A socialist epidemiologist on the pandemic

 

George Davey Smith, who is Professor of Clinical Epidemiology at Bristol University, talked with Martin Thomas from Solidarity about the pandemic.

So far, and short of a vaccine, we seem to have only ancient measures to control the spread of the pandemic: physical distancing on different levels, and hygiene. Have we learned anything about variants of those measures which may be both effective and sustainable long-term?

An important thing that’s becoming clear is that this is very likely going to become the fifth endemic coronavirus. There are four seasonal coronaviruses (sCoVs) that cause symptoms of the common cold and occasionally more serious illness, and might even bring forward deaths among the elderly.

That the four previous coronaviruses are endemic doesn’t mean that they are prevalent all the time. They tend to arrive seasonally. Levels of infection go up and down. There are levels of immunity and cross-immunity [immunity from one particular sCoV that provides some protection against another], so infection rates from any particular one can remain low for several years, until increasing again.

Any strategy for Covid has to start with the fact that this virus is likely to become endemic. One of the first scientific papers on this virus, coming out of China at the end of February, showed that 80% of infections were not being picked up as cases, being asymptomatic or being accompanied by mild common symptoms that would not reach medical attention. With that level of asymptomatic transmission it is extremely difficult to eliminate a virus.

Even when a vaccine arrives — if one arrives — over a quarter of people in the USA say they won’t take it, with the equivalent figure in Britain being not far behind; and it is unlikely to be available widely enough in poorer countries. One of the more promising vaccine projects is for a vaccine which doesn’t stop transmission, but makes you less ill if you’re infected, which of course won’t eliminate the virus; and viral characteristics are likely to change, as with flu, leading to constant vaccine modification.

The Independent SAGE proposal of a “zero Covid” aim is, in my view, unrealistic. On 14 July David King, the chair of Independent SAGE, said: “I don’t believe schools should be reopened until we’re reached zero Covid”. This completely fails to recognise the huge and lasting adverse consequences of interrupted education, especially for the poorest in society.

The Independent SAGE model is New Zealand. But that’s two remote islands where you could have literally no one arriving on a plane for a while. And even there the long-term sustainability of “zero Covid” is not certain.

There has been talk of ‘shielding’ old and frail people through the pandemic until they are protected by chains of infection in the wider population becoming short and rare. In fact many countries have failed drastically, and have had higher rates of infection in concentrations of old and frail people, in care homes for example, than in the general population. Can we change that for the future?

One of the things we know about the virus is that it is extremely selective as regards the people who become sickest, with age being the strongest influence by far. The average age of deaths is over 80. In the early days in Britain there was extreme anxiety about personal risk across all age groups. In public communication, there was a failure to convey the very high degree of uncertainty in all the scientific estimates there is.

Figures were reported as if they were definite — such as the oft-repeated predication there would be 510,000 deaths from Covid — when there was (and remains) considerable uncertainty.

Seasonal influenzas have on occasion led to more deaths than Covid in younger age groups, and even with the four previous endemic coronaviruses there have been outbreaks associated with deaths in care homes. Boris Johnson and others have conveyed that this is the worst health crisis of anyone’s lifetime; but the 1968 flu epidemic (let alone the 1951 or 1957 outbreaks) was more deadly if you make allowance for the fact that life expectancy has increased greatly since then and there are many more frail 90-year olds today.

The tragedy with the care homes in Britain was that by 23 March it was well know how age-selective the virus is. The shielding of institutions where a large proportion of people are susceptible should have been done properly. Indeed, the predictors of which care homes have done badly are the obvious ones: casualised staff, inadequate PPE, and so on.

Germany has done better, and one factor there is that care homes are better in Germany, and a higher proportion of people at any particular age are living in them. In the USA, one of the reason why the Bronx has done terribly is that on the East Coast of the USA, if you want a cheap care home, the Bronx is the place to look.

Hospital-acquired infection has also been important. People have died who acquired their infection with the virus in hospital.

I think the lockdown-easing will continue for now with a decline in cases and deaths, but we may well see an increase again in October. Endemic coronaviruses are generally seasonal. We won’t know about this one until it has run through a whole year, but the decline in infections in Europe in April and after could reflect seasonality as well as lockdowns. Trump’s asinine declaration that the sun was going to burn Covid away has understandably led to a reluctance to talk about seasonal influences.

It is obvious we should use the time for preparation. Denmark has a law to provide for safe accommodation for quarantining people who have been infected and are in crowded housing. China, Taiwan, and South Korea have done the same, moving people to hotels or to hospitals.

Testing, tracing, and isolation is very difficult to make complete when we have such a high level of asymptomatic infection. Masking can be useful, but I don’t know if it’s workable to make mask-wearing completely mandatory whenever you’re out of doors.

The basic public health message has to be: to stop infection growing, you have to limit contacts which may bring transmission, even though you can’t have 100% protection for every encounter.

What about the work of Adam Kucharski and others suggesting that the infection spreads in a very uneven way, the bulk in ‘clusters’?

Everyone loves the language of “the superspreader”, but heterogeneity of transmission probabilities can come from varying susceptibilities, perhaps due to pre-existing partial immunity, of people to a given dose of the virus. That may be behind Adam Kucharski’s finding that 80% of transmission comes from just 10% of potentially-transmitting encounters.

What about ‘clustering’ on a larger scale? Britain is unusual in a relatively even spread of infection across the country, while in the Netherlands, say, infection has been highly concentrated in North Brabant and relatively low in Rotterdam and Amsterdam?

Yes, there is no Parliamentary constituency now which has had no deaths. The relatively even spread may be to do with patterns around the time of entry of the virus into Britain. We had lots of students moving around the country for the start of term then, and lots of people returning from holidays in Italy or France. Thus in Norway it’s been suggested that one reason it has done well is that people go ski-ing within the country, rather than abroad.

Camilla Stoltenberg, the director of the National Public Health Institute in Norway, says that the decision to shut the schools there was probably wrong. Studies have consistently shown that children, as well as being much less likely to suffer badly, appear not to transmit the virus as much as adults. Given all the bad consequences of a long period with schools closed, I don’t think the schools should have been shut in Britain. But once they’re shut you can’t easily undo that.

What about the risk of students taking infection from school into households with elderly or frail people?

One of the key issues in preparing for September and October is to try to get a coherent shielding strategy, with good assessments of risks for individuals and what can be done about them. It can’t just be self-assessment, because over 50% of people who have died from Covid-19 in care homes already had dementia; but of course for institutions like care homes, the whole institution needs shielding.

Several universities are planning to have PCR testing in place for returning students and all new arrivals. Obviously that testing will only give a snapshot, but it could help with efforts to reduce the number of infections. Institutions may also develop their own phone apps to detect possibly infectious contacts. To work well, the operation has to be locally based.

Current estimates are that rates of having been infected are still low, even in the countries hardest-hit by the virus, so the chances of a contribution to infection control from even moderately widespread immunities in the general population, gained from having gone through infection rather than from a vaccine, are small. What do you think?

We don’t know, but the general consensus is that in Britain, for example, the proportion who have responded to an infection is likely to be higher than the 5% estimated from antibody tests. Other aspects of immunity — known as cell-mediated immunity — can provide some protection in the absence of detectable antibodies. In the largest serology study in the world, in Spain, individuals who had clearly been infected did not have detectable antibodies at a later date. The crucial question is the extent to which there may be varying levels of susceptibility due to pre-existing sCoVs and prior SARS-CoV-2 exposure.

I doubt that there is a single explanation of the different outcomes so far in different countries. A big part may have been played by chance events in different countries at the early stage. People debated a lot about why Edinburgh did much worse than Glasgow with HIV, but it now looks like it was mainly a matter of early arrivals of cases before it was clear what was happening and interventions were in place.

In the short term, it is unlikely that lockdowns have had no effect. Some countries have done very well so far with no or very limited school closures, for example, but probably how lockdowns have been implemented will have influenced both the on-target (Covid-ameliorating) and off-target (adverse effects on people’s lives) outcomes.

A very large percentage of deaths in Britain has been in care homes, and a lot must be put down to the mess-ups by the government and how care homes and hospitals have been inadequately protected and managed. But, for example, a contribution to why Britain has done so badly (though by no means the full explanation) may be that Britain has had very mild flu seasons for the last two years.

If you compare across Europe, one predictor of Covid-19 deaths is low flu deaths in the last two years. It may be that high flu deaths in recent years leave you with a smaller surviving frail elderly population vulnerable to Covid-19. It’s unclear what the final conclusion will be about this, but indicates the complexity of research in this area.

Overall, it is still too soon to start doing fine-scale league tables of which countries have done worse and which better. We haven’t had a complete cycle of the virus yet, and we haven’t even started to observe the long-term consequences of both viral infection or the social responses to it.

*  This interview also appears in the current issue of Solidarity and on the Workers Liberty website.