Method Study 2

35 Figure 6 Illustration of the four main categories of key concepts of carers’ support needs and the interaction between them (Paper I).

These four categories were consistent across carer types and care settings, but not associated with the degree of dependency of the person cared for nor the cultural and socioeconomic context. For further elaboration of results, see Paper I.

36

professionals to support each other in the data collection process. First, focus groups with professionals were conducted, and second, focus groups with carers were conducted, allowing carers to reflect upon the perspectives articulated by the professionals and to authenticate their relevance. Last, individual interviews were conducted to enable investigation of the more personal and sensitive experiences relevant to support needs.

A rigorous process of collecting and analysing data to answer the research question is described in the following.

15.1.2 Participants

To ensure identification and selection of participants with knowledge and experience of the phenomenon, purposeful sampling was used (97). To obtain an information-rich range in variation of participants, predefined inclusion criteria had to be decided from the beginning. Knowledge gained from mapping and synthesising literature of carers’ support needs in Study 1 was used to decide the inclusion criteria for the current study (see Table 3). Hence, recruitment was conducted in one rural and one urban municipality in Denmark. Including participants from both rural and urban settings was important because carers’ support needs may be influenced by accessibility to services and availability of specialist providers of health and social care (98). Alongside recruiting participants based on experience of the phenomenon, purposeful sampling constitutes a pragmatic recruitment strategy that takes into consideration availability and willingness of participants who are able to reflect and express themselves (97).

Table 3 Inclusion criteria for participants in Study 2

Interview type Inclusion criteria

Focus groups, professionals Working in dementia social care or health care

Focus groups, carers ≥ 18 years, provides help on a regular basis because of a personal

relationship rather than financial compensation to a closely related person who has received a dementia diagnosis, able to communicate in Danish Individual interviews, carers

15.1.2.1 Criteria for carers

Variation was sought among carers with respect to the following criteria: sex, co-habitation, progression of dementia and relationship to the person with dementia. These criteria were chosen based on previous studies indicate their importance (53,99,100). In particular, carers feeling isolated due to high dependency or supervision of the person with dementia has shown to be important of carers’ experience of support needs (101). Also, relationship quality may be important for carer motivation, and spouses have been seen to experience higher relationship quality and find more meaning in the caring role than other carers (44).

Carers were recruited by key professionals in each municipality. This strategy was chosen to promote a

37

heterogeneous composition of participants, and because these professionals were in a unique position to engage carers who may not have volunteered on their own. An example of this is carers of a person with dementia in the later stages of the disease who could be assumed to have less energy to participate in research. Therefore, key professionals were asked to prioritise recruitment of carers who may not volunteer by themselves.

15.1.2.2 Criteria for professionals

Diversity was also sought regarding experience and educational background among professionals. Various professions are represented in dementia care with educational backgrounds ranging from upper secondary school to a Master’s degree (15,60). The level of education and specialisation within dementia care may affect the quality of care and support provided by professionals (60). The different educational experience and clinical expertise and knowledge may impact professionals’ views on carers’ support needs.

15.1.3 Data collection 15.1.3.1 Settings

Several factors were considered to ensure the best set-up for conducting the interviews, as carers may find themselves in a vulnerable position, and professionals may be hesitant to compromise themselves in front of unknown colleagues or a superior.

Focus groups were held in local meeting facilities in each municipality. Accessibility for participants was important, so carers were offered financial compensation for transportation to guarantee their opportunity to participate. For professionals, reducing time spent on transportation was important given their busy schedules. The focus groups lasted a maximum of 2 hours, which allowed enough time to welcome people and for a discussion to evolve naturally without a feeling of time pressure. Coffee, tea and cake were offered to create a welcoming and relaxed environment. An experienced moderator (THC) and co-moderator (HKK) conducted the focus groups. The co-moderator’s role was to facilitate the discussion and make sure that all participants’ voices were heard (102). The role of the co-moderator was to observe the discussion, take notes and follow up on issues that were passed over too lightly.

In the individual interviews, participants had the choice of the interview being conducted in their own home or at a local meeting facility. They were given this choice to ensure a feeling of safety in the interview setting, because taking part in an individual interview may feel intimidating for some. Interviews were conducted by the PhD student (THC). All interviews were recorded and transcribed verbatim afterwards.

15.1.3.2 Focus groups

The composition of groups was important. To enable a good group dynamic, eight participants were preferred (102,103). However, as few as five participants were deemed enough to constitute a focus group,

38

as a smaller group rather than a larger group can sometimes lead to intense involvement and contribute to unique knowledge (103). At least two focus groups of professionals and carers were required to adequately sample the complexity of carers’ support needs, respectively. To ensure healthy functioning of the groups and a reasonable comparison of issues raised by participants, some segmenting of the group of carers was necessary to achieve homogeneity (103). Carers’ ability to see resemblance to their daily life when caring for a person with dementia was chosen as an important criterion, and carers were categorised based on whether or not they co-habitated with the person with dementia. The daily life of a carer living with a person with dementia has obvious restrictions, and carers may more easily recognise themselves in a person living a life resembling theirs.

The strategy of purposive sampling was used for recruitment of carers and professionals, who were unfamiliar with one another. This was desired as participants may feel anonymous and speak freely when thinking that the information shared is kept confidential from colleagues, family and friends (103).

15.1.3.3 Individual interviews

Carers in the individual interviews were identified to complement carer participants in the focus groups.

Therefore, interview participants were recruited after completion of the focus groups. Also, carers

reluctant to participate in focus groups were given the opportunity to give expression to their experience of carers’ support needs in interviews which ensured more intimate surroundings (102). The adequate

number of participants was determined by a repeated process of reflections: first, does preliminary analysis of focus groups provide sufficient variation of carers’ experience of their support needs to answer the research question? Second, do individual interviews contribute as much variation as expected after conducting the first three interviews? Due to the specific sampling strategy and that the interviewer was experienced in achieving a well-focused interview with a good dialogue, a provisional number of five to ten participants was assumed to be enough. However, before finalising the data collection, a preliminary analysis was undertaken to consider if the data from the interviews showed adequate relevance and depth for the research question (104).

15.1.3.4 Interview guide

An interview guide was produced to ensure richness and consistency in the data collection across all interviews. The guide functioned as a memory aid that was organised around key topics assisting the moderator to facilitate a discussion (103). Through brain-storming (102), the PhD student along with one co-supervisor (HKK) formed ideas for how to organise the interviews, so that the guide could be used in the context of focus groups and individual interviews (see interview guide in Appendix 1 and Appendix 2). The guide was organised into three sections. Firstly, an introductory question where participants were asked to present themselves. Secondly, to enrich the discussion in the interviews, an activity of sorting and choosing

39

text cards most important to each participant was chosen (103,105) and thirdly, a final question to close the discussion and to give the moderator a sense of what was most important in the discussion.

The text cards used for discussion of key topics were developed based on the results of the scoping review in Study 1 (Paper I). This review of existing knowledge of carers’ support needs provided an overview of important topics that should be explored in the interviews. Sixteen cards were produced in accordance with the generic categories synthesised in Study 1 (see Table 3 Paper I) including four blank cards to

encourage participants to generate new topics for discussion (see Appendix 3 for examples of text cards). In the activity part of the interview, participants were given 8 minutes to go through the topics described on the text cards, remove the cards not important to them and to choose the cards representing the topics most important to them. Afterwards, participants were given the chance to elaborate on their choices. In the focus groups, the moderator invited participants to engage in the discussion and speak their mind or comment on the other participants’ choices. In the individual interviews, the interviewer facilitated discussion by asking carers to reflect on multiple perspectives of their choices.

15.1.4 Data analysis

Inductive content analysis (91) was used to clarify the specific support needs of carers. This method made it possible to condense the various statements of support needs into content-related categories. Choosing this approach allows a broad description of the phenomenon of carers’ support needs. Also, no specific theory describes carers’ support needs, and an inductive approach allows for a combination of specific data into a general statement. Due to the overall aim of developing a questionnaire to assess carers’ support needs, manifest interpretation of the content was chosen (90). To generate items in a questionnaire, a clear description of the construct to be measured is necessary (2) and preferably, items should be phrased using the target population’s own words. By using manifest interpretation, it was possible to stay very close to the transcribed text and to describe what carers and professionals actually said about support needs (106).

When performing inductive content analysis, several phases are undertaken, which are described below.

15.1.4.1 Preparation

The transcribed material of all focus groups and individual interviews was chosen as the unit of analysis. A thorough reading of all the material to become familiar with the content initiated the groundwork for conducting the subsequent phases in the analysis process (91,106).

15.1.4.2 Organising

In this phase, open coding, creating categories and abstraction were conducted (91) using NVIVO 11 (http://www.qsrinternational.com/nvivo/nvivo-products/nvivo-11-for-windows) to help organise the data.

Firstly, open coding was done by identifying units of meaning in the text and providing them with a heading

40

that intuitively described the unit (106). This process was undertaken repeatedly, because as coding progressed, new details in the text were recognised, thus calling for an examination of the text more than once. Secondly, categories were created by grouping higher orders of headings. When grouping content, patterns across all of the material were sought (91), and all data had to fit into at least one group and no data were allowed to fit under more than one heading (106). Thirdly, abstraction was conducted by generating sub, generic and main categories describing carers’ support needs (for an example of the abstraction process, see Table 2, Paper II). Sub-categories were created based on the grouped content and further grouped into generic, and finally main, categories. Abstraction continued until additional grouping of content no longer made sense and would have detracted from the meaning of the specific category (107). The organising phase was conducted in discussion with the main supervisor (HHL) and co-supervisor (HKK), although only the PhD student took part in the open coding.

15.1.4.3 Reporting

Presenting the results of the content analysis depends on the depth of interpretation of the data (106).

Summarising the results of the manifest content analysis should explain the categories of support needs at a descriptive level, using quotations from the data material (91). Also, an easy overview in tabular form of the sub, generic and main categories should be presented (Figure 7) (106).

15.1.5 Ethical considerations

The study complied with the General Data Protection Regulations and was registered with the Danish Data Protection Agency (2015-57-0016-020a). The study followed the principles of the Declaration of Helsinki (108), however in agreement with Danish law, no ethics approval was required (109). All participants gave their informed written consent and all information was stored in a secure locker or server, and will be destroyed when the study is finished.

When interviewing carers and professionals, special consideration was given to ensuring the anonymity of the people spoken about by the participants. Further, in the focus group and interview situations, it was likely that sensitive information could be shared about a person not present, who would not be in a position to speak for themselves. Therefore, a respectful and well-intentioned tone was asked of all participants at the beginning of the interviews and especially in the focus groups, with participants being reminded of this if the tone changed.

An important consideration before conducting the interviews was that topics might appear about carers’

support that participants had not thought of before or been able to express. The process of reflection might also cause carers to subsequently request additional care and support from professionals that service providers could not fulfil. Therefore, participants were informed when consenting to participate that the

41

aim of this study was only to gather information on carers’ support needs to help improve services for carers in the future (see information letter to participants Appendix 4).