Sage Journals: Discover world-class research

Abstract

This article focuses on key roles that the ill-defined concept of ‘public benefit’ plays in accessing the public health data held by the UK’s National Health Service. Using the concept of the ‘trade-off fallacy’, this article argues that current data access and governance structures, based on particular construals of public benefit in the context of public health data, largely negate the possibility of effective control by individuals over future uses of personal health data. This generates a health data version of the trade-off fallacy that enables widespread involvement of commercial actors in personal data, despite public concerns over commercial involvement in, and potential exploitation of, public health data. The article suggests that, despite ostensibly robust regulatory and governance structures, this publicly held data is potentially subject to similar logics of accumulation as seen elsewhere in the digital economy, highlighting the inadequacies of current data regulatory frameworks in the digital era.

Keywords

Public health data digital economy digital regulation public benefits public interest political economy trade-off fallacy digital governance Big Data

Introduction

Healthcare is a key industry that is contributing substantially to the global growth in data (Business Wire, 2018), and which has seen the entry of major digital technology companies in recent years seeking to pair health data with their abilities in data analytics. Against this backdrop, this article explores the mechanisms for accessing patient health data held by the UK’s National Health Service (NHS). As the largest single integrated healthcare provider in the world, its stock of patient data holds substantial value for health research and its market value has been estimated as several billion pounds (Spence, 2019). Access to such a resource would greatly enable companies to develop products that may be applied in healthcare markets globally. Yet access to this data source is also highly regulated by multiple legislative and regulatory measures that aim to provide individuals with substantial protection of and autonomy over their personal data, and which allow access to data solely for ‘public interest’ purposes.

This article employs a critical political economy lens to explore the interplay between ‘public interest’ research objectives, the UK regulatory and governance frameworks, and digital technology companies’ commercial interests in health in the case of access to NHS-held patient health data. Its core argument is that the ill-defined concept of public benefit enables the concept to be harnessed to facilitate widespread commercial access to patient health data, despite public concerns towards such involvement. As a result, this source of data is potentially subject to logics of data accumulation similar to those seen elsewhere in the digital economy.

Following the view that (big) data¹ should be studied within its socio-political contexts (Beer, 2017), this article construes of health data as a distinct source of (big) data, and connects trends in its use with the emerging literature from critical sociology, political economy of communication, and legal and ethics studies, that focuses on the socio-political impacts arising from how digitisation and digital infrastructures increasingly mediate the ways we live and work – a literature that is coalescing around what may be viewed as the political economy of the digital.² This body of work identifies several trends that increasing digitisation, data collection, and digital mediation of our lives are contributing to, including: de-democratisation, through fragmenting the public sphere (Gandy, 2017; Pariser, 2011; Vaidhyanathan, 2011; Wu, 2016); new forms of classification that entrench inequality and discrimination (Eubanks, 2018; Fourcade and Healy, 2017; Gandy, 2009; O’Neil, 2016); and new ways of enacting social control (Yeung, 2017; Zuboff, 2019). Meanwhile, the dominance of a few tech giants in digital expertise and infrastructure is contributing to new patterns of global inequalities (United Nations Conference on Trade and Development, 2019). In health, concerns have been raised over the potential ceding of democratic control and accountability as private actors become increasingly involved in public health data (Green, 2019). As such, while this article engages with health literatures on public interest and the UK data regulatory framework to pursue its analysis, it is undergirded by a political economy approach orientated towards the wider digital sector.

The article begins with an introduction to commercial interest in health data (part 1). It then moves (in parts 2 and 3) to analyse access to NHS-held public health data for research, using the concept of the ‘trade-off fallacy’ (Turow et al., 2015) to explore the extent to which the mechanisms that facilitate the sharing of this data may exhibit similarities with data accumulation trends in the wider digital economy. Originally developed in relation to consumer data collected by digital companies, the ‘trade-off fallacy’ investigates the notion that consumers engage in a rational cost–benefit analysis (a ‘trade-off’) when they give up their personal data online in exchange for benefits, such as online services and discounts. Turow et al. (2015) argue that the notion of an empowered consumer, who understands the benefits and costs of giving up their data, is in fact a misrepresentation that ‘give[s] policymakers false justifications for allowing the collection and use of all kinds of consumer data often in ways that the public find objectionable’ (p. 3).

Taking ‘public benefit’³ as the benefit that public health data is ‘exchanged for’, and which is being used to justify increased sharing of patient data, this article suggests that the health data-sharing context shares key features with the wider commercial digital economy. Part 2 explores the legitimating roles that the concept of public interest plays in moral arguments and current governance structures for the reuse of NHS-held patient health data. However, a key difference in the UK health data context is the integral role played by the public health body which holds the data; unlike typical consumer interactions with commercial firms, in which individuals directly consent to share their data (or allow their data to be ‘tracked’) through end user licence agreements, a data-holding public body effectively acts as an agent between a commercial firm and patients whose data they hold. In the case of health data there is also a moral ‘social obligation’ dimension, beyond an individual’s immediate benefit, to data sharing.

Part 3 interrogates the possibility that a rational, cost–benefit analysis ‘trade-off’ occurs in the context of public health data. Drawing on the literature on social harms arising from the digital economy, it argues that such a cost–benefit calculus is not possible, due to the ill-defined nature of the concept of public benefits, and the inherently speculative nature of Big Data analytics.

Part 1: Commercial interests in health data

Digital health has witnessed keen commercial interest in recent years with new venture capital investments of billions annually (Lovett, 2019; Thorne, 2019). Commercial activities range from consumer wearable devices such as Fitbits, to using AI to develop new pharmaceuticals, to collaborative ventures between digital tech companies and public health providers. At their core, these activities boil down to the same logic as elsewhere in the digital economy: acquire masses of data and apply analytics to it.

Drawing on Srnicek’s (2017) analysis of the digital economy, this section suggests that commercial interests in health data can be largely subsumed into two connected interests: companies’ desire to secure strategic infrastructural position, and the potential for product development through mining the data collected from digital platforms. While new products may not be developed immediately, for the largest tech companies entering the health sector likely forms part of a long-term strategy to expand access to data – and potentially secure exclusive access, thus increasing their proprietary ‘digital enclosures’ (Andrejevic, 2007). In this way, while Apple may claim publicly that they make no money directly from their freely available ResearchKit tool (Lashinsky, 2017) – which provides a platform on Apple’s iPhones and watches for medical studies to be conducted – by placing this tool in the public domain, this ensures high-quality research and related data is channelled through Apple’s proprietary devices and platforms.⁴ In doing so, these companies act as health data ‘prospectors’ (Powles and Hodson, 2017).

Srnicek (2017) argues that digital platform companies aim to increase market share, so as to secure strategic positions as an ‘essential infrastructure’ (p. 63). In this role they undergird other economic sectors by operating as a kind of (ideally monopolistic) utility. These platforms enable other organisations to outsource their IT needs, including servers, data storage, computing power, software development tools and analytics. An example is the Swiss pharmaceutical company Novartis’ recent collaboration with Microsoft to apply AI to all areas of their business from finance to manufacturing (Neville, 2020).⁵ In late 2019, Google announced a deal with Ascension, which runs 2600 hospitals in the US, whereby Ascension will begin utilising Google’s cloud data storage service and business applications known as G Suite (Copeland, 2019),⁶ and Google can access Ascension’s patient data without needing to notify doctors or patients.⁷

Google has previously provided a platform service to host patient data and conduct analytics in its ventures with the NHS (Powles and Hodson, 2017); however, the NHS has since developed its own platform, NHS App – along with a new NHSX division – to support increased digitised delivery of healthcare (Department of Health and Social Care, 2019). NHSX’s CEO Matthew Gould (2019) writes that the aim is for NHS App to provide a ‘thin platform’ that then allows others to create other applications (apps) on top of it. This approach ‘means not competing against the market and resisting the urge to build or commission everything ourselves’ – in other words, it embraces the participation of commercial and other actors, opening up access to patient health data as it does so. Other public health systems are also seeking partnerships on digital health technologies. In Canada, Health Canada is establishing a new Digital Health Review Division that aims to expedite product approval reviews for app developers, and ‘to adapt to rapidly changing technologies in digital health’ (Government of Canada, 2018).

Through monopolising the data collected, platform and app owners can enjoy (exclusive) capabilities to generate proprietary products or other services based on that data; in health this may include algorithms that help to develop new drugs and other medical services. Google’s collaboration – through its DeepMind subsidiary – with London’s Moorfields Eye Hospital, that used AI on eye scans to detect early symptoms of sight loss, has resulted in one commercial AI-based product being prepared for market (Murgia, 2019). It is not clear that such products are the main draw for commercial tech giants. Industry commentators suggest that Google DeepMind’s forays with Moorfields and the Royal Free Hospital Trust likely form part of a longer-term business strategy for Google – specifically, acting as test cases to build credibility in Google’s AI abilities (Murgia, 2019; Powles and Hodson, 2017).

However, such forays may still lead to a profitable side-business in proprietary products (whether in AI-developed diagnostic tools or pharmaceuticals). In the case of Google’s product developed with Moorfields, the hospital will be able to use it for free for the first five years, but no cost details are available for other health providers or after this five-year period (Murgia, 2019). It needs to be emphasised that the type and quality of data used is critical for optimal machine learning efficacy: ‘no matter how good the learning algorithm is, it is only as good as the data it gets’ (Domingos, 2017: 45). In other words, this potential revenue stream cannot flow without access to good-quality data. As a single integrated health provider, the NHS links together many diverse types of health data from across 70 million patients’ entire lives.⁸ Commercial arrangements made by public health systems such as the NHS thus raise concerns over the mass transfer of health data to commercial technology giants (Powles and Hodson, 2017) and the transformation of healthcare (van Dijck and Poell, 2016).⁹

Part 2: Health data and public benefit

Collection of data, including personal data, is core to the digital economy’s operations. Turow et al. (2015) explore the argument that consumers engage in a rational cost–benefit analysis (a ‘trade-off’) when they give up their personal data in exchange for benefits such as discounts and online services; this ‘trade-off’ legitimises collection of personal data by digital companies. Turow et al. conclude that ‘the trade-off argument’, which depicts an empowered consumer who understands the benefits and costs of giving up their data, is a misrepresentation. The trade-off fallacy thus highlights obfuscatory practices used by institutions to legitimate widespread collection of personal data. Citing Pollach’s analysis of websites’ data use notices (‘privacy policies’), Draper and Turow (2019) argue that these obfuscatory practices enable the parties behind the privacy policies to obtain data they would otherwise not have access to ‘if users were fully informed about data handling practices’ (p. 1831). This echoes arguments made elsewhere concerning the logic of accumulation in the digital economy, where ‘the strategy has been to collect data, then apologise and roll back programs if there is an uproar’ (Srnicek, 2017: 101; see also Vaidhyanathan, 2011; Zuboff, 2019).

This article construes of ‘public benefit’ as the incentive that public health data is ‘exchanged for’, and the concept that is being deployed to justify increased sharing of patient health data, particularly with commercial actors. This section explores the legitimising roles that the concept of ‘public interest’ plays in moral arguments and in legislative mechanisms for gaining access to patient health data. It argues that, given the concept’s ability to override other protections such as patient consent, this concept acts as de facto gatekeeper to patient data.

The moral, ‘civic duty’ appeal

Unlike other types of Big Data, health data is widely acknowledged as having the potential to benefit wider society if researchers are able to share, link and reuse them in data analytics. While health scholars have noted that a healthy population ‘is a public good’ (Green, 2019), recently the WHO’s chief scientist has claimed that health data itself is a public good (Wall, 2019). Some have suggested that individuals have an ethical duty to share their health information for research on the basis of ‘solidarity’ or ‘social contract’ arguments (Ballantyne and Schaefer, 2018; Prainsack and Buyx, 2017; Wellcome Trust, 2015). This perspective is shared by researchers of AI and Big Data, including those with concerns about widespread data collection but nevertheless broadly believe that health data should be shared in order to reap society-wide benefits (see O’Neil, 2016; Schneier, 2015).

There is little doubt that substantial benefits may stand to be gained from applying analytics to large patient health datasets. An abundance of business briefings, government reports and peer-reviewed publications discuss possible and documented real-world applications, ranging from clinical decision support and disease surveillance, to population health management, costs efficiencies, improving clinical trial design, and improved quality of care (see e.g., Auffray et al., 2016; Islam et al., 2018; Kruse et al., 2016; Raghupathi and Ragupathi, 2014). A 2013 McKinsey report speculated that unleashing the value of health data could reduce US healthcare costs by $300–$450 billion, equivalent to 12–17% of its $2.6 trillion annual costs (Groves et al., 2013).¹⁰ This is not to say that all applications of health data analytics will succeed and all these aspirations will be realised; the demise of IBM’s Watson (Strickland, 2019) and Google FluTrends (Butler, 2013; Lazer et al., 2014) are prominent AI failures in health that serve as reminders against technofundamentalism, or ‘the faith that technology can redeem all of our sins and fix all of our problems’ (Vaidhyanathan, 2011: 77). However, this does not deny that many (lower profile but still valuable) benefits are possible.¹¹

UK legal frameworks and governance for public health data

There are several protective barriers that researchers, including those working with commercial digital organisations, must overcome in order to gain access to NHS health data, in contrast to, for example, health-related data generated by personal consumer devices. Approval must be sought through dedicated legal frameworks and data governance structures, which are designed to allow access¹² to data solely for purposes ‘in the public interest’; the workings of these structures are delineated below.

In the UK, the legal framework for accessing NHS patient data consists of the Data Protection Act 2018, the EU General Data Protection Regulation (GDPR), and the Common Law of Confidentiality (Medical Research Council (MRC), 2018). Access to identifiable patient data for research is usually achieved by obtaining explicit (verbal or written) consent. Provisions through Article 89 of the GDPR allow specific derogations (exemptions) concerning the reuse of data for scientific research; such derogations in effect enable a wide range of reuse of patient health data for research purposes based on a one-time informed consent by individuals (Starkbaum and Felt, 2018). In the UK such uses must be ‘in the public interest’ (Global Alliance for Genomics and Health, 2019).

Access to health data without consent is possible in England and Wales through Section 251 of the NHS Act 2006 and its current regulations (usually referred to as ‘Section 251’) (MRC, 2018). Through this process, the Common Law of Confidentiality is temporarily set aside for the specific purpose applied for, although responsibilities resulting from the Data Protection Act are still applicable (e.g. the obligation to be ‘lawful, fair and transparent’). Similar access without patient consent is available in Scotland and Northern Ireland; in Scotland, ‘interpretation of confidentiality law allows the disclosure of confidential patient information to support good quality research when this is deemed to be in the public interest’ (MRC, 2018, emphases added).

The Health Research Authority’s Confidentiality Advisory Group makes approval decisions on whether access through Section 251 is appropriate, based on ‘public interest’. The final decision to release the requested data is made by a Caldicott Guardian, that is, ‘a senior person within an NHS organisation responsible for protecting the confidentiality and enabling appropriate sharing of confidential patient information’ (MRC, 2018) who chooses to allow or deny access.¹³

While in practice it can be difficult for researchers to know exactly how to navigate these different pathways to access patient data (Ford et al., 2019), the principles for access are clear. As seen above, Section 251 approval is ultimately dependent on demonstrating a ‘public interest’ for use of the data. Explicit consent itself ‘is designed to convince data subjects and public stakeholders of a pre-determined public good in research’ (Sexton et al., 2018), and reuse of patient data based on Article 89 of the GDPR is likewise grounded in ‘public interest’. As such, the basis for each type of permitted access is in each case grounded in the appeal to public interest. This has been understood as a social licence, or the notion that those conducting research have a social legitimacy, which can be withdrawn if those conducting the work are seen to not be fulfilling the conditions for the social licence (Carter et al., 2015).¹⁴ Furthermore, the Nuffield Council on Bioethics (NCoB, 2015) advocates for patients to be involved ‘as collaborators in the whole system' in the development of any data initiatives that use patient data, due to a principle of ‘[expressing] respect for them as persons who have morally significant interests’ over the use of this data (p. 91).

Part 3: Exploring the possibility of a health data ‘trade-off’

From the previous section, an implicit trade-off is derived, which is that, in the context of NHS patients and health research, health data is ‘exchanged’ for public benefits – or, the potential public benefits that stand to be gained justify the sharing and reuse of patient health data.

This section explores the possibility that a rational, cost-benefit analysis ‘trade-off’ is occurring. Two features of this possible trade-off are explored: (i) that the public is willing to allow their personal health data to be shared with commercial actors for the purposes of health research; (ii) a rational cost–benefit analysis is occurring in this exchange. The below analysis suggests neither of these hold true. Subsequently the public health data-sharing context shares two similarities with the trade-off fallacy: obfuscatory processes that serve to legitimise data transfer to commercial actors (qua research partners); and commercial actors’ acquired ability to appropriate the value of the data, as a result of the data transfer.

Interrogating public willingness to trade access to health data

A large body of research has investigated public perspectives towards commercial involvement in healthcare. It finds that, while the public approve of health research in general, they frequently distrust commercial involvement (see e.g. Clemence et al., 2013; Critchley and Nicol, 2009). In systematic reviews that evaluate studies on public perspectives towards the sharing of personal health data (Aitken et al., 2016; Stockdale et al., 2019), concerns over commercial involvement are a prominent theme. Hill et al. (2013) found that while patients approve of using patient data for health research, there was widespread suspicion of commercial uses. Likewise, a 2013 Wellcome Trust/CM Insight (2013) study concluded that, while sharing of patient data within the NHS was widely accepted, there was a strong sense that this data should not be shared with bodies outside of the NHS, ‘especially not with private companies such as employers, insurance providers and drug manufacturers’ (p. 3). Elsewhere, the British Medical Association (BMA, 2015) found that patients saw benefits for using their health data for research, such as supporting research, improving care and public health, however patients’ two biggest concerns were abuse of the data by private companies outside of NHS, and privatisation. In Canada, which also has a publicly provided health system, an Ontario-based study found general support for research based on linked administrative health data (with some conditions); however, there was mixed and more negative reaction when there is private sector involvement (Paprica et al., 2019).¹⁵ The UK’s Nuffield Council of Bioethics (2015) also notes a lack of public support for biomedical research once for-profit private companies are involved.

Despite this documented aversion, this body of research has also – seemingly paradoxically – suggested that the public does not outright reject all commercial involvement. The recognition that commercial involvement may sometimes be required to bring about public benefits leads to some tentative (conditional) acceptance of commercial involvement (Aitken et al., 2016), what has been referred to as a ‘necessary evil’ (Grant et al., 2013). On this alternate view – that the public may accept (limited) commercial involvement in health research – there may be a case that the sharing of data with commercial partners for research is justified.

This highlights the need to understand the specific questions driving these studies on patient views. Starkbaum and Felt (2018) have argued that an alliance of interests, made up of actors who share a strong interest in ensuring as much data sharing and access as possible, has resulted in a paradigm shift in the discourse from protecting privacy to promoting public health (see also van Dijck, 2014). Such shifts can be seen in the activities of the Wellcome Trust (2015), who explicitly advise researchers to ‘[s]et the conceptual framework to control the debate’ over the benefits from sharing data and ‘change the general language of debate’ from ‘default-closed to default-open’.

An illustrative case is the Wellcome Trust/Ipsos MORI’s (2016) ‘The One-Way Mirror’ report. The study takes as its starting premise the public aversion to the sharing of personal data for research (particularly with private companies): in its introduction, the report cites a 2014 ONS/ESRC-funded study, which found the public felt a lack of control over their own data and a feeling that data reuse was an invasion of privacy, and a 2014 study by the Royal Statistical Society that found in principle most people feel that sharing of health records with private companies should not happen. The 2016 Wellcome Trust/Ipsos MORI study thus highlights its findings (that 54% of people supported sharing of health data with commercial organisations for the specific purposes of health research; and that, faced with the prospect of losing out on research that otherwise could not happen, 61% of people will opt for commercial involvement) vis-à-vis these earlier studies by the ONS/ESRC and Royal Statistical Society. Within the report however, it acknowledges, ‘in principle, [workshop participants] would prefer the NHS to retain all its functions in-house rather than allowing private sector involvement’ (2016: 36).

The possibility arises that a misrepresentation of public attitudes – the acceptance of commercial involvement in research – may be occurring that is being used to justify uses of data the public in principle finds objectionable. While researchers are not marketers, they similarly have a motivation to encourage the public to share and allow broad reuse of their data, so that this data can be readily available for research.¹⁶

Can an adequate cost–benefit analysis be claimed?

A ‘trade-off’ is legitimised by the notion of a rational cost–benefit calculus on the part of the agent who chooses to give access to personal data in return for benefits. However, a key difference in the UK public health context is the integral role played by the data-holding public body, such as an NHS trust, which effectively acts as an agent or trustee between a commercial firm and patients whose data they hold. A ‘trade-off’ may be argued to have occurred at two points: first when an individual chooses to consent to their data being used in health research, second, when a governing body judges that the proposed data access is serving the public interest.

While the concept of ‘public benefit’ has been heavily employed to lobby for expanding the reuse of patient data in research (Starkbaum and Felt, 2018), health scholars and charities have argued that the concept itself remains ill-defined (Aitken et al., 2018; Scott et al., 2018; van Dijck and Poell, 2016).¹⁷ Without widely agreed or established delineations of ‘public benefits’ in the context of health, it is difficult to determine what would be included in any cost–benefit analysis – or to otherwise engage in informed debate on legitimate uses of health data. Public health and data resources have also both been described as public goods – derived from economics, a ‘public good’ is a service or product with two criteria: it is non-rivalous and non-excludable, such as clean air or national defence. However, the concept is often used more loosely in the context of healthcare to mean ‘the goods, activities or services involved are provided for public benefit’ (NCoB, 2015), irrespective of whether proprietary commercial interests are involved.

The below analysis draws on the literature on harms from the digital economy to argue that patients do not engage in a rational cost–benefit calculus to allow commercial organisations to access data; and that attempts to conduct rational cost–benefit analysis likely does not appropriately take into account the risks and distribution of economic benefits – rendering any such calculus insufficient.

Inadequacy of cost–benefit analysis: Patients

Aitken et al. (2018) argue that understandings of ‘public benefits’ should be grounded in, and align with, public views and values, including any economic benefits from projects that share or reuse patient data, whether commercial actors are involved or not.¹⁸ This stance aligns with the notion of researchers’ ‘social licence’ (Carter et al., 2015). Due to the NHS’s status as the world’s largest single integrated system of patient health records, the issue of who stands to benefit (and how) from this ‘asset’ is arguably a keen object of public interest. However, the research on UK patient views indicates that the public are often unaware of possible secondary uses of their health data (see e.g. BMA, 2015). In their workshops on patients’ views of the possible benefits from reuse of personal health data, Aitken et al. (2018) note, ‘no one spoke of societal benefits in terms of economic benefit’, despite the European Commission’s description of data as ‘the lifeblood of the global economy’ (p. 10).

This suggests it is disingenuous to argue that a rational cost–benefit calculus is occurring on the part of patients. Although the patient views literature has suggested that the public may tentatively accept (consent to) commercial involvement ‘for research purposes’, this belies the nature of the regulatory mechanisms, broad consent and data analytics. While consent ‘is designed to convince data subjects and public stakeholders of a pre-determined public good in research’ (Sexton et al., 2018), the key attribute (and problem) of broad consent is that future uses (and subsequent benefits) of data are as-yet unknown. In health research, this is the explicit aim of broad consent: i.e., to open up all possible routes to develop new treatments or costs efficiencies.

There are clear affinities between health research objectives and the profit-seeking strategies of Big Data. In Big Data analytics, ‘potential uses of the information are basically limited only by one’s ingenuity’ (Mayer-Schönberger and Cukier, 2013: 96). Yet analytics open up data for all ends, not just ‘noble’ ones, because investigations based on Big Data analytics are ‘structurally speculative’ and we cannot know beforehand what patterns might be unearthed; in fact the ‘real prize’ is the generation of unintuitable correlations that you would not have even thought of beforehand (Andrejevic and Gates, 2014: 187; see also Domingos, 2017). In this new paradigm, data collectors do not want to be limited in how they use the data: ‘function creep is not ancillary—the creep is the function’ (Andrejevic and Gates, 2014: 189). So while a health research team might, say, seek insights to help pinpoint interventions within a broader objective of tackling health inequalities, for its commercial partners, the same project may generate insights that support revenue-generating ends.

Discussing the inadequacy of consent by itself to control uses of one’s data in the digital age, Solove (2013) notes that ‘[c]onsent legitimizes nearly any form of collection, use, or disclosure of personal data’ (p. 1880). Barocas and Nissenbaum (2014) likewise argue that if narrowness of purpose is not built-in (such as prescribing the recipients and principles of sharing data), then due to future unknown uses of personal data, so-called ‘informed consent’ effectively equates to ‘a blank cheque’ (p. 59). It is important to note here that Big Data analytics work by making inferences (about patterns and correlations) that form the basis of predictions; however, this capacity to detect subtle correlations – ‘often the very thing about Big Data that generates the most excitement’ – is ‘the same feature that renders the traditional protections afforded by anonymity … much less effective’ (2014: 56).

While in the case of NHS-held health data there are multiple legislative and regulatory measures that aim to provide individuals with substantial protection of and autonomy over their personal health data, the above demonstrates that this protection is not as robust as initially appears. In particular, while the requirement that access to data is granted solely for purposes ‘in the public interest’ appears restrictive, in practice this potentially allows expansive uses of patient data when data analytics are involved.

Inadequacy of cost–benefit analysis: Regulatory bodies and governance

The possibility of reconciling tensions between different interests and possible risks in accessing patient health data ultimately depends on governance processes. As discussed above, it is the responsibility of governance bodies and committees to determine whether a specific application to access patient data serves the ‘public interest’.

In absence of widely agreed delineations of ‘public benefits’ in the context of health, this section explores what factors may be conceived as public interests or public benefits by bodies that govern access to patient data. The published literature that has examined benefits of using data analytics in health mostly considers these benefits in terms of clinical and administrative decision-making (Islam et al., 2018), i.e. improving direct care and health-system improvements. This aligns with benefits from health data identified by the UK’s Academy of Medical Sciences: identifying causes of disease; identifying effective treatment; monitoring public health; protecting patients and the public (in terms of safety of medicines, vaccines, or in relation to environmental issues); and evaluation of health services (Home Affairs Committee, 2008). However, improving direct care and system efficiencies may not be the only factors that are, or should be, considered in any cost–benefit analysis of granting access to public health data – and these do not begin to consider any potential risks from reuse of health data.

In a 2015 report, the UK’s Nuffield Council of Bioethics sought to identify the range of normative and empirical benefits and risks from the sharing, linking and reusing NHS patient health data – directly considering these in the context of ‘public interest’. Due to the Council’s position as a premier independent bioethics advisory organisation that has considerable influence on public policy and bioethical debate internationally (Chan and Harris, 2006), the benefits and risks they identified are summarised here and critically interrogated using a political economy perspective.

The Nuffield Council report views the public interest as being served in three dimensions by the increased sharing, linking and reuse of public health data: efficiencies in healthcare (e.g. reducing administrative costs); new knowledge and evidence base (from which to develop new treatments, improve care, and inform policy); and innovation, particularly through private-sector collaboration, to generate economic growth. The report emphasises that the key benefits being pursued through reuse and sharing of health data ultimately derive from ‘exploiting the value’ of the data (NCoB, 2015), and that innovation and economic growth benefits were determined by the government to be a priority.

There is a gap, seemingly, between what features in the published literature and thus promotes public legitimacy of data sharing – namely efficiencies in healthcare systems and improved evidence base, both leading to improved direct care – and the government priority of stimulating economic growth through maximising value from NHS data, particularly through private-sector collaboration.¹⁹ Acknowledging the lack of public support for biomedical research when for-profit private companies are involved, the Nuffield Council of Bioethics (2015) report notes with frankness: ‘[t]he current reality of medical research is that it relies upon clinical and commercial research collaborations and partnerships to develop innovations for the health care system’ (pp. 90–91). This highlights a potential gap between public understandings of public benefits from health data, as illustrated in the literature on public views on public health data, and understandings of public benefits from the perspective of governance policymakers.

The range of possible harms from the sharing and reuse of health data are subsumed with three categories in the report: misuse of data (leading to harms from detriment to health, loss of privacy, financial loss, reputational damage, stigmatisation and psychological distress);²⁰ potential discrimination (including targeted advertising, differential pricing that compounds social disadvantage, and discrimination in insurance and employment); and ‘state surveillance’ of citizens. The report recognises that the first two categories view harms through the perspective of individual rights, which is the definition of harm used by the ICO; this excludes many incidents that would likely be considered harmful by data subjects (2015: 39).

This is not to say that individual rights-based harms are not significant; however, other types of harms, such as collective and cumulative harms, cannot be registered.²¹ In contrast to the Nuffield report, within the literature on harms deriving from the digital tracking (surveillance) of individuals there is an emphasis on wider societal harms. This literature highlights the inability of current regulatory frameworks to capture the societal harms arising from the collection and exploitation of data in the digital economy; drawing on this literature, key limitations are identified in the Nuffield Council’s analysis of benefits and risks.

Popescu et al. (2017) argue that regulatory frameworks that focus on individual harms, particularly qua individual privacy violations, are limited: ‘the rational calculus of privacy at individual levels is usually skewed because the benefits from more disclosure … are immediate and observable, whereas the potential harms from privacy violations, when articulated, are neither easily quantifiable, nor indeed causally tied to a specific instance of data collection’. A recent report jointly undertaken by the British Academy and the Royal Society (2017) points to such potential governance weaknesses in the UK landscape, arguing: ‘there is a clear case for [the creation of] a single body to provide effective stewardship of the data governance landscape’ (p. 58), and which can thus take stock of broader trends and ‘anticipate the future consequences of today’s decisions’ (p. 60). This echoes views from an earlier report by the Nuffield Council into then-emerging new biotechnologies, which suggests that a key consideration for governance is the potential for new technologies ‘to affect social relations and to shape the conditions of common life in non-trivial ways, potentially changing the future options available to all in ways that may favour only some’ (NCoB, 2012: XX, emphases added).

In construing of discrimination as occurring from one-off events, as the Nuffield report does, this diminishes the recognition of substantial cumulative impacts arising from individual ‘legitimate’ uses of profiling (Gandy, 2009; O’Neil, 2016). And while ‘state surveillance’ is noted as a category of harm in the Nuffield report, the report does not flesh out the concept (it simply references the Snowden revelations). While surveillance itself is a broad phenomenon,²² ‘Snowden’-style surveillance typically refers to covert surveillance conducted for reasons such as identifying terrorism (see e.g., Schneier, 2015); from a health perspective, a ‘state surveillance’ concern may be the facilitation of biopolitics through health promotion strategies based on tracking health-related data (Lupton, 2012).

Within the Nuffield report, ‘processing [of] data without consent’ is noted as arguably the most contentious of the harms identified. ‘In a scoping review commissioned to inform the Nuffield Council’s report, this harm is taken as ”the use or proposed use of health or biomedical data against the stated objections of the individual(s); without fair notice; without obtaining consent; or, in cases where consent was refused” (Laurie et al. 2014: 116).’ The authors note that these would include cases where individuals did not wish the use of the data but the use was nevertheless lawful. Yet, as the above analysis on the mechanisms to access patient health data shows, the processing of data without consent is still possible in the name of ‘public interest’.

In their work on distributive harms from the digital economy, Popescu et al. (2017) further identify exploitation as a key class of harm (where individuals are either unable to fairly benefit from the economic value of their data, or are not in a position to trade it). This arises from the potential disparity in ability to exploit the value of data when one party occupies a ‘structurally weaker’ position (see also Andrejevic, 2014). In the context of personal health data, individuals (and public health systems in poorer countries) are less able to benefit from the value of the data, due to lack of expertise in data analytics and the technical data infrastructure. The largest commercial tech giants already have a monopoly on both infrastructure and expertise in data analytics (Zuboff, 2019). Worryingly for the public sector, tech giants are increasingly gutting life science researchers from top public research institutions (McBride and Vance, 2019), while one investigation found that UK universities had lost a third of their leading machine learning and AI experts to Silicon Valley’s tech firms alone (Boland, 2018). Meanwhile, the top four tech companies not only dominate their sector but are the world’s most valuable companies by far (PwC, 2019).

Others have noted that roll-out and application of AI technologies in healthcare will be built on top of the existing (unequal) landscape of healthcare delivery, potentially perpetuating healthcare inequities due to who has the necessary infrastructure and who does not (Green, 2019; Panch et al., 2019; Sharon, 2016). This disparity over resources is not insignificant. Wilbanks and Topol (2016) note, ‘the market value of Internet-enabled devices that collect and analyse health and fitness data, connect medical devices and streamline patient care and medical research is estimated to exceed US$163 billion by 2020 […] Such a tsunami of growth does not lend itself to ethically minded decision-making focused on maximizing the long-term benefits to citizens’. The danger of commercial actors exploiting the data they hold for gain has been highlighted by a McKinsey report into the impact of Big Data on healthcare:

For example, owners of MRI machines, looking to amortize fixed costs across more patients, could be more proactive in identifying underserved patients and disease areas. If they use the relevant data to convincingly market their services, regardless of clinical need, patients could end up pursuing and receiving unnecessary MRIs. Taken to an extreme, this strategy could ultimately destroy healthcare value, since payors would be spending more on patient MRIs but patient outcomes would not necessarily improve. We see such risks as real and possible unavoidable. (Groves et al., 2013)

The (im)possibility of an adequate cost–benefit analysis

As the literature on harms from the digital economy emphasises, it is difficult to make distributive risks tangible. The weaker relative position of individuals and the public compounds the difficulties in understanding non-immediate harms.

The above analysis highlights gaps between public versus government or researchers’ understandings of ‘public benefit’, particularly with respect to potential economic gains. The failure to delineate the possible benefits and harms from the sharing and reuse of publicly held data, while appealing to the same concept of public benefit in order to promote the sharing of this data, results in obfuscatory practices of data collection. Far from an unwillingness to reveal the exact purpose of a specific research project, a key issue is that big analytics cannot specify an end objective in advance; the precise virtue being harnessed in health data research is data analytics’ open-ended quality. Given the policy priorities of deriving economic benefit through exploiting the value of this data, particularly in collaboration with commercial companies, the result is that the term ‘for the specific purpose of research’ does not preclude the possibility of commercial gains ensuing from these research collaborations (as seen in the case of Google’s research ventures with the NHS).

The ‘trade-off fallacy’ study (2015) concludes that consumers experience ‘futility’ in their attempts to control how their personal data is used. It is not clear yet that UK patients feel a sense of futility with respect to their personal health data; however, this is a strongly negative prospect. Instances where the public belatedly discover that their health data has been used or shared in ways they find unacceptable can substantially erode public trust and increase overall cynicism in both the healthcare system and public research (Carter et al., 2015).

Conclusion

Understanding what is meant by ‘public benefit’ stands to be an area of contestation as the ‘boundaries of acceptability’ around the uses of public health data are negotiated (Aitken et al., 2018). This analysis aims to contribute to this issue through exploring the role ‘public benefit’ plays in facilitating access to NHS-held patient health data in the UK. It finds that, despite the legislative and regulatory measures to protect and prevent misuse of personal data, and promote individuals’ control over their personal data, the ostensive protections provided by these measures are largely negated by construals of public benefit that legitimate broad access.

The article argues that while research uses of health data lean on the social legitimacy of ‘public benefit’, by leaving this concept (and the means through which research ends are achieved) nebulous, the situation points towards the generation of a health data version of the ‘trade-off fallacy’. In this scenario, despite public aversion to commercial involvement and fears over ensuring harms, in accepting the social need to improve public health through increased sharing and reuse of their personal data, individuals are tied into accepting any subsequent commercial involvement and financial benefit that may derive from the pursuit of important ‘public benefits’. The range of possible economic benefits remains ambiguous – potentially leading to outcomes that may be counter to public values and interests. This is not to suggest any malintent is being pursued by governance bodies and committees; rather, if governance aims to assess each case on its intended health benefits and the harms directly deriving from that single case (particularly through the prism of individual rights), then this remit limits the ability to take other harms, such as cumulative effects, into consideration.

This article seeks to connect these issues in uses of public health data to the wider digital economy. While Turow et al.’s original formulation of the trade-off fallacy was developed to generate new empirical research on consumers’ views regarding sharing data, this article instead works through existing empirical studies of patient attitudes to explore potential similarities in how personal data is being shared in different spheres of the data economy; in doing so it aims to extend the deployment of the concept, in a novel way, to a new area outside of the individual consumer–corporation relationship. In particular, despite ostensively robust and hard-fought protections through legal frameworks, such as the 2018 GDPR which is widely regarded as among the most robust data laws globally, public health data is likely still subject to the same conditions of data accumulation elsewhere in the digital economy, involving corporate strategies of obfuscation and the inability of individuals to adequately control their data. This identifies particular challenges for regulators and policymakers, who are increasingly recognising their inability to regulate effectively in the digital era, given the rapid convergence in telecommunications, markets and digital platforms that is blurring traditional boundaries between markets, consumers and producers (Organisation for Economic Co-operation and Development, 2019). In contrast to the relationship between consumers and corporations, a ‘health data trade-off fallacy’ raises questions about appropriate data stewardship by public bodies that have a remit of ‘public interest’, and shifting relations between individuals, government bodies and the corporate sector. While consumers can seek to withhold their data from companies they disapprove of by avoiding their consumer services, it is patently not viable for individuals to avoid healthcare. Combined with the increasing use of health data and reliance on digital infrastructures for healthcare delivery, this raises multiple concerns regarding accountability, democratic control, and new social inequities that need to be addressed in regulation and governance frameworks.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Sarah Cheung

Notes

References

Aitken

de St. Jorre

Pagliari

, et al. (2016) Public responses to the sharing and linkage of health data for research purposes: A systematic review and thematic synthesis of qualitative studies. BMC Medical Ethics 17: 73.

Aitken

Porteous

Creamer

, et al. (2018) Who benefits and how? Public expectations of public benefits from data-intensive health research. Big Data & Society 5(2). Available at: https://journals.sagepub.com/doi/full/10.1177/2053951718816724

Andrejevic

(2014) The big data divide. International Journal of Communication 8: 1673–1689.

Andrejevic

(2007) Surveillance in the digital enclosure. The Communication Review 10: 297–317.

Andrejevic

and Gates

(2014) Big data surveillance. Surveillance & Society 14(2): 185–196.

Auffray

Balling

Barroso

, et al. (2016) Making sense of big data in health research: Towards an EU action plan. Genome Medicine 8: 71.

Ballantyne

and Schaefer

(2018) Consent and the ethical duty to participate in health data research. Journal of Medical Ethics 44: 392–396.

Barocas

and Nissenbaum

(2014) Big data’s end run around anonymity and consent. In: Lane

, et al. (eds) Privacy, Big Data and the Public Good: Frameworks to Engagement. chapter 2. New York, NY: Cambridge University Press.

Barrell

(2019) Digital health isn’t about big data – It’s about people. Pharmaphorum, 1 July. Available at: https://pharmaphorum.com/views-analysis-digital/digital-health-isnt-about-big-data-its-about-people/ (accessed 3 September 2019).

10.

Beer

(2017) The social power of algorithms. Information, Communication & Society 21(1): 1–13.

11.

Bennett

(2011) In defence of privacy: The concept and the regime. Surveillance & Society 8(4): 485–496.

12.

Boland

(2018) Britain faces an AI brain drain as tech giants raid top universities. The Telegraph, 2 September. Available at: https://www.telegraph.co.uk/technology/2018/09/02/britain-faces-artificial-intelligence-brain-drain/ (accessed 13 September 2019).

13.

Boyd

and Crawford

(2012) Critical questions for big data. Information, Communication & Society 15(5): 662–679.

14.

British Academy and the Royal Society (2017) Data Management and Use: Governance in the 21st Century. London: The British Academy and the Royal Society.

15.

British Medical Association (BMA) (2015) Secondary Uses of Healthcare Data Public Workshop Debrief. Available at: https://www.bma.org.uk/collective-voice/policy-and-research/ethics/secondary-uses-of-data/public-workshop (accessed 11 July 2019).

16.

Business Wire (2018, 27 November) Seagate Launches New Data-Readiness Index Revealing Impact Across Four Global Industries as 30 Percent of Data Forecasted to Be Real-Time by 2025. Available at: https://www.businesswire.com/news/home/20181126005585/en/Seagate-Launches-New-Data-Readiness-Index-Revealing-Impact (accessed 12 September 2019).

17.

Butler

(2013) When Google got flu wrong: US outbreak foxes a leading web-based method for tracking seasonal flu. Nature 494: 155–156.

18.

Cabinet Office (2017) Guidance: Research Code of Practice and Accreditation Criteria. Department for Digital, Culture, Media and Sport. Available at: https://www.gov.uk/government/publications/digital-economy-act-2017-part-5-codes-of-practice/research-code-of-practice-and-accreditation-criteria (accessed 20 September 2019).

19.

Carter, Laurie and Dixon-Woods. (2015) The social licence for research: Why care.data ran into trouble. Journal of Medical Ethics 41: 404–409.

20.

Centers for Medicare and Medicaid Services (n.d.) National Health Expenditure Fact Sheet. Available at: https://www.cms.gov/research-statistics-data-and-systems/statistics-trends-and-reports/nationalhealthexpenddata/nhe-fact-sheet.html (accessed 3 September 2019).

21.

Chan

and Harris

(2006) The Nuffield Council on Bioethics: An Ethical Review of Publications. Available at: http://nuffieldbioethics.org/wp-content/uploads/2014/06/Nuffield-ethics-review-final.pdf (accessed 28 September 2019).

22.

Clemence

Gilby

Shah

, et al. (2013) Wellcome Trust monitor wave 2: Tracking public views on science, research and science education. Available at: http://www.wellcome.ac.uk/stellent/groups/corporatesite/@msh_grants/documents/web_document/wtp053113.pdf (accessed 11 July 2019).

23.

Copeland

(2019) Google’s ‘Project Nightingale’ Gathers Personal Health Data on Millions of Americans. Wsj.com, 11 November. Available at: https://www.wsj.com/articles/google-s-secret-project-nightingale-gathers-personal-health-data-on-millions-of-americans-11573496790 (accessed 2 December 2019).

24.

Critchley

and Nicol

(2009) Understanding the impact of commercialization on public support for scientific research. Public Understanding of Science (Bristol, England) 20(3): 347–366.

25.

Dandekar

(1990) Surveillance, Power, and Modernity. New York, NY: St. Martin’s Press.

26.

Department of Health and Social Care (2019). NHSX: new joint organisation for digital, data and technology. Gov.uk, 19 February. Available at: https://www.gov.uk/government/news/nhsx-new-joint-organisation-for-digital-data-and-technology (accessed 4 July 2019).

27.

Devlin

(2017) UK needs to act urgently to secure NHS data for British public, report warns. The Guardian, 30 August. Available at: https://www.theguardian.com/science/2017/aug/30/uk-needs-to-act-urgently-to-secure-nhs-data-for-british-public-report-warns (accessed 24 July 2019).

28.

Domingos

(2017) The Master Algorithm. London: Penguin Books.

29.

Draper

and Turow

(2019) The corporate cultivation of digital resignation. New Media & Society 21(8):1824–1839.

30.

Eubanks

(2018) Automating Inequality: How High-Tech Tools Profile, Police, and Punish the Poor. New York, NY: St. Martin’s Press.

31.

Feiner

(2019) Google to acquire Fitbit, valuing the smartwatch maker at about $2.1 billion. CNBC, 1 November. Available at: https://www.cnbc.com/2019/11/01/google-to-acquire-fitbit-valuing-the-smartwatch-maker-at-about-2point1-billion.html (accessed 2 November 2019).

32.

Ford

Boyd

Bowles

JKF

, et al. (2019) Our data, our society, our health: A vision for inclusive and transparent health data science in the United Kingdom and beyond. Learning Health Systems 3(3). Available at: https://onlinelibrary.wiley.com/doi/full/10.1002/lrh2.10191

33.

Foucault

(1995) Discipline and Punish. A Sheridan (Trans). New York, NY: Vintage Books.

34.

Fourcade

and Healy

(2017) Seeing like a market. Socio-Economic Review 15(1): 9–29.

35.

Gandy

Jr (2009) Coming to Terms with Chance: Engaging Rational Discrimination and Cumulative Disadvantage. Farnham / Burlington, VT: Ashgate.

36.

Gandy

Jr (2017) Surveillance and the formation of. Public Policy. Surveillance & Society 15(1): 158–171.

37.

Garcia

(2019) Google’s ‘Project Nightingale’ center of federal inquiry. CNN, 15 November. Available at: https://www.cnn.com/2019/11/12/tech/google-project-nightingale-federal-inquiry/index.html (accessed 1 December 2019).

38.

Giddens

(1990) The Consequences of Modernity. Stanford, CA: Stanford University Press.

39.

Global Alliance for Genomics and Health (2019) GDPR Brief: How is Article 89 implemented across the EU/EEA? GA4GH , 3 June. Available at: https://www.ga4gh.org/news/how-is-article-89-implemented-across-the-eu-eea/ (accessed 25 September 2019).

40.

Gould M (2019) The NHS App: a platform for innovation. Technology in the NHS blog, gov.uk 31 May. https://healthtech.blog.gov.uk/2019/05/31/the-nhs-app-a-platform-for-innovation/ (accessed 4 July 2019).

41.

Government of Canada (2018, 10 April) Notice: Health Canada’s Approach to Digital Health Technologies. Available at: https://www.canada.ca/en/health-canada/services/drugs-health-products/medical-devices/activities/announcements/notice-digital-health-technologies.html (accessed 3 July 2019).

42.

Grant, Ure, Nicolsonet al. (2013) Acceptability and perceived barriers and facilitators to creating a national research register to enable ‘direct to patient’ enrolment into research: The Scottish Health Research Register (SHARE). BMC Health Services Research 13: 422.

43.

Green

(2019) Time to interrogate corporate interests in public health?. Critical Public Health 29(3): 257–259.

44.

Groves

Kayyali

Knott

, et al. (2013) The ‘Big Data’ Revolution in Healthcare: Accelerating Value and Innovation. Center for US Health System Reform; Business Technology Office. New York, NY: McKinsey.

45.

Hill

Turner

Martin

, et al. (2013) “Let’s get the best quality research we can”: Public awareness and acceptance of consent to use existing data in health research: a systematic review and qualitative study. BMC Medical Research Methodology 13: 72.

46.

Home Affairs Committee (2008) A Surveillance Society? Fifth Report of Session 2007–08, Vol. 1. Available at: https://publications.parliament.uk/pa/cm200708/cmselect/cmhaff/58/58i.pdf (accessed 2 October 2019).

47.

Institute of Chartered Accountants in England and Wales (ICAEW) (2012) Acting in the Public Interest: A Framework for Analysis. London: ICAEW.

48.

Ipsos MORI/Wellcome Trust (2016) The One-Way Mirror: Public attitudes to commercial access to health data. Available at: https://wellcome.ac.uk/sites/default/files/public-attitudes-to-commercial-access-to-health-data-wellcome-mar16.pdf (accessed 13 September 2019).

49.

Islam

Hasan

Wang

, et al. (2018) A systematic review on healthcare analytics. Healthcare 6(2). Available at: https://www.mdpi.com/2227-9032/6/2/54.

50.

Kitchin

(2014) Big Data, new epistemologies and paradigm shifts. Big Data & Society 1(1). Available at: https://journals.sagepub.com/doi/10.1177/2053951714528481

51.

Kitchin

and McArdle

(2016) What makes Big Data, Big Data? Exploring the ontological characteristics of 26 datasets. Big Data & Society 3. Available at: https://journals.sagepub.com/doi/full/10.1177/2053951716631130.

52.

Knapp

(2019) This Startup Used AI to Design a Drug in 21 Days. Forbes, 2 September. Available at: https://www.forbes.com/sites/alexknapp/2019/09/02/this-startup-used-ai-to-design-a-drug-in-21-days/#151387702594 (accessed 2 September 2019).

53.

Kruse

Goswamy

Raval

, et al. (2016) Challenges and opportunities of big data in health care: A systematic review. JMIR Medical Informatics 4(4): e38.

54.

Lashinsky

(2017) Tim Cook on how Apple champions the environment, education, and health care. Fortune, 21 September. Available at: http://fortune.com/2017/09/11/apple-tim-cook-education-health-care/ (accessed 29 December 2019).

55.

Laurie, G., K.H. Jones, L. Stevens and C. Dobbs. 2014. A Review of Evidence Relating to Harm Resulting from Uses of Health and Biomedical Data. http://nuffieldbioethics.org/wp-content/uploads/A-Review-of-Evidence-Relating-to-Harms-Resulting-from-Uses-of-Health-and-Biomedical-Data-FINAL.pdf (accessed Sep 14, 2019).

56.

Lazer

Kennedy

King

, et al. (2014) The parable of Google flu: Traps in big data analysis. Science (New York, N.Y.) 343(6176): 1203–1205.

57.

Lovett

(2019) Verily rakes in $1B for global development, potential acquisitions. Mobi Health News, 4 January. Available at: https://www.mobihealthnews.com/content/verily-rakes-1b-global-development-potential-acquisitions (accessed 3 September 2019).

58.

Lupton

(2012) M-health and health promotion: The digital cyborg and surveillance society. Social Theory and Health 10: 229–244.

59.

McBride

and Vance

(2019) Apple, Google, and Facebook Are Raiding Animal Research Labs. Bloomberg, 18 June. Available at: https://www.bloomberg.com/news/features/2019-06-18/apple-google-and-facebook-are-raiding-animal-research-labs (accessed 19 June 2019).

60.

McGoey

(2015) No Such Thing as a Free Gift. London: Verso.

61.

McGrail

, et al. (2018) Position statement on population data science. International Journal for Population Data Science 3(1): 4.

62.

Mayer-Schönberger

and Cukier

(2013) Big Data: A Revolution that will Transform How We Live, Work, and Think. Boston, MA: Houghton Mifflin Harcourt.

63.

Mazzucato

(2018) The Value of Everything: Making and Taking in the Global Economy. New York, NY: Public Affairs.

64.

Medical Research Council (MRC) (2018) MRC ethics series: Using information about people in health research. Available at: https://mrc.ukri.org/documents/pdf/using-information-about-people-in-health-research/ (accessed 19 September 2019).

65.

Murgia

(2019) Artificial intelligence group DeepMind readies first commercial product. ft.com, 30 March. Available at: https://www.ft.com/content/0e099914-514a-11e9-9c76-bf4a0ce37d49 (accessed 3 August 2019).

66.

Neville S (2020) Drug companies hooked on data. Financial Times, 27 Jan, p. 20.

67.

NHS Digital (n.d.) National data opt-out. NHS Digital. Available at: https://digital.nhs.uk/services/national-data-opt-out (accessed 19 September 2019).

68.

Nogrady

(2019) Artificial Intelligence Shakes Up Drug Discovery. The Scientist, 1 May. Available at: https://www.the-scientist.com/bio-business/artificial-intelligence-shakes-up-drug-discovery-65787 (accessed 29 August 2019).

69.

Nuffield Council on Bioethics (NCoB) (2012) Emerging Biotechnologies: Technology, Choice and the Public Good. London: Nuffield Council on Bioethics.

70.

Nuffield Council on Bioethics (NCoB) (2015) The Collection, Linking and Use of Biomedical Research and Healthcare Ethical Issues. London: Nuffield Council on Bioethics.

71.

O’Neil

(2016) Weapons of Math Destruction. New York, NY: Penguin Random House.

72.

Organisation for Economic Co-operation and Development (2019) Regulatory effectiveness in the era of digitalisation. Available at: http://www.oecd.org/gov/regulatory-policy/Regulatory-effectiveness-in-the-era-of-digitalisation.pdf (accessed 12 December 2019).

73.

Paavola

(2018) Amazon moves into healthcare. Becker’s Hospital Review, 20 December. Available at: https://www.beckershospitalreview.com/healthcare-information-technology/amazon-moves-into-healthcare-a-2018-timeline.html (accessed 31 December 2019).

74.

Panch

Pearson-Stuttard

Greaves

, et al. (2019) Artificial intelligence: Opportunities and risks for public health. The Lancet Digital Health 1(1): e13–e14.

75.

Paprica

Nunes de Melo

and Schull

(2019) Social licence and the general public’s attitudes toward research based on linked administrative health data: a qualitative study. CMAJ Open 7(1): E40–E46.

76.

Pariser

(2011) The Filter Bubble: What the Internet is Hiding from You. New York, NY: Penguin.

77.

Pasquale

(2015) The Black Box Society: The Secret Algorithms that Control Money and Information. Cambridge, MA: Harvard University Press.

78.

Patil, Lu, Saunderset al. (2016) Public preferences for electronic health data storage, access, and sharing – Evidence from a pan-European survey. Journal of the American Medical Informatics Association 23(6): 1096–1106.

79.

Popescu

Baruh

Messaris

, et al. (2017) Consumer surveillance and distributive privacy harms in the age of big data. In: Digital Media. Transformations in Human Communication. 2nd ed. New York, NY: Peter Lang, pp. 313–327.

80.

Powles

and Hodson

(2017) Google DeepMind and healthcare in an age of algorithms. Health and Technology 7(4): 351–367.

81.

Prainsack

and Buyx

(2017) Solidarity in Biomedicine and Beyond. Cambridge: Cambridge University Press.

82.

PwC (2019) Global Top 100 companies by market capitalisation, July 2019

83.

Raghupathi

and Raghupathi

(2014) Big data analytics in healthcare: promise and potential. Health Information Science and Systems 2: 3.

84.

Rule

(1973) Private Lives and Public Surveillance: Social Control in the Computer Age. New York, NY: Schocken Books.

85.

Sadowski

(2019) When data is Capital: Datafication, accumulation, and extraction. Big Data & Society. Avaialbel at: https://journals.sagepub.com/doi/full/10.1177/2053951718820549.

86.

Schneider

Sarnak

Squires

, et al. (2017) Mirror, Mirror 2017: International Comparison Reflects Flaws and Opportunities for Better U.S. Health Care. New York, NY: Commonwealth Fund.

87.

Schneier

(2015) Data and Goliath: The Hidden Battles to Collect Your Data and Control Your World. New York, NY / London: W.W. Norton and Company.

88.

Scott

Burall

Perrin

, et al. (2018) Data for Public Benefit: Balancing the Risks and Benefits of Data Sharing. Dunfermline: Carnegie UK Trust.

89.

Sexton

Shepherd

Duke-Williams

, et al. (2018) The role and nature of consent in government administrative data. Big Data & Society 5(2). Available at: https://journals.sagepub.com/doi/full/10.1177/2053951718819560.

90.

Sharon

(2016) The Googlization of health research: From disruptive innovation to disruptive ethics. Personalized Medicine 13(6): 563–574.

91.

Solove

(2013) Privacy self-management and the consent dilemma. Harvard Law Review 126: 1880.

92.

Spence

(2019) How can we place a value on health care data? Ernst & Young. Available at: https://www.ey.com/en_gl/life-sciences/how-we-can-place-a-value-on-health-care-data (accessed 27 September 2019).

93.

Srnicek

(2017) Platform Capitalism. Cambridge: Polity Press.

94.

Starkbaum

and Felt

(2018) Negotiating the reuse of health-data: Research, Big Data, and the European data protection regulation. Big Data & Society 6(2). Available at: https://journals.sagepub.com/doi/10.1177/2053951719862594.

95.

Stockdale, Cassell and Ford. (2019) “Giving something back”: A systematic review and ethical enquiry into public views on the use of patient data for research in the United Kingdom and the republic of Ireland. Wellcome Open Research 3: 6.

96.

Strickland

(2019) How IBM Watson Overpromised and Underdelivered on AI Health Care. IEEE Spectrum, 2 April. Available at: https://spectrum.ieee.org/biomedical/diagnostics/how-ibm-watson-overpromised-and-underdelivered-on-ai-health-care (accessed 29 November 2019).

97.

Thorne

(2019) Digital health investments hit an all-time high of $8.6B in 2018. Geekwire, 11 January. Available at: https://www.geekwire.com/2019/digital-health-investments-hit-time-high-8-6b-2018/ (accessed 3 September 2019).

98.

Turow

Hennessy

and Draper

(2015) The tradeoff fallacy: How marketers are misrepresenting American consumers and opening them up to exploitation. Annenburg School for Communication, University of Pennsylvania, USA.

99.

United Nations Conference on Trade and Development (2019) Digital Economy Report 2019. Value Creation and Capture: Implications for Developing Countries. New York, NY: United Nations Publications.

100.

Vaidhyanathan

(2011) The Googlization of Everything (and Why We Should Worry). Berkeley: University of California Press.

101.

van Dijck

(2014) Datafication, dataism and dataveillance: Big Data between scientific paradigm and ideology. Surveillance & Society 12(2): 197–208.

102.

van Dijck

and Poell

(2016) Understanding the promises and premises of online health platforms. Big Data & Society. Available at: https://journals.sagepub.com/doi/10.1177/2053951716654173

103.

Wall

(2019) Are you happy to share your health data to benefit others?. BBC News, 28 June. Available at: https://www.bbc.com/news/business-48784205 (accessed 28 June 2019).

104.

Wellcome Trust (2015) Enabling Data Linkage to Maximise the Value of Public Health Research Data. London: Wellcome Trust.

105.

Wellcome Trust/CM Insight (2013) Summary Report of Qualitative Research into Public Attitudes to Personal Data and Linking Personal Data. Available at: https://wellcome.ac.uk/sites/default/files/wtp053205_0.pdf (accessed 3 September 2019).

106.

Wellcome Trust/Ipsos MORI (2016) The One-Way Mirror: Public attitudes to commercial access to health data. Available at: https://wellcome.ac.uk/sites/default/files/public-attitudes-to-commercial-access-to-health-data-wellcome-mar16.pdf (accessed 13 September 2019).

107.

Wilbanks

and Topol

(2016) Stop the privatization of health data. Nature 535(7612): 345–348.

108.

(2016) The Attention Merchants: The Epic Scramble to Get Inside Our Heads. New York, NY: Alfred A. Knopf.

109.

Yeung

(2017) ‘Hypernudge’: Big Data as a mode of regulation by design. Information, Communication & Society 21(1): 118–136.

110.

Zuboff

(2019) The Age of Surveillance Capitalism. New York, NY: Public Affairs.

Disambiguating the benefits and risks from public health data in the digital economy

Abstract

Keywords

Introduction

Part 1: Commercial interests in health data

Part 2: Health data and public benefit

The moral, ‘civic duty’ appeal

UK legal frameworks and governance for public health data

Part 3: Exploring the possibility of a health data ‘trade-off’

Interrogating public willingness to trade access to health data

Can an adequate cost–benefit analysis be claimed?

Inadequacy of cost–benefit analysis: Patients

Inadequacy of cost–benefit analysis: Regulatory bodies and governance

The (im)possibility of an adequate cost–benefit analysis

Conclusion

Footnotes

Declaration of conflicting interests

Funding

ORCID iD

Notes

References