Sampling and recruitment

Eight case study sites demonstrating exemplars of one of four models of provision were approached and invited to take part in the study (Table 5 details the case studies, visits, interviews and timescales for each site). WP2 identified two core groups of older PWLD, those who have greater physical and mental health needs than their peers and those who have recently been, or are about to be, discharged from long-term institutional care – our sample did not include anyone from the latter group. We did, however, include participants who would be characterised as having severe learning disabilities. We included an additional case study site (NCH_2) identified during proposal development as an exemplar in the development of a new provision for older PWLD and complex needs in an inner-city area. The delay in the development of the site led us to treat it as distinct from the other case studies.

TABLE 5

Case studies for each model of provision

Access to each case study was negotiated at provider level and then with local managers and support workers. Relevant introductions from contacts developed during WP2 were made and meetings arranged with the provider chief executive offices (CEOs) and, subsequently, other staff members to build relationships and identify potential participants. It was important that staff fully understood the point of the research and what participation would involve. It also allowed the team to find out more about the support offered. Discussions included thinking about the best ways to recruit and engage with participants and to include people who lack capacity via gaining consent by proxy from family members.

Provider DOP_1 asked to delay participation due to staff changeovers and joined the fieldwork 3 months later. Provider ISL_3 was unable to continue to commit the resources to participate because of staff sickness. It withdrew from the study, rejoining 7 months later. This resulted in fewer PWLD participants from this provider (n = 3). One provider declined to take part as the CEO felt that the organisation was not able, in pandemic conditions, to provide excellent support. A replacement provider was selected from the shortlist.

For formal interviews, a framework was produced to sample diversity including age, gender, ethnicity, family context and length of time living or working in the provision. In practice, we were limited to a sample defined by the people living and working at each of the sites, which were a mix of rural, small town, and urban, and our sample was White British. Ages ranged between 40 and 70 years.

Developing co-research training materials and delivering co-researcher training

It was important to involve PWLD and family carers in a form of ‘co-operative experiential inquiry’,¹¹² conducting the research with rather than on PWLD.¹¹³ We co-designed training resources to facilitate the involvement of PWLD and family carers as ethnographers and to support academic researchers to facilitate co-research. We were mindful that people had key skills of observation and listening and wanted to draw these out without imposing constraints on conventional means of research.¹¹⁴

This work was supported by engagement funding from the NIHR School for Primary Care Research and is reported in Mikulak et al.¹¹⁵ We recruited a small team comprising academics (n = 3), a Public Involvement co-ordinator and family carer, PWLD (n = 2), and two further family carers. The resources, developed through online workshops and an ethnographic exercise, are publicly available at www.mmu.ac.uk/research/research-centres/hpac/projects/growing-older-planning-ahead.

Ethics

Informed consent was sought from all participants who were deemed to have capacity to consent. Participants with learning disabilities were given Easy Read information sheets about the study and Easy Read consent forms. They were encouraged to complete these with the researcher and a trusted person such as a family member or support team member. The researchers went through the consent form with the participants (and, where appropriate, with their trusted person) and made sure participants had a chance to ask questions and understood their right to withdraw consent at any point, before the consent form was signed on two copies by the participants and the researcher. Participants retained one copy of the consent form. Consent for participants who were deemed to not have capacity was sought through a consultee. In all cases when this process was used, the consultee was a parent of the participant, had an intimate knowledge of and regular contact with the participant and deemed the person’s participation to be in their best interest.

Data generation

The move of co-principal investigator Ryan to Manchester Metropolitan University in May 2021 necessitated an ethics amendment and subsequent delay in starting the fieldwork. In the interim, the time was used to engage with the sites online. We learnt that considerable input and encouragement of support staff, which included making clear the point and importance of the research, was necessary to ensure engagement, particularly given staff shortages and other constraints.

Longitudinal observations, interviews and documentary analysis were used to understand how people were supported in their everyday lives, how they negotiated ageing, and how ‘BTCO’ were understood and ameliorated. In total, researchers spent between 12 and 25 days, including some weekends and evenings, at sites within the four models of care. Co-researchers were present at around 30% of the visits. While originally structured to take place an average of 2 days a week, the geographical location of the sites, and COVID-19-related disruption, meant that visits were organised flexibly, working around last-minute changes to plans caused by illness or staff absences. The number of visits was balanced against their duration (fewer visits meant the visits were longer).

Detailed field notes were kept that included a description of the setting, interactions between people, objects and the environment, emotions, impressions and reflections. Interviews with PWLD, their family members, support workers, provider managers and commissioners were conducted for each site allowing us to develop our understanding of the conditions necessary to provide excellent support (Table 6). An interview schedule was used (see Report Supplementary Material 9), although the process was flexible, allowing participants to introduce issues they felt were important. Interviews, at a participant’s home or online/by telephone, lasted 20 minutes to 2 hours, were recorded with permission and were transcribed verbatim.

TABLE 6

Work package 3: engagement with participants

Independent supported living

Based on WP2 (see Chapter 3), three providers offering ISL services were selected: a social enterprise provider in the north of England (ISL_1); a large, not-for-profit national organisation (ISL_2); and a small, limited company provider operating in the south (ISL_3). An add-on site (ISL_4), a not-for-profit charity, was identified that provided exemplary care and support for older PWLD in a deprived area of northern England. Support ranged from 24 hours of one-to-one support with two-to-one support for going out, to a few hours of one-to-one support per week. In-person ethnographic visits lasted 2–6 hours and revolved around participants’ timetables, with researcher(s) taking part in activities or shadowing (arts and crafts, drama, walks in town/city and in nature, shopping, outings for coffee, going to the cinema, picnics and drives, going to a day centre) and spending time in participants’ homes. Most visits took place on weekdays, including evenings; however, a minimum of one weekend visit was included for each provider. Online meetings took place between in-person visits, during weekdays, and lasted between 15 and 30 minutes.

Day centres

One provider (DOP_1), a charity operating community hubs in the Midlands, was included. Participants were recruited from two hubs with older PWLD. Of the six participants (three from each hub), four lived alone (one person paid for a few hours of support and one person had a few hours of support from the provider), one lived with their parents and one person lived next door to their sister.

In-person visits lasted 3–7 hours during hub opening hours (9 a.m.–4 p.m.), with researcher(s) taking part in and/or shadowing activities (arts, crafts, cooking, board games, computer games, chair yoga) and shadowing participants on outings. The hubs closed at weekends.

Shared Lives

Shared Lives is a model in which people with a wide range of support needs, including PWLD, autistic people, young care leavers and people with dementia, live in their local community with families. Two exemplary schemes were selected from WP2: one based in a northern city (SL_1) and one in the south-west (SL_2). Participants used respite Shared Lives or lived full time with families, and visits included spending time at the Shared Lives home or going out with the Shared Lives carer and participants.

Nursing care home

Nursing and Residential Care Home_1 is a nursing care home in a large market town in northern England. The service provides support to 4 households of 6 PWLD and additional nursing care needs (24 people in total).

Ethnographic observations focused on three participants. Visits to the site lasted approximately 3–4 hours, including mornings, afternoons, evenings and weekends. The stories of two men who had died in the nursing care home before the research took place were included in interviews with their relatives and the staff team.

The co-research team

We recruited and trained a team of co-researchers, PWLD (n = 8) and family carers (n = 1) using the co-designed training resource discussed above. The aim was for 50% of the visits to sites to be with a co-researcher other than the two nursing home sites. In practice, this was not possible, due to individual preferences of participants, inaccessibility of certain sites, and availability of co-researchers.

Observations from co-researchers were documented in various ways. The team’s preference where possible was for the researcher and co-researcher to go somewhere quiet to talk about the day and the researcher to take notes. This was not always possible and a debrief was sometimes held online, or by telephone, later in the day, or the following day. Three informal online meetings were held with the co-research team, which generated additional observations about the sites.

Co-researchers brought their knowledge and expertise to the research. For example, one researcher drew on her experiences as a wheelchair user to comment on the lives and home of two participants who used wheelchairs. Others were able to reflect on their experiences as quality checkers for the CQC. Co-researchers local to the site area also brought expert knowledge of local places, venues and self-advocacy.

Policy documentary review

Relevant policy documents were requested from each site. These focused on how the organisations support PWLD and ‘BTCO’ as well as on the design and delivery of support around growing older and EOLC. ‘BTCO’-related documents included those explicitly focused on ‘BTCO’ and those relating to reducing restrictive practice and physical interventions, and mental capacity. For sites without EOLC documents, documents relating to health and well-being were reviewed. See Report Supplementary Material 10 for included documents.

Data analysis

Data analysis was conducted alongside fieldwork, with researchers comprehensively familiarising themselves with the observation notes, interview transcripts and policy documents. These data sets were entered into NVivo (QSR International, Warrington, UK) and thematically analysed. The research team open-coded data, meeting to discuss developing coding frameworks. Coding categories were flexible, and an iterative approach to this first stage of analysis incorporated alternating between transcripts, notes, codes and categories as new codes was identified. A more conceptual analysis involved selecting particularly rich and relevant categories, reading and re-reading the data, generating themes, making links, identifying patterns and thinking about the data in a less linear and descriptive way. Meaning, action and process were used as an interpretive analytic lens to understand how participants understand and make sense of their lives, their actions and their interactions. Our aim was to capture the micro-detail of how participants live their lives and engage with other people at home and in the community.

The policy documents were also analysed using NVivo. In addition to the content, we reviewed their intended audience and authorship, references to research and external standards, accessibility of language, and the process of reviewing and updating them. Existing policies were aimed predominantly at staff. The documents were up to date, with frequent revisions and version history stated.

We did not plan to involve co-researchers in data analysis because of the quantity of textual data involved.¹¹⁶ However, insights generated during fieldwork led us to revise this. Two analysis workshops were held with two co-researchers and a support worker, which involved close audio examination of data extracts where participants talked about ‘BTCO’ (Mikulak et al., in preparation).

Nursing Care Home 2: the commissioning and development of a new inner-city service

Nursing Care Home 2 (NCH_2) was identified during the original proposal design development; however, the opening of this service was delayed. Following discussions with the project team and Advisory Groups, we agreed to continue to research NCH_2 to generate insights into the process of planning a new service specifically for this group. In addition, NCH_2 enabled us to consider the potential benefits and challenges involved in commissioning an inner-city service for older PWLD and ‘BTCO’, in the light of key policy drivers to keep people within their local communities.^{23,80,102,117,118}