This article is also available in
Chief Statisticians from across the world are leading the response of National Statistical Systems to the data challenges of the COVID-19 pandemic, even in the context of major disruptions in day-to-day statistical operations. Sir Ian Diamond, UK’s National Statistician, shares in an interview how the Office of National Statistics of the United Kingdom is innovating and working together with the family of National Statistic Institutes around the globe to provide timely and reliable data to monitor and contain the spread of the disease and its socio-economic impacts, and to inform the design of effective recovery policies. Here you can find the video recording and a slightly edited transcript of the interview.
What is the role of ONS in coordinating national COVID-19 data collection efforts in the UK?
The role of ONS is to lead the government’s statistical service. Right across government, ONS works very closely with different arms of government to make sure that there is coordination in the data that are prepared for COVID-19.
Particularly, we’ve worked closely with the Department of Health and Social Care over the course of the epidemic to improve the reporting of daily deaths data. We’ve worked very hard with the Treasury, and the Department of Business Energy and Industrial Strategy to ensure that the economic data are being brought together quickly and efficiently to the national effort. And I think as a National Statistical Institute we have to step up and provide a coordinating role, and that’s something we’ve been very proud to do.
What are the key national sources of data for monitoring the pandemic?
We have worked very hard on this and I’m going to start with some new things that we’ve done, but I’ll also mention the way that we have adapted some of our business as usual.
On the new front, we’ve stepped up a new business survey that went out very, very quickly to over 16,000 businesses and which aims to find out the way they’re reacting to the economic aspects of the pandemic. We still have a weekly social survey which enables us to understand how people are responding to some of the non-financial interventions that the government has put into place and how they’re feeling about the data.
We have a very new, very large survey that is collecting swabs from people to measure the prevalence of the virus in the population and collecting blood to measure the level antibodies in the population. By going back to people over time, this will allow us not only to estimate prevalence, but also the rates of transmission.
We have really worked hard in our Data Science campus to use some very innovative ways of collecting data. For example, on footfall at major stations following lockdown to look at the impact of that kind of lockdown. So, we have innovated and brought very quickly a really large set of new data sources that the government can use to be able to plan.
But at the same time, have worked very, very hard to make some of our “business as usual” statistics, particularly our deaths data, as timely and as useful as possible for government. We produced deaths data based on death registrations with a lag of about 10 days. So, we’ve worked with our colleagues in the Department of Health to get more timely data, which may not have the full coverage but can help plan the government’s response to the pandemic. And we’ve used those deaths data to link with census data to understand the impact of the virus on the people in different communities or with different demographic or ethnic characteristics and bringing those out regularly has really helped the government to plan.
Our labor force survey remains incredibly important in planning, and I’ll turn to that a little later. And we really have worked hard to make sure that our measurement of GDP and inflation are available, ready and accurate as the government plans its way out of the economic challenges that the pandemic has brought.
How is ONS ensuring the continuity and quality of key statistical operations in the face of restrictions to conduct field operations?
We, along with many other statistical institutes, have had to stand down our face-to-face field work. That has meant that we have had to very quickly move to telephone data collection for our labor force survey. The good news is that we knew something about the biases that already existed, so we’re able to adjust for that. However, it’s very, very important to do that very, very quickly. Migration statistics were needing to work very quickly towards use administrative of data. So, in many ways, some of the challenges that we have faced are things that we would have wanted to do very, very quickly. In addition, when it comes to measuring inflation, we have used things like web scraping of prices rather than face-to-face data collection. So, we have been quickly and, I hope, effectively changing some of the ways in which we collect data to ensure that we are able to provide the data that our government needs.
How does the ONS strike the right balance between timeliness and quality of data in the current emergency situation?
I think this is a really good question. Government needs timely data; there is no question about that. But equally, government does not need wrong information to make poor policy. So, while you can talk about quality against timeliness, you absolutely need really to be thinking about what can you do with quality in the short term, as well as, perhaps, improving on that in the long term.
So, perhaps the best example of this might be on deaths data, where we know that our deaths registration data will be the most comprehensive, but it comes with a 10-day lag. So, we’ve worked the Department of Health and Social Care to be able to produce daily data of all the deaths where there was a positive test for COVID-19. There will be some cases that are not in those data—for example, where the practitioner has not got a positive test but feels that the cause of death was COVID-19. But we believe that the data that we do have are very close to good and therefore good for timely reporting.
We’re also looking very hard at some fast economic indicators that we could bring over the next few months to enable government to understand what is happening in the economy while waiting for, if you like, the monthly GDP data. What is ONS doing to ensure that relevant data and information on the spread and impact of COVID-19 is open and readily accessible to the public and policymakers, while at the same time ensuring privacy and confidentiality of personal information?
It is incredibly important that we are open with the data that we use. And we have developed over the last few years a secure research service that allows researchers to access data against an ethical agreement when the research project that they wish to do is demonstrably in the public interest. We have worked very hard during the pandemic to enable researchers, for the first time, to access this secure research service from their homes and this is, I think, a major breakthrough.
I talked earlier about a very large study that we are doing which is enabling us to monitor COVID-19 prevalence and the level of antibodies in the population. The protocol for this study was made open and available before the data were collected. There are regular announcements and publications of the results, and the data are available in real time in the secure research service for secondary analysis. This is just an example; there are many more.
What tools or platforms are being used to communicate information related to COVID-19 to users, including the general public?
The list is limitless. Clearly, we have our regular publications, and a number of ad hoc publications which bring data to the public very, very quickly. We are also feeding in to government policy on a regular basis, so the government’s dashboards have a lot of ONS data. In addition, we use different means of communications, such as social media, to get to people. And we are regularly doing interviews, in television, on radio and in other fora, to make sure that our results are communicated in a way that gives the maximum exposure to the population.
Is ONS following and/or adapting any practices in data collection either here in the UK or globally for data comparability across countries?
This is a very important challenge and one of the questions that is always asked is, is there a [ranking] table showing the challenge faced by different countries from COVID-19? And my answer is that, sadly, it is not possible to have any kind of coherent [ranking] table. You can see which countries are most affected, and which are perhaps less affected. But you can’t really make a [ranking] table for two reasons. One, because what is reported is not as alike as one might like, as some countries do things based on samples, other countries have different timings, etc., and being able to compare the number of deaths is actually very difficult. Secondly, we know that COVID-19 is impacted by, for example, demographic factors, so that age structure of the population is important. And we know that it is likely to be transmitted more in inner cities where there is high density of population, so the degree of urbanization of the population is going to be important. One really does need to be very careful in making international comparisons.
The other thing which moves slightly away from that, but it’s still incredibly important, is to recognize that there will be deaths, not directly from COVID-19, but which are a result of it. And to capture them one needs a way of using all-cause mortality excess deaths. And again, I think there is real potential opportunities to do that, but there remain challenges in for the two reasons that I just described.
In your view, what are the most promising innovations in terms of the use of new data sources and new methods to address the need for detailed, timely and high-quality information on the pandemic? Can you tell us more about how ONS Data Science Campus, for example, is contributing in this regard?
I think there are enormous opportunities at all times to innovate, and I don’t think one should wait for a crisis like a pandemic to innovate. That’s why we have a Data Science Campus, where we can, for example, use web scraping to understand the number of people who are asking for Google Maps around places with particularly high footfall. That is important. We have used web scraping as well to identify the availability across the country of the sort of goods that people buy in a crisis. For example, we knew at the beginning of the lockdown that there was great difficulty buying hand sanitizer anywhere, and we could point to where it was possible to buy it. We’ve taken that basket of goods and followed their prices over time and—given that are our Prime Minister has just had a baby—, the good news for him is that we can say that nappies have reduced in price. So, we were able to follow all these kinds of things very much.
The Data Science Campus has an enormous role to play. I would also say that it is not just the Data Science Campus that has elevated data innovation, being absolutely brilliant, but also some of the innovations around very quick reporting. I’ve mentioned our social survey. We can stop data intake on a Friday at four o’clock, and we have results delivered to the government on Monday by midday. That kind of speed of innovation, of delivery, is incredibly important at a time like this, and requires innovation right across the organization, as does—to go back to the very large prevalence survey—the fact that those data are going into the secure research service in real time. It takes an enormous amount of effort from across the office to innovate. To do it at such pace, and when we’re all working from home as well, makes this a really exciting and innovative time.
What are your views on how new solutions developed in this phase, can become our new normal? Is there something we’re learning from this crisis and the way national statistical offices are adapting to it?
I think there are so many new normals. Let me just start giving one: Alongside many other National Statistics Offices across the world we have moved very quickly to work from home. That brings with it both opportunities and challenges. You’re ensuring people’s wellbeing when they’re working, and perhaps on their own, while they’re trying to manage caring responsibilities with children being at home as well, making sure that everybody feels that they are contributing—these are things that we absolutely have to worry about as the National Statistics office.
But at the same time, some of the things that we have learned will become a new normal. I believe, for example, that we need to be thinking very much about more flexible work in the future, perhaps more than we have in the past. We can work much more, perhaps on a basis of trust with our colleagues. And those kinds of changes in the way we work are, it seems to me, really positive. At the same time, we have really upped our speed in which we’ve used some administrative data. That’s a new normal that we need to seize and if we can stand up things very quickly in a crisis, why can’t we do things like that all the time? Now, I recognize that some people have been working incredibly hard and we need to build things into a sustainable way for all our teams, but at the same time, we’ve demonstrated that we can do things at pace with quality and that are incredibly relevant to government decision-making. Those are the things that we can’t possibly afford to lose.
What are ONS’ plans for the post COVID-19 period, including for addressing the new data needs to understand the impact on the economy and all other sectors of society?
I think the first thing to say is that it is a very brave person who says “post COVID-19”, because some of the coronaviruses that exist in the population have been around for over 200 years. So it is extremely likely, in my view, that we will move into a time where we are controlling COVID-19, where it is in the population but at a very low level, and part of the control will be through treatments and vaccinations. However, it may be some time before we get there. In the meantime, we absolutely need to be working on ways of identifying and acting on outbreaks very quickly. ONS has been part of the government’s strategies for using data and data science to identify outbreaks, and it will continue so to do.
We equally know that the economy is going to take a very big hit from the government actions, the right government actions, to impact on COVID-19. And given that plan, we need to work tirelessly to make this recession a steep V-shaped recession, if it’s at all possible. The last thing we need is an L-shaped recession, because an L-shaped recession would push many people into unemployment and poverty, and that would result in negative effects on health and increased mortality. So there is an enormous amount that the government knows it needs to do to make sure we come out of this economic challenge quickly.
And ONS has to be at the heart of that through fast economic indicators, through ensuring that our measurements of unemployment and the skill bases in the country are quick and early. Our job is to react to changes in the economy and measure those changes. That’s why a National Statistical Institute is always in a state of change. And, so ONS will be reacting very, very quickly to the kinds of decisions that are needed to provide data, to inform government policy and to provide data on the changing economy, and to help government steer us on a course out of the hits that the economy is going to take.
How are you adapting statistical capacity development efforts to the new situation? For example, how can developed countries continue providing technical assistance and collaborating with statistical offices in developing countries?
I figure it’s incredibly important that right across the globe we are supporting each other, sharing best practices and working to make sure that we are supporting and informing all our governments. ONS has a number of partners with countries, including in various low-income countries. For example, we met recently with our partners from Rwanda, from Ghana and from Kenya for a discussion around what we were all doing to support our governments through this COVID-19 pandemic, how we might innovate and how we might support each other. As we go through the pandemic, we will continue to share best practices and support each other, getting back to helping people with the next round of censuses or other important statistical priorities. We are a global family of National Statistic Institutes. It is critical in my view that we work together for the global good.
Sir Ian is the UK’s National Statistician, Head of the Government Statistical Service (GSS) and Chief Executive of the UK Statistics Authority. He provides overall leadership for the Office for National Statistics (ONS) and the statistics profession across government. Sir Ian advises ministers, the Cabinet Secretary and senior officials on the production, dissemination and use of statistics across government.
Professor Sir Ian David Diamond, FBA, FRSE, FAcSS is the former Principal and Vice-Chancellor of the University of Aberdeen. His previous roles include Chief Executive of the Economic and Social Research Council, Chair of the Research Councils UK Executive Group, Deputy Vice-Chancellor at the University of Southampton and Chairman of the Social Security Advisory Committee.