Article Text

Download PDFPDF

The perils of a broad approach to public interest in health data research: a response to Ballantyne and Schaefer
  1. Norah Grewal,
  2. Ainsley J Newson
  1. Faculty of Medicine and Health, Sydney Health Ethics, The University of Sydney, Sydney, New South Wales, Australia
  1. Correspondence to Norah Grewal, Faculty of Medicine and Health, Sydney Health Ethics, The University of Sydney, Sydney, NSW 2006, Australia; norah.grewal{at}sydney.edu.au

Abstract

The law often calls on the concept of public interest for assistance. Privacy law makes use of this concept in several ways, including to justify consent waivers for secondary research on health information. Because the law sees information privacy as a means for individuals to control their personal information, consent can only be set aside in special circumstances. Ballantyne and Schaefer argue that only public interest, and only a broad conception of public interest, can do the special ‘normative justificatory work’ to override consent requirements. Other, similar-sounding concepts, such as public benefit, public good and social value, also provide useful services. But none more so than public interest. In fact, they argue, public interest is the superior concept precisely because it can capture those concepts as well as a range of other interests. Our response focuses on this claim. We argue their strategy is not as promising as it might first seem. Ballantyne and Schaefer construe the important role that public interest plays in this context as their endpoint. They claim that unless the concept is open and content-rich, it will lose some of its importance. But by refusing to place limits around it, their inquiry leads us back to a catch-all concept that lacks clear focus or meaning. In reply, we argue that, for practically minded theorists, a narrow conception of public interest is more useful. Further, the narrowing of public interest in this context can be achieved by first analysing it in its legal, rather than ethical, sense.

  • informed consent
  • public policy
  • regulation
  • research ethics
  • law

Statistics from Altmetric.com

Introduction

Ballantyne and Schaefer discuss several problems that can arise when there is a lack of clarity over a key concept—in this case, public interest.1 More specifically, they consider how public interest should be conceptualised when it comes to invoking consent waivers for research using identifiable health data.1 The concept is important because research ethics committees (RECs) need a strong justification to override the requirement of consent. Overriding consent involves an ‘intrinsic privacy violation’ in limiting an individual’s claim over their health information.2

Only the concept of public interest can provide the requisite ‘normative justificatory force’.1 Notably, Ballantyne and Schaefer choose to conceptualise public interest broadly. They contend that this breadth is a key reason to prefer public interest over its alternatives—it enables both the positive and negative implications of the proposed research to be captured, as well as a wide range of other ethical considerations. Throughout their paper, they develop and defend some of the attributes that such a concept of public interest needs to fulfil its important normative justificatory function.

We agree with Ballantyne and Schaefer that the concept of public interest needs work and welcome their in-depth analysis. But in this response, we wish to suggest that their paper does not achieve its aim. That is, the conceptual footing for public interest in health data research remains boundless. Instead, we contend that a narrow conception of public interest—bounded by its use in law—will be more viably operationalised by RECs.

Conceptual clarity and public interest

Ballantyne and Schaefer argue that too many similar terms, used interchangeably, sometimes even within the same regulatory document, risks a conceptual vacuum. This theoretical uncertainty also has practical consequences. RECs must sift through these overlapping terms to determine whether granting a consent waiver is ethically justifiable. This lack of clarity both gives rise to differing interpretations and uncertainty within RECs and endangers public trust in the process of ethics review. Without conceptual clarity, RECs risk being inconsistent and lacking transparency in their decisions.

Thus, the first task to alleviate the burden on RECs is to choose the right term and adopt it uniformly. Ballantyne and Schaefer start by explaining the differences between four commonly used concepts: public interest, public good, social value and public benefit. They go on to claim that public interest is preferable, for two reasons. First, there are deficiencies with the alternatives. Public good is too ambiguous; public benefit is too restrictive; and social value is (counterintuitively) too narrow.

The second and key reason is that public interest ‘better captures the core normative force justifying consent waivers’.1 In other words, public interest is the only term strong enough to do the important work of overriding the requirement for consent. Only a broad concept can ensure the waiver of consent is ethically justifiable, because it can capture a wide range of ethical considerations. Thus, the concept of public interest can easily capture some of the features of both public good and public benefit. Other interests it encompasses include:

‘…important freedoms and securities often called common goods: civic liberties, freedom of expression, privacy, proper administration of government, justice, national security, public health and safety, and economic well-being.’1

Further, Ballantyne and Schaefer suggest that public interest can and should go beyond well-being, welfare and health, to also allow considerations of ‘welfare, distribution, human rights, respect’.1 This is as well as accommodating ‘diverse values, commitments and circumstances’.1 Other considerations it can make room for include the ‘broad implications of allowing or preventing the [proposed] research’.1 They also argue that what constitutes the ‘public’ in public interest should be construed broadly, to include all potentially impacted individuals and groups and even ‘humanity at large’.1

They argue that when used to invoke consent waivers, a public interest criterion can and should involve weighing both the positive and negative elements of a research study. The public interest is met if the positive elements outweigh the negative, even if there is a prima facie setback to the public interest through, for example, over-riding privacy protections. It is this process of weighing and trading often competing ethical considerations that is vital and what sets public interest apart from its alternatives.

Broad conceptualisations detract from conceptual clarity

We agree with Ballantyne and Schaefer that the concept of public interest lacks clarity. But in their attempt to drive it in a practically useful direction, they have neglected some of the necessary work. Such analysis, we argue, involves more than choosing breadth. We suggest another, more thorough way to develop the concept below. But first, we consider why they might have settled on their approach.

Ballantyne and Schaefer insist on a broad conception of public interest because, in our view, they take the concept’s importance as their endpoint. They seem to argue that:

P1: The concept of public interest needs to be broad enough to include a wide range of (often competing) ‘interests, rights, risks and benefits’.1

P2: If the concept of public interest were not broad, then it would lose its importance.

P3: Public interest performs important ‘normative justificatory work’.

C: Therefore, public interest must be conceptualised broadly.

We agree that part of the task at hand is to figure out how important it is to meet the public interest before granting a consent waiver, in view of other considerations like autonomy. But conceptual clarity may remain elusive when a concept is filled with ‘interests, rights, risks and benefits’ until it reaches a preferred level of ‘normative justificatory’ importance. Indeed, Ballantyne and Schaefer’s proposal does not really differ from how public interest is currently used, where the concept seems to mean whatever the relevant regulation wants it to mean. That is, it is precisely the broadness of public interest in its current form that makes it so unclear and open to manipulation or varying interpretation. Instead, we propose that only a narrow conception of public interest can achieve the ‘clarity, transparency and consistency’ they desire.1

Further, Ballantyne and Schaefer’s account of public interest underplays the role of the law.3 Perhaps this is beyond the scope of their argument. Indeed, they note that their task is to encourage conceptual rather than regulatory harmonisation in ethical discourse and REC practice. But the requirement to obtain consent before handling health information is often a legal one. In turn, the authority to waive that requirement also often comes from legislation and guidelines, as shown in table 1 of their paper.

The link between ethical guidance and legal requirements is key. To take Australia as an example, table 1 quotes a part of the National Statement2 ,4 that interacts with other, legally binding National Health and Medical Research Council guidelines.3 So, if a ‘permitted health situation’ applies under the federal Privacy Act,4 ,5 the relevant Australian REC must also follow guidelines that the federal Information Commissioner has agreed are in the public interest.3 ,6 These guidelines themselves ‘provide a mechanism for weighing the public interest’,3 when consent is impracticable.

Many countries have data protection legislation that includes similar research provisions. Ballantyne and Schaefer’s account of the normative basis for consent waivers mostly aligns with the approach to information privacy that underpins such legislation. But in our view, there is more to it than their explanation allows. The role of public interest as an over-riding justification must be understood in its legal sense too. By not discussing it in this sense, Ballantyne and Schaefer underestimate what is at stake. It is not only that handling health information for certain purposes without consent is morally impermissible.7 Such handling is also prohibited by law. When public interest performs its ‘normative justificatory work’, it often does so as the core of an exemption to otherwise unlawful actions. If left ambiguous, the concept makes it difficult for RECs to ensure consent waivers are lawful. In other words, their decisions can also attract legal consequences.

This is not another reason, however, to fortify the concept by overloading it. If the goal is to make the concept workable, then a more complete account would yield a narrow conception. This approach might include analysing paradigmatic legal cases (from several areas of law); interviewing key stakeholders to glean what public interest means in their practice; and critically comparing competing accounts and historical uses. Such work could also be useful when governments invoke inconsistent conceptualisations of public interest.8

In their final paragraph, Ballantyne and Schaefer state the next task is to ‘operationalise’ the concept of public interest by developing a framework for use by RECs.9 We suggest that getting to a framework will be difficult where the concept is as broad as the one they put forward, and where so much of its operationalisation seems to be left in the hands of RECs.

For those seeking to recover a concept of public interest that is useful, the authors’ conception leaves us where we began. We arrive back at the same boundless concept, but it is now conceived of as so important that reining it in may be less achievable than before.

Ethics statements

References

Footnotes

  • Twitter @biomedethics

  • Contributors NG conceived the idea for the response and provided an initial outline of the argument. NG and AJN collaboratively wrote and edited the draft. NG and AJN approved the final version.

  • Funding This paper is funded by The Australian Genomics Health Alliance (Australian Genomics), a project funded by a National Health and Medical Research Council (NHMRC) Targeted Call for Research grant (GNT1113531).

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; internally peer reviewed.

  • In focusing on identifiable health data, they claim that research with large volumes of such data is now at the point where specific consent is impractical to obtain and that data is difficult, if not impossible, to anonymise. For the purposes of this Response we will proceed on these same premises, although they are not uncontroversial.

  • This notion is underpinned by a popular conception of information privacy. While discussing this in depth is beyond the scope of this Response, we note this notion encompasses an assumption that certain types of information, including health information are intrinsically personal and so should be protected by privacy regulations. Unless the relevant individual consents to waive their privacy, handling their personal information is wrong – unless the circumstances also warrant overriding the requirement for consent.

  • It’s not that they overlook the law, or political institutions for that matter. For example, they do note that many governments are currently reviewing research ethics standards in response to changes in data-related technology.

  • See paragraphs 2.3.9–2.3.12 of the National Statement2 on consent waivers.

  • Briefly, for example, if the entity is an ‘organisation’, obtaining consent is ‘impracticable’, handling ‘health information’ is ‘necessary’ for research ‘relevant to public health or public safety’ and so on: s 16B(3) (as it relates to sch 1 cl 6.2(d)).

  • The ‘Section 95A Guidelines’3 specify that the National Statement2 provides ethical guidelines, while the s 95A Guidelines state what is needed to comply with the Privacy Act 1988 (Cth).4 They also state that if both an ethical guideline and a legal requirement apply, the legal requirement will prevail.

  • We accept their proposition for the sake of this Response but note that it could be challenged.

  • For example, the Australian government has recently proposed new data sharing legislation,5 tentatively named the Data Availability and Transparency Act. The Discussion Paper released in late 2019 proposes a public interest test for sharing public sector data, which rests on a consequentialist and utilitarian account of the concept of public benefit – where public benefit means ‘economic well-being’. Therefore, because it falls within the scope of economic well-being, even solely profit-driven data sharing would be considered to meet the public interest.

  • We consider a framework to be a practical tool with the primary role to aid deliberation. It must clearly specify relevant values and is ideally drafted as a series of questions. Another key feature of frameworks is that they should not require any in-depth theoretical knowledge to apply, as this knowledge has already been used to develop the framework. This approach has been discussed in depth by Angus Dawson.6

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

Other content recommended for you