Etikk i praksis. Nordic Journal of Applied Ethics (2019), 13(1), 55-80
http://dx.doi.org/10.5324/eip.v13i1.2525


Responsibility for Assistive Technologies: Product Assessment Frameworks and Responsible Research and Innovation


Erik Thorstensen

Work Research Institute, Oslo and Akershus University College of Applied Sciences, erikth@oslomet.no

The approach to innovations known as Responsible research and innovation (RRI) aims to move the innovation system towards creating products that strive to realize social values along with economic benefits. This paper discusses the systematic assessment of assistive technologies in order for them to meet the aims expressed in RRI. A central issue in the discussion is how to facilitate an integration of insights from the discourse on RRI with more established assessment approaches such as Health Technology Assessment (HTA). Based on the literature on existing socio-ethical assessment tools, I investigate how these tools can be combined with HTA and how they can add perspectives from RRI that might increase the socio-ethical value of assistive technologies. Through a discussion on how to understand RRI, HTA, assessment and integration, I suggest a list of four possible approaches that have the potential to be applied as assessment approaches that integrate insights from RRI and HTA. These are then evaluated on their ability to address issues that have emerged from a literature review on RRI and assistive technologies, on empirical studies in this technology field and on their product focus. In conclusion, I argue that the Ethical Impact Assessment, the Socratic approach, the Ethical Matrix, and the HTA Core Model seem to be the most promising methodologies, but that these need adjustments to cover substantive themes from RRI.

Keywords: energy scenarios, justice, future generations, capabilities approach
 

Introduction

Worldwide we see a proliferation of assistive technologies in hospitals, institutions and homes, ranging from high-tech robots to low-tech analogue alarms. There are high hopes for assistive technologies in policy circles as part of an approach that might empower people with disabilities of all ages to reside at home longer, as opposed to living in different kinds of institutions. In the last 25 years, researchers, policy-makers and health professionals have especially focused on older adults in order to accommodate smart homes for this group (Thygesen 2009). The vision behind this practice has been to increase older adults’ independence and self-esteem, while reducing both health care expenses and an increased number of employees in health care delivery. In addition, several technology firms are launching solutions promoted as meeting these policy objectives. In the field of assistive technologies, we then encounter both policy pull and technology push, as von Schomberg (2011) calls them. Seen from a socio-ethical perspective, concerns over assistive technologies relate to issues such as how care is perceived, experienced and given (Roberts & Mort 2009), the possible medicalization of homes (Hofmann 2013), focus on caregivers at the expense of users (Topo 2008), and possible conflicts between empowerment of older adults and the larger goal of addressing ageing as a demographic challenge (Pols & Willems 2011).


The Assisted Living project is a research project using insights from Responsible Research and Innovation (RRI) in order to research and develop assistive technologies. The project consists of a facility for independent living, an SME for developing assistive technologies and smart home solutions, technology assessors, and researchers in technology assessment, ethics, ICT, nursing, and occupational therapy. We compare the assistive technology developed in the RRI-based project to an existing assistive technology developed in a project that is not based on RRI. The intention is to obtain information on the possible influence of process on product.


In the health technology field, Health Technology Assessment (HTA) is a common denominator for different policy advice formats on novel products or solutions. Consequently, this approach has high legitimacy and high policy relevance. In order to make RRI relevant and legitimate for decisions regarding health policy, I will explore how different assessment frameworks used for socio-ethical assessments of health technologies reflect the core components of RRI. A number of such assessment dimensions have been identified in the current context in a previous paper (Thorstensen 2017b). I will then focus on evaluating assessment approaches, but this study does not claim to exhaust all the possible socio-ethical approaches that might be combined with HTA and also resonate well with RRI.


This paper argues that Ethical Impact Assessment (ETIA) (Wright 2011),1 the Socratic approach (Hofmann 2005b; Hofmann, Droste, Oortwijn, Cleemput, & Sacchini 2014), the HTA Core Model (Lampe et al. 2009), and the Ethical Matrix (Kaiser & Forsberg 2001) are relevant candidates for assessing assistive technologies in line with HTA and the central thinking in RRI. In order to reach this conclusion, this paper begins with a general introduction into RRI and product assessment, followed by an explication of what a product assessment methodology of assistive technologies in line with RRI should contain. Based on literature that assesses health technologies and RRI and assistive technologies, I present the possible candidate methodologies and relate these to the aforementioned criteria.


Responsible Research and Innovation

Responsible Research and Innovation (RRI) is a term used for research policy, for a cross-sectorial approach to science and innovation governance, and for an intended practical approach to research and innovation (Ribeiro, Smith, & Millar 2016). For the Assisted Living project, the last term of the phrase RRI applies. When conceived of as a method for innovation, Assisted Living project utilizes a conception of RRI as containing:

 

  1. A specific focus on addressing significant societal needs and challenges
  2. A research and development process that actively engages and responds to a range of stakeholders
  3. A concerted effort to anticipate potential problems, identify alternatives, and reflect on underlying values, and
  4. A willingness from relevant actors to act and adapt according to 1–3 (Wickson & Forsberg 2015: 1164)

 

The Assisted Living project understands itself as responding to the “Ageing society” as a “societal need or challenge.” The justification for using RRI in assistive technologies lies in the connection between RRI and these societal challenges (von Schomberg 2013).2 In addition, all four points above constitute the normative foundation for establishing how to assess a product in line with RRI.


Part of the novelty in the Assisted Living project, is the use of RRI as a guiding principle for doing research and innovation in practice. RRI has a novel normative approach to research and innovation in its study of the moral purpose of innovations. This approach aims to combine central value aspects in societies with the social and economic benefits of research and innovation through close links with industry (Pacifico Silva, Lehoux, Miller, & Denis 2018). Different approaches and RRI both include affected parties, users, stakeholders and policy-makers. However, they differ somewhat in that von Schomberg (2011) has tried to explicate how to conceptualize these values whereas Owen, Macnaghten, and Stilgoe (2012) tend to emphasize that these values should be sought among the participants, and operationalized towards application in research (Owen 2014). Wickson and Forsberg (2015) strive to combine these two approaches.3 In the European Union, RRI has been presented in relation to the research policy areas (the so-called “keys”) of public engagement, open access, gender, ethics, and science education (and earlier also governance).


Product assessment

A more specific meaning of the term “product assessment” is conditioned by background and discipline. In an early contribution to RRI, Stilgoe et al. mention some questions that are relevant for a product assessment:

 

·     How will the risks and benefits be distributed?

·     What other impacts can we anticipate?

·     How might these change in the future?

·     What don’t we know about?

·     What might we never know about? (Stilgoe, Owen, & Macnaghten 2013: 1570)


von Schomberg (2011) argues from a different perspective that in RRI, products should be evaluated through their ethical acceptability, sustainability and social desirability. When it comes to assistive technologies for the elderly, products must have acceptable functioning and design in line with the users’ and central stakeholders’ values and preferences, how they relate to quality of life, and general ethical principles. Central to the overall idea of RRI is to reflect on the purpose of research and innovation (Owen et al. 2012; Stahl & Coeckelbergh 2016).
  
The concept of a “product” needs further clarification.
4 My understanding follows Brey (2012), who suggests three relevant levels for an ethical analysis of technology: 1) the technology level, 2) the artifact level and 3) the application level. With solutions at the application level, Brey refers to “the concrete use of a technological artifact or procedure for a particular purpose or in a particular context, or a specific configuration of an artifact to enable it to be used in a certain way” (Brey 2012: 311). In the present case, “application” would thus refer to an artifact being used by elderly individuals with diminished cognitive capacity in their homes in order to perform certain tasks or sets of tasks. Brey’s distinction between technology, artifact, and application makes it possible to distinguish different aspects of product assessment and to differentiate it from Technology Assessment – which has as its focus Brey’s “technology” level. I follow Brey in this article, and by “product” I refer to Brey’s application level that includes a specific purpose and the configuration surrounding the application in order for it to achieve this purpose.
  
Recently, researchers from the Responsible Industry project presented an approach where they argue that effective RRI should be assessed according to “certain identifiable consequences” (Stahl et al. 2017). These consequences are derived from the RRI “keys” of the European Commission (2012) as well as from an expert group on suitable policy indicators for RRI (Strand et al. 2015). Stahl et al. have employed the keys together with the process criteria from RRI, as listed above, to create a project self-assessment tool to monitor the RRI quality of a project or an organisation (ORBIT 2017). The ORBIT tool and similar RRI process evaluation tools fall outside the scope of the meaning of product assessment in this article.
 

Health Technology Assessment
The historical background for HTA is similar to that of the different forms of Technology Assessment (TA). They both originated in and through the same impulses that created the U.S. Congressional Office of Technology Assessment (OTA) in 1972 and with the same focus on efficiency, understood as the cost-effectiveness of a given health intervention. Traditionally, pharmaceuticals, vaccines and medical equipment have been easier to assess through HTA, while different types of health care practices have been more challenging to assess due to their interlinked and complex nature (Banta 2003). Adding to the complexity, studies reveal a lack of clear definitions in the literature on how value is or should be defined (Antoñanzas, Terkola, & Postma 2016).


Currently, the INAHTA (International Network of Agencies for Health Technology Assessment) defines HTA as follows: “Technology assessment in health care is a multidisciplinary field of policy analysis. It studies the medical, social, ethical, and economic implications of development, diffusion, and use of health technology” (INAHTA 2016). This expands the scope of the original use to include social and ethical aspects in a systematic manner. HTA still has as its main purpose to assist in decisions related to policy. Now, the policy context can be at any level, including at the low level of a hospital or the higher level of a municipality assessing new interventions.


With respect to RRI and HTA, Cuijpers and van Lente (2015) present a strong case for how HTAs do not allow for different sets of logic or values. In a similar vein, Moors and Peine (2016) argue that HTA assumes a perspective that downplays solutions enhancing individual agency and promotes medicalization. The integration of the social and ethical dimensions has suffered from being analyzed independently of the other epistemological dimensions in the HTA. As a remedy to this dissociation, suggestions have been put forth that mainly serve to add complexity to the procedure while subsuming social and ethical issues under general effectiveness research. (Refolo et al. 2016). Another dimension that has rendered this integration problematic has been that the specific methods have been immature (Hofmann 2005a).
The EU-funded project Integrate HTA suggests including assessments of socio-cultural dimensions into HTA, in addition to or in parallel with an ethical assessment (Lysdahl et al. 2016; Mozygemba et al. 2016). Including other normative issues seems to be in line with the move from ELSA (Ethical, legal and social aspects) to RRI. Since HTA is by definition an interdisciplinary endeavour and since RRI revolves around continuously discussing and reflecting on assumptions, inherent values and responses, the preferable mode of integration in the product assessment approach and the HTA would seem to be those of coordination and of interaction, as defined by Hofmann et al.:

 

·     Coordination: Ethics is still an independent part of the HTA, but the role and weight might differ depending on contextual factors. The results from the other parts of the HTA will influence the input to the ethical parts, but not the methodological choices of the ethicists.

·     Interaction: Ethics and other disciplines will be in a continuous exchange of viewpoints and results with possible redefinitions of the policy question, methods, and relevant comparative cases for all disciplines. (Hofmann et al. 2015: 131-132)

 

This means that there should be some form of interaction between an assessment method based on socio-ethical concerns and monetary and health concerns in an HTA.

Indicators for choosing an approach

In an earlier study based on a review of the RRI and assistive technologies literature as well as dialogues with stakeholders in the current project, I documented that a range of values ought to be taken into account when developing assistive technologies (Thorstensen 2017b).5 I will now recapitulate the central findings from this article. A central concern is that a product assessment approach used in an RRI project should be able to incorporate these points.


The good life:

A central issue for the assessment of assistive technologies is that the stakeholders are included in thinking about and reflecting upon what constitutes a benefit in the particular case. These benefits should not be reduced to utilitarian concerns but should also include social relations, values such as exercise and amusement, and reflect modifications to the lived environment through devices and a novel type of care. Including the stakeholders’ own understanding of benefit can be carried out as a part of the assessment (when interviews or other qualitative approaches are used) or in phases prior to the actual assessment.


Risks and benefits before use:

The above concerns should be used in a solid and complete pre-trial (i.e. ex ante) testing of a product. Such testing should also encompass privacy, safety, security, information collection and sharing as well as a thorough review of the evidence base of the product. Assessments of the product’s economic, social and ecological sustainability ought to constitute a part of this evidence base. However, this criterion will be disregarded for the present purpose, since we are striving for an assessment method that is applicable both to products made in an RRI process and other products (without any possibility to conduct an ex ante assessment), where an assessor or an assessing body lacks information about the process.


Distribution of risks and benefits:
An assessment should include considerations about the distribution of risks and benefits. In particular, one should be aware that assistive technologies might entail a transfer of the risks towards the elderly. If such a transfer seems to be the case, alternatives to the solution should be explored or the transfer needs to be justified in terms of other types of benefits.

Distribution of responsibilities:

As with the distribution of risks and benefits, the distribution of responsibilities regarding the product should be well understood, but also open for discussion. This relates particularly to the division of responsibility among the different operators in the service provisions connected to a product.


Training:

A final concern is that the assessment also investigate how users are introduced to the product, since health professionals and users alike identify didactical approaches (or their absence) as key aspects in the process.

  
These criteria overlap to some extent with the previously referenced product assessment question posed by Stilgoe et al. (2013), who mention the distribution of risks and benefits. The dissimilarities, however, have to do with future impacts and changes in such impacts. As Cuijpers, van Lente, Boenink, and Moors (2014) point out, the future orientation inherent to RRI demands a responsibility towards an imagined or assumed future. In order to fulfill this demand, I believe that an assessment approach would also need to include an exploration into contingency, in order to assess plausible changes (Stilgoe et al. 2013). Recently, Pacifico Silva et al. (2018) investigated Responsible innovation in Health (RiH) in a policy setting, and they argue that this should encompass five value domains consisting of nine dimensions.6 Their addition to the themes above are the health relevance, the organizational dimension and frugality. Health relevance relates to the innovation’s intent to address neglected or under-prioritized disease or adding to incremental benefits for the more affluent. The organizational dimension studies how or if the producer – inspired by the literature on social entrepreneurship – seeks to provide additional values to society. The frugality dimension is rooted in the literature on health economics showing the pressures on the health system by novel technologies and focuses on whether greater value is produced through fewer resources. Regarding the RRI qualities of the frameworks, I use frugality, social entrepreneurship and health relevance as parameters based on the consideration of the frameworks suggested by Pacifico Silva et al. (2018)


The following criteria are central in analyzing how to address the quality dimension of ethical assessment approaches: comprehensiveness and a broad inclusion of values, user-friendliness and transparency (Beekman et al. 2006; Forsberg, Shelley-Egan, Thorstensen, Landeweerd, & Hofmann 2017; Kaiser, Millar, Thorstensen, & Tomkins 2007). In practical terms, this translates to an assessment that includes a broad range of relevant values and other ethical aspects. The ability to practically operationalize these criteria is key, so that other persons (users, stakeholders, decision-makers or the public) can understand how conclusions follow from the relevant values and other ethical aspects.7
 
 
The values proposed by the desired assessment approaches should at least encompass a good life, risks and benefits before use, risks and benefits in use, distribution of risks and benefits, distribution of responsibilities, and training within a fixed and transparent framework that allows for case-relevant and socio-ethically relevant aspects and arguments. Stilgoe et al. (2013) supply the criteria of other possible impacts, future changes of impacts and systematic inquiry into ignorance, as well the purpose of the innovation. From Stahl et al.’s (2017) substantive list of product assessment criteria I retain gender/equality and diversity, open access, social justice/inclusion, sustainability and science education – and here ethics should be added for systematic purposes. Based on Pacifico Silva et al. (2018), I use health relevance, frugality and social entrepreneurship as criteria. In terms of the process criteria of governance and public engagement from the European Commission’s six “keys,” I believe that governance is covered by distribution of risks and benefits and distribution of responsibilities, while public engagement is covered by comprehensiveness and a broad inclusion of values, since these presume the inclusion of a variety of perspectives and viewpoints.


Discussion of assessment approaches

Several scholars have recently studied in detail some of the possible candidate frameworks (Forsberg et al. 2017) and their relation to HTA (Hofmann et al. 2015; Lysdahl et al. 2016). I will use their work as a point of departure in addition to my work in the literature review for the Assisted Living project to present the possible frameworks that may be suitable to apply as an RRI product assessment framework. I have selected the frameworks and methodologies discussed here based on whether they appeared in the reviewed literature or if they have been discussed in relation to HTA and socio-ethical issues.8 Some of the approaches reviewed by Assasi et al. (2014) propose a variety of a literature reviews, but these will not be considered as they constitute what Hofmann calls an “add-on approach.”9


According to Hofmann et al. (2015) and Lysdahl (2016), the possible interactive and coordinated approaches between HTA and socio-ethical assessments are Social Shaping of Technology (SST), Axiological (Socratic, EUnetHTA), Constructive Technology Assessment, Interactive Technology Assessment (iTA), and the Ethical Matrix. These authors also mention utilitarianism, discourse ethics and wide reflective equilibrium as interactive ethical approaches. However, these approaches would not qualify as frameworks in the sense that they provide practical guidelines for addressing ethical quandaries. Since utilitarianism is a theory of normative ethics, discourse ethics is a way of justifying and legitimizing moral judgements, while
wide reflective equilibrium is an approach that might well be used in order to analyze or to develop coherence between principles and judgements (Forsberg 2007c). Of the methods or frameworks discussed in the mentioned literature review (Thorstensen 2017b), only the Ethical Impact Assessment has the potential to focus on an application (Forsberg et al. 2017), whereas the remaining ones are mainly used in planning and/or interventions.10


Constructive Technology Assessment (CTA) is not a form of product assessment, but an approach to identify societal impacts of technology developments early, in order to make these impacts as beneficial as possible (Schot & Rip 1997).
Interactive Technology Assessment (iTA) shares the same purpose as CTA and does not include interaction between stakeholders, but is limited to interviews and analysis (Reuzel 2004). Social shaping of technology (SST) is explained by Lysdahl et al. to be a position that sees

 

technology as the product of societal processes within industry, research institutes, governmental bodies, and society at large, rather than an independent artefact that has a certain, measurable impact on its target. Therefore it is important to understand the engagement and strategies of various actors, and the way various problems are defined and resolved (2016: 64).

 

SST’s strongest asset is its ability to create novel insights as to the interaction between society and technology, which is invaluable. The challenge of this approach is that it lacks structured resources to guide an assessment towards specific topics, which limits its comprehensiveness and transparency (Forsberg et al. 2017). Consequently, the approaches I will address in more detail in this paper are the HTA Core model, the Socratic Approach, Ethical Impact Assessment and the Ethical Matrix.

The HTA Core model