Providing ethics advice in a pandemic, in theory and in practice: A taxonomy of ethics advice

The pandemic significantly raised the stakes for the translation of bioethics insights into policy. The novelty, range and sheer quantity of the ethical problems that needed to be addressed urgently within public policy were unprecedented and required high ‐ bandwidth two ‐ way transfer of insights between academic bioethics and policy. Countries such as the United Kingdom, which do not have a National Ethics Committee, faced particular challenges in how to facilitate this. This paper takes as a case study the brief career of the Ethics Advisory Board (EAB) for the NHS Covid ‐ 19 App, which shows both the difficulty and the political complexity of policy ‐ relevant bioethics in a pandemic and how this was exacerbated by the transience and informality of the structures through which ethics advice was delivered. It analyses how and why, after EAB's demise, the Westminster government increasingly sought to either take its ethics advice in private or to evade ethical scrutiny of its policies altogether. In reflecting on EAB, and these later ethics advice contexts, the article provides a novel framework for analysing ethics advice within democracies, defining four idealised stances: the pure ethicist, the advocate, the ethics arbiter and the critical friend.

These novel circumstances required high-bandwidth two-way transfer of insights between academic bioethics and policy. Inspired by the idea of translational medicine, which focuses on improving the processes by which insights generated at the laboratory bench lead to benefits for patients, we describe the attempt to improve the transfer of insights between theoretical ethics and real-world ethical decision-making as translational bioethics. 2 High-quality translational bioethics requires attention to context: it cannot be presumed that the factors relevant to translation of insights from more theoretical to more practical contexts will be uniform. Academic bioethicists in states such as the United Kingdom, in which ethics advice for policy had previously been ad hoc and informal, faced different translational challenges from those in states such as Germany or Switzerland, which had long-standing and well-entrenched National Ethics Committees.
This paper is grounded in the context of ethics advice in the United Kingdom and develops as a case study the brief career of the Ethics Advisory Board (EAB) for the NHS Covid-19 App. This case study shows both the difficulty and the political complexity of policy-relevant bioethics in a pandemic and how this was exacerbated by the transience and informality of the structures through which ethics advice was delivered.
The paper then analyses how and why after EAB's demise, the Westminster 3 government increasingly sought to either take its ethics advice in private or to evade ethical scrutiny of its policies altogether and how this made the task for translational bioethics even more difficult. It argues that this closed approach to translational bioethics was itself unethical, falling short in ways that are neither explained by the inherent difficulties of applying ethics frameworks to public policy nor the additional difficulties introduced by the pandemic. While the empirical analysis focuses on a particular geographical context, the paper builds from this to a typology that has implications for ethics advice for policy in liberal democracies more generally. The typology distinguishes four idealised roles: the pure ethicist, the advocate, the ethics arbiter and the critical friend. We examine why public institutions prize ethics arbiters and critical friends and why problems occur when there is disagreement or confusion about which role individuals are adopting.  Political accountability is thus clear, but responsibility for the ethical justifiability of policies that are developed and implemented is less clear. Unless steps are taken to avoid this, responsibility for ethical justifiability may fall uncomfortably between the two stools of politicians (who own the high-level policy direction, but may not fully understand the detail) and public servants (who understand the detail, but may have limited ability to ameliorate ethical problems that are revealed by the policy's effects in practice).
This raises an important set of questions for academic bioethicists who want to contribute to ethical deliberation within policymaking. This challenge is distinct from, and significantly more difficult than, the challenge for academic scientists who want to influence the policy process through the provision of scientific evidence. Scientific evidence on the effectiveness of different means to politicians' ends appears to be much easier to assimilate to and harness within politicians' priorities than expert ethical scrutiny. This is because expert ethical scrutiny may challenge politicians' ability to frame the policy issues and construct publics and may also raise questions of democratic legitimacy by stepping into the territory of 'political judgement'.
The normative structure of democracy itself places some limits on the ways in which ethical values should be brought into the policy process. We can distinguish between two potential models for the bioethicist in public policy, distinguished by how the bioethicist situates themselves relative to broader public values. First, bioethicists could take their role to be to argue for their own value judgements within the decision-making process, regardless of the extent to which these value judgements are widely shared within society as a whole. Where this model is used as the basis for policy- to be done in response to a specific policy question and (2) whether the advice is delivered from the perspective of the public interest (see Table 1).
How to provide ethical scrutiny of public policy that is both democratically legitimate and robust has been addressed at an institutional level in many countries by the establishment of a National Ethics Committee. Such bodies have a democratic remit to perform ethical scrutiny and provide ethics advice, and the executive may also have a legal duty to seek this advice and have regard to it. 9 This provides a way of bringing ethical scrutiny into the heart of government policymaking, while also ensuring democratic  Statistics and the government's Data Ethics Framework. 16 Articulating a value framework that expresses and constrains the otherwise indeterminate idea of the public interest will be particularly important where the stakes are high, and the decisions to be taken will lead to losers as well as winners.

| ETHICS ADVICE AND THE PUBLIC INTEREST
To the extent that a government department or a public body has such an ethical framework, there would usually be a strong presumption that those who provide it with ethics advice should do so in ways that draw on and are compatible with this framework. One implication is that bioethicists who are asked to provide ethics advice to politicians or public bodies are often asked to answer questions that have been defined by others (perhaps in a way that seems too narrow) and to do so with respect to a set of values that they might, in other contexts, wish to critique. For example, a government's value framework may presuppose that supporting its domestic life sciences industry is of vital strategic importance. This may profoundly shape the questions that it seeks ethical advice on in healthcare resource allocation, perhaps focusing only on how to make access to expensive novel cancer drugs fairly available within the public system, rather T A B L E 1 A typology of ethics advice for public policy. The account that follows draws on material that has since been placed into the public domain via EAB's minutes and final report, 20 subsequent published interviews with panel members, 21 media reports and a range of contemporaneous notes.
When EAB was first convened, the government's view seemed to be that it would be the app rather than anything else that would bring Both the decision to adopt a centralised model and to trigger alerts on the basis of self-reported symptoms were ethically controversial, and both were dropped before the app finally launched. 24 While these decisions about app design followed from the government's strategy at the time, it was obvious to informed commentators that the strategy itself was very different from that recommended by the WHO, and different also from that followed in countries in which Covid had been brought under control. The WHOrecommended approach was to focus on virological testing, manual contact tracing and isolation. 25 The Westminster government's strategy presupposed that the outbreak could be brought under control without widespread testing or manual contact tracing through the use of the app alone. This created a degree of tension about the nature and scope of the ethics advice being sought.
As was discussed in EAB's first meetings, the app would be able to play the role envisaged for it in the strategy only if a number of demanding, but unproven, assumptions all held simultaneously: (1) the vast majority of those who had phones capable of running the app downloaded it, (2) the app worked well at a technical level in detecting the presence and proximity of other phones, (3) those who downloaded the app continued to interact with and to use the app over a period of months and (4) a high percentage of those who received alerts asking them to self-isolate complied. 26 However, given that EAB were led to believe that the app 'would be the leading pillar of a national strategy for Covid recovery, and would be rolled out nationally within a few weeks', 27 advising that the strategy as a whole was deeply flawed would not only have been deeply unwelcome to DHSC but was also outside of EAB's terms of reference.
EAB's initial analysis and conversations made clear that maintaining high levels of public trust in the project was a necessary condition for the success of the strategy. Without sufficient levels of trust, not enough people would be willing either to download the app, to continue using it, or to act on its recommendations.
Maintaining trust would in turn require that the project communicate in a way that established its trustworthiness. The app's proposed reliance on self-reported symptoms rather than test results created difficulties for trustworthiness, and was picked out as central not only by EAB but also by the initial focus groups run with members of the public. 28 Moreover, DHSC was far too slow to publicly articulate a rationale for why a centralised model was being adopted rather than a decentralised model. This allowed a narrative to take hold that a disproportionate amount of data would be collected-a narrative that was articulated ably and energetically by advocates for decentralised models and proved corrosive of public trust. 29 EAB was dissolved by the government in July 2020, after a number of high-profile leaks from its meetings, and some EAB members giving interviews to the press indicating that they were not happy with the accuracy of the information provided to the Board or with various aspects of the app design and testing. 30 As EAB's final report explained, the strategy for controlling Covid changed from a 'tech project' to a 'large-scale health project, in which tens of thousands of people are employed' 31 in the 3 months between EAB's first and last meeting. The failure of the initial strategy allowed a realignment to what could perhaps have been accepted initially, namely, that what was required was an integrated public health strategy, which was deeply integrated into broader NHS infrastructure. The app was then reframed as a potentially useful add-on to the Test and Trace strategy that was being led by testing and manual contact tracing. 32 When EAB was dissolved, there was no firm commitment to launch the app, but merely an aspiration that it would be launched for the winter if its value proposition could be made sufficiently strong. The app itself was then completely redesigned by a different team, using a decentralised model that was made possible by an API developed in common by Apple and Google, and was formally launched at the end of September 2020. 33 The belated pivot to a decentralised model, after two months of arguing 22 EAB, op. cit. note 20, at p.10. 23 Ibid: 10. 24 The decision to use self-reporting was driven by lack of testing capacity. It was ethically controversial, as what little research there was indicated that individuals were highly unreliable judges of whether they had Covid-19. Moreover, self-reporting meant that individuals could cause others to be required to self-isolate by reporting symptoms maliciously or frivolously. Given that those who were being required to self-isolate would be aware of the unreliability of the information that led to the alert, EAB thought it likely that compliance might be low. Use of a centralised model was ethically controversial, as contact tracing could also be done effectively in a privacy-preserving manner through a decentralised model. So the centralised approach was vulnerable to the charge that it collected far more data than was necessary. One key reason that the centralised model was proposed, which was not publicly acknowledged, was that only a centralised approach would allow the high levels of noise introduced by self-reporting to be reduced to acceptable levels  April 2020, reminding him that 'Engaging with, and accounting in a transparent way to, all sections of society for the decisions that are taken is essential for trust and trustworthiness'. 46 The letter was not answered.

| RECOMMENDATIONS AND CONCLUSIONS
Our analysis has revealed two fundamental dimensions on which translational bioethics for policy can differ, which, when combined, give us four ideal types. 49 The first dimension is whether someone aims to provide an analysis from their own perspective or from the perspective of the public interest. As we saw earlier, thinking from the perspective of the public interest will require the bioethicist not just to think from their own perspective but also that of other citizens, and may require them to accept certain legitimate constraints on the kinds of reasons that are deemed relevant.
The second dimension is whether the analysis aims to give an ethical recommendation about what is to be done in response to a specific policy question. Ethical analysis might stop short of making a specific policy recommendation for various reasons. In many cases in philosophy and bioethics, analysis aims to answer an ethical question about an idealised scenario or to establish an abstract principle, and will remain silent about the implications for real-world decisionmaking. 50 In other cases-as we saw with MEAG-an individual or group may seek to give an account of the ethical considerations that are relevant to a decision, without recommending what should be done.
Together, the two dimensions provide four ideal types: 1. The Pure Ethicist. The pure ethicist does not take on a role in public policy formation. They believe that the task of high-quality ethics research is best served in isolation from the contingent constraints of public policy: pure ethicists do not need to be immersed in how public policy operates in order to carry out good research. Hence, it is never a defect of ethical research that it has no utility or is carried out in a manner that is detached from or does not aim to engage with public policy.
2. The Advocate. The advocate seeks to engage decision-making processes with particular outcomes in mind. On this approach, the best or main role and value of ethics research lie in providing arguments for particular political causes. NGOs, whose role is to advocate for the interests of a particular group, frequently adopt this role, but it is one that may also be adopted by scholar- We put these four roles forward as ideal types, in awareness that some interventions may fall between the four roles, and that the same individual or committee may act in more than one of these roles in different contexts. Nonetheless, distinguishing these roles clarifies the terrain. Common challenges include when the same individual acts in more than one of these roles in respect of the same issue, and where there is unclarity about the role that a person or committee is being asked to play with their ethics input into a policy process.
Things are unlikely to turn out well unless the policymakers seeking ethics advice, and committees providing it, agree about whether the committee's role is that of an Ethics Arbiter, or of a Critical Friend and whether being a member of the committee is compatible with simultaneously adopting an advocacy role outside of the committee in respect of the same issues. EAB's terms of reference suggested that it was asked to provide critical friendship, but in practice, things were different. As with MEAG, it appeared that some policymakers would have been far more comfortable with EAB adopting an Ethics Arbiter role. Some EAB members were prominent privacy advocates in their roles outside of EAB and had argued publicly against the centralised approach to data collection that was at the time a key element of the app's design. This may have made it more difficult for the committee to maintain policymakers' confidence that it could perform the role of an ethics arbiter.
After the demise of EAB, the Westminster government's response was marked by how little independent ethical scrutiny and challenge, of any of our four ideal types, was incorporated into decision-making processes. Where ethical analysis was sought, it was, as we saw with MEAG, assigned an ethics arbiter role and the substance of the analysis remained largely private. In other cases, such as the consultation on mandatory vaccination of NHS workers, what was presented as public consultation was obviously tokenistic. uk. However, this dissemination is often haphazard and is unlikely to occur in a timeframe necessary for aiding a pandemic response. never makes a difference in public policy or that ethical scrutiny will inevitably be reduced to ethics-washing or worse. 53