to control the virus through suppression and not simply managing its spread". It goes on to say that:
We detect ambivalence in the government's strategic response, with some advisers promoting the idea of simply "flattening the curve" or ensuring the NHS is not overwhelmed. We find this attitude counter-productive and potentially dangerous. Without suppression, we shall inevitably see a more rapid return of local epidemics resulting in more deaths and potentially further partial or national lockdowns, with the economic costs that will incur.
The group wants the government to "refocus its ambition" on ensuring sufficient public health and health system capacities to ensure that we can identify, isolate, test and treat all cases, and to trace and quarantine contacts.

It also says that the government must develop a clear quarantine and messaging policy which takes account of the diversity of experiences of our population, variations in household structures, and with appropriate quarantine facilities in the community. This, it adds, should be accompanied by real time high quality detailed data about the epidemic in each local authority and ward area.

I'm sorry here that the group itself is muddying the waters. There is no "epidemic" when it comes to each local authority and ward area. The epidemic is and will remain a statistical artefact, defined by the expectations of what future infection levels might be.

Here then is the rub. Without a vaccine to give us long-term protection, the suppression process is not a one-off task. Case numbers can be brought down to "acceptable" levels (that being a movable feast), so that the disease is brought down from its current epidemic level to become endemic – a background, fairly static level.

But this will not remain static, without continuous and active intervention by the public health services. This we see currently in Wuhan, where plans are being drawn up to test the entire population of 11 million people for Covid-19, in order to prevent a recrudescence of the disease.

This is possibly an extravagant and unnecessary response, and we are told that several senior health officials are indicating that testing the entire city would be unfeasible and costly. But it points to a new normal, where activities to suppress infection become a background part of our daily lives.

Says the Independent Sage Group, this requires a virus control system "which has long term sustainability". It must, the group says, "be built into an enhanced public health protection system, taking advantage of the primary and secondary health care system, but also incorporating locality-based integration (integrated Care Systems) including local government and social care, and crucially with community participation".

For the government, this presents it with a considerable political dilemma. It must recognise that the progressive centralisation of public health services over that last decades has not worked and that the system, as presently structured, is not fit for purpose. It must thus reverse the tide of centralisation and restore local structures, powers and responsibilities.

In terms, though, I see this as more problematical for Johnson administration than the immediate measures to control the epidemic. Throughout its progress, it has exercised central control, micromanaging measures to the nth degree. Somehow, I cannot see this government loosening its grip, in which case I see its plans doomed to failure.

And that is very much the view of Devi Sridhar, chair of global public health at the University of Edinburgh. She is not part of the Independent Sage Group, but sees in our current path a disastrous ending.

Lifting lockdown without the public health infrastructure in place to contain the virus, she says, will allow Covid-19 to spread through the population unchecked. The result could be a Darwinian culling of the elderly and vulnerable, and an individual gamble for those exposed to the virus. Predictably, she asserts, this should be avoided at all costs.

Just for once, though, we might give a thought to Ambrose Evans-Pritchard in the Telegraph. He cites a Covid cardiologist at a top London hospital, who declares, of the government's performance, that, "Basically, every mistake that could have been made, was made".

But if so much has gone wrong in the past, and some might be the errors of previous administrations, where Jeremy Hunt admitted to parliament on Monday that, "It is now clear that a major blind spot in the approach taken in Europe and America was our focus on pandemic flu rather than pandemic coronaviruses".

But the sins of the past might be nothing compared to what we face in the future. The mistakes made now will decide that future, when the ripples on that pond could end up being more like a tsunami.