Follow the science, just not that one! Inequality, bias and UK science advice.

The long awaited public inquiry to examine the UK’s preparedness and response to the Covid-19 pandemic took a positive step forward on Friday 11th March. Inquiry chair, Baroness Heather Hallett who was appointed in December 2021, launched a public consultation on the Inquiry’s [Draft] Terms of Reference which will set out the scope of the committee's investigations. The draft terms include some important questions and areas of focus, including the communication and use of evidence, protecting the clinically vulnerable and how decisions were made. In our chapter for Being Human During Covid-19, we address some of these issues, thinking through how epidemiological models were used in the formulation of science advice, and how that advice was then used to justify policy.

In this post, we focus on a part of this story that has gone comparatively underreported within the avalanche of media coverage on Covid science advice, but that we argue should be front and centre within the inquiry: that the biases that help to maintain social inequalities may also be present in science advice, and how we can help to address this by rethinking the use of models for policy.

All in it together?

Although the pandemic has been felt by all, its effects have not been felt equally across society. When well-known links between inequalities and poor health re-emerged in hospital data in April 2020, existing biases went unchallenged as the government failed to act. Prompted by signals emerging from hospital patient data, SAGE assigned the science advisory group “COVID-19 Clinical Information Network (CO-CIN)”, with investigating the disparities in COVID-19 case numbers and outcomes across ethnic groups. The reports that followed from CO-CIN and Aala et al, confirmed the disproportionate impact of COVID-19 among Black, South Asian and minority ethnic patients. Aala et al., suggest a range of causal factors in need of further investigation including comorbidities and social deprivation, which are consistently higher in minority ethnic populations, and the wider spread of COVID-19 in ethnically diverse regions. However, minutes from SAGE meeting 29 reveal an aversion to acknowledging inequalities such as deprivation as important, focusing instead on comorbidities, namely diabetes, as the sole explanation for unequal outcomes despite the data adjusting for comorbidities and still finding ethnicity to be a significant independent risk factor for South Asian patients. 

By explaining the disproportionate impact of COVID-19 on marginalised groups through comorbidities, to the exclusion of other suggested risk factors and the evidence base on the social determinants of health reflects pre-existing biases present in public discourses which are thought to produce and sustain health inequalities. For example, narratives in which health inequalities are blamed on patients’ lifestyle choices or cultural makeup have become normalised , despite research confirming that comorbidities, like Type 2 Diabetes Mellitus, are strong indicators for broader inequalities sufferers may be living with. Intended or not, use of a comorbidity centred narrative legitimised the presence of unequal COVID-19 outcomes for advisers who rapidly moved on to other areas of discussions and enabled ministers to invisablise inequalities within a public narrative of being ‘all in it together’, rather than attempting to address them. To date no direct political action to address the exacerbation of inequalities has emerged, even following Public Health England’s disparities report published in June 2020, as well as a distinct absence of inequalities as a variable for models to account for. We are not in a position to know the thought processes of SAGE participants during the 29th meeting, but the processes by which evidence is included or excluded from assessments should be a central element of the inquiry, with a particular focus on how these processes reflect pre-existing biases. 

Expert accountability

A common criticism of SAGE, particularly in the early months of the pandemic, was over its perceived lack of independence, to the extent that ex-Government Chief Scientific Adviser Sir David King launched the rival Independent SAGE that trumpeted its transparency and distance from government. However, the events we describe above occurred entirely within the science advice system, apparently prior to the involvement of politicians. So the issue runs far deeper than misplaced concerns over the permeability of the science-politics ‘boundary’. Biases are inevitable, particularly in narrowly-constructed groups of science advisors where perspectives from marginalised publics and “marginal” academic disciplines remain woefully underrepresented. The challenge is to mitigate these biases by including a wider range of expertise within science advice, including the valuable knowledge a diverse range of publics could bring to the table.

Opening up science advice to greater public scrutiny and challenge in ‘real time’ could enable new policy questions to be asked and reframe problems to prevent inequalities from being an afterthought. Margaret Thatcher’s oft-quoted statement that “advisers advise, ministers decide” has proved to be somewhat misleading. Expert advisers also make many decisions about the direction of their research, the questions they address and the ways in which evidence is assessed and presented. These decisions are not made according to some abstract scientific authority, but according to the values - and biases - that experts hold as human beings. This simple truth obligates experts to be accountable not only to their own scientific communities, but also to wider publics as a matter of democratic necessity. Accountability which, as we argue in the chapter, must also extend to the tools scientists use to ‘do’ science, particularly science created in close proximity to political power, such as epidemiological models.

These models could be used more creatively as part of public discussions about policy. Rather than their role during Covid, as black boxed arbiters of ‘gold standard’ truth, we suggest that they could be used to generate public discussions about the values and normative positions that underpin the models, which risk being overlooked in policy deliberations. What input could and should wider publics have shaping the questions that models are used to answer, such as the potential for Covid to further exacerbate inequalities? Such an approach to science advice would have to be more comfortable with public reasoning than private consensus-making. This cosmopolitan approach to the use of science for policy moves beyond familiar arguments of ‘ building public trust in science’ to consider how we can create space to discuss and incorporate public values, concerns and visions of the public good. From climate change, the Mackay Carbon Calculator provides a glimpse of what a cosmopolitan approach could look like, opening up energy models for public use in a way that can spark discussion over policy priorities.

We are hopeful that the UK’s impending public inquiry will be the venue to address many of the issues we raise, such as accountability, transparency and inequalities. In doing so, the inquiry should incorporate diverse perspectives on the pandemic as a means to comprehensively scrutinise the relationship between science and policy, including, as outlined in the terms of reference, the experiences of families most affected by the ongoing crisis.

50% off (£45.99 £22.99) when you order from Bristol University Press. Use code POB 50 by midnight on 1 October 2022 at: https://bristoluniversitypress.co.uk/being-human-during-covid-19