A robotic shower system

Ethical questions play a crucial role in technological and also in robotic development and is even more important in the healthcare sector dealing with vulnerable groups. In order to identify the most important ethical issues the MEESTAR was utilized as an analytical instrument for discourse.

Fig. 3 shows the MEESTAR with its different dimensions.

The MEESTAR is based on three axes. Manzeschke et al. suggested seven ethical dimensions which are presented on the X‑axis (care, autonomy, safety, justice, privacy, participation and self-conception; [15]). The Y‑axis presents the four assessment-stages how the ethical dimensions can be perceived (harmless until the use should be opposed from an ethical viewpoint). The Z‑axis represents three different perspectives (individual, organizational, social) to discuss the ethical dimensions. For each ethical dimension two up to four key questions are provided [15]. For an ethical evaluation MEESTAR supports discussions with content and topics. Although MEESTAR was developed for age-appropriate technologies (i. e. mainly sensor-based monitoring systems) it was assumed that these and other ethical dimensions could play a crucial role for the robotic shower system; however, an assumption was that the ethical dimensions of MEESTAR were not equally important to the system and that there might be other ethical issues crucial to the system. The approach was adapted in order to identify further essential ethical issues.

Different methods are applied throughout the I‑SUPPORT project. In the beginning of the project interviews with 30 PU and 15 SU took place analyzing the requirements for the shower system and discussing renderings to gain first impressions on ethical questions. Based on these results, four focus groups were organized to discuss ethical issues with respect to the I‑SUPPORT system utilizing MEESTAR. The purpose of the first three focus groups was to assess important crucial ethical dimensions with respect to a robotic shower system as well as to gain feedback on the advanced rendering and mock-ups for a more detailed feedback to technical partners. Of the focus groups two were active in the Independent Living Centre of Frankfurt UAS¹: one with PU living independently in their private home and one with SU. The work of the third focus group with PU and SU took place in a nursing care home where the PU were residents and SU were nursing care staff of the care home. The reasons for this location were that the robotic system is also meant to be for institutional settings, such as nursing care homes; PU in care homes are usually more mobility restricted and have experienced more functional loss compared to PU able to come to the Independent Living Centre. The assumption was to obtain a deeper insight into the requirements of this environment and the impact on the I‑SUPPORT system. This was also a reason for the composition of this focus group with PU and SU as they could contribute from their different perspectives of daily practice. After the experiences in the three focus groups the research team decided to have a workshop on the MEESTAR model carried out by one of the developers. As a consequence, the team decided to undertake an additional fourth focus group with PU in a nursing care home. The aim was to utilize an adapted MEESTAR focussing particularly on the individual perspective of PU as this seemed to be most relevant at that stage of the technological development.

All participants in the focus groups gave informed consent. The focus groups were voice recorded. In order to analyze the collected data, audio recordings were transcribed. The qualitative analysis followed Mayring’s process of summarizing, explication and structuring [17]. Data derived from the transcript of each focus group was coded based on thematic categories (e.g. ethical dimensions) utilizing the MAXQDA software (VERBI Software, Consult – Sozialforschung GmbH, Berlin, Germany) [14].

Table 1 gives an overview of the composition of the 23 participants in the focus groups.

The schedule of the focus groups started with an introductory round, a presentation on the project and an introduction of the mock-up and the rendering (Figs. 1 and 2, [5, 12]). In the first three focus groups this was followed by a discussion on benefits, desirable functions, operability and usability requirements. The importance of the different ethical dimensions of the MEESTAR system was then discussed, additional concerns and values recorded and participants assessed from their perspective the three most important dimensions of MEESTAR with adhesive dots, which were then added together to give a total assessment. The goal of the fourth focus group with PU was to discuss the meaning of the ethical dimensions for the robotic shower system and assess the potential stage of danger/harmlessness from their point of view.

One of the surprising results was that PU, i.e. people with functional loss, have a positive attitude towards a robotic shower system if it enables them to take care of their personal hygiene. It was considered to be better to use technology instead of a human caregiver because of the preferred privacy of this intimate process. Autonomy was equated with a maximum of decision-making and freedom of action. The robotic shower system should provide a variety of choices, different settings and individualized programs according to user preferences, e. g. both user groups emphasized the possibility of taking a shower when and for how long they wished.

OK, I say people who are still very independent, they can of course also choose the time when to shower. So there isn’t a fixed time set: now I have to take a shower. I can also have a shower in 2 hours time (Transcription mixed group PU: 223–223).

These arguments have an underlying additional ethical theme: choice. The possibility to choose is strongly related to autonomy [8]. Often choosing out of a range of options is one of the last parts of autonomy when experiencing functional loss. This also includes the choice to decide whether you want to use a robotic shower system as well as design and functional aspects e. g. the choice of colors or different auxilliaries, such as sponges and washcloths.

Although professional caregivers rated autonomy high, they saw themselves in a multifold dilemma: taking care of the personal hygiene process does not only mean that caregivers bring a person to the bathroom, undress, shower, dry, and dress the care recipient. This process has veiled aspects which seem to be essential in professional care. These aspects are A) communication, B) promotion of still existing resources of the care recipient and C) monitoring health changes.

The ethical dimension safety mainly referred to the safe use of the robotic shower system. For SU it was just as equally important as autonomy. The discussion revealed a dichotomy: on the one hand the role as autonomy enhancer for the PU, on the other hand as autonomy restrictor because of the feeling and being responsible for the safety of the PU. The SU also emphasized that to feel safe is highly important and features for the system could be the possibility to trigger alarms or integrating sensors that recognize falls. At no time should the robotic system jeopardize PU; therefore, an emergency stop is highly important. Although, PU did not prioritize safety with their allocation of sticky dots, the discussion in all focus groups revealed safety as being extremely important for them and a sine qua non which concerns not only the system but also the shower environment. The SU rated the ethical dimension justice high and relating mainly to distributive justice. They pointed out that these technologies should be provided to anybody who needs it, and not only for those who can afford it. Other concerns were that the robotic shower system might contribute to savings in staff.

The dimension self-conception (highly rated by PU) was discussed with respect to usability requirements (easy to operate, a feeling of happiness when you are able to control the robotic shower system) and non-stigmatizing features of the robotic system (e. g. friendly colors, appealing background noise, and ambience).

Participants in all focus groups also discussed human-robot and human-human relationships. The opinion of PU and SU differed. Some of the PU would prefer to use the robotic shower system (in order to be independent and have privacy). Other PU prefer to be supported by a human being. A personal relationship and talking during the showering process can play an important role, which was also confirmed by SU who stressed the importance of that aspect especially in home care. Worries with respect to the human-robot relationship were fears of losing the job and concerns on the nature of future work if robots do the work. There seemed to be a unanimous view that relationship work should not be omitted and time saved by technology should be used for essential care tasks which are currently lacking because of scarce staff resources. These results are comparable to the interviews with PU and SU [5].

So I am missing the interpersonal relationship. From my point of view, this is what constitutes care. I completely miss this. As assistive device, sure positive. It is a good idea, but I’m missing the communication or […] on the other hand I know also from my friends, who work in ambulatory/domiciliary services that clients prefer to talk for 5 min instead of getting the back washed. So that is a conflict for me. (Transcription mixed group SU: 406–406).

Another issue of human-robot interaction is that the user is the master of the robotic shower system, i.e. the user can control the robotic system at any time and not vice versa.

Prerequisite will be, of course that I control this thing and not the thing me. Sort of I take the lead and not the machine. This is indeed a very important factor. (Transcription SU: 276–276).

The discourse-oriented approach provided valuable information on core functions of the shower process. The discussions during the focus groups provided not only multiple ideas for the functionalities and desirable features of the system, but also aspects to be considered by the socio-technical arrangements if the robotic shower system is commercially available.

The project iToilet [16, 19] develops an ICT-enhanced toilet system supporting older adults and their care giver in this basic ADL. Comparable to I‑SUPPORT a user-centered approach is applied and PU, SU and other stakeholders are involved in the technological developmental process. Recent publications [18, 19] focussed on user requirements. Ethical issues are considered in the study design of the pilots and as a rationale for the project [18]. Here the ethical dimensions autonomy/independence, empowerment and dignity are mentioned. Similar to I‑SUPPORT, PU and SU are not hesitant to discuss requirements from their perspectives despite the subject being surrounded by even more taboos [18, 19].

In I‑SUPPORT the ethical dimensions were explained in simple language to participants. Facilitators got the impression that these definitions blurred during the course of the focus groups as this is not everyday language. For future work, recourse of everyday language seems to be appropriate, especially if user groups have vision or hearing impairments, are non-academics or not native speakers. The 3‑h time frame was too tiring for some of the participants, even more so for those who lived in a nursing care home. Thus, results reflect the subjective normative assessment of the participants. For the technological development, the insights into the causes and the new ideas created have been valuable.

Future focus groups should take these into account and develop approaches which consider easy to understand language and time constraints as well as often limited (research) resources. For I‑SUPPORT a structured assessment instrument will be developed and piloted in the formative and summative evaluations of the different prototypes by PU, SU and other stakeholders.

A systematic review of argument-based ethics literature on care robots in care of the aged revealed four main ethical approaches [22]: the deontological approach focussing on human reasoning on what is good or should be done. Values described are autonomy and dignity, deception and truth, social isolation and connectedness. The principlist approach as the practical translation of the deontological approach discusses four principles: respect for autonomy, beneficience, non-maleficience and justice. The objective list approach puts forward several “capabilities” or “goods” that can be reached or supported by care practices. The care ethical approaches look at the special care relationship between caregivers and care receivers and widen it to a contextual and political level. In this context MEESTAR could be subsumed to the deontological and principlist approaches, which are viewed as a too restricted approach as “The ethical landscape pertains to human moral agents seeing robot technology merely as a collection of neutral instruments that can or cannot be used to promote the well-being of older adults. Consequently, these approaches seem to lead to an ethical assessment of care robots instead of an open ethical reflection about their use” [22]. Weber defies to assign MEESTAR to a theoretical concept and suggests its use as a discoursive ethical approach [23]. The questions arises whether the way ethical assessment instruments are utilized might contribute to more openness and to possibilities to develop answers on issues, such as what is good care and what impact do “care-robots” have on the concept of care and on society. This question cannot be conclusively answered and here more research is needed also taking into account the different types of robot technology. The approach to include PU and SU from scratch of technological development at least gives the opportunity to get insights into their perceptions and the reasons for that.