Examining the ethical and social issues of health technology design through the public appraisal of prospective scenarios: a study protocol describing a multimedia-based deliberative method

Our empirical work focuses on three thematic areas where innovations are already emerging and which will alter significantly the moral landscape of future generations [6, 14]: enhancement technologies in teenagers, preventive interventions for genetically ‘at risks’ adults and ageing in a high-tech world. Applying a mixed-methods study design, our specific objectives are to:

In order to generate in-depth findings that can inform health innovation design and implementation, we adopt a Science and Technology Studies (STS) perspective that recognizes that technologies, as well as their designers and users, are value-laden. STS scholars argue that society and technology are necessarily co-constitutive [15], influencing each other’s evolution and setting the stage for changes to unfold in both spheres. Philosophers of technology add to this observation that morality also evolves through new sociotechnical practices [6]. For instance, abortion had to be socially and morally legitimized before prenatal screening could emerge as a viable innovation. Stem cell research would not be as widespread today but for the implementation of in vitro fertilization (IVF) techniques that generate ‘surplus’ embryos. While moral concerns about the status of embryos have pushed scientists to find alternative ways of producing stem cells, such concerns have also been significantly transformed by the use of IVF.

Accordingly, the STS perspective sheds light on the reciprocal relationships between values, society, and technology. This is why we choose to examine the perspectives of public since they constitute the social fabric through which innovations take shape and are implemented. Our research is also grounded in the work of Boltanski and Thévenot [16] on how judgments are formed and articulated in an attempt to settle socially shared situations. From an empirical standpoint, we adopt an empirical ethics approach that seeks to understand not only how actors themselves mobilize values, but also ‘the implicit normativity within facts and technologies’ ([17]: 70). This approach relies on social science methods to generate insights into value-laden social practices, contrasting empirical observations with theoretical frameworks [18]. Below, we summarize three bodies of knowledge that informed our study.

Taking stock of the prospective methods that have been proposed since the 1960s, a Dutch team [6, 31] developed an ethics research method building on sociotechnical scenarios [32]. The goal is to support reflective deliberations about technological change that may take shape in the ‘more distant future’ (10 to 30 years). Scenarios usually take a narrative form, providing a plausible, intelligible story about a putative future. Scenarios may pursue different purposes (research or policymaking), are constructed through different processes (relying on informal versus formal knowledge) and possess different characteristics (simple/few variables versus complex/interconnected events). For Boenink, ‘scenarios are descriptions of possible futures’ that are ‘historically informed speculations’ which seek to foster the self-reflexivity of participants ([6]: 6)^c.

Given our own focus on technology design, we expect prospective scenarios to also spark creativity: participants may reframe or emphasize desirable/undesirable goals and features that have hitherto been marginal in the context of current healthcare systems [33]. Davies and Powell [34], who reviewed alternative approaches to knowledge transfer and exchange—such as storytelling, the arts, or immersive learning—argue that the use of fiction draws attention to the significance of ‘language and metaphor in shaping how ideas are shared and received’ and may help social research better draw in experience, emotion, and appeals to identity. Such elements are pivotal to the ways in which individuals express deeply held beliefs [35]. Narratives ‘permit the exploration of difficult issues in a non-threatening form’, they ‘help individuals and groups address the affective dimensions of being’ and they ‘open up multiple perspectives’ that can lead to concrete changes ([34]: 6–7). For instance, adopting a structured process of facilitated discussions called ‘experience-based design’ in which clinical staff and patients shared their stories and experiences, Bate and Robert [36] showed how it enabled solving quality and safety issues by having staff and patients co-design particular healthcare services. Our study thus posits that prospective scenarios enable individuals to envision and relate to historically informed potential futures, possibly fostering through an immersion in fiction their creativity and reflexivity about the practical and moral implications of technological and social developments in healthcare.

While research on public involvement is steadily growing, one assumption is that participants may choose to revise their position if they are presented with new evidence or arguments [37]. Without deliberation, however, knowledge cannot circulate, challenge normative assumptions, and eventually transform judgments [38]. Because one becomes aware of taken for granted moral norms when they ‘are no longer able to provide satisfactory responses to new problems’ ([39]: 5), our prospective scenarios will expose participants to usability issues (e.g., how over-, under- or misuse of technology may affect safety, autonomy, mobility, etc.) and normative challenges to which they are likely to relate both affectively and rationally.

By examining a two-week long online forum hosted for the residents of the New York City area on how to rebuild the former World Trade Center, Black [40] showed how ‘dialogic moments’ that enable open-endedness and human connection helped explore more fully the complexities of participants’ commitments and values. For Black, dialogue occurs through interaction that triggers ‘momentary experiences of profound mutual awareness of the other person’ ([40]: 95). Such moments can address conflicts based on moral difference, since participants may realize how they are tied to one another and always part of a larger group. When participants are made accountable to other actors [16], dialogue may change their views: ‘Each speaker incorporates and reinterprets the other’s contributions in his or her own way […] the process of trying to convince others may alter one’s own mode of expression but also the reasons one finds convincing’ ([12]: 58). Further, Stromer-Galley and Muhlberger ([41]: 187), in surveying 179 residents of Pittsburgh who deliberated online about solutions to the problem of underused schools, found that ‘expressions of disagreement do not generally harm [participants’] evaluations’ of deliberative processes. They thus recommend that public deliberations ‘experiment with enforced devil’s advocacy’ since it ‘may be useful for more fully exploring all sides of an issue’. Similarly, Walmsley [33], who examined a citizen panel whose task was to design a biobank, underscored how ‘persistent disagreements’ among participants pushed them to be more creative in order to find solutions that could eventually reconcile their varying principles. Overall, when discussing prospective scenarios and considering unsuspected moral challenges, productive group disagreements and dialogical moments may help disclose, clarify, and challenge participants’ normative assumptions and judgments.

We chose to structure our prospective scenarios around three thematic areas (not specific diseases or technologies), not only to maximize the ability of participants to identify with the stories being told, but also to enable setting the putative technical and social changes within a broad moral landscape. Together, the thematic areas help address a large spectrum of usability and ethical issues, while each emphasizes the needs and preferences of different social groups across the life course: the use of enhancement technologies in teenagers; preventive interventions for genetically ‘at risk’ adults; and ageing in a high-tech world.

These thematic areas were chosen by considering the requirements for conducting a solid and instructive comparative empirical ethics analysis, including relevance, diversity, coherence, and feasibility [42]. First, these themes address topics that are already capturing the attention of policymakers and attentive publics because the real-life implications of such developments may have to be dealt with in their lifetime. Second, the themes offer enough diversity to explore the subtleties, prejudices and nuances by which participants may ponder normative issues in different contexts, for different human beings (i.e., increasing teenagers’ performance versus offsetting elderly people’s frailty). Third, the common thread behind these themes, which will increase the coherence of our comparative analyses, is that they all address the technological redefinition of ‘normal’ cognitive and physical states and processes, and the growing emphasis on the ability to exercise agency over one’s body. Finally, examining three themes will provide a more solid contribution to knowledge, while keeping the empirical work and the participants’ tasks reasonable.As indicated in Figure 1, throughout our study, we will receive feedback from a committee comprised of members with expertise in: family medicine, engineering, nursing, pediatric psychiatry, bioethics, geriatric care, genetics, and public engagement (see Acknowledgments). So far, the committee members have helped us narrow down the thematic areas and have commented on drafts scenarios (Step one), they pre-tested our online forum, providing advice regarding its structure and content (Step two), and will also criticize our preliminary findings so we can develop more solid empirical analyses (Step 4).

Following the three-step method of Boenink et al. ^d, we first performed a literature review to document the current moral landscape for each of the thematic areas, gathering quantitative and qualitative published research on ethical aspects of health technology [43]. We paid attention to broad categories of values, such as ‘terminal values (goals or objectives), procedural values (means and process for achieving the goal), or substantive values (criteria justifying decisions and actions for goal achievement)’ ([44]: 61). The observations from this literature informed the kinds of personal and societal dilemmas that could be illustrated in our prospective scenarios.

Second, for each thematic area, we identified how plausible technoscientific developments may interact with the moral landscape and pose regulatory or policy challenges. Here, our team turned to another type of documentation that guided our reflection for choosing the technologies to be featured in each scenario. The Health Canada Science and Technology Foresight Workshop [45] identified close to thirty trends and drivers in five domains (social, technology, economy, ethics/values, environment), several of which were relevant to our thematic areas. Also, the UK Sigma Scan Horizon Scanning initiative [46], which compiled 2,000 sources and interviewed 300 experts, ‘created a cross-sectoral information basis for all foresight activities’ ([47]: 2). Sigma Scan is managed by Foresight, which reports directly to the British Government’s Chief Scientific Adviser and the Cabinet Office. There are more than 270 policy briefs, each rated in terms of its ‘impact’, ‘likelihood’, and ‘controversy’. A preliminary search had identified 51 relevant briefs classified under the following categories: healthcare (6), values (7), body and mind sciences (25), and nanotechnologies (13). These briefs helped our team to identify not only emerging technologies but also convergence across domains that various scientific developments enable (i.e., neurosciences, nanomaterials, epigenetics, renewable energy, robotics, big data analysis, and information technology).

Third, integrating the literature review with these already developed foresight briefs, our research team iteratively drafted and refined prospective scenarios (an initial four-page document was created for each thematic area). From a technological standpoint, our aim was to generate scenarios that were somewhat audacious and controversial, yet empirically plausible [47]. From a sociological standpoint, our goal was to develop compelling and engaging scenarios that would draw participants’ attention to the concrete ways in which society, morality and technology influence each other. To facilitate establishing resonance with a large audience, we chose to create short stories, since ‘part of what makes stories powerful is that they are able to display values and worldviews that are typically not talked about explicitly’ ([40]: 105). A story usually pivots around the ‘quest’ of its key protagonist, which forms the ‘story theme’ [48]. Stories are made up of ‘a series of temporally ordered clauses that provide the orientation, complicating action, and resolution of the story’, while fully formed narratives may also include ‘a coda, which signifies the end of the story’ ([40]: 100). Aligned with the deliberative aim of our online forum, we chose to invite participants to develop and share their own coda.

Table 1 offers a summary of the technologies we have ‘invented’ for developing a coherent and manageable but open-ended representation of each thematic area and their corresponding story themes. For each thematic area, we devised a collective dilemma taking place in 2030 and a personal dilemma arising ten years later. Both bring forward conflicts and normative issues, but the personal dilemma focuses on the specific quests and challenges of one main character (e.g., a teenager, a young adult, and an elderly person). Designed to generate comments by participants in the online forum, each text depicting a dilemma is around 500 words (e.g., a format adapted for on-screen reading). Table 2 summarizes ‘tips for writing a good story’ [48] that guided our writing process. For achieving balance between the benefits and risks illustrated in the scenarios, we used the feedback of our expert committee and the patterns of ethical argumentation described by Swierstra and Rip [39], which are summarized in Table 3. These patterns pinpointed various positions that protagonists in our stories could defend.

The use of audiovisual material is a particularly effective means for enabling informed deliberations among participants and to help articulate their perspectives, because it facilitates the grasping of technoscientific principles with which they are not familiar and increases their ‘reactivity’ by putting ‘images at work’ ([49]: 866). According to Evans and Plows [11], when a specific topic is under consideration, members of the public may have no expertise, interactional expertise (able to converse intelligently about a given topic), or contributory expertise (able to contribute knowledge to a given field). When developing our scenarios, we assumed that participants’ expertise may fall into any of these three categories. As such, the technical and clinical issues were described assuming that participants did not hold expertise.

By choosing to develop video clips, we put forward an ‘audiovisual, elicitation-based’ data collection strategy [49] that aims to enrich the data to be gathered without, however, imposing a pre-determined reading of the issues at play. When using audiovisual-based elicitation methodologies, participants are asked to respond to what they have seen/heard, to comment or contest the expressed claims, and to explain whether or not it alters their understanding of the issues. The video clips were developed with both the face-to-face workshops and the online forum in mind. We created for each thematic area a three-minute video clip, using a combination of static and dynamic images, including actors and voice-over narrative. To guide the pre-production, production, and post-production stages of the video clips, we used a set of tools (scripts, storyboards, sequence-by-sequence shooting plan) that we had applied in a previous project and described elsewhere [50]. The aim of each video clip was to describe the prospective technologies—providing answers to questions such as ‘how does it work’ and ‘what does it do’—and to illustrate the context in which the technologies would be used by 2040.

While knowledge on Internet-based research is still growing [51], we found one study that compared face-to-face and online focus groups with participants affected by colon cancer. Campbell et al. ([52]: 100) observed that while similar themes emerged from both, online groups may attract participants who ‘would be physically unable to attend a face-to-face focus group’ and those ‘who feel reluctant to discuss sensitive health issues in person’. The authors note, however, that the need to type responses may lead participants to shorten or omit certain comments, an issue that might have been exacerbated by their chosen synchronous format. Black [40] argues that storytelling is likely to be fostered in asynchronous online posting because participants can take the time to respond without interruption, express their ideas or tell their stories more completely than if they were in a face-to-face interaction. Thus, the added value of gathering data through two types of deliberative environment is that it increases participant diversity (because of varying time/space constraints) and it enables the examination of variations in how normative claims are articulated and shared (oral, immediate, and public versus written, delayed, and anonymous).

While these two environments call for distinct data collection strategies, an integrated recruitment strategy will be deployed for both. We adhere to the definition of Burns et al. ([10]: 184) for whom ‘the simplest and most useful definition of the public is every person in society’.^g We will foster the participation of individuals from all walks of life; the only exclusion criteria we intend to apply is for persons who are fully trained healthcare professionals. Multiple recruitment tools and strategies will thus be deployed in parallel by reaching out to local organizations that cater to three age groups: young adults (18 to 25); adults (30 to 55); and people over 60 years old. The goal will be to reach participants who may share an interest in social and technology issues, but who may do so from a large range of normative perspectives and reasoning processes. For instance, we will target groups that organize reading clubs, conferences, cultural events, or training activities (for young entrepreneurs, occupational-based networking, or retired people). To ascertain the interest and relevance of each organization serving as an intermediary, a study recruiter (a communication expert who will also act as the workshop/forum moderator) will reach out to them by phone or in person meeting. Then, an electronic invitation letter will be circulated through various newsletters, websites, and social media. The invitation will provide the coordinates of our workshop/forum moderator, our study website and information regarding the Health Research Ethics Board approval. Through a phone conversation, the moderator will gather basic demographic and socioeconomic information about each potential participant. From the pool of interested participants, four groups of 10 members will be assembled using a reasoned sampling technique [42]. We will aim at forming one group of young adults, one group of adults, one group of people over 60 years old, and one mixed group. The goal will be to obtain sufficient homogeneity in socioeconomic status to ensure that the exchanges are comfortable and lively, yet diversified [53]. Those who are not available at the day and time set for the workshop will be invited to participate in the online forum. Akin to a snowballing strategy, participants will be asked to share the invitation to the forum with friends and relatives.

Morgan [54] argues that the ability to observe the extent and nature of participants’ agreement and disagreement is a unique strength of focus group research. Although methodologists generally advocate seeking homogeneity in group composition around age, gender or occupational status, deliberative focus groups bring together participants who can engage in more challenging discussions [53]. Hollander [55] underscores that the context-dependent data generated through group discussions is socially situated. This perspective emphasizes the larger social structures within which discussions take place, as well as the various motivations and expectations of participants and the position(s) they represent as they contribute to ongoing discussions [56]. The group moderator is pivotal in facilitating open and detailed exchange with and among participants. Crossley even points out that a moderator may deliberately try to trigger a strong reaction, in order to obtain ‘valuable insights into how people may actually respond’ to health interventions ([57]: 1479).

To support an in-depth discussion of our three prospective technologies, we will host four 3.5-hour deliberative workshops. Once each participant has introduced her- or himself, the video related to the first thematic area will be shown and then each participant will be asked to share with the group two or three features of the technology he/she sees as desirable as well as two or three undesirable features. A group discussion will ensue focusing on potential ways to improve the technology. The same structure will be applied to the other two thematic areas, with about one hour per theme and a 15-minute break (refreshments and snacks will be served). Forty-five Canadian dollars will be given to each participant at the end of the workshop in recognition of their participation.

Our forum will be hosted on a login/password-secured page and facilitated by the same moderator as the workshops. Defining a fixed timeframe in asynchronous online forums creates momentum, fostering a more intense participation. Our forum will thus run over a four-week period, starting after the fourth workshop will have taken place. Participants will be invited to read and agree to a consent form, to view a brief animation explaining the study, to read the dilemmas, to view the videos and to respond to open-ended questions to kick-start a collective debate. Participants will be able to return to the forum whenever they wish and comment on each other’s comments. A cheque of forty-five Canadian dollars will be sent by regular mail to each participant at the end of the forum.

Our analyses will rely on three main data sources: observation; deliberative content; and a self-administered participant survey. Table 4 indicates how this mixed-method strategy capitalizes on the respective strengths of the datasets that will be generated. While the deliberative content represents a rich source of data for documenting the process and outcome variables related to all three research objectives (ROs), observation provides additional insights into the processes underlying RO.1 and RO.3. The survey, by documenting the views of all participants, including those who may occupy the less ‘discursive place’ in a group, will add nuance to our qualitative findings and support a structured assessment of the prospective method (RO.3).

The workshops and online forum will be moderated by someone with experience in group processes and directly observed by someone trained in qualitative research. The observer will take notes and the discussion will be recorded and transcribed verbatim (4 × 3.5 hrs = 16 hrs of audiotaped discussion). Our observation technique is ‘non-participant’ because the observer plays no instrumental role in the group discussion, although the observer has a certain influence simply by being present [58]. By contrast, the moderator is a participant-observer because this person plays a pivotal and overt role in the workshop discussions and is active in the online forum. From a qualitative research standpoint, this represents an empirical opportunity to examine more broadly the context in which the data collection process unfolds and its influence on the phenomenon [56, 58].

For the workshops, detailed observation notes will be recorded in two steps. First, during the event, the observer will record short notes on a pre-structured form describing the characteristics of the interactions between participants (e.g., climate, turn taking, flow/intensity of interactions). Second, the day after the workshop, an annotated summary of the deliberations will be written, integrating and structuring the notes recorded on the form, and exploring analytical insights. Observations of the online forum dynamics will be complemented by navigation data. For instance, the number of page visits per participant, the number of comments per participant and the ratings of the comments will be downloaded into a spreadsheet. Using a Wordpress platform, we will follow the Journal of Medical Internet Research checklist, which covers methodological and technical issues, including ethics (i.e., consent, privacy, security) and calculation of ‘response’ rates (e.g., the view, participation and completion rates of unique visitors, based on the use of cookies and log file analysis) [59].

Verbatim transcripts of the workshops and the content of the online forum will be integrated into the qualitative indexing software QDA Miner. They will be analyzed using a mixed coding strategy, involving a set of predetermined codes based on the list of variables described in Table 4 and a set of empirically-induced codes. For the workshops and the online forum, both intra- and inter-group analyses will be performed (see below).

All participants will be asked to complete an online survey when the forum closes. Table 5 lists the open- and close-ended items of the survey. Given that our recruitment strategy emphasizes diversity, and as qualitative research standards of rigour suggest, the survey will also enable the gathering of demographic and socioeconomic data, contributing to an in-depth description of our participants [60]. A specific set of items to compare the two deliberative environments is reserved for participants who will have attended one workshop. With the survey data and given the number of respondents (limited to 40 in the workshops and expected to be around 65 in the forum), descriptive statistics will be compiled (distribution of frequencies, measures of central tendency). To explore whether different patterns of responses are observed in socioeconomic sub-groups or across thematic areas, analysis of variance (ANOVA) will be performed.

Multiple data studies require a careful management of data collection and analytic processes in order to maximize the cross-fertilization of data sources throughout the study [61]. To integrate the qualitative and quantitative findings, we will use analytical tools suggested by Miles and Huberman [62], including comparative matrixes to identify patterns based on the similarities/differences observed across thematic areas, deliberative environments, participant characteristics, and chronologies, which are particularly helpful to examine conversational flows and argumentative ‘turning points’. During such exploration of the empirical material, constant comparative analysis forces one to search for empirical observations that may refute one’s emerging interpretation [63].

An inductive approach to the interpretation of our findings will prevail since the principles of qualitative research predominate in our mixed-method study design [42]. These analyses will be guided by our research objectives. For RO.1, to analyze the knowledge and normative assumptions from which participants appraise and reach ‘situated arrangements’ [16] about the value of sociotechnical changes across the three thematic areas, we will use a template that we developed for focus group research [56]. This analytical template brings forward how epistemological content and group processes influence each other, taking the moderator’s influence into account. Similarities and differences between the groups and the two deliberative environments will be examined systematically through comparative matrixes. To identify usability and ethical issues (RO.2) as required for an empirical ethics analysis, our attention will focus not only on the participants’ views, but also on the issues that are omitted or reframed. Here, comparisons across thematic areas will also be emphasized, using the scholarly literature and our expert committee to strengthen our analyses. Finally, assessing whether the participants’ reflexivity, creativity, and critical thinking have been fostered by the prospective method (RO.3) requires identifying whether and how the processes (including videos and short stories) may have supported participants’ ability to envision, describe, and make critical observations about design assumptions and features, and articulate their own alternative visions of the future.

We will host a debriefing session with the expert committee at step four once preliminary results are available. The goal will be to consolidate our analyses by having these distal collaborators ask new questions and suggest competing explanations in order to avoid the phenomenon of ‘collective persuasion’ and to identify neglected dimensions. Overall, we will apply recognized strategies for ensuring the rigour of qualitative research [60]. First, seeking to increase the credibility of our findings (a qualitative criterion similar to internal validity) and using the guideline developed by Datta [61], we will iteratively revise and adapt our study design when needed and pre-test the workshop/forum guide and survey. Second, we will provide as much information as possible regarding the participants’ characteristics and the context in which the deliberations unfold. This, coupled with the debriefing session on preliminary results with the expert committee, will contribute to defining which aspects of our findings may be transferable to other settings. Third, we will make explicit how our own assumptions regarding design features may influence data collection and analyses. Fourth, we will keep a detailed fieldwork diary, support our analyses with ‘raw’ empirical material, and facilitate access to our instruments and database (procedural accountability).

While steps one and two were completed in January 2014, steps three and four should be completed by the end of 2015. We expect the transcription of the 16 hours of workshop deliberations, the downloading of the online comments and navigation data, the cleaning of the exit survey data, and data analysis and integration to take up to 18 months.