Emotional Safety in the Context of Dementia: A Multiperspective Qualitative Study

2.1 Study design and setting

A descriptive qualitative study was conducted based on semi-structured guided interviews, focus groups, guided feedback, and data collected through participatory approaches. A predominately inductive content analysis design was applied [12]. The qualitative study was performed in North-Rhine-Westphalia, Germany and included participants from different communities. People in an early stage of dementia, relatives of PlwD, and public stakeholders were included to ensure a multiperspective and comprehensive view of emotional safety in the context of dementia.

2.2 Sampling design, recruitment, and retention

A qualitative research design with a criterion-based convenience and snowball sampling was chosen [14]. The inclusion criteria for all target groups were as follows: 18 years of age or older, living in North Rhine-Westphalia, Germany, and authority to provide informed consent for study participation. Further-more, people at an early stage of dementia (self-re-ported or confirmed diagnosis), relatives of PlwD, and public stakeholders who encountered dementia in their personal and/or professional backgrounds were included.

The participants were recruited between November 2017 and March 2019 predominately by the team of the non-statutory welfare association and partially by the research team. The number of interviews was calculated by considering a community approach, and data saturation was expected to be re-ached [15]. To obtain a more typical sample, we recruited PlwD from community services rather than dementia research centers, although advantages were expected based on Cridland et al. (2016) [16].

The participants were recruited through personal contact, flyers, newspaper articles, and participatory strategies (e.g., communication boards in market-places and thematic exchanges at breakfast meetings). The participants, especially the PlwD, were approached face-to-face or by telephone or mail to establish a relationship of trust between the parties participating in the project to ensure an understanding of the phenomenon of interest and knowledgeability and enhance participant retention [16, 17]. The team members of the non-statutory welfare association invited interested parties to engage in voluntary conversations regarding the project objectives and topics in person or over the telephone.

The following three different approaches were used to confirm the diagnosis of early dementia of PlwD: self-reported dementia, self-reported diagno-sis of dementia, and/or a transmitted document of diagnosis. A self-reported diagnosis was verified by the staff of the local non-statutory welfare associ-ation. Each PlwD was offered an voluntary Mini-Mental State Examination (MMSE). Performing an MMSE was not an inclusion criterion.

2.3 Participants

Our study is based on a balanced and comprehensive sample design involving a total of 32 interviewed participants (Table 1). Fourteen semi-structured individual interviews with PlwD (N = 5), relatives of PlwD (N = 7), and public stakeholders (N = 4), including two dyadic interviews (PlwD were attended by their relatives), were conducted. None of the participants of the individual interviews participated twice in interviews. All PlwD had the opportunity to participate in all types of interviews. One interviewed PlwD also participated in a focus group. Three focus groups involving 17 participants, including PlwD (N = 2), relatives (N = 4), and public stakeholders (N = 11), were conducted. One participant represented the perspective of a public stakeholder and the perspective of a person with a suspected diagnosis of dementia. Three participants participated as public stakeholders but were also relatives of people with neurodegenerative diseases (two with dementia and one with another disease). Data saturation was reached; thus, the researchers assumed that further data collection would lead to similar results [15].

The individual and dyadic interviews lasted from 18 to 46 minutes (mean 30 minutes), and the focus group interviews lasted from 64 to 75 minutes (mean 69 minutes). Women (n = 23) comprised the majority of the participants (N = 32). The overall mean was 66 years (N = 31) with a standard deviation of 12.2 years. The participants had a mean of 1.7 children (N = 32), while the PlwD (N = 6) had a mean of 1.5 children. Most participants were married or lived in civil partnerships, had an average school degree with a vocational degree, were workers in services, sales, or management, and received a pension or were retired. One person had a migration background. The educational degree of the PlwD (females and males, each n = 3) was a vocational (school or company) degree (n = 3) and a university or college degree (n = 3). Four PlwD provided a communicated verified diagnosis of early dementia, and two PlwD provided a documented diagnosis of dementia. Three of the six PlwD agreed to participate in a voluntary MMSE and had MMSE scores ranging from 10 to 30, with a mean score of 22.0 and a standard deviation of 10.6.

The feedback meeting (N = 15) lasted for 60 minutes. The participants (n = 12 females and n = 3 males) that were present participated in all types of interviews as follows: PlwD (N = 3), relatives (N = 6), and public stakeholders (N = 6).

Two researchers (SK and SB) conducted the individual interviews, dyadic interviews, and focus group interviews between August 2018 and April 2019. Due to the different data collection methods (triangulation), we aimed to obtain a reliable picture of emotional safety in the context of dementia. Semi-structured guided individual face-to-face interviews were conducted with each target group. If requested by the PlwD and their relatives, we conducted dyadic interviews, which have been shown to provide support and strengthen feelings of safety [18]. The semi-structured guided focus groups were based on a scoping focus group design with an ”individualistic social psychology perspective” [19] to consider the individual view of each participant. In the feedback group, communicative validation of the results was performed, enabling a meta-perspective of all results of all participant groups. Two moderators (SK and SB) presented the generated main and subdimensions to the participants for verification or adjustment as necessary.

All interviews and focus groups were audio-recorded and observed with field notes recorded by another research team member [20]. The interviews were transcribed by an external transcription office according to published guidelines [21, 22]. According to Pesonen and colleagues (2011), the presence of a third party (who was not personally interviewed) was allowed in selected interviews if the participant expressed a desire for support. To ensure ongoing consent, the participants were asked to provide their consent to participate before each interview. The interviews started with the researcher providing an introduction and explaining the rationale and aim of the interview. In general, only a few questions guided the interviews, and communication techniques, such as paraphrasing and requests to narrate, were applied to gain as much information as possible regarding the themes. After the interviews, the PlwD were asked to voluntarily complete the MMSE [23]. Structural data (e.g., concerning age and education) were collected separately by a structural questionnaire before the interviews by the team from the non-statutory welfare association. This team collected participatory data by recording field notes using a predefined form during their meetings.

The participants were free to choose the time and location of the data collection [17]. The interviews occurred during the day in a room familiar to the participants, the project buildings, other public meeting places, the participants’ home, or the participants’ workplace.

To support the group of PlwD during the focus group interviews and avoid opinion leadership, the introduction question was posed to all participants separately. Each participant had the opportunity to present his or her opinion. To support the participants, especially the people with cognitive impairments, during the interviews, cards were used as stimulus materials [19] and, in the case of dementia, as a re-minder. The cards contained the following topics related to emotional safety, which could be ide-ntified from previously collected data and a systematic review [1]: “everyday tasks”, “places”, “aids/ technology”, “personal contacts”, “leisure time (e.g., trips, holidays, and sports)”, “work”, “home”, “traffic”, and “dementia diagnosis”. Furthermore, a card with a free text field was included. The cards were multiplied by four and laid in a structure manner in front of the participants on each side of the table. The participants were free to choose a number of cards with their preferred topics. The selection could be changed during the course of the interview or reference could be made to new cards. The participants were informed that they could discuss with each other.

2.5 Interview guides

The interview guides (see Supplementary Tables 1 and 2) were developed following the “SPSS method”, which includes four steps, namely, “(S) collecting questions in relation to the phenomenon of interest”, “(P) checking against the state of research and openness”, “(S) sorting in relation to the sequence and/or content”, and “(S) subsuming topics in relation to prompts, probes and key-words” (translation SK) [24]; these steps were supervised by the research team and pretested. All guided interviews were subdivided into the following three parts: introduction, main part, and closing. The introduction (individual and dyadic interviews) contained an initial question regarding a usual day in the interviewees’ everyday lives and a following question regarding non-daily activities. Following the previously mentioned activities, the next question addressed the feeling of safety/lack of safety in the mentioned situations. The second question asked the participants what is needed to feel safe. The closing part included a question regarding whether the interviewees wanted to discuss anything else related to the feeling of safety.

The introduction to the focus groups started with a warm-up question regarding interest in the topic of feeling safe in the context of dementia. The main part allowed an open discussion regarding feeling safe and a second question regarding what would be recommended to enable a feeling of safety. At the end, each participant was asked whether he or she had anything else important related to the feeling of safety to share.

2.6 Data analysis and synthesis

The data analysis was based on a predominately inductive content analysis approach [25] using MAXQDA 11 (VERBI Software, 2014). Inductive analyses allow open coding and the development of categories on different data levels. Additionally, deductive subcategories were guided by the definition of emotional safety and the research question; these subcategories included “needs”, “conditions”, “strategies”, “influencing factors”, and “constructs related to emotional safety”.

First, all data of each transcript were examined. The analysis of the transcript sections was performed according to our research question. In each target group, we analyzed the participants’ explicit statements referring to emotional safety and implicit statements that were not directly related to emotional safety but described in the interview to avoid conformation bias and identify phenomena related to emotional safety in the context of dementia. Second, the categories were developed and concretized during repeated analyses of these sections separately. Third, the categories were analyzed according to the five deductive subcategories described above. In the final step, we compared and analyzed the content of the subcategories among the categories and among the perspectives. We analyzed the data from the interviews (SB, SK and CB) and participatory approach (FW and SB) separately. All interviews were initially coded by one coder (SB). Two senior researchers supervised the coding procedures (SK), checked all coding (SK and CB), and provided coding recommendations (SK and CB). The data synthesis (SK and SB) included an analysis of the categories while considering the different perspectives of the participants, an analysis of the common and varying dimensions and an analysis of the contradictory, complementary and interrelating findings. The analysis team (SK, SB, and CB) discussed the categories and data synthesis and made adaptations in case of need.

2.7 Ethical approval

The EMSIDE model project was approved by the ethics committee of the German Society of Nursing Science (project reference number: 17-013).

3.1 Dimensions related to emotional safety

In total, five clusters of dimensions, 16 main dimensions and 373 subdimensions were developed. We were able to show that the conditions, strategies, and influencing factors (related to the diagnosis of dementia and psychosocial, biographical, physical, and economical factors) were allocated to the following five clusters: “social togetherness”, “personal condition”, “health”, “physical environment”, and “society”. Emotional safety in the context of dementia was described by all participant groups in relation to at least one other construct (e.g., well-being). The clusters comprise the 16 main dimensions as follows:

“social togetherness” (“perceived communality”, “social interaction”, “support”, and “use of dementia services”);

The results related to the main dimensions were classified based on the following deductively defined subcategories: a) “conditions and strategies” (Table 2), b) “influencing factors” (Table 3), and c) “constructs related to emotional safety”.

The comparison of the explicit and implicit data that addressed the topic of emotional safety from the interviews and participative data showed no new main dimensions. Additionally, the data from the feedback group showed no new main dimensions. However, the feedback group substantially contributed by reflecting upon all perspectives of the participant groups on a meta-level and in the development of the overarching dimension “living and learning in relations”. However, it cannot be ruled out that further single subdimensions could be developed based on further interview data depending on the individual participants or other settings.

The overarching dimension resulting from all explicit data and feedback data, i.e., “living and learning in relations”, comprises the following two aspects: 1) the interaction of people in the context of dementia and people’s relation to the topic of dementia, and 2) the (common) learning of all participants. All developed clusters contained the dimensions of 1) interaction of people in daily life. Furthermore, all clusters contained the terms 2) “learning”, “education”, “training”, “competences”, “knowledge”, or “experiences”. These two core aspects in the context of dementia, i.e., “living and learning in relations”, were emphasized as the most important aspects in the feedback group. Additionally, the individual needs of each person and individuality of his or her situation were central in the feedback group. For example, the participants of the feedback group stressed that structure also needs a degree of flexibility and should be provided by considering and understanding individual needs.

To develop solutions and strengthen the feeling of safety, one participant suggested that individuals need to cooperate and develop a situational need- and resource-oriented “emotional safety thinking” (F2). Another participant suggested that “good togetherness” can provide emotional safety to all individuals (FB) as the feeling of safety of one person (e.g., formal caregiver) can affect the feeling of safety of another person (e.g., PlwD) (FB).

In total, 236 subdimensions of conditions and strategies were developed (Table 2). In all five clusters, we observed several common, complementary, contradictory, and varying conditions and strategies from a multiperspective point of view. Not all main dimensions were addressed by all participant groups. We also observed that the content was interrelated.

3.3 Common conditions and strategies of the participants and content interrelations

The cluster “social togetherness” contained the most frequently mentioned conditions and strategies (n = 101) of all clusters and was addressed by all participant groups. The main dimensions “perceived communality” (n = 28), “support” (n = 27), and “use of dementia services” (n = 11) were addressed by each group of participants. The main dimension “social interaction” (n = 35) was among the largest main dimensions containing most of the categories. Although there were no subcategories based on the (dyadic) interviews with PlwD in this dimension, both the conditions and strategies of the other dimensions contained aspects related to learning as a com-monality. The conditions were related to needed “knowledge”, e.g., regarding symptoms, course of the disease, dementia-related behavior, and “experience”. Other common conditions were related to personal traits (e.g., positive attitudes) and communicative, psychosocial and professional skills (e.g., empathy and consideration of appropriate communication and behavior). A dimension of “continuous learning” as a condition could be developed based on the focus group data. Related to such conditions, we identified strategies containing “education” and “training”, e.g., “addressing stressful situations, awareness, sympathy, and understanding challenging behavior” and/or “appropriate behavior addressing PlwD by participating in meetings” and/or the creation of a need-oriented environment for relationship building. Further strategies related to “social interaction” addressed appropriate communication and performance regarding “relationship building” and the creation of “(...) a feeling of being accepted” among PlwD. In this context, the relatives and public stakeholders assumed that the emotional safety of others, e.g., of the relative, is also a condition for one’s own emotional safety, e.g., of the PlwD.

Additionally, the other clusters contained aspects of learning as follows: the cluster “personal condition” contained “knowing and safeguarding basic human needs”; the cluster “health” contained “learning to cope with disease”; the cluster “physical en-vironment” contained “using learned safety behavior and physical safety aids”; and the cluster “society” contained “education (society, starting with children)”. Specifically, in the feedback group, “learning” was the core condition emphasized and was chosen as the most important by all participants.

In the cluster “social togetherness”, another large main dimension identified was “perceived communality”, which covered all perspectives of the target groups. The conditions were predominantly related to relationship and/or relations, e.g., “feeling of to-getherness”, “close relationships” and “living in relations”. Here, the strategies were also interrelated with the conditions as expressed by “PlwD contact”, “common activities”, and/or “taking care (of each other)”. Additionally, this aspect was discussed in the feedback group and appeared to be chosen as a core dimension of the results. It was stressed that “living in relations” extends beyond relationship. Importantly, the “feeling of togetherness” is a condition of emotional safety, and the perspective of society regarding the topic of dementia considering not only “knowledge” but also “acceptance” and “respect” was stressed in terms of “living in relations”. Although the cluster “society” contained fewer categories and not all individual perspectives were addressed, we identified “education” in the perspective of the public stakeholders and focus groups. The strategies included public relations, “professional presentation of the topic of dementia” and “education of society (starting with children)” with the aim to learn about the topic of dementia in a manner that shows the positive sides. The main dimension “fundamental awareness” (n = 2) included one condition related to public competences that can be improved through education. In this context, the main dimension “normality of the topic of dementia” (n = 8) was stressed by relatives in terms of “open communication” and the “presence of the topic of dementia in society”.

Furthermore, “support” was another dimension addressed in all perspectives. From the perspective of PlwD and public stakeholders, “support”, as a main dimension, includes the willingness of PlwD to obtain support and related factors (e.g., trust). The relatives and focus group participants mainly referred to conditions of care and daily routines (e.g., assurance to (quickly) obtain additional support and be well cared for and knowledge of where to receive help). The strategies ranged from need-oriented assistance in daily life (e.g., personal, technical and cognitive) to the development of a support network. The “use of dementia services” as a main dimension is need-oriented and contains strategies related to information, communication, interaction, activities, training, therapy, and support for PlwD and/or relatives.

Additionally, both the clusters “personal condition” and “health”, which also contained nearly half of the categories, were addressed by all participant groups. The following three main dimensions of the two clusters were addressed by all participant groups: “feeling of structure” (n = 20), “perceived familiarity” (n = 18), and perceived changes (n = 30). The main dimension “feeling of structure” comprised knowledge and awareness of past and future structures (e.g., routines).

The strategies were related to the identification and creation of supporting circumstances for existing structures and the implementation of new structures. Structures and knowledge allow safety. Additionally, the feedback group emphasized that a “feeling of structure” also should allow flexibility, e.g., rigid time structure, in terms of individual needs. Additionally, “learning [to address] new structures” was highly important. The conditions of the dimension “perceived familiarity” were related to the physical and psychosocial environment and structure in everyday life. All strategies concentrated on maintaining this familiarity (e.g., through regular contacts).

Many conditions and strategies mentioned in the different participant groups can be considered complementary; for example, in the main dimension “safety” (n = 5) (cluster “physical environment”), the “technical aids internal and external (monitoring, emergency)” and “adaption of physical environment (considering abilities, avoiding risks)” can be considered complementary. However, there were also intrapersonal complementary conditions. For example, in the cluster “health”, from the perspective of the PlwD, the dimension “resilience (coping)” (n = 5) included the complementary dimensions of “developing new self-esteem” and “using their diagnosis for own needs to obtain further resources”, which are both conditions described as complementing each other but also can build on one another.

3.4 Contradictory conditions and strategies of the participants in content interrelations

The main dimension “communication of diagnosis” (cluster “health”, n = 7) was not addressed by the public stakeholders; however, implicit statements were made regarding this topic. The dimension “communication of diagnosis” reveals surprisingly contradictory strategies between perspectives, i.e., within a participant group and/or within a person. Interestingly, “masking the disease” but also, in contrast, “openness about the disease” were mentioned as strategies used to increase the feeling of safety. One PlwD mentioned positive experiences with openness and the lack of knowledge of friends and public stakeholders (e.g., employees) (ID 04). The participants reported that they constantly experienced inner conflicts.

This person referred to the process of the “internalization” of addressing the diagnosis of dementia, which may be an indication of the person’s coping process; however, this process was not further defined (ID 04).

Another person with dementia stated that the knowledge of friends regarding the diagnosis and the course of the disease provides a feeling of safety and is highly relevant (ID 03). In the interview, it becomes clear that the person was more active in coping with the diagnosis with the help of a social support organization as follows:

In addition, it must be considered that this person is approximately 20 years younger than the person described above. This finding indicates that the characteristics and topics of several clusters can be interrelated. This finding also shows that the cluster “health” is related to the dimensions of the cluster “social togetherness” in terms of “social interaction”.

The relatives described that openness simplifies everyday situations (ID 24). However, the relatives also do not want to be confronted with stigmatization in society (ID 02). The relatives are aware of the potentially contradictory opinion of their relatives with dementia (ID 24).

Further contrary strategies can be identified in terms of staying at home and self-efficacy. A trade-off exists between the decision to leave PlwD in their own environment and the decision to change their housing situation. Such decisions include the awareness that PlwD do not feel safe when participating in public life (ID 03). However, they also need to know that retreating into domesticity is helpful in the short term (ID 03, 04) but leads to problems in the long term (ID 04).

In the context of self-efficacy, a trade-off exists among control by others (F2), the uncertainty PlwD feel in everyday life and the positive feeling PlwD experience when making achievements (ID 03, 02, 40). Strategies that show a compromise were also mentioned (ID 03, 21).

3.5 Conditions and strategies from the perspective of PlwD and content variety

We observed that not all main dimensions could be addressed by the subdimensions in all participant groups. Specifically, differences were observed between the PlwD and the other participants. No subdimensions in the main dimension “social interaction” could be specified for the PlwD. Additionally, the “process of diagnosis” (n = 6) was only addressed in the focus group, and the cluster “society” could not be coded among the PlwD. However, some dimensions were related specifically to the PlwD. In the cluster “personal condition”, the main dimension “home as a space for personal preferences” (n = 5) was only addressed by the PlwD and in the focus group. The PlwD stressed the importance of the “opportunity to live according to own preferences” and “time flexibility to suit own preferences”. Considering the other participant groups, “home as a space for safety” (n = 18) was not addressed by the public stakeholders, and “home as a space for personal preferences” was not addressed by the relatives.

A prominent main category in the cluster “health” was the main dimension “perceived changes” (n = 30). Here, more strategies than conditions could be described, and most strategies were provided by the PlwD. Interestingly, the conditions showed a poor variety, exclusively emphasized the inner perspective of PlwD in terms of psychological, philosophical and theological reflections and concentrated on “personal attitudes”, “increased trust in God”, and “perception of fully functional human sense”.

Although no outer conditions were mentioned, a large variety of strategies were used to address perceived chances in the context of dementia from several perspectives. Some strategies could be labeled proactive in terms of “planning situations”, “adapting structures by considering the needs of PlwD” and/ or “hiding impairments”, and some strategies could be labeled passive in terms of “avoidance” and/ or “staying at home”. The PlwD mentioned both proactive and passive strategies.

The main dimension “process of diagnosis” (cluster “health”) showed less variety in the content and was almost exclusively addressed by the participants in all focus groups but surprisingly not by the PlwD, relatives, and public stakeholders in the individual (dyadic) interviews; this dimension contains the condition “contact with competent professionals” and an interrelated strategy, e.g., “dementia-specific education for professionals”. Here, the aspect of learning was also observed. More variety was observed at the micro, meso and macro levels as follows: one strategy addressed “self-diagnosis”, one strategy mentioned professional contact and one strategy referred to a higher level of “politics” in general.

3.6 Factors influencing emotional safety in the context of dementia

Factors (n = 137 subdimensions) were identified as possible conditions influencing emotional safety in a complex way. All groups of participants described several factors, such as dementia-related, psychosocial, biographical, physical, and economic factors. In general, the influencing factors related to dementia included “stage/symptoms of dementia” (ID 01, 03, 24, 26, F1, F2, F3, FB), “symptom recognition” (ID 02), “the time since diagnosis” (F1), “confrontation with impairments/symptoms” (ID 01, 03), “knowledge (regarding dementia symptoms/diagnosis)” (ID 04, 24, F2), and “challenging behavior of PlwD” (F2).

The participants mentioned psychosocial factors, e.g., “personal attitudes” (ID 02, 04, 21, 22, 25, 40, F1, F2, FB), “fear of negative reactions from others” (ID 03, 04, 40), “negative picture of dementia” (ID 02), “stigmatization” (ID F2), “perceived anonymity” (ID F2), and “inter-role conflicts” (ID 24, 26). In the context of the dimension “perceived changes”, the “interplay of psychosocial and dementia-related changes” could be identified. Aspects of support, e.g., “missing support” (ID 24, F2) or “rejection of support” (ID 21), were also important. Regarding biological factors, “characteristics of PlwD and others” (ID 02, 21, 22, 25, 26, 40, F1, F2, F3, FB), “common history” (ID 02), “parental upbringing” (ID 40), and “loss of former life” (ID 21) were influencing factors. The physical factors “environmental changes” (ID F2, F3), “physical distance to relatives” (ID 26, 39), and “missing technical solutions for PlwD” (ID F2) were highly important. Economic aspects were mentioned in general but not in a dementia-specific context as follows:

3.7 Constructs related to emotional safety

The following common constructs related to emotional safety at the psychological, psychosocial, and physical levels were reported by all or most participant groups: participation (ID 01, 02, 03, 21, 22, 24, 26, 37, 38, 39, 40, F2, F3), fear (ID 02, 03, 04, 21, 22, 24, 26, 37, 38, 39, F2, F3, FB), feeling of loneliness/togetherness (ID 21, 22, 24, 26, 37, 38, 39, 40, F2, FB), well-being (ID 03, 24, 25, 40, 41, 42, F1, F2, F3), stress (ID 21, 25, 26), and physical safety (ID 02, 21, 26, 37, 38, F2). In addition, the PlwD addressed the following constructs: harmony (ID 01), being calmer (ID 03), happiness (ID 03), self-efficacy and self-control (ID 01, 03, 40), feeling threatened (ID 37, 38), mobility (ID 37, 38), boredom (ID 37, 38), nervousness (ID 03), and being able to learn and give up tasks (ID 01, 03). The relatives reported the following additional constructs: feeling of being in good hands (ID 26), openness (ID 02, 24), being accepted (ID 02, 24, 25), being reserved (ID 02), being avoidant (ID 02), sorrow (ID 02, 24, 26), depression (ID 25), aggression (ID 25), panic (ID 26), desperation (ID 26), comfort (ID 02), and helplessness (ID 26). The public stakeholders reported the following additional constructs: feeling complete (ID 22), sleeping well (ID 21), experiencing a loss of control (ID 21), overload (ID 21), being proud (ID 22), and living in conflict (ID 21). Based on the data of the focus groups, the following additional constructs were identified: underload and feeling ashamed (ID F1).

4.1 Constructs related to emotional safety

Finally, we obtained complex insight into common constructs related to emotional safety at the psycho-logical, psychosocial, and physical levels, e.g., participation, fear, feeling of loneliness/togetherness, well-being, stress, and physical safety. In a concep-tual analysis of the feeling of safety among pat-ients during non-dementia-specific inpatient hospitalization, Mollon (2014) described that the main consequences resulting from being in a state of emo-tional safety are “control”, “hope”, and feeling “re-laxed or calm” [35]. In our results, we also observe the experience of loss of control, self-control, and a feeling of being calmer.

4.2 Strengths and limitations

Although we were able to obtain a multiperspective and comprehensive picture of emotional safety in the context of dementia, there were some limitations regarding our sample design, data collection, and coding procedures. The PlwD and other participants were recruited from community services to obtain a more representative sample and consider different experiences with relevant actors in the local context. The sample of the individual interviews was balanced; however, a lower number of PlwD participated in the focus groups and feedback group due concerns related to having dementia, although beforehand, several efforts were exerted to support the PlwD. However, one PlwD exclusively participated in the focus group accompanied by his/or her relative. The recruitment was based on a criterion-based convenience and snowball-based sample. Additionally, in some cases, it was difficult to differentiate between the perspectives of the participants because the public stakeholders could also be relatives.

The aim of the study, i.e., to explore emotional safety, was explained to the interviewees several times, and we addressed the topic explicitly, which could have influenced the interviewees’ answers in this direction. However, a strength of our study is that we analyzed the differentiated data regarding what was mentioned explicitly in relation to emotional safety and what was mentioned implicitly in the interviews. We observed that the findings agreed in principle, and the findings were verified in a feedback group. In addition, we compared our results using participatory data and found that the results were similar but on different levels of abstraction.

The interviews and feedback group were performed by two interviewers. To ensure the interview quality, all interviews were performed based on a standardized procedure, and the first interviews were pretested, supervised, and discussed.

The initial coding was performed by one coder. Therefore, the reliability of the results could be res-tricted. However, the coding procedure was supervised, and all codes were repeatedly checked against the transcription sections by a second coder, dis-cussed and adapted. The data synthesis was performed by two researchers. Finally, a third researcher checked the coding trees and provided recommendations for adaption. We observed that not all main dimensions contained subdimensions in all participant groups. We discussed this fact and observed that similar subdimensions were coded in other dimensions; thus, the subdimensions that best fit the meaning in the context of the interview section and interview overall were considered.

Although generalizability of the findings is limited due to the qualitative design, we were able to reach data saturation, based on the results from different data sources including data from the participatory approach.

4.3 Conclusion

Our study highlights the relevance of the topic of emotional safety in the context of dementia considering care provision, services, and research from a multiperspective point of view. We found that emotional safety is a crucial need of PlwD, relatives of PlwD, and public stakeholders. The results allow more differentiated empirical picture of people’s needs for emotional safety in the context of dementia. We used a multiperspective point of view to show that there were several common, complementary, contradictory, and varying conditions and strategies that were related to individual situations, needs, and influencing factors. Considering “social togetherness”, “health”, “personal condition”, “society”, and the “physical environment” in light of “living and learning in relations” is required for care provision and services that respond to the need for emotional safety in the context of dementia and the development of interventions. In developing and choosing strategies to enhance emotional safety, complex and individual situations at the micro, meso, and macro levels in each target group should be considered in relation to each other. In consequence, care provision and service should be more integrative and be based on a consideration of emotional safety in the context of dementia. PlwD feel safer when they are seen as meaningful and esteemed persons who are learning and living in relations of their living environments, including the care processes, services, and society. The developed dimensions allow a new, structured way to find solutions for achieving emotional safety in defined settings, as well as a consideration of steadily changing social and physical settings. These solutions should include (collective) learning on all levels without the loss of individuality.

Further research should focus on the investigation of the relevance of the developed dimensions in other health care settings, e.g., stationary settings or digital settings (e.g., telehealth physician – PlwD communication) to determine how the needs of PlwD, relatives of PlwD, and public stakeholders are considered and how living and learning in relations has been or can be realized to improve emotional safety. It should be investigated how these dimensions and the way of thinking about care provision and services can best be implemented in or transferred to different settings also considering aspects of patient safety.

The proposed dimensions could be used to develop a framework for measuring emotional safety in the context of dementia. A dimension-based modular instrument could be developed to capture multiple perspectives in a structured way to determine needs, conditions, and targeted strategies.