Skip to main content

Covid-19: Transparency and Accuracy of Statistics

Volume 807: debated on Monday 9 November 2020

Private Notice Question

Asked by

To ask Her Majesty’s Government, further to the statement by the Office for Statistics Regulation on the transparency of data related to COVID-19, published on 5 November, and reports that charts on projected daily deaths from COVID-19 have been reissued, what assessment they have made of such reports; and what steps they are taking to review (1) the transparency and accuracy of statistics relating to COVID-19, and (2) the decisions that are based on such statistics.

My Lords, the Office for Statistics Regulation is 100% right: the best use of data and statistics is critical in this unprecedented time. All slides and data from press conferences are published on GOV.UK, normally at the time of the press conference. The Government are committed to transparency to build public trust throughout the pandemic; that is why we publish data, the modelling used and any revisions as part of this process.

I thank my noble friend. Does he realise that this rebuke from the statistics regulator is unprecedented, as is the unparalleled series of errors, dubious charts, outdated data and failed projections? It would be bad enough if those errors were random, but they all point in the same direction: alarmism justifying a lockdown. It is invariably a symptom of groupthink when sincere people—I have been there myself—become so wedded to a prediction or policy that they uncritically accept evidence that supports it and ignore facts that challenge it. Will he heed the warning of the great Professor Feynman: when you convert even the finest scientists into policy advocates, you risk ending up with what he called cargo cult science? Should we not leave advocacy to politicians and ask scientists for balanced advice?

My Lords, we are grateful to the Office for Statistics Regulation for its challenge; its points were perfectly reasonable and we take them on board completely. However, I reject the characterisation made by my noble friend and his suggestion that the modelling is either political or erroneous in some way. I remind him that, in January, the modelling showed that the epidemic in China was considerably larger than anything reported at the time. In February and March, we used data from the “Diamond Princess” and elsewhere to show how the threat of Covid was much larger than had previously been understood. In March, we showed that the epidemic in the UK was doubling every three to four days, allowing us to make the difficult decision to lock down. Throughout the spring, the modelling demonstrated that half the UK had not been infected, as previously thought.

In mid-September, the modelling showed that we were at the start of a second wave, despite those who said that there was no evidence of it. It also showed that the uptick in cases involving younger people would spread to older adults and, as a result, into healthcare. Most recently, the six-week projections of SPI-M that were produced throughout October, based on contemporary trends, have been remarkably accurate at assessing the trajectory of hospital admissions and deaths.

My Lords, the fact is that the public have lost trust in scientists and science. They lost trust in government long ago. Is it not time for the Government to ask the Royal Society to carry out a thorough check and review of every statistic released by SAGE or any other government adviser so that we can be sure that the statistics are presented properly, are sound and are not exaggerated so as to mislead the public?

My Lords, I am grateful to the Royal Society for its involvement in much of the work that we are discussing; it is a key contributor to some of the scientific thinking and modelling. As for public support, I remind the noble Lord that there is enormous public support for the measures introduced by the Government: in fact, more people think that our measures have not gone far enough than support them.

My Lords, the same level of regional and cluster detail is needed for the Covid-19 status of residents in care homes as for those in the NHS. When do we expect to have this level of detail for care homes? Can the Minister tell the House how many people are currently resident in English care homes with Covid-19 and what level of confidence the Government have in official statistics on that subject?

My Lords, the noble Baroness is right that, statistically, care homes present a unique challenge. There are more than 15,000 care homes, many of which are not plugged into day-to-day statistical canvassing and, therefore, knowing exactly what happens in every care home every day is a particularly large challenge. However, we have thrown an enormous amount of resources at that problem, and our understanding of the care home situation in relation to Covid is much better than it was. The precise statistics she asks for today are not at my fingertips, but I would be glad to write to her with a number.

My Lords, would the Minister agree that it is an extremely serious matter when the statistics authority criticises government advisers’ use of statistics? If the public are to accept lockdown and all the restrictions involved, they need to have confidence in the statistics and that they are not speculative. Would the Minister agree that the graph with four scenarios for daily deaths from Covid, rising to 4,000 a day—a rate that exceeds that of Brazil, which has three times our population—should never have been shown at the Prime Minister’s press conference? If he does not agree with that, why was it subsequently modified?

My Lords, my noble friend is entirely right: statistics are critical and very important to public trust. No one takes them more seriously than this Government. However, I remind him that it was not the statistics that the Office for Statistics Regulation expressed concern about; it was about material being used in press conferences that has not been published at the press conferences as they happened. That was a function of the speed at which that press conference was turned around, but he is entirely right that that chart had a presentational error in it, which was corrected. It was published as a result of the publication of the data behind it. I reassure him that the data upon which decisions were made and the data that went into the central case of that chart was correct, and the fact that we have changed it demonstrates that we are committed to transparency in all these matters.

My Lords, my question is about when Her Majesty’s Government will make accessible communication a priority. The Prime Minister’s press conference was like a scientific symposium, except that the slides were presented too quickly, with too much information. It felt as if we were being blinded by science. Does the Minister agree that providing information that is accessible to all viewers would be a more effective public education strategy, and that that means using everyday language?

On the manner in which the information was delivered, I take the noble Baroness’s comments completely on board. While it is not my role to be in charge of the presentation of No. 10 presentations, I think a lot of people would agree with her that there were a lot of slides, which were very detailed and not all formatted for the TV screen. However, we are trying our hardest to share with the public as much of the insight and science as we possibly can, and we are trying to hit that balance between too little and too much information. We are trying to publish data as soon as it can be reasonably verified. There will be some scratchiness around the edges on that, and I take the noble Baroness’s points about last Saturday completely on board. However, the commitment to transparency and open debate on these issues is sincere.

When will the Government start sharing data and having meaningful discussions before decisions are taken? Given that public confidence in these decisions is crucial for them to work, will the Government start working with the opposition parties, which they expect to—and which have—supported the lockdowns and proposals, as Keir Starmer and others have been offering for months?

The noble Baroness is right to pinpoint the sharing of data as being very important, and we have been as open and transparent as we can be. We publish an enormous amount of data. Just before this debate, I tweeted three of the main portals to the data, which there is not only an unprecedented quantity of but which is more up to date than could reasonably have been expected a few months ago, when such data was not available. Some of these decisions are made extremely quickly because the data changes so quickly. Sometimes, one believes that we are on track for one thing, and then the virus changes course and we have to change our policies accordingly. That is simply a fact of the challenge of fighting this virus: speed is of the essence, and sometimes it has been extremely difficult to do the kinds of consultation that the noble Baroness quite reasonably describes.

My Lords, my noble friend the Minister just said that the data changes quickly. Does he accept that all three datasets published towards the end of last week on reliable information on the number of positive cases, which is to say those of the Office for National Statistics, the government dashboard and King's College London’s COVID symptom study, all point to the second wave having already peaked and being on the way down, unlike the faulty models used to justify lockdown last weekend? Does he accept that this gives the Government every reason to pause the decision to impose a national lockdown and reconsider it?

The noble Viscount and I have corresponded on this matter. I do not accept that they suggested that the number of admissions was on the way down. Undoubtedly, the rate of increase has decelerated, but a lengthening doubling time is not the same as a halving time. The doubling time for hospital admissions was eight days at the start of September, 14 at the start of October and 20 days at the end of October. That is a slowing down of the increase, but it is not the same as a decrease.

My Lords, can the Minister tell us what analysis the Government have made of the root cause and relationship to lockdowns of some 20,000 excess deaths unrelated to Covid-19 during the first wave of the pandemic, including suicide, bacterial sepsis, cardiovascular disease, cancer and other conditions? Should the Government not be publishing regular analyses of the overall harm and not just contested projections of 4,000 deaths each day or the other figures that the noble Lord, Lord Lilley, and the noble Viscount, Lord Ridley, have just mentioned?

The noble Lord makes a very good point. We do publish data on excess deaths, which is available on the PHE website, but he is entirely right: we are deeply concerned not just about the Covid deaths but the impact of Covid on others who may be seeking to access the healthcare system. That is why we made the very hard and tough decision to lock down before the NHS was put under too much pressure, and it is why we have made the commitment to keep the NHS open during this lockdown in order to manage down that excess deaths figure to which he refers.

My Lords, given the leak, there must have been an element of rush for the Saturday press conference: 4 pm became 7 pm; I suspect a degree of panic. The leaker—and Gove, Hancock and Cummings must be suspects—is the cause of this. The refusal to accept supplementary questions at the press conference also contributed, because no sources were given, and the small print showed that the information was from early October. Therefore, in future, please could we have a Spiegelhalter kitemark on graphs at press conferences?

We all respect the word of Professor Spiegelhalter, who is a great man, but we have instead the Office for National Statistics and the publication of the graphs and the data behind them. I would be glad to send to the noble Lord a link, both to the slides and the data behind them, so that he can check them out for himself.

Communications and the manner of briefings set a benchmark of standards in these challenging times. So why are the Government insensitive to the needs of those who are hard of hearing or sight by not making the Prime Minister’s and other briefings appropriate?

I will take on board the comments of the noble Viscount. We try to make our briefings as accessible as possible. The point he made is perfectly reasonable. Let me look into whether there is more we could be doing and talk to the stakeholders involved about whether we should be doing more.

Data should always make trust greater. Covid-19 appears to have more of an effect on BAME communities than their white counterparts. What further research are the Government carrying out in relation to the data to find out the reasons why?

The noble Lord is entirely right. The concerns we have for disadvantaged groups and those of an ethnic background are deep and sincere. That is why we have a large programme of work, sponsored by the NIHR, looking into a variety of different research projects to understand the behaviour of the virus and why it hits certain groups particularly hard.

Could the Minister clarify whether or not Professor Neil Ferguson, who has given such misleading forecasts, was involved in the preparation of the charts and graphs used on 31 October? Not only were they out of date, they were so inaccurate that the question arises whether those involved in the preparation of the material paused to consider if what they had produced might be badly misleading.

My Lords, I am afraid I do not know the precise roles of individual academics in the preparation of those charts. I am happy to go back to the department to see if I can find out, and will reply to the noble Lord.