A new public health approach to data: why we need data solidarity
BMJ 2024; 386 doi: https://doi.org/10.1136/bmj.q2076 (Published 23 September 2024) Cite this as: BMJ 2024;386:q2076
The current approach to governing digital data is stuck in the paper age. Regulations such as the European Union’s European Health Data Space (EHDS) have been designed to protect primary data subjects—the people who provide their data—and especially sensitive information such as their medical data. In digital societies, however, the risks and benefits of digital practices, ranging from data creation to secondary data use, routinely affect a much wider range of people than the people that the data come from. Data from one group can be used to benefit or discriminate against another group. Digital practices are also embedded in stark power asymmetries, between platform providers and users, and within and across countries.
To tackle these challenges we need an approach that goes beyond merely giving individuals more control over their own data, but also enabling communities and other groups to exercise meaningful control over how data are used, by whom and for whose benefit. Just as in public health where we understand that individual health rights need to be complemented by public health measures, our focus must be to fully harness the benefits of digital data and to effectively protect people from harm.
Data solidarity tackles the public value of data use
A Lancet and Financial Times Commission underlined the need to apply public health values and approaches to governing digital health futures.1 Making solidarity-based data governance (in short: data solidarity) practical is one priority for following through on the Commission’s recommendations.2 In order to help promote a better understanding of data solidarity and its relationship to other governance approaches, we have developed a glossary of terms.3 To support a structured assessment of the public value of data use, an online tool has been developed which is openly available and can be used by anyone—patients and other citizens, corporations, or public bodies—to better understand the likely public value of specific instances of data use.4 By applying the tool, users can see what they could do to minimise risks and enhance benefits, especially for marginalised or underserved groups.
Applying data solidarity can prevent harm
Data solidarity, had it been used, could have helped to prevent some recent data scandals. For example, in 2015, the Royal Free NHS Trust in England passed 1.6m patient records on to a Google-owned AI company, arguing that the company would create a tool to help detect kidney failure.5 A public value assessment before the event would have shown that the use of data without the meaningful consent of patients, and without any plans to share commercial benefits with the communities that the data came from, should not have gone ahead. Similarly, the tragedies caused by the use of algorithms by public authorities—such as in the case of the Dutch child allowance scandal6 and the Robodebt case in Australia,7 both of which cost human lives—could have been prevented with better collective (i.e. civil society) scrutiny of the data use by public authorities, as well as harm mitigation mechanisms in place.
Three pillars of data solidarity
We suggest focusing on three pillars of data solidarity.28 Firstly, we need to make data use easier when it promises to bring great benefits without posing significant risks to people or communities. This can be done by easing regulatory burdens or by providing public funding. Secondly, we need to actively prevent harm by prohibiting data uses that pose high risks and ensure effective enforcement. As not all harm can be prevented, there is also the need for mitigation strategies that support all people who have experienced harm from data use, independent of whether any law was broken.9 Thirdly, commercial profits from data use should be shared with the people and communities that have enabled the data use in the first place. Commercial profits can be shared via taxation, or various other forms of benefit sharing.
Risk does not lie in data types but in data use
Data solidarity requires two important shifts compared to the status quo. Firstly, data solidarity does not assume that risk lies in data types, but instead in data uses. Types of data use that entail few risks and are likely to yield considerable benefits for people (and not only for companies) should be treated differently from data uses that do the opposite. Secondly, in addition to giving people control over data individually, data solidarity makes use of collective instruments of control and oversight, including data commons and stronger use of laws to support data use with high public value and outlaw harmful data use.
Public value assessments carried out by organisations or corporations to see who is likely to benefit from their data use, and who is likely to carry the cost, would help to make data use more democratic—instead of allowing governments or powerful corporations to claim that everything they do is in the “public interest.” This supports the strengthening of a new type of active digital health citizenship.
Footnotes
Competing interests: We have read and understood BMJ policy and declare that we have no competing interests.
Provenance and peer review: not commissioned, not externally peer reviewed.