Covid-19: Forecasting and Modelling

Maggie Throup Excerpts

Tuesday 18th January 2022 Westminster Hall

[The Parliamentary Under-Secretary of State for Health and Social Care (Maggie Throup)]

It is a pleasure to serve under your chairmanship, Sir Edward. I thank my hon. Friend the Member for Isle of Wight (Bob Seely) for introducing today’s debate, and all hon. Members for their thoughtful and wide-ranging contributions.

Throughout the pandemic, we have been supported by world-leading scientists, epidemiologists and modellers, many of whom have worked around the clock, often without being paid for their contributions. During the fast-moving and uncertain pandemic, that support has been critical to ensuring that the Government have access to the latest and most reliable scientific advice.

The UK is very fortunate to have such strong academic expertise to drawn upon. Without the tireless work of our country’s scientists both within and outside SAGE, the UK would have been left in the dark at many key moments over the past two years. Scientific advice from disciplines ranging from immunology through to behavioural science have all played a role. Virologists have helped us to understand how different variants behave in the body, while clinicians have advised on higher risks to different patient groups. Similarly, epidemiology and infection disease modelling help us to understand the spread of covid-19 across the population, and the impact it might have.

It is important, however, to remember that such modelling is a tool to enable Ministers to make evidence-based decisions. Modelling provides a good way of understanding the range of possible futures that the pandemic might have in store for us; a good way of identifying what will determine which of those future we could face; and a good way of exploring how different policies, rules and guidance could determine which of those futures we could face.

[Greg Smith]

Will the Minister give way?

[Maggie Throup]

I am conscious that I need to leave time at the end, but I will endeavour to get through my speech and take interventions.

It is not, however, and never can be, a crystal ball, regardless of who is doing the modelling. Models cannot perfectly predict the future, and modellers would not claim they do so. Contrary to how they may be presented in the media, modelling outputs are not forecasts, nor do they focus only on the most pessimistic outcomes. Model advice to Government is not simply a single line on a graph.

There is always uncertainty when looking into the future: uncertainty from potential policy changes, the emergence of new variants, or people’s behaviour and mixing and the changes that that brings. Central to modelling advice is an assessment of this uncertainty, what factors drive the uncertainty and how the results might change if the model’s inputs and assumptions change as new evidence emerges. As such, the modellers look at a wide range of possibilities and assumptions in order to advise policy makers on principles, not to attempt to say exactly what will happen..

[Mr Baker]

I am grateful to the Minister for giving way. She heard what I said about my conversation with the Prime Minister—it is, of course, a true account of what happened. The reality is that the Prime Minister was shown a terrifying model that subsequently proved to be wildly incorrect, but he took away freedoms from tens of millions of people on that basis. The Minister must surely agree that that does not accord with the very sensible words that she is saying. That is not what actually happened. The Prime Minister was bounced on the basis of profoundly wrong models.

--- Later in debate ---

[Maggie Throup]

I appreciate the point my hon. Friend is making, and I will come to a point that shows that models are just models; they are not predictions. Yes, they are sometimes proved wrong, but that is for different reasons. It could be that people change their behaviour as a result of the information that they get.

One example that I was about to come on to is a model in December that considered a range of assumptions for omicron’s intrinsic severity, ranging from between 10% and 100% of delta’s, in addition to the additional reductions in severity that vaccines and prior infection provide. Fortunately, we now know that severity is not at the upper end of this range, and models have been updated to reflect the evidence as it emerges. It is quite right that as new evidence emerges, models are adjusted to take that into consideration.

My hon. Friends the Members for Isle of Wight and for Penistone and Stocksbridge (Miriam Cates) raised concerns about the Government’s reliance on modelling advice, both more generally and from individual modelling groups. I want to reassure hon. Members that encouraging a diverse range of opinions, views and interpretations of the data is all part of the process. SPI-M-O and SAGE do not rely on just one model or group but look at advice from a number of independent and world-leading institutions. Robust scientific challenge has been vital to the quality of SAGE advice, with modelling papers regularly released online and the methodology and underlying assumptions clearly laid out for everyone to challenge and bring forward other evidence—it is all out in the open. Sir Patrick Vallance has said:

“No scientist would ever claim, in this fast-changing and unpredictable pandemic, to have a monopoly of wisdom on what happens next.”

As the chief medical officer has emphasised, hard data on what is actually happening to patients and to the population as a whole is an essential part of the advice given.

Modelling is a helpful tool, but it must be considered alongside what is happening to real people at home, in schools or in hospital beds. As SAGE has been so visible and transparent in its advice, some people may think that it is the only form of advice to the Government, but this is not the case. Modelling and other advice from SAGE has been invaluable during the Government’s response to covid-19, but it is only one of the many issues we consider. Modelling helps us to understand the possible risks from the spread of covid-19 but, ultimately, this needs to be balanced against other health, economic and societal impacts.

A number of hon. Members questioned the accuracy of modelling forecasts from SAGE. I would like to reiterate that such modelling outputs are scenarios, not predictions and forecasts. As such, comparisons between past scenarios and what happened in reality should be made with caution. Comparisons must be made on a like-for-like basis, and often they are not.

[Sir Edward Leigh (in the Chair)]

Can you leave a minute at the end, Minister?

[Maggie Throup]

Yes, I will.

To be frank, what we are doing in many cases is comparing apples and pears. Nevertheless, past modelling has proved remarkably accurate in many cases.

My hon. Friend the Member for Newcastle-under-Lyme (Aaron Bell) said that lessons must be learned, and lessons will be learned. The hon. Member for Putney (Fleur Anderson) mentioned that as well.

In closing, I would like to take this opportunity to emphasise just how appreciative we should be, and are, to the scientists, academics and Government advisers for all their hard work over the last two years. It was fitting to see this rewarded in the new year’s honours list. Finally, I would like to thank hon. Members again for their participation in today’s debate and the opportunity to discuss the matter further.