Nickel et al. (2023) makes a case for ethically rehabilitating the notion of digital health empowerment, a notion that, as we documented there, has come under intense scrutiny by many ethicists and social scientists. De Proost and Grey (2024) are largely sympathetic to this rehabilitation effort. They recognize and applaud the aim of investing empowerment with realistic ethical content, linking it with established ethical values in the health domain such as autonomy and justice. However, they also present two important critiques of our approach. The purpose of the present article is to reply to these critiques.

De Proost & Grey’s first critique is that our article does not commit to a notion of relational autonomy. This is problematic in their view because relational autonomy is needed to understand structural effects of power relations on individual choice, a methodological perspective that emerges from feminist theory. They want to avoid a situation in which health inequalities are entrenched “while at the same time protecting the autonomy of those who are most powerful and privileged” (ibid., 4). Presumably, what the authors mean here is not that the health-related autonomy of the powerful and privileged could be violated without good reason, but rather that their autonomy is much less likely to be threatened by existing power relations, and requires fewer structural initiatives in order to be adequately protected.

Relational autonomy theorists emphasize that our autonomy is shaped and constrained by the social relationships and environments in which we live. However, the development and exercise of autonomy can be hindered in environments characterized by disadvantages or harmful social relationships. Relational theorists are particularly concerned with exposing the effects of social relations of domination, oppression, and exclusion on individuals’ self-conceptions, opportunities, capacities, and choices (Mackenzie, 2010). Under such conditions, individuals are restricted in their ability to exercise autonomy and develop the skills and competencies for independent decision-making.

To a large extent, we are sympathetic to this point. Theories of relational autonomy offer some powerful criticisms of the classic Western enlightenment conception of autonomy that focuses on the isolated, purely rational (male) decision maker whose capacity for choice is impervious to outside influence (Lee, 2023). However, one need not adopt a relational autonomy view in order to take into account that people are situated, vulnerable, irrational at times, and that their choices are deeply influenced by their relationships with others. Critiques of relational autonomy include feminist critiques, like those of Khader (2020).

While De Proost and Grey (2024) argue for a greater emphasis on relational autonomy and the capabilities approach to support digital health empowerment, we find the applicability of relational autonomy to practical and scalable interventions problematic. There are critical questions about the feasibility of implementing scalable digital health solutions rooted in a relational understanding of autonomy. Attempting to reshape complex social structures to enhance relational autonomy is a huge task. If we require it as a precondition for digital health empowerment, this may be too great a barrier. Designing and implementing practical and effective digital health interventions that empower historically disadvantaged or disempowered individuals, if effective, should count as genuinely empowering. And as we explicitly argue, the further empowerment of the already-privileged still counts as empowerment, even though it may be less morally important to pursue for reasons of justice.

We deliberately engage with the traditional bioethical concept of autonomy as put forward by Beauchamp and Childress (2019) so as to make our argumentation attractive to a wide range of readers. However, we do not endorse the view that autonomy outweighs justice, nor the view that the autonomy of the powerful is somehow more important or determinate than the autonomy of the disempowered. As we write, we wish to understand empowerment as “embodying a somewhat open range of values” such as autonomy and justice, where technology can be linked with these values “under some (but not all) reasonable interpretations of these complex concepts” (Nickel et al. 2024, 9). We opt for this “normative interpretation” approach instead of engaging in conceptual engineering. Our approach leaves open the possibility of understanding autonomy relationally, but does not “engineer” empowerment or autonomy to mean this.

Now let us turn to the second critique. De Proost & Grey, 2024 argue for adopting the capability approach (CA) as a way of understanding digital health empowerment. On the CA, digital health empowerment would be understood as a matter of expanding health capabilities. There are two possible ways of understanding this intervention. One is to say that digital health empowerment essentially involves health capabilities as interpreted by the capability approach. Another is to say that digital health empowerment is compatible with the capability approach as one way of unpacking the idea of empowerment.

We agree with the compatibility claim whole-heartedly. The CA is an intuitive, important and influential family of theoretical views with special relevance for health economics, policy evaluation, and even for creating design specifications (See for example Oosterlaken, 2012 and Jacobs, 2020). We recognize the fundamental importance of the core insight that when measuring well-being we should count not just people’s preference-satisfaction or functional assessments, but also their capabilities in a wide range of important areas of human activity and value. The specific notion of health capabilities has a foothold in empirical literature and deserves to be highlighted more. Currently the CA literature is still developing in this domain and has not been applied widely outside the realm of health economics and health policy. In addition, the view has variations, partly because it is tailored empirically to those capabilities that are found culturally important in different settings (Sen, 1999), but more importantly because there are different understandings of how broadly we should understand the boundaries of health capabilities in relation to capabilities more broadly. Most researchers understand health capabilities as being any capabilities within the CA approach that are affected by health and health outcomes: not just “life” and “bodily health” but also “senses, imagination and thought,” “other species,” and “play” (Simon et al., 2013).

This latter point does raise concerns for seeing the CA as an essential way of understanding the capabilities involved in digital health empowerment. In our paper, we emphasized two main capabilities as part of empowerment: the capability to access health care, and the capability to understand and apply health information. But for obvious reasons, even if we see these two capabilities as central to digital health empowerment, that is not the same as adopting the CA. For these two limited capabilities are neither necessary or sufficient for obtaining the whole package of capabilities from the CA. They can be accepted even by those who reject the CA as a view of health-related welfare. There is even a risk that by defining health capabilities too widely, almost any effective digital intervention could count as digital health empowerment. For example, an intervention that enables capabilities related to “other species” or “play” might technically count as furthering digital health empowerment, even though it is not obviously health-related.

Returning to de Proost & Grey, another way to understand their criticism is that digital health empowerment essentially involves health capabilities as interpreted by the CA. We hesitate to endorse this essential-involvement claim for the reason just mentioned. More broadly, we do not wish to alienate from our approach those who take a different view of health-related welfare. It’s worth taking a step back to consider the intent of our article. We make reference to a prior literature on empowerment in health promotion that has been somewhat lost in more recent technology-focused discussions of digital health empowerment. The capability approach is not mentioned explicitly in this prior health promotion literature. Even the word “capability” does not appear in some of its key publications. In sum, then, although we find it an excellent project to link the notion of health capabilities with the literature on health empowerment more generally, our decision not to endorse the CA view is a considered one.

As a closing note, we wish to acknowledge the importance of de Proost & Grey’s critiques as an attempt to further spell out the normative commitments of digital health empowerment in a positive and realistic sense. From a philosophical point of view, it is often attractive to offer more detail rather than less, and to connect concepts such as empowerment, justice and autonomy to one another by enriching the content of all three. However, in the interdisciplinary field of digital health empowerment, we also see benefit in a more open and minimal approach that sees these concepts as distinct value constructs that can come apart, perhaps allowing wider empirical scope to investigate their contingent relationships in different domains.