University of Cambridge
Nearly every day there are new and often compelling statistics on COVID-19. Whether it is the seven-day rolling average of new infections, hospitalization rates, excess deaths, or vaccine doses administered, the general public has been transfixed by numbers as this pandemic unfolds.
With so many statistics available, the communication and interpretation of COVID-19 numbers is critical. Professor Sir David Spiegelhalter is a renowned statistician and Chair of the Winton Centre for Risk and Evidence Communication at the University of Cambridge. He has been communicating and explaining Coronavirus statistics with his colleague Anthony Masters, a statistical ambassador for the Royal Statistical Society, in a weekly column in The Guardian. The duo has a book coming out this September called COVID by Numbers, Making Sense of the Pandemic with Data, in which they analyze and interpret these statistics.
HVP Editor Kristen Jill Abboud spoke with Spiegelhalter recently about what the numbers tell us and how they have helped inform the public.
An edited version of the conversation appears below.
There has been so much statistical data generated about COVID. What has the role of statisticians been in responding to the pandemic?
It’s been a big time for statistics, right from the beginning. We need to distinguish between epidemiological modelling—making projections about what might happen, which are often very important in terms of decision making—and what a statistician does, which is looking at the data, at the actual things you are measuring and observing. I think statisticians have done—and I would say this, wouldn’t I—an extraordinarily good job and have never been so popular or in demand.
Statisticians have a particularly important role in that we tend not to have opinions about what should be done or about policies, and we don’t speculate about what’s going to happen either. We just try to interpret the data and point out what it doesn’t say and what conclusions we might be able to draw with some caution, and to help people understand that. That has provoked, to some extent, a conflict with the media because the media love issuing blame and asking people to speculate. What I’ve found since March 2020 is that the media has gotten better at accepting that this isn’t what statisticians do, and the general public loves it because they don’t necessarily want scientists telling them what to think.
Do you think the general public now has a better grasp of statistics as a result of COVID?
It’s difficult to say and that’s why we’ve written a book as a resource to explain a lot of this. We don’t criticize what’s been done. Our only criticism is of bad communication of statistics and people who are trying to distort those statistics to make inappropriate claims. Certainly, the appetite for good information has grown and the media coverage has improved. From my perspective, the data literacy among the general public has also improved, but I haven’t got any statistics to back that up (laughs). This is perhaps one good thing that has come out of the pandemic. It is clear that the public can be provided with complex messages if they are presented well and told well by a trusted communicator.
Was vaccine efficacy one of the complex messages?
I have spent more time explaining vaccine efficacy than almost anything else. I hadn’t realized there were so many ways in which this could be misunderstood. That issue and diagnostic test accuracy and false positives and predictive value of a positive test are also complicated messages.
It’s interesting because I’ve spent most of my career arguing against the use of relative risks and instead using absolute risk for communication because that’s what tells people what their actual risk is—out of 1,000 people this is what we’d expect to happen, rather than the notorious example of if you eat that bacon sandwich, your risk of developing cancer goes up by 20%, or something like that. But when we come to vaccines, suddenly we’re talking about relative risks—a 95% reduction in relative risk is the important factor because the absolute risk varies so much from context to context. If you’re in a place where there is no virus, then you can expect a 95% reduction from zero. It is much more reasonable to talk about relative risks in this context.
Does explaining vaccine efficacy clearly help reduce skepticism?
It takes more than a fact argument to convince people, but you need the facts. Trustworthy communication is something we’ve been emphasizing. It means not trying to persuade somebody to do what you think is right but presenting the potential benefits and harms of actions. There’s a lot of discussion about how to counter misinformation, and the first step is to find out where the problem is. Usually, it’s not just a factual issue—those are easy to address in some sense. But, of course straight misinformation needs to be countered because you need to preempt misunderstanding.
Why has it been so difficult to determine whether we will reach herd immunity against SARS-CoV-2?
The basic herd immunity equation is very simple, but it’s hugely more complicated than that. With this virus, most people don’t infect anyone at all, and some people infect a huge number of people, so the distribution of how many people one person will infect varies greatly. And this affects how you determine herd immunity or community immunity, as do many other factors as well, including things like the Delta variant, which is so much more transmissible. These are all figures calculated based on averages in a context that is enormously more complicated.
People joke about statisticians being obsessed with averages, but we’re actually obsessed with variability. And the variability in this disease—how it affects people, how infectious they are—is just staggering.
What COVID statistic has surprised you most?
The overwhelming effect of age on risk. The risk from COVID doubles for every six years older you are. This virus takes whatever weakness you have and multiplies it. We like to say COVID is a bully that picks on weakness.
Interview by Kristen Jill Abboud