There is an increasing consensus that UC and VSD are necessary requirements for ethically sustainable development of assistive and rehabilitation technology [7, 8, 13]. However, little analysis is available on the prerequisites of successful adoption of such approaches. Based on the inherent goals and objectives of UC and VSD described above, we argue that four basic normative requirements are necessary for the successful implementation of ethical NART.
Minimization of power imbalances
Both UC and VSD presuppose the minimization of power imbalances in decision-making and a certain degree of inclusiveness and democratization in the design process. This shift in the location of power across the technology design continuum is best achieved through a goal-oriented cooperation among designers, developers and end-users. This principle implies that in order to be involved on an equal footing in the design process, all stakeholders should be incentivized to share common goals that could be pursued through coordinated and cooperative efforts. In fact, in absence of common goals or even in presence of mutually conflicting objectives between different stakeholders (e.g. designers vs users), no successful cooperation within the UC and VSD framework is likely to occur. An example of conflicting objectives between different stakeholders is the observation that designers and developers often prioritize the effectiveness of a new technology whereas users often prioritize usability. Effectiveness refers to the accuracy and completeness with which end-users can achieve certain goals in a certain environment. Usability is the easiness and extent to which a technology can be used by users to effectively achieve these goals. This discrepancy between effectiveness and usability has been particularly investigated in the context of assistive BCI, one of the technologies featured in the CYBATHLON 2016 [1]. For example, a review of BCIs as access pathways for people with severe disabilities has shown that most current prototypes are developed with focus on speed and accuracy instead of usability [39]. These conflicts of objectives can have detrimental consequences for rehabilitation as they could concur in the phenomenon of technology abandonment. This refers to the fact that users of an available assistive or rehabilitation technology might stop using it after an initial phase, a phenomenon that is particularly common with technologies for home use. Scherer has reported that about one third of all assistive technologies are abandoned, and many others might continue to be used sub-optimally due to unease and discomfort. As she states: “we have no information about the number of people who continue to use devices they are unhappy or uncomfortable with because they cannot abandon them without facing more severe consequences” [40]. In addition, the absence of common objectives among different stakeholders involved in the design and development of assistive and rehabilitation technologies is likely to cause the so-called “problem of many hands” [41]. This problem denotes the risk that in complex process where multiple stakeholders are actively involved errors can be made although no class of stakeholders acted in an explicitly reckless or negligent way.
To overcome this problem, there is a need for harmonizing the objectives of all relevant stakeholders involved in the design process through an iterative and dialogic confrontation. This could be achieved by creating cooperative scenarios where all stakeholders are incentivized to pursue a common goal or objective.
Compliance with biomedical ethics
The second requirement for the successful implementation of ethical assistive technology in rehabilitation is compliance and coherence with biomedical ethics. NARTs are integral part of biomedicine and biotechnology. Nonetheless, their degree of ethical scrutiny by biomedical ethicists is often lower compared to other domains of biomedicine and biotechnology such as pharmacological interventions. This is probably due to many factors including the relative novelty of NART, a less stratified history of misuse and different risk-related perceptions among professionals.
We argue that successful technology development via UC and VSD presupposes the compliance with biomedical ethics. As we said before, this requirement can be fulfilled through compliance with multiple approaches and values in biomedical ethics such as utilitarianism, Kantianism or virtue ethics. Among others, one viable and, according to some, easy-to-implement approach is principlism, a practical approach for ethical decision-making that focuses on four common-ground moral principles: beneficence, non-maleficence, autonomy and justice. Research shows that the principlist approach has the largest circulation among health professionals and the highest prevalence in ethics curricula for health science students [42, 43]. This fact could, ceteris paribus, guarantee better acceptance and easier implementation among health professionals. However, it is important to highlight that, at any rate, referring to any specific ethical theory in a predetermined manner risks to preempt normative input from users. Therefore, it is important that, at any rate, ethical theories or principles are chosen based on the needs and values of users, and adapted to these needs and values through an iterative and flexible process. In other words, the investigation of the users’ needs and values should determine which ethical content is most suitable for a certain technology in a certain patient population, not vice versa.
Principlism, uses a “common morality” approach and “mid-level” prima facie principles: beneficence, non-maleficence, respect for autonomy and justice [31]. Beneficence is the promotion of the wellbeing of people with disability through the successful implementation of assistive and rehabilitation technology. As we have seen above, the field of assistive and rehabilitation technology urges a broad concept of beneficence that is not only focused on the effectiveness of new technologies but also on their usability.
Non-maleficence is the principle of preventing or minimizing harms associated with the use of assistive and rehabilitation technology. This principle is promoted through the implementation of safeguards for the safe and secure use such as the precautionary approach, namely the idea that technologies whose consequences are difficult to predict should be first investigated in a safe setting [19]. Neurorehabilitation experts have tried to systematize the principle of non-maleficence in relation to robot-assisted neurorehabilitation [44]. Their model is based on the postulation of three fundamental laws called the laws of neurorobotics in rehabilitation, a re-elaboration of Asimov’s laws of robotics [45]:
-
(I)
A robot for neurorehabilitation may not injure a patient or allow a patient to come to harm.
-
(II)
A robot must obey the orders given it by therapists, except where such orders would conflict with the First Law.
-
(III)
A robot must adapt its behavior to patients’ abilities in a transparent manner as long as this does not conflict with the First or Second Law.
The first law postulates that rehabilitation robotics should be safe not only in terms of movement, but also from other medical points of view. This can be achieved by designing new products in accordance with international standards such as ISO 13482:2014 [46] and through careful consideration of unintended harms, where harm is understood as any “possible damage to patients” including discomfort and time spent on ineffective rehabilitation. The second law postulates that assistive technologies should not replace therapists, but rather complement existing treatment options. Therapists should always be on the loop of robot-assisted rehabilitation and maintain a position of control in relation to the adjustment of technological parameters, the avoidance of harmful compensation strategies and identification of trade-offs between rehabilitative goals and the psychological dimension of patients. Risks of reduced control over technological parameters such as is the discrepancy between the desired and actual values of some parameters of the electromechanical Gait Trainer [47] should be prevented. At the same time, based on the third law, automatic features and artificial intelligence might be used to support rehabilitation therapists by performing all the control changes required for a successful therapy.
The principle of respect for personal autonomy, as stated above, should not be seen exclusively as the promotion of decisional and executional autonomy, but of self-realization as well. To achieve that, UC approaches should not only involve the active participation of end-users and investigate their perceptions only in relation to quantitative parameters such as effectiveness and usability, but should proactively incorporate user-driven ethical and psychological factors in product design. Given the requirements of context-sensitive design, this attempt to “materializing morality” [48] through assistive technology should be dependent on the specific context and environment of end-users.
Finally, justice is the principle of biomedical ethics that requires assistive technologies to be fairly accessible to users, affordable across various socioeconomic classes, and evenly distributed across rehabilitation clinics in various world regions. While this principle can be incorporated into product design by favoring scalable, low-cost and pervasive technologies, yet design alone might be insufficient. In addition to that, justice-promoting policies should be pursued at various levels of health-technology regulation. Reimbursement policies and State incentives have been advocated elsewhere as possible justice-promoting regulatory interventions [27].
Translationality
The third requirement is translationality. In fact, the ethical goal of maximizing wellbeing for all individuals with disability through the use of NART is highly dependent on the process of translating research from the designing lab to the rehabilitation center. In order to maximize the societal benefits of NART, we need to ensure that new technologies actually reach the patients or population for whom they are intended and are implemented correctly [49]. Slow or incomplete translation across bench, bedside and community ─ which the European Society for Translational Medicine calls the “three main pillars” ─ is likely to reduce the beneficial impact of assistive technology on the global healthcare system. According to the Institute of Medicine’s Clinical Research Roundtable, two distinct phases in the translational process are in particular need of improvement: the first translational block (T1) prevents basic research findings from being tested in a clinical setting; the second translational block (T2) prevents proven interventions from becoming standard practice.Footnote 4
Social awareness
Finally, the fourth requirement is raising social awareness and favoring knowledge dissemination across society. The public is often skeptical or reluctant regarding the use of new technologies because of lacking knowledge on the technology and its applications [50]. Sociologists have identified historical patterns and dynamics of opposition to technological innovation. For example, Juma has explored the multi-layered dimensions of socio-political resistance to various types of technological innovation including biomedical technology. These include established social norms, financial considerations, health implications, social disruption, as well as prejudices or human ignorance [50]. Patterns of resistance to new technologies have also been observed in the specific context of healthcare technology [51]. This opposition seems to be particularly significant in relation to technologies that operate in proximity to the human body such as wearable devices and neural prosthetics. A 2014 Pew survey showed that 53% of Americans think it would be a bad thing if “most people wear implants or other devices that constantly show them information about the world around them.” In contrast, just over one third (37%) think this would be “a change for the better” [52]. Since many NARTs operate in close proximity or direct physical contact with patients, and have invasive or non-invasive connections with the human nervous system, they are likely to be affected by these negative public perceptions.
The media, a major catalyzer of attention and knowledge on novel technological possibilities, have started only recently to properly cover the domain of neuroengineering, assistive and rehabilitation technology. Concurrently, since NARTs are still in an initial phase of the technology life cycle, their pervasive implementation might still be limited by enduring habits of health professionals, financial limitations and issues of resource allocation or conservative managerial decisions ─all phenomena that have already been observed in other sectors of healthcare technology [53,54,55]. If improving the effectiveness, usability and ethical potential of assistive technology is the grand challenge for neuroengineering, raising social awareness is the corresponding societal challenge. It is worth stressing that these requirements should not be seen as values per se, but as conditions of possibility for the consideration and incorporation of values through UC and VSD (see Fig. 2). In fact, we hypothesize that UC and VSD approaches cannot be properly implemented if: (i) major power imbalances persist, (ii) biomedical ethics is ignored, (iii) prototypes are not adequately translated into viable products for users and (iv) there is a lack of social awareness about the clinical benefits. However, we recognize that this causal relationship can be bidirectional as: (i) the four normative requirements enable UC & VSD, but, in parallel, (ii) the adequate realization of UC & VSD guarantees the fulfillment of the four normative requirements.