This afternoon, the Science and Technology Committee is taking evidence from experts on social distancing, contact tracing and the coronavirus ‘R’ value, which is the number of people someone with coronavirus goes on to infect. A value of greater than one could mean the virus spreads exponentially.
We’ll be covering the committee hearing live on this blog, flagging up major developments. The session is split into two parts: at 2:30pm, MPs will question Professor Sir Ian Diamond, the UK’s national statistician, and John Edmunds, professor of infectious disease modelling at the London School of Hygiene and Tropical Medicine. Both are part of the group of scientists advising the government on its response to the pandemic.
At 3:30pm, the panel will hear from Professor David Peters, chair of the department of international health at John Hopkins University, and Sir David Spiegelhalter, Winton Professor of the public understanding of risk at the University of Cambridge.
Updates will appear below.
4:34pm – Clark, wrapping up, asked whether the government should take a “differentiated approach” to lifting lockdown, based on the specific risks to certain groups (such as age groups).
Spiegelhalter says shielding is an example of that, but it would be better to have a “more sophisticated” approach, with more groups differentiated. People could “self-allocate” between these group, he says, depending on their circumstances and “risk appetite”.
People would also need to consider the risk to others, not just themselves, he adds. For example, the risk to schoolchildren is neglibible, but they might be able to pass the virus to older family members.
However, emphasising that different groups have different risks is important, he says.
That’s the end of the hearing.
4:25pm – Spiegelhalter suggests that a “really substantial” proportion of 3,000 “excess deaths” a week that have not been attributed to Covid-19 (currently “unexplained”) can be attributed to the disruption of the health service during the pandemic.
Other countries are slower to bring out data on excess deaths, he says, which makes specific comparisons difficult.
4:20pm – The experts are asked how certain we can be that earlier testing, and restrictions, have helped suppress coronavirus.
Peters says that, with caveats, we can say that it has helped. He points to South Korea as one example. But he would be retiscent to compare the UK and Germany, for example.
Spiegelhalter says you can “observe” that lower rates of deaths are linked to certain policies. But he is not willing to say it would have turned out differently in the UK if different, specific policies were followed.
4:15pm – Spiegelhalter says it’s “extraordinary” that we still do not know basic facts about coronavirus in the UK, such as how many people have it.
He says an ONS survey (see 2:35pm) will be helpful, but that it has come very late.
Other countries have done more testing, earlier, and therefore have better data, such as Germany.
Peters agrees that early testing and tracing is important, and creates more data to inform policy decisions.
4:10pm – Western nations seem to be more affected than those in the East. Do we know of any reasons for that?
Peters says that it’s difficult to say: some countries have very limited data. In Africa, for example “we’re not going to know what happens, ever”, because the data won’t be there.
He says we’ve seen that “social trust” and adherence to public policy is important (although he does not mention specific countries). He re-emphasises that data will be limited in poorer parts of the world.
4:05pm – Peters is asked whether countries with a lower first wave of infections might have a higher second wave.
He said that’s “feasible”. It depends on when you lift restrictions, and how quickly you can respond to changes. In the 1918 flu pandemic, countries with lower first peaks had higher second waves. “Depending on how policies are implemented, you could well see different waves,” he says.
4:01pm – Clark says the risk to young people from Covid-19 is very low. How can that influence policy decisions?
Spiegelhalter says policies should be based on those relative risks, which would mean different policies for different age groups. Other factors could move you up or down the scale, such as sex or ethnicity, he says. This could create a “traffic light”-style system for policy.
“The idea of stratify and shield is a very powerful idea,” he says.
3:56pm – Spiegelhalter says the risk of dying from Covid-19 doubles with every seven years of age. There have only been two deaths under the age of 15, he says. The average age of people going into intensive care is 60, but the median age for women dying of Covid-19 is 84, he says.
Men have double the risk of dying of Covid-19 as men do, he says.
3:53pm – Clark asks Spiegelhalter about an article he wrote in the Guardian, which has been cited by politicians, including Boris Johnson, when they advise against comparisons between different countries.
Spiegelhalter says politicians are misinterpretating his words: you can, of course, make international comparisons, he says. When you see big differences in outcomes between countries (he mentions Denmark’s low mortality rate”), you can draw conclusions. You must of course take into account the context of each country, such as the age of the population, he says.
He advised against making crude comparisons using “flawed outcome measures”.
3:50pm – Peters says one of the key things for international comparisons is to look at breakdowns by age, sex and co-morbidities. You can see “big differences” between countries, which makes it difficult to, for example, compare crude mortality rates.
You can compare mortality rates between specific groups (one age group, for example), to assess different countries’ responses.
3:47pm – Clark asks Peters about how useful it is to compare data from different countries during the pandemic. Can lessons be learned in that way?
Spiegelhalter says it is useful, and that looking at data from other countries can help analyse containment measures. In cases where countries have enacted “aggressive” lockdowns, it has worked, he says.
You can also analyse “packages” of interventions, he says. Overall, on restrictions: “earlier, rather than later, more rather than less,” is better, he says.
3:38pm – Diamond once again praises the “remarkable pace” of UK statisticians in working on the coronavirus.
That’s the end of the first part of the session. After a short break, the committee will question Professor David Peters, chair of the department of international health at John Hopkins University, and Sir David Spiegelhalter, Winton Professor of the public understanding of risk at the University of Cambridge.
3:35pm – On contact tracing, Edmunds says being able to trace new cases would be “probably impossible” at the moment because the numbers are so large, so incidence has to come “right down” for contact tracing to be feasible.
It “will play a role” going forward, but he “wouldn’t want to rely on it alone”. We will need other social distancing measures in place to stop the epidemic spiralling out of control, he says.
How low is “right down”, he is asked?
That’s an operational question, not an epidemiological question, he says.
3:28pm – Edmunds is asked how big an impact individual restrictions have had on the R value. He said it’s very difficult to tell, because of way those measures were introduced in quick succession.
Edmunds says it is “quite uncertain” what the impact of “social bubbles” – the idea that two households might be able to meet up as a self-contained unit – would have on the spread of the virus.
Diamond and Edmunds both say they have not seen the government’s plans for exiting lockdown, which Boris Johnson plans to announce on Sunday.
3:23pm – Clark asks whether it makes sense to have a single national value for R, given that the reproduction rate varies in different places and settings (care homes, etc).
Diamond says there are multiple values for R. He calls for more work to understand the epidemic over time in care homes, because it “is not going to go away in the short term”.
Edmunds says, strictly speaking, there is one R value, and that the reproduction number in different settings impact one another (because people move between different settings).
However, it is important to get separate R values for different settings so that different policies can be applied, if necessary.
The data in hospitals is, unfortunately, “not really good enough” for scientists to estimate R, which will vary between different hospitals.
“Care homes are even worse,” he says. There is currently insufficient data to estimate the R value in care homes.
3:20pm – Diamond says the ONS study can work in tandem with the React study, led by Imperial College London. The React study is testing 100,000 randomly selected people for Covid-19, as well as, later, for antibodies. “I’m of the view we need a multiple set of data sources,” he says, adding that the studies are complementary.
3:15pm – Edmunds explains that all data sources for calculating R have their own biases. Death rates are important, but lag behind infections, so they are not good for measuring rapid changes in R. Measuring changes in behaviour, for example, gives a quicker estimate.
One problem with the data is that children were not initially surveyed. From now on, scientists will include children.
Diamond adds that R needs to be seen in context of the incidence of the disease. If there are very few cases, and R is 1, the number of cases will remain stable: ie, very low.
3:10pm – Edmunds says the R value is reviewed around twice a week. Various different groups are calculating R, which reflects the changing incidence of coronavirus infections, and these groups meet to try to agree on a number. This process is led by a sub-group of the Scientific Advisory Group for Emergencies (Sage).
3:06pm – Clark asks Edmunds what the current “R” value in the country is. The R value, or reproduction rate, is the number of people that someone with coronavirus goes on to infect. The government has stressed the number must remain below one.
Edmunds says that R is currently somewhere between 0.75 and 1, adding that it’s “just below one”.
It varies regionally: some regions are “closer to one”, and London looks to have the lowest number, around 0.75, he says.
The R value has actually gone up over the past few weeks, he says, and that it was probably around 0.6 or 0.7. But that doesn’t reflect the reproduction rate in the community – it’s more a product of the fact that measurements are taking into account “ongoing” outbreaks in care homes and hospitals. “More of what we’re seeing when we’re measuring [R] is incidents going on in these settings,” he says.
The lockdown has worked for bringing the R value coming down in the community, he says.
3:02pm – Diamond says that, currently, the ONS is linking death registrations to census data in order to record the ethnicity of those that are dying. It would “make our work a little easier” if ethnicity were recorded on death certificates, he says.
He would like to see deaths registered electronically quicker than they currently are, he says. Deaths can be registered even if the cause of death is not certain, because post-mortems can be conducted after the death is registered.
2:57pm – Diamond is asked whether part of his role, and the role of the national statistics offices, is to keep the government honest. He says the government has been kept “informed” by data from the national statistics service.
Government has been “veracious” in its appetite to be informed, and have had to make “very difficult decisions”. The ONS has statisticians embedded with government to ensure ministers’ information, and announcements, are “statistically sound”.
2:50pm – On Covid-19 deaths, Diamond says that “longer term”, he expects “quite a number” of deaths to be attributed indirectly to Covid-19, such as for people who die with cancer because they are not going to hospitals for screening.
He also says that a recession will impact death rates. The UK is set for a “significant decline” in our GDP, and if our recovery is slow, then poverty and unemployment will increase. That will cause a “reduction in healthy lives”, and potentially more deaths.
He says the number of “excess deaths” – the number of deaths over and above the normal average for this time of year – is the highest since records began in the 1990s. Normally, deaths peak in the winter months. “To see them in the middle of a sunny April is absolutely sobering,” he says.
2:45pm – Diamond says the first results from the ONS survey mentioned above will be available next Thursday, 14 May. From that point, results will be published every Monday and Thursday.
Clark asks whether the study could’ve happened earlier. Diamond says the ONS was asked to commence work on April 17, and the turnaround has been “one of the most rapid” he’s ever seen.
Clark says that, perhaps the request to the ONS could’ve come earlier. Diamond days the ONS “responded to the first request we got”.
2:35pm – Committee chair Greg Clark asks Diamond what the best way to measure both the incidence of Covid-19 and deaths from the disease is.
Diamond says a lot of the work on prevalence so far has been based on modelling. In the future, we will start to get “data-led measures” of national prevelance through an Office for National Statistics (ONS) survey. The survey is “in the field” right now, and takes swabs from a nationally representative sample of households. Some blood samples are taken too, so that scientists can measure antibodies.
This process will give a better measure of prevalence.
Researchers will return to households one week later, giving estimates of the rate of transmission. In future, we can get estimates of the prevalence of both the disease and antibodies in the population.
That is phase one. In phase two, the study will expand, allowing us to get estimates of the prevalence of the population twice a week. We will get national and regional estimates, as well as for sub-populations, such as ethnic minority groups or healthcare workers.
This will give us a “good, empirical knowledge” of the disease over the next year, Diamond says.