Doing co-creation: power and critique in the development of a European health data infrastructure

ABSTRACT

This paper engages with the translation of the abstract concept of Responsible Research and Innovation (RRI) into practice. Our investigation is situated within a large-scale European project seeking to develop a citizen-centered digital health data platform. We specifically engage with co-creation practices and their potential for creating and maintaining inclusive environments generative of critique and empowerment. Critique for us means questioning practices and the modes of thought they rest on, both of ourselves and participants in the co-creation process. We combine a narrative analysis of EU funding calls framing the project with a reflection on our experiences as STS scholars during the co-creation process and on how citizens formulated their critique regarding the health data infrastructure. We end discussing the possibilities and limits of such a co-creation approach and the spaces and modes of critique made possible and, thus, how it lives up to the promises of responsible innovation.

KEYWORDS:

• Infrastructure
• co-creation
• digital health
• modes of critique
• frictions

Introduction

This paper engages with how the abstract concept of Responsible Research and Innovation (RRI) (Felt 2018; Guston and Stilgoe 2017; Owen, von Schomberg, and Macnaghten 2021; Stilgoe, Owen, and Macnaghten 2013) is translated into the realm of practice by means of co-creation approaches (Deserti, Real, and Schmittinger 2022; Leino and Puumala 2020). It contributes to a growing debate of how to make innovations more context-sensitive, inclusive, just, and sustainable through engaging with users throughout the process of conceptualising, developing, and realising them. It also shows how important it is to see concepts such as RRI and their operationalisation as highly context-sensitive, with digital health being a fast-developing and highly sensitive field in which such approaches are deployed.

Digital health has become a buzzword (Bensaude Vincent 2014) in the European context, promising to assure the future of health and health care, which are under growing pressures. We witness increasing efforts to develop new health-related digital infrastructures in order to make citizens’ data mobile between places of production, collection, and use. This paper draws on work done in the large-scale Horizon 2020 project Smart4Health (authors being the social science partners) which developed, tested, and validated a health data platform prototype for the European Electronic Health Record exchange, enabling citizens to digitally collect, organise, and share their health data. The project, following a responsible innovation logic and employing a citizen-centred co-creation approach, promised to empower citizens through their active inclusion in the development.

Here, we specifically engage with co-creation practices and their potential for creating and maintaining inclusive environments that are generative of critique and empowerment. Taking inspiration from Foucault’s (1988, 154–155) reflections, we see critique not as ‘a matter of saying that things are not right as they are’, but rather as ‘a matter of pointing out on what kinds of assumptions, what kinds of familiar, unchallenged, unconsidered modes of thought the practices that we accept rest’. For us, critique means ‘to show that things are not as self-evident as one believed’ and to make them questionable. This goes hand in hand with an RRI approach that ‘inevitably force[s] consideration of questions of power’ (Stilgoe, Owen, and Macnaghten 2013, 1572). Concretely, given the Smart4Health project’s emphasis on individual benefits in terms of citizens’ empowerment our attention focuses on the presence of power differentials encountered nonetheless.

When addressing empowerment, we take power to be relational (Law 1991) – not as inherent to an individual or institution, but enacted in social practice and mediated, and potentially reinforced, by sociotechnical infrastructures in development (Bowker et al. 2010). In the paper at hand, we delve into different relationalities and practices comprising the co-creation process within the Smart4Health project. We ask what needs consideration to foster impactful, critical practices of citizen-centred co-creation in a long-term, large-scale transdisciplinary research and innovation action in the area of digital health – particularly to avoid the establishment of an infrastructure with potentially disempowering effects.

Starting off by revisiting debates that are key to understanding the issues at stake, we briefly address the key dimensions of responsible innovation, reflect on the diverse challenges identified with regard to co-creation as one form of operationalisation, engage with the impact of the projectification (Maylor et al. 2006; Torka 2009) of funding innovations, and specify our symmetrical approach to infrastructural co-creation. After presenting the data and methods, we proceed with our analysis in three steps. First, we trace the space for co-creation established by the funding scheme, the corresponding inscribed logic and the prevailing value regimes of which the Smart4Health project is part. Second, analysing our experiences in the field, we inquire the practices that enabled co-creation in the project. Third, we engage with diverse forms of critique in the process of co-creation as articulated through different kinds of frictions. We end with a reflection on the possibilities and limits of such a co-creation approach and thus of living up to the promises of responsible innovation.

Co-creating innovation: operationalising responsible innovation in a projectified environment

Responsible innovation and co-creation

The framework of Responsible Research and Innovation (RRI) was introduced into the European policy realm about a decade ago (Owen, von Schomberg, and Macnaghten 2021) with the promise of making innovation processes more transparent and inclusive. The integration of societal actors should ensure both the sustainability and societal desirability of solutions (von Schomberg 2011). Inviting different groups of stakeholders to engage in processes of innovation should open-up the possibility of asking the normative questions of ‘what counts as an ‘improvement’ through innovation and for whom (von Schomberg et al. 2013, 54). While the concept of RRI has gained traction, its concrete meaning and the translation of its dimensions into practice remain open to context-specific interpretations. Stilgoe, Owen, and Macnaghten (2013) identify anticipation, inclusion, reflection, and responsiveness as the four key dimensions in the process of responsible innovation, while the European Commission chose a different set of perspectives, mentioning ethics, gender equality, science education, societal engagement, and open access. Taking an RRI approach should however allow a much broader assessment of the value of innovations (Felt 2018), shifting attention from the (market) value of innovation, or a simple assessment of potential harm, to the values that are embedded in and realised through innovations (Felt 2017b). While RRI is seen as a European concept and researchers have rightly asked what it means for it to travel to other regions of the world (Doezema et al. 2019), even within the European context, RRI can unfold very differently depending on cultural context, the domain of innovation in which it is applied, and how societal actors feel affected by change. Accordingly, our case of building a prototype of a citizen-centred health data platform that works across European borders has been a unique challenge, as various dimensions such as variants of existing infrastructures, health disparities, and different health care cultures needed consideration.

Within the RRI discourse, co-creation has been highlighted as an approach that would allow for the concept to be put into practice. Co-creation has a long history, initially coming from marketing research (Prahalad and Ramaswamy 2004), a background which has certainly left traces in the way it is referred to in the European policy realm. While consumers/customers are seen as important actors in the creation of valuable innovations, the aim of co-creation is often to improve acceptability on markets and thus also to harness the economic potential of innovations. More recently, several analyses of both the proliferation of the notion of RRI and co-creation in Europe (Macq, Tancoigne, and Strasser 2020; Ruess, Müller, and Pfotenhauer 2023) have offered valuable insights into the history of the concept in the policy domain. They stress the fact that not only do both concepts remain vague and open to diverging interpretations, but also that details of realisation often focus more on advancing specific visions of innovation rather that questioning the direction in which innovations are being developed. These observations raise the question under what conditions co-creation has the potential to open up the range of possibilities for future users to shape the form and direction of innovation.

Projectification, infrastructuring, and power relations

While the idea of co-creation is being experimented with in many settings and is under continuous negotiation, in our case co-creation of a digital infrastructure happens within the framework of a European research project. For responsible innovation this projectification has consequences. It needs to be organised in an environment structured by roadmaps, timelines, deliverables, and key performance indicators, which makes a more open-ended co-creation approach difficult, as this would require space and time for reflection, reorientation, and re-timing of processes leading to the envisaged innovation. Being closely committed to the value of efficiency and working in ‘the iron cage of project rationality’ (Maylor et al. 2006) makes it easy to close down critique in the name of the project logic and related time pressures (Felt 2017a). Indeed, the project logic always comes with a specific timescape (Adam 1998) that defines timing, tempo, duration, timeframes, sequence, and rhythms, and with this, which synchronicities become possible. This demands a complex choreography of alignments of work rhythms of the different actors involved and defines when windows of opportunities open up and (have to) close down.

We also need to consider that co-creation in our case takes place in the context of what is not only a major infrastructuring effort, but one which spans different national contexts. As we know from infrastructure studies, any new sociotechnical infrastructure ‘links with [pre-existing] conventions of practices’; it does not grow ‘de novo’ but has to ‘[wrestle] with the “inertia of the installed base” and inherits strengths and limitations from that base’ (Star and Ruhleder 1996, 113). This requires reflecting on pre-existing information infrastructures, their strength and vulnerabilities, and the embedded tacit visions and desires (Larkin 2013); taking into account existing practices, power relations, social and cultural disparities, and persistent inequalities (Erikainen et al. 2019); and considering how the new digital health infrastructure might generate new or reify existing ones.

Power is enacted not just in the manifestation of infrastructures, the relations they privilege, and the position they assign to actors. There are power dimensions in who gets to articulate infrastructures’ initial justification, who gets to shape them further, and with what consequences. The RRI dimension of inclusion draws our attention to questions of power and who is given a voice in the evolution of any research and innovation process – and who is not. It calls for different forms of engagement in defining problems, identifying assumptions in framing, and suggesting, as well as assessing, solutions and reminds us that we need to look for those who, for whatever reason, might be excluded from participatory processes.

Understanding the multiple formats and agents of critique

Criticism and disagreement are crucial dimensions in and of participatory processes and can (and should) be taken as constitutive and performative (Irwin, Jensen, and Jones 2013). A productive co-creation process needs to take critique seriously in addition to collecting contributions regarding the functionality and performance of the infrastructure. Co-creating an infrastructure needs to carefully consider those that leave and opt out of participation and use, as this can also be an expression of critique. Just as non-use is not always grounded in deprivation or inequality, the adoption of a technology does not have to be the norm – the inclusion of non-users in development processes, considering them as ‘actors who might influence the shape of the world’ (Wyatt 2003, 78), is therefore crucial. Use and non-use can be indicative of a position taken towards the broader infrastructure and the social conditions that a technology is embedded in or connected to (Satchell and Dourish 2009).

We strive for a mode of co-creation that has the potential to integrate dissensus and counter-participation (Tironi 2018) and that conceptualises critique and disengagement as productive, specifically as an entry point to grasp infrastructural issues. Inspired by Tironi and Valderrama’s (2021) mobilisation of frictions (Tsing 2005), we take a symmetrical approach to infrastructural co-creation, attending to critique voiced by those that do and do not use the platform prototype as well as those that do and do not (or no longer) choose to participate in co-creation.

Frictions are the result of bodies or actors having divergent aims, meeting in ‘awkward and unstable encounters’ (Tironi and Valderrama 2021, 195). Frictions indicate energy-laden moments, where things are ‘set (…) in motion’ (ibid.), potentially disrupting a relation between user and technology. A focus on frictions in co-creation sensitises us to articulations of power within the infrastructure under development, ‘remind[ing] us that heterogeneous and unequal encounters can lead to new arrangement of culture and power’ (Tsing 2005). Analysing frictions delivers important insights into infrastructural issues as they can point to situations where a relation cannot be made or breaks, where people feel like something is not for them. Such moments should not be reframed as individual challenges, a framing that would disempower (non)users by shifting responsibility for infrastructural effects onto the individual.

Material and method

The Horizon 2020-funded Smart4Health project (2019–2023) brings together 19 consortium partners from various European countries (Portugal, Germany, Luxembourg, Belgium, the Netherlands, Italy, Austria) and one from the US. Over the project duration, our task has been to develop, implement and run a co-creation environment to gradually develop user requirements based on engagements with citizen users and professional users, and to empirically assess the degree of implementation in several different use cases in Germany, Portugal, Luxembourg, the Netherlands, and Italy.

The concept co-creation environment underlines the spatiotemporal dimension of our approach, as it spanned all phases of prototype development and various institutional and cultural settings. It, thus, had to be flexible yet robust, and allow for situated in-depth engagements and more longitudinal (re-)engagements. As social scientists, we had the role of knowledge brokers, discussing with our project partners insights from the engagements in diverse arenas and with diverse participants.

Obviously, co-creative work does not start from scratch. Not only were we confronted with pre-existing software solutions and the specific know-how of actors involved in the project, but we also had to learn that technical developers have their own visions of how to give voice to society in their developments (recall the relation of co-creation to market research). Of the different formats of co-creation, we decided to specifically focus on aspects of value co-creation approaches and experience-based co-design,¹ approaches which challenge and potentially disrupt power differentials – both within the co-creation process itself as well as within the sociotechnical infrastructure to be developed.

Our analysis is based on a broad set of materials, allowing us to reflect on the spaces and modes of critique that were made available to citizens by both the project architecture and the innovation trajectory sketched by policy actors. Concretely, we build the analysis on three pillars. First, we analyse a set of EU policy documents in the field of digital health related to the project as well as documents outlining standards, practices, and normativities of co-creation within EU funded projects. Second, we reflect on our engagement in the project and the rhythms and practices of co-creation as well as the negotiations involved in making critique productive. Third, we draw on the material generated in multi-method co-creation engagement exercises (interviews, group discussions, scenario-based prototype testing).² Input was also provided by our project partners to capture moments of critique and articulations of perceived (dis)empowerment when using the platform (frictions). We do all this symmetrically for users, non-users, and no-longer-users. Non-users appear in our co-creation processes in several forms: (1) as participants in the co-creation exercise who see themselves as potential non-users of the infrastructure were it not to meet certain expectations, (2) as personas constructed by participants in the co-creation process who they see as excluded through the implementation of a specific version of a digital infrastructure, and (3) as written descriptions by our partners in the use cases abtou why potential users resisted platform use. All these positions are key to consider when building such an infrastructure.

Our analysis follows a constructivist Grounded Theory approach (Charmaz 2014), staying open to the theoretical possibilities the material provided us with and specifically attending to practices. While we went through the three sets of material separately, the comparative thrust of Grounded Theory allowed us to make our analyses productive across our material, for instance, by using the insights from the analysis of policy documents as an additional lens for the analysis of interviews and group discussions and vice versa.

Findings

In what follows, we look into the multi-sited and multi-layered translations from the policy ideals of responsible innovation, over the co-creative work practices possible in a large-scale transdisciplinary EU project, to how citizens formulated their critique, their ‘it could be otherwise’, with regard to the infrastructure.

Making space for co-creation

To trace the spaces for co-creation opened up by EU funded projects, we analysed the ‘scripts’ embedded in the respective calls (Akrich 1992; Felt et al. 2016).

Investigating these EC calls is important as they open up or close down spaces of participation as the underlying funding scripts present a specific vision of science-society relations, the issues and challenges to be addressed, the actions to be taken, and the different stakeholders to be involved. We focused on a relevant set of EC calls in the field of digital health issued within the framework of the Horizon 2020 funding programme (2016–2020), specifically those explicitly referring to developing digital health solutions for/with citizens, centring co-creation, public participation, and societal issues. This allows for the capturing of the discursive frame that co-shaped the formation of the Smart4Health project we were part of. In what follows, we will outline three tightly coupled narrative strands we identified.

Making problem-solution packages

Call texts in the digital health domain commonly start with a diagnostic narrative (Penkler, Felder, and Felt 2015) sketching quite a dramatic picture of healthcare systems. This frequently comes with a reference to the ‘ageing of the population together with the rising burden of chronic conditions (incl. mental diseases) and multi-morbidity’ that would call for urgent action (SC1-DTH-02-2020).³ We thus encounter a risk narrative pointing to developments potentially compromising the healthcare services provided to society at large as well as related responsibilities. The response is twofold: individuals are called to act in preventive ways, to take responsibility for their health, and through this, for the healthcare system more broadly. In turn, they are promised a clear improvement: ‘service delivery around people’s needs for health and social care’ (SC1-DTH-02-2020).

Yet, we encounter a second group of narratives also conveying a sense of urgency, this time drawing on a rhetoric of missed opportunities. Calls employing this framing present the ‘large amounts of valuable health data’ as ‘untapped sources of growth’ that could easily be used to support ‘research and better healthcare’. Such narratives, however, also include references to potential hurdles of a technical but also a social nature, pointing to the need of the support of citizens ready to share their data. Improvements are then called for on both fronts in order to ‘fully unlock these sources of value’ (SC1-DTH-08-2018) while being staged as a time-sensitive issue.

Given the feeling of urgency present in both problem narratives, pathways to solutions are generally also delivered. They are both technical, mainly focusing on digital solutions such as health data platforms or apps, and social, underlining the need to engage with citizens as future users through co-creation or other forms of engagement. Pursuing the suggested solutions paints a bright future, not only for citizens, but also institutions, governments, and the EU overall. Promises encompass, for instance, ‘better care […], less inequality, better health promotion, better disease prevention’ (SC1-DTH-13-2020), effectiveness, efficiency, increased productivity, and the generation of value.

What we know from studies of the relation of problem framing to potential solutions is that the ‘quality of solutions to perceived social problems depends on the way they [problems] are framed. If a problem is framed too narrowly, too broadly, or wrongly, the solution will suffer from the same defects’ (Jasanoff 2003). Those with the power to define what constitutes a problem to be addressed thus have a huge responsibility as they shape potential solutions. Simultaneously, since it is difficult to object to solutions that attempt to counter the breakdown of the health care system, this representation restrains the agency stakeholders – in this case citizens – feel they have within the project. As not only their own wellbeing, but also that of entire nations or even Europe depends on their readiness to see the risk, citizens perceive responsibility shifted onto them and thus tend to feel that questioning the problem framing is not an option.

Creating data opportunities

As the projects addressed by these calls are focusing on large-scale endeavours, they always also address a range of anticipated data-related challenges that need to be tackled. Such challenges are based in the fragmentation of the knowledge and practices of diverse health care actors, the lack of uniform standards or communication paths when it comes to health-related data, and the lack of collaboration or readiness of different stakeholder groups to comply to common rules. The following call illustrates clearly the multidimensionality of the data challenges:

Large amounts of valuable health data are generated and collected during and between citizens’ medical examinations across Europe. However, opportunities to reuse these data for research and better healthcare are often missed because health data continue to be confined in data silos, often not matching semantic standards, quality needs and safe data exchange techniques. With 24 official languages spoken across EU Member States, the EU eHealth interoperability task is even more daunting. (SC1-DTH-08-2018)

Consequently, prescribed solutions strongly focus on facilitating the alignment of frameworks, practices, or technologies. This is to be attained through the implementation of specific technological infrastructures aimed at responding to challenges that go beyond merely technological domains (e.g. Electronic Health Record (EHR), ICT solutions, big data). In line with the development of the European digital market, the aim is to realise a ‘harmonised’ European framework (e.g. through standards for data exchange, policy frameworks, and methodologies), facilitating interoperability. Furthermore, when developing prototypes of solutions, outcomes should also be ‘scalable’ and ‘extensible’, e.g. having the potential to be ‘integrated in a federated platform structure [that is] constantly accessible and portable within any other Member States of the EU’ (SC1-DTH-08-2018). We thus see both the time limits imposed on the project and the expectation that the outcomes will continue beyond the project duration.

Given these aims, the space of co-creation has clear limitations, needing to keep in line with these overarching aims. This aspect is essential to consider from the angle of empowerment and critique: since the project outcomes are meant to serve as foundations for further innovations, anything that follows from them builds upon the roles and power relations woven into the project through the call scripts.

Defining citizens and their place

Call texts are also engaged in ‘making publics’ (Felt and Fochler 2010) for whom these data infrastructures should create advantages. In many ways, members of society are treated like an undifferentiated group of people, all of whom are expected to benefit from the health-related innovations – either through the use of a data platform or other digital devices, the health-related knowledge that is generated and communicated to them, or the resulting prescribed practices such as preventive behaviour.

Only ‘vulnerable population groups’ (CO-CREATION-07-2017), are identified as publics needing specific attention. They are seen as potentially able to yield valuable input for the development of measures leading to collective well-being. However, this framing partly connotes a sense of optionality. Using phrasings like ‘taking into account’ or ‘careful consideration should be given’ (SC1-DTH-13-2020), we can sense at least some readiness to engage with them. Yet, this often comes without any further explanation of the form or extent of these interactions and renders citizens an optional resource that can but must not be drawn upon.

Conversely, considering the solutions discussed above to the calls’ framings, citizens are more than mere passive entities along this innovation pathway. Rather, certain things need to be delivered by them for the calls’ objectives to be attained. This, for instance, applies to the framing of citizens’ data as an untapped resource that can be re-used and recontextualised for novel infrastructures, services, or policies to bolster innovation and growth and should thus be readily provided by them. Correspondingly, gathering new knowledge on/with citizens, e.g. through co-creation or other forms of participation (never specified to any degree), is also presented as a way of identifying further sources and opportunities for innovation. This implies that the actors to be involved also actively have to adjust their practices to achieve the prescribed objectives. This is highlighted by phrasings referring the need for citizens and other stakeholder to ‘take up’, ‘commit’, or ‘buy into’ the envisioned solutions or suggested strategies for achieving them, as is visible in the following extract:

Sustainable behaviour change refers to efforts to change people’s personal habits to prevent disease, stimulate healthy people to monitor their health parameters and thus lowering the risk of developing (chronic) conditions. (SC1-DTH-02-2020)

Concretely, we see a shift of the responsibility for preventing disease to the individual, requiring them to adjust their current health practices and adopt new ones. This also holds true for our project consortium, which is supposed to actively put in ‘efforts’ and ‘stimulate’ citizens to adopt healthy habits.

The framing is also found in the calls’ use of the concept of ‘empowerment’ (e.g. SC1-DTH-10-2019-2020), which is commonly presented as the aim and desired outcome of the projects. However, it is never defined and no indicators are given as to when this state would be achieved. Sometimes there is more specification, e.g. that ‘communication (towards citizens, patients and healthcare providers) [would] allow citizen/patient empowerment’ (SC1-DTH-08-2018).

Here, the projects funded in the calls are again meant to provide citizens with some sort of framework, i.e. a digital infrastructure, policy framework, or knowledge on healthy behaviour, that would enable them to take up the role they are assigned by the call. Thus, the use of ‘empowerment’ rather implies nudging citizens into a specific, desired behaviour. Strikingly, the only concrete definition of ‘empowerment’ as provided in a call’s footnote, suggests exactly that:

Empowering citizens refers, among others, to enhancing their self-management, raising health literacy, involving people through co-production of care and supporting informal carers. (SC1-DTH-13-2020)

Empowering citizens thus implies the delegation of work onto them, individualising the systemic problem.

Overall, in the calls analysed, the concept of empowerment and co-creation are both wide open for diverse interpretations – they are in ‘a conceptual and methodological limbo’, to quote Acuña Mora and co-authors (2022) – allowing those carrying out the projects to interpret them in a rather flexible manner.

Arranging co-creation in practice

How do these policy framings translate into co-creation practice? To start, it is important to underline that an infrastructure is not a clearly delimited technological artefact. Rather, it is a complex entity consisting of many different elements to be aligned continuously throughout a long process such as in the case of the co-creation of the Smart4Health prototype. For the co-creation process, this meant that different life worlds (professional and personal), loyalties, and temporalities had to find suitable arrangements in order to develop a sustainable solution. Furthermore, the co-creation process had to be adaptive to limiting factors and interests that materialised along the way, such as the COVID-19 pandemic necessitating virtual settings for many forms of engagement. In the following, we will reflect on the work that went into synchronising the different environments and the imaginations of a ‘good process’ when it comes to moments of intervention and rhythms of development, before turning to the challenges of co-creating an infrastructure.

Aligning rhythms of co-creation

As the process of co-creation in Smart4Health would stretch over many years, we developed a design with alternating loops of engagements with (potential) users and consortium partners. We thus provided insights to technical partners trying to work within their project timelines and simultaneously worked with partners in the field to conduct participant recruitment. We imagined an ideal flow of four such waves of user engagement, with consortium internal co-creation workshops at the end of each wave. However, we quickly realised that the alignment of multiple temporalities and work practices that fed into and drew from the overall co-creation process was anything but straightforward. Continuous alignments were necessary to ensure that input from the engagement exercises were delivered ‘in time’ to be folded into the technological development processes and it was crucial to have participants who would be ready to engage with us when it came to certain aspects.

Aligning several temporalities is not only about agreeing on collaborative schedules but must in the first place also involve managing expectations about the pace of progress in terms of co-dependent tasks and activities that each involved partner leads. Yet, working under projectified conditions meant that the allocation of efforts to partners and of timeframes for completing and reporting on them could lead to the situation where some partners were under pressure to advance, while others were required to proceed with diligence and not to rush things, specifically when working with citizens. Therefore, while all project partners would in principle subscribe to the idea of co-creation, the different work realities required individual work-arounds. The following examples provide some insights into how such a performative dimension can materialise.

One narrative encountered in the field could be labelled ‘starting from scratch’ and refers to the developers’ claim that they would be fully open to user needs identified in the co-creation process. This tacitly denies the fact that infrastructures never grow ‘de novo’ (Star and Ruhleder 1996) and moves to the background that certain building blocks of the technology/infrastructure under (further) development were (as stated in the proposal) expected to be already brought into and used in the project. Such discrepancies between promissory narrative and project reality can cause trouble in collaboration when narratives are instrumentalized to exert pressure to produce user requirements while choices were actually already closed off given the installed base of the infrastructure.

Challenging temporalities in project work were thus used to justify introducing pre-defined choices and re-using previously made developments for the sake of efficiency. While this is neither uncommon nor problematic per se, their short- or long-term effects on the project may turn out to be critical. First, this re-use risks undermining the overall participatory objective inscribed in co-creation for the sake of proceeding quickly. This is particularly critical in a project that abides by the RRI principles with the aim of being inclusive and addressing diversity adequately. Secondly, it is key to know whether or not a development is still reasonably open for changes, as user input to already closed down options can lead to important dysfunctionalities and user dissatisfaction in the co-creation process.

Furthermore, we came across a pragmatic and efficiency-driven approach captured by the phrase of ‘having to start somewhere’. This implies the necessity to define a starting point for the collective endeavour and to agree on which aspects of the development are to be focused on first, and which, later. This can become a problem, if solutions developed first – that is at a moment when input collected from users has been scant – are regarded as closed and only minor adaptations are possible. For example, building such early solutions (a minimally viable product) only on so-called ‘happy cases’, i.e. data literate (prospective) users of the technology, potentially later creates an environment in which less digitally-literate users are considered outliers; their needs and concerns will therefore only be integrated in the form of incremental, add-on improvements of the technological design. Thus, the need to abide by the tight project temporality can also be instrumentalized as reason to declare developments that would address more diverse users as being ‘too late’ or requiring ‘too much effort’ and ultimately ‘out of scope’ for the project period. Hence, when aspects get marginalised, the probability of them being addressed in an adequate and timely manner is low, even if they are of critical importance (Goedhart et al. 2021).

Infrastructural co-creation

The collaborative co-creation work in the project was based on the premise that citizens are able to articulate needs they have with regard to a health data infrastructure to be created and that these can be translated into user requirements. However, this articulation of needs has to come at a point in time when the platform is no more than a vision, a vague but also complex entity with which they have had little to no experience with so far. The systematic collection and careful consideration of user requirements, or so the assumption goes, will then lead to the development of a health data infrastructure that will be valuable, broadly accepted, and used, thus allows citizens to become empowered agents regarding their own health.

We, first, address the dimension of needs articulation and individual empowerment within infrastructural co-creation. The co-creation of smaller-scale innovations can generally draw on users who are well-versed in outlining their needs, as they feel able to assess in advance what is at stake or recognise which problems they are already confronted with. However, in the co-creation of a complex sociotechnical infrastructure in the domain of digital health and thus, of infrastructural change, things are different. The ‘object’ at the core of co-creation was both a complex digital infrastructure and one that would change health-related practices. Use thus had to be understood as an evolving process that goes through several stages (encounter, experience, integration). This meant that not only would the immediate effects need to be considered, but the potential long-term impacts of living with/in such an infrastructure as well.

As a result, the co-creation of the infrastructure warranted the specification of different kinds of requirements. Some (1) addressed the performance of very concrete features and specific functionalities, while (2) others were requirements that spoke more to the performance of the infrastructure and addressed its sociotechnical dimension as a whole. As agreed with our technical partners, user requirements were translated into a language they were acquainted to working with, producing statements that always clarified the actor making the claim and specified what they want to be able to do and their objective (e.g. ‘As a citizen doing … , I want to … in order to … ’). Such formulations seem to stress the citizens’ role in articulating needs and concerns, but also underline the idea of individual empowerment, rendering the relational dimensions of power and empowerment invisible.

When looking at the implementation of user requirements we found an interesting divergence. Those addressing technical issues were taken up rather straightforwardly: they were put on a timeline, prioritised, or their realisation discarded due to technical issues. However, it was far less clear who would take care of the second kind of requirements, i.e. those that speak to the sociotechnical dimension of the functionalities that make the infrastructure work, and when this would happen. In our co-creation work, we saw that these requirements and any critique formulated through them were routinely defined by the technical partners as not within their scope of (technical) responsibility and taking care of them was transferred to ‘others’. However, it was exactly these dimensions of the platform prototype that would be key for users to engage with in the first place and would offer insights regarding the (powerful) relations the infrastructure could or should enable. We thus developed concrete strategies to collect such articulations and better understand potential instances of infrastructural (dis)empowerment.

Second, we asked what the notion of careful consideration would mean in our spaces of co-creation. By design, the specification, implementation, and validation of user requirements was a locus of continuous negotiation. While it was clear that not all articulated needs could be taken care of within the scope of prototype development, we implemented a process that allowed us to trace and document the decisions taken concerning user requirements: the collaborative Performance Accountability Table (PAccT). It has four ‘spaces’, each of which is devoted to a possible work status: requirement formulated, integration done or not done, validation of the requirements that were implemented, and documentation of the choices within the consortium. In the end, the aim was to bring all user requirements into the documentation space, including an explanation of their (non)implementation status. Through this process, we ensured transparency and accountability by making explicit that ways/degrees of implementation were indeed deliberate decisions based on careful considerations.

These PAccT engagements brought to the fore different closure mechanisms in the consideration processes that also underline the negotiations of the identity of the platform prototype. Sometimes, the partners would define a requirement as out of the scope of the project, thus making cuts where the prototypical infrastructure would end and what citizens could expect of it. This had the effect of allowing input from the co-creation process to retain its value, but postponing its integration. This mechanism is closely related to the practice addressed above, i.e. claiming that it is necessary to start somewhere.

Overall, tracing the path of user requirements through the project space thus provided us with important insights as to where critique emerges, how it is made mobile, and where, when, and how it can actually make a difference.

Mobilising frictions for co-creating an infrastructure

Three modes of articulating critique

In our co-creation work, we encountered three modes of articulating critique. First, there were moments where citizens fully identified with the development. They perceived it as valuable, subscribed to the project idea and the platform logic, and welcomed the setting in which the co-creation was happening. In doing so, some addressed the inescapability of such a development – ‘it’s something you just have to get used to’ – and voiced their readiness to adapt as they actually ‘like the idea of the project’. Critique thus always stayed within the confines of the project logic, taking the form of suggestions for incremental improvements.

In the second mode of critique, citizens questioned inscribed users and forms of use, while retaining a collaborative position and offering critique that would need to be taken into account for the platform to become valuable to them. They provided – so their hope – solutions for misalignments, e.g. adapting the platform to better fit health practices they conceptualised as normal. In other instances, citizens voiced critique of scenarios of continued use by referring to more infrequent use practices. Using and not using the platform were thus not seen as distinct states, but dimensions of a dynamic and situated process:

I’m […] currently not preoccupied with my health situation because I’m healthy. Only when I was pregnant, I was preoccupied and interested to read my reports.

Participants would also identify specific user groups, underlining the general usefulness and value of the development not necessarily for themselves, but for others. Projecting use onto older people was a recurring mode, given their assumed non-proficiency with digital infrastructures and practices, and their presumed need for spokespersons in the co-creation setting. Doing so allowed our engagement partners to provide critique from a co-creative position of advocacy.

Both modes were instrumentally productive in closing gaps between platform implementations and use practices, while not questioning the infrastructure per se. They were thus easier to communicate within the consortium because they were seen as being supportive of the overall idea.

However, there was a third mode of critique that sometimes quite fundamentally questioned the infrastructure, with citizens openly disengaging from the platform itself and/or active participation in its development. An interview partner, for example, recalled that for some of her colleagues, the registration process ‘was too big a hurdle’, which eventually meant that ‘we lost them […] along the way’.

Experienced frictions can trigger feelings of frustration, anger, or estrangement, hampering engagement or stopping it altogether. Retreating from the user status within the project is frequently interpreted as taking a non-productive position. We, however, took another approach, understanding ‘leaving’, ‘disengaging’, and ‘giving up’ as performative acts and moments of productive critique, and thus as valuable articulations of infrastructural co-creation. Collecting frictions experienced by responsive project participants who resisted subscribing to the idea of becoming a regular platform user but nonetheless assessed the platform with us and by citizens who retreated from both platform use and project participation provided us access to a fuller set of potential issues that need attention in order to not infrastructure disempowerment.

Working with frictions

What insights did this approach – taking all three modes and the frictions they articulate as infrastructural critique – produce?

First, we saw frictions emerge between citizens’ expectations and the uncertainties they encountered regarding platform processes and data flows. This points to the relevance of data infrastructure literacy (Gray, Gerlitz, and Bounegru 2018) for the degree of destabilisation such frictions can produce. For some participants, even though they experienced frictions, the complexity of some of the processes remained understandable:

I’m used to it […]. Because I know it, I know it’s necessary and [laughs] of course it’s a long process, […] but that also protects you. No effort, no protection.

While many tolerated a degree of uncertainty without translating a friction into a personal problem, for others this was not possible. One interview partner recalled participants with little to no digital literacy who ‘took it personally’ when encountering frictions. They did not assign responsibility for failure to the way the infrastructure works but saw it as their own failure and as a consequence, disengaged. A high level of data infrastructure literacy thus helps reflect on the responsibility for problems, and not take the blame for things that do not function.

Second, finding one’s place within a new sociotechnical infrastructure requires alignment between the platform and citizens’ own material health (data) practices, which not all citizens are equally equipped to do. When it came to health care professionals’ (HCP) alignment with the data infrastructure, citizens were well aware that they work under rather tight time regimes. Many showed understanding for them not engaging with a digital tool that might at least initially produce frictions in the process of dealing with numerous patients.

Furthermore, users reflected on their everyday experiences and pointed to the fact that platform access could be hampered due to the material conditions at their local HCP. It could simply be the case that ‘the internet would not work’ and they would not get a signal. Unequal distribution of (high-speed) access to data is clearly identified as an issue disrupting the ideal of such a data infrastructure. The availability of an appropriate device and the capacity to independently navigate this infrastructure was also questioned. Pointing to the fact that ‘there are a lot of people that don’t have money to have a good smartphone’, participants underlined that existing inequalities would potentially be reinforced by digital health infrastructures. The multi-layered nature of the challenge was explained to us in the context of a company participating in the project:

We have […] relatively many employees who don’t have their own email address. If at all, it’s a collective account for the whole family. Or with women, it is the case that men only have an e-mail address, which in turn has discouraged many. […] They [are] just not digitally affine at all […]; they have […] no smartphone, they write text messages.

These issues point to a crucial friction in the context of dominant narratives of continuous availability of and accessibility to data and of individuals being empowered to easily navigate (health) data infrastructures.

Thirdly, co-creation participants reflected on the amount of data work they would be ready to do. Several participants reported that they are ‘not very orderly’ with their health data, simply ‘lumping together whatever they got in connection to the doctor or hospital’. They would thus cherish a vision of automation that would support their data collection. However, in reality, not all health data a person has collected over a lifetime is digitally available and, most importantly, not all health care providers are connected to the platform or have their institutional data infrastructures ready for connection and automated data transfer. A project collaborator reported.

She seemed extremely frustrated and upset that she had to do so many steps. When an error message appeared […] she was unwilling to continue dealing with it. She said she didn’t want to spend any more time on it and stopped participating in the project.

As this quote shows, in order to have complete health data collection available, citizens realised that they would have to engage in data work themselves and be ready to encounter hurdles, which turned out to be a breaking point for some.

Questioning the empowerment narrative

Our analysis of frictions also shows that citizens questioned the vision of empowerment by reflecting on the shifts in responsibilities towards them. This included the responsibility for the completeness of the data collection. Participants pointed to the possibility of data not making it into the platform or only with difficulty, as their format would not be compatible with that of the platform. Some speculated that HCPs might not be connected because of technical limitations or lack of interest, thus not supporting their data collection.

Concerns were also voiced about the responsibility for the accuracy of the data. Just because data has made it into the platform does not mean that it represents what has actually happened, a participant noted.

Imagine that the doctor prescribed me some medication, but I decided for some reason to not take it. (…) It would be wrong information in the end.

If a prescribed medication is not ingested, they argued, because the platform would not differentiate between giving and taking medication, this would lead to inaccurate information being recorded that a responsible user would have to rectify.

Furthermore, HCPs were seen as crucial actors when it comes to the empowering function of the platform. Some citizens were hesitant regarding their ability to make their HCP use the platform, as ‘practices are full’ and doctors are overburdened. The conclusion was clear: ‘if I know he doesn’t like it, he doesn’t have time for it, I wouldn’t share it [my data] much either’. Thus, a customer-based understanding of patients, who through their choices can cause institutional change, misconstrues the prevailing power relations and dependencies between actors in the healthcare setting.

In summary, articulations of frictions are powerful forms of critique providing insights into moments where the platform is contested. This can refer to both its concrete realisation within the project as well as its wider infrastructural embeddedness and future vision. When the platform, as part of the wider health care infrastructure, produces frictions, it creates work for users of the infrastructure. Empowerment here means taking over the work necessary for the infrastructure to function. Some would have to put in more work than others, potentially exacerbating marginalised positions and thus, creating disempowerment.

Discussion and conclusions

In this paper we have unpacked what it means to try to follow the ideals of responsible innovation as guiding principles in organising the long-term transdisciplinary co-creation process of a health data platform. The aim was – as spelled out in the project proposal – to empower EU citizens to become active participants in managing their own health. What can we learn from our co-creation experiences when it comes to the space given to and the kind of critique in, of, and for responsible innovation? And what does this teach us about possibilities and limitations of co-creation (and with it, responsible innovation) in the framework of projects building large-scale digital health infrastructures?

We started out by elaborating on our understanding of critique as addressing the tacit, unconsidered, and unchallenged modes of thought that guide the processes of imagining, designing, and building a digital health infrastructure. We emphasised that we see this as closely related to issues of power and empowerment, power always being an outcome of relations that could in the future be mediated through such a data platform.

In what follows, we bring together our observations. Specifically, we draw attention to the fact that even though the importance of including citizens in the innovation process is highlighted in the policy realm, the reality of co-creation is already pre-framed through the funding scripts, the very format of the project, and the fact that we were engaged in co-creating a sociotechnical infrastructure in a highly sensitive domain – health.

What are these unconsidered modes of thought that deeply framed the outcome of co-creation and shaped the space we could create for future users to have a voice in the process? To start, we encountered a powerful diagnostic narrative at the policy level, reflected in the funding script: a society at risk due to its own success and increased longevity, with only one way forward – digital innovation. The powerful ‘triple-win’ or ‘triple-opportunity’⁴ narrative envisions citizens, the healthcare system, and economic/innovation actors all profiting when unconditionally embracing digital health solutions. Unquestioned digital solutionism can be seen here at work (Morozov 2013). This dominant problem-solution package was mirrored in the perspectives of our partners building the digital infrastructure, but could also be traced in the accounts of the (non)users who quite frequently saw ‘digital progress’ as quasi-unavoidable. Digital technologies were thus seen as driving societal developments, and ‘reasonable citizens’, as those who would willingly comply. This impacted co-creation and with it, responsible innovation, in important ways: it allowed for some adaptations to user needs, but without too much disruption or more profound questioning.

The framing power of the project as the dominant organisational form of doing research and innovation was also considerable. We experienced that it is not only a space of opportunities that different actors can inhabit for a time and that offers the necessary resources to realise a potential solution to an identified problem – in our case, through co-creation. The project actually already defines and thus stabilises what is to be seen as ‘the problem to be solved’ right from the start, and often comes with huge promises, all of which is non-negotiable. In our case, the digital health data platform was predetermined as being an important part of the solution to contemporary challenges in health care, an assumption that allowed for the prioritisation of the digital over the social dimensions in finding solutions and, in the same move, the redistribution of responsibilities (and related work) from institutional actors to individuals.

We also witnessed that the project logic comes with specific temporalities, leading to difficulties in synchronising the work rhythms of co-creation and technological development. While we initially envisaged moving in waves towards incremental improvement in sync with technological implementations, this turned out to be quite difficult. While this could be seen as a minor obstacle, we want to argue with Innerarity (2012) that being able to ‘synchronize with other people’ and to ‘make decisions at the appropriate moment’ is essential for both empowerment and inclusion in defining and implementing user requirements. Thus, being in-time becomes an important ‘locus of social opportunities’ (ibid.). This difficulty in alignment is increased by the promises made by a project application, which come in a format ordered by roadmaps, timelines, deliverables, and key performance indicators, an excellent example of what Maylor and co-authors (2006) call ‘the iron cage of project rationality’. This means that if we take responsible innovation seriously and see co-creation as a means to this end, we have to revisit the temporal modalities in which all this is embedded. We have to consider who can afford to participate in co-creation activities, and leave participants more time to reflect, revisit, and re-question, and to see how, in the long run, relations of (dis)empowerment might emerge as collateral realities (Law, Rubio, and Baert 2011) of well-intended infrastructural projects. This might not follow the efficiency narrative, but would definitely allow for infrastructures that are more inclusive and socially robust.

Yet, while these hurdles just described definitely shaped and partly also limited the co-creation environment, we nevertheless encountered a variety of forms and levels of critique. Some critique was directly voiced by citizens participating in the co-creation activities and could be translated rather straightforwardly into user requirements. But we have also shown that critique materialised indirectly as well, such as through frictions, diverse forms of resistance against prescriptions imposed by the system, and demanding clarifications. Here, the issue remains if these forms of critique can be made ‘audible’ to those involved in the design and development of the technological features of the platform, or if critique has to come in the standard mode or format to feed into co-creation processes.

Furthermore, we underlined the importance of a symmetrical approach that makes space for different kinds of (non)users. What does this mean? It is key to not only attend to critique voiced by those who seem to be keen users of the platform and who are ready to fully embrace the framing narrative. We also need to listen to those who think that such a platform ‘is not for them’, who do not use the platform prototype, or who decide to no longer do so. Taking responsible innovation seriously means being attentive to these kinds of ‘users’, as they more readily question the problem framing and not only contribute to solutions that fit the problem framing.

Finally, the co-creation of infrastructures – even if only a prototype – requires specific attention to power relations, as infrastructures always play a stabilising role. We thus argue for focusing specifically on potentially disempowering relations in the development process, with regard to not only those who engage with, but also those who opt out of using infrastructures. Such an approach allows for the critical questioning of the naturalisation of infrastructures, a perspective which would otherwise treat non-use as an issue to be addressed by education or nudging, allowing infrastructural power to be exercised. Indeed, despite serious efforts to empower citizens through infrastructural co-creation, both empowerment and co-creation run the danger of remaining empty signifiers that allow for the (re-)inscription and perpetuation of unequal power relations. Incorporating citizens into processes of infrastructuring sensitive domains such as health thus always also risks making them collaborators in infrastructuring disempowerment, scripting disadvantages, tacitly supporting epistemic privileges and unevenly distributed labour/profit and risk/benefit. Making responsible innovation and co-creation powerful tools for creating more inclusive health data infrastructures in Europe thus starts with making space for a much broader meaning of critique.

Acknowledgements

We would like to thank all our project partners in Smart4Health for their readiness to engage with us all along the project as well as all the citizens and health care professionals offering their time and insights when participating in the co-creation activities. Parts of this paper were also presented at the EASST 2020 and the EASST/4S 2022 which helped sharpen some of our arguments. We would also like to express our appreciation of the comments from two anonymous reviewers who helped us to sharpen some of our arguments.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Horizon 2020 Framework Programme: [grant number 826117].

Notes

1 Two Smart4Health project deliverables offer detailed insights into the social science project framing and the methodological approach: (Felt, Öchsner, and Rae 2019a; 2019b). For insights into the needs and concerns of citizens voiced in the co-creation process: (Felt, Öchsner, and Rae 2022).

2 Participation was voluntary. We obtained informed consent from all participants in the co-creation engagement. All direct references have been anonymised.

3 These codes refer to policy documents listed in temporal order at the end of the paper.

4 EU Parliament, 2015. Silver Economy as Opportunity, 7; EU Commission, 2017. Blueprint Digital Transformation, 8.