Introduction
For the second phase of the study, in-depth case studies were undertaken to examine how social distancing and isolation of residents were being implemented in CHs for older people. This involved individual interviews with CH staff, residents and family members, and the collection of social distancing and isolation policies and protocols and routinely collected CH data. When interviewing staff members, we asked about their experiences of social distancing and isolation and their understanding of resident experiences, based upon what residents had reported to them and their own observations during the pandemic. When interviewing residents and family members, we asked about their own specific experiences of social distancing and isolation. We also conducted focus groups with senior health and care leaders in England and national-level stakeholders to understand their experiences of developing and applying policy for CHs, and how they responded to the resulting challenges.
Individual interviews: method
Research team
The interviews were carried out by four members of the research team: SP, SS, AD and JF. All interviewers are established academic researchers with experience in qualitative interviewing and a background in health and/or social care research. The interviewer and participant had no relationship before the interview, as all recruitment was carried out by the CH manager or project champion at the participating case study sites.
Recruitment
Six CHs in England were recruited for the study. Care homes were invited purposively, using a sampling frame designed to maximise variability in terms of size of the CH, geographical location, Care Quality Commission (CQC) rating, registration (nursing, residential or dual registration), ownership and incidence of COVID-19. The pandemic experience for the CH sector has impacted on CHs being research ready; the team worked hard over a prolonged period to recruit two case study sites with a ‘requires improvement’ CQC; this was not successful. We also managed a key issue around recruitment of participants from black and minority ethnic (BAME) groups. At meetings of the Study Steering Committee and Data Monitoring and Ethics Committee, we discussed these challenges and there was agreement that we should prioritise diversity of participants and not focus on recruiting CHs with a CQC rating of requires improvement. This revised plan was shared with NIHR and we were granted permission to proceed with this revised plan.
Recruiting CHs began with an initial meeting with interested provider organisation representatives, who had been sent the study information via existing contacts and networks of the research team. Following this meeting, the provider representative nominated a CH that met the criteria and would have the capacity to participate in the study. A further meeting(s) took place between the manager of the nominated CH and the researchers to provide additional details about the research and involvement of the CH. All meetings took place remotely using Microsoft Teams.
Care home managers were asked to nominate a ‘project champion’ to be the point of contact within each home to help facilitate the research, which was conducted entirely remotely due to COVID-19 restrictions on visiting care facilities. The project champion was required to be a member of staff who knew staff and residents well, and who had the capacity in their role to help with the recruitment and interview process. It was undertaken by staff members with different roles in each home, including the CH manager, deputy manager, well-being co-ordinator, activity co-ordinator and administrator. The project champions were briefed about the study and guided by research team members throughout the process; this included us working closely with CHs to try and increase diversity of participants. Potential interview participants were nominated by the CH manager in collaboration with a member of the research team and invited to participate by the project champion, using the paper copies of the study information sheets and consent forms sent to the home by the research team. The information sheets and consent forms were tailored to each participant group; resident and relative documents were produced in an easy-read format, following guidance from the Dementia Engagement and Empowerment Project (DEEP) and feedback from our PPI group members.
The purposive sample of participants required from each home included staff (n = 5), residents (n = 3) and relatives or friends of residents (n = 3). For CHs with nursing or dual registration, the staff sample consisted of the manager/deputy (n = 1), RNs (n = 2) and care staff (n = 2). For the CHs without nursing (residential), the sample included the manager/deputy (n = 1) and senior care workers or care workers (n = 4). Inclusion criteria for staff included being permanent staff (i.e. not agency) and having worked at the home during the pandemic. Staff were purposively recruited to ensure a range of age, gender, ethnicity and time in service. Resident participants were also recruited purposively, to ensure a range of genders, ethnicities and different health and care needs. Inclusion criteria were that residents must be over 65 years old and have the capacity to consent. Residents were asked during the consent process if they would like to nominate a friend or family member to participate in the research. If they or the invited family member declined or there was no nomination, the CH manager/project champion was asked to recruit a family member or friend of a non-participating resident. The project champion was responsible for collecting informed consent from participants and sending scanned copies of the completed forms to the researchers ahead of the interview. All participants were given the option to have a phone or video call with the researcher before giving signed consent, to ask any questions or talk through the research process; however, none chose to do so.
As part of the consent process, participants were asked if they were happy for the interview to be video-recorded so that excerpts from interviews could be used to create a short film as part of an evidence-based resource for CHs. The information sheet explained that a television production company technician would be present for the remote interview to ensure that there were no issues with recording. The technician could be asked to leave the call at any point should the participant wish. The option for being video-recorded was voluntary. If a person declined, they were asked whether the interview could be audio-recorded or, if preferred, only written notes to be taken. Participants were also given the option to have their face pixelated in the final video if they were happy to be video-recorded but wanted to maintain anonymity. In total, 12 participants (5 staff, 4 residents and 3 family members) chose to be audio-recorded but not video-recorded, and none opted for written notes only. The remaining participants all agreed to be video-recorded.
Setting
All resident and staff interviews took place at the CH, using an iPad sent to the home by the research team. Interviews were carried out in either the resident’s room or a quiet place in the CH such as the manager’s office, visitors’ room or hair salon when not in use. Relatives were given the option of doing their interview in their own home using their own device (e.g. smartphone, laptop, tablet, telephone). One family member chose to be interviewed at home and all others were carried out at the CH, complying with requirements for visitors.
The iPad was set up for each interview by the project champion and positioned so that the participant could see the researcher on the screen. Interviews that were being video-recorded were carried out using VMix, a secure video-call service hosted online and accessed by the technical team at KMTV. KMTV were responsible for making both the audio- and video-recordings for those participants who were video-recorded and for forwarding the audio files to the research team. Interviews that were audio-recorded only took place on Microsoft Teams. All participants were asked if they would like the project champion to be present during their interview and 17 participants (7 staff, 5 residents and 5 family members) chose to have the project champion present.
Data collection
Interviews at CHs were conducted non-simultaneously, and interviews at one CH were generally completed before interviews at another CH began. This approach to data collection meant that data were collected at different times for different CHs. These time periods were as follows: Care Home 1, February and March 2021; Care Home 2, March and April 2021; Care Home 3, April and May 2021; Care Home 4, June and July 2021; Care Home 5, August and September 2021; Care Home 6, October to December 2021. Interviews were semistructured, with a separate schedule of questions for each participant type. Interview schedules were developed by the research team and reviewed by the PPI group, and by a CH manager and resident from a non-participating CH.
Additional prompts were added to the schedules following the initial interviews of each participant type in the first CH and were agreed upon by the research team. The schedule of questions was shared with each participant at the point of recruitment to the study. Immediately before the interview began, the researcher checked consent, reminded that participation was voluntary and that the interview could be paused or stopped at any time, and gave the participant the opportunity to ask any questions. A demographic form was also completed before the recording began. For resident participants, these were collected with permission from residents and the care manager (e.g. about their primary health needs, length of time living in the CH, age group, gender, ethnic group). For families/friends participating in the study, demographic data included the nature of their relationship to residents, age group, gender and ethnicity. Staff participants were asked for their role title, length of time in the current role, length of time working in the CH sector, age group, gender and ethnic group. Once started, interviews lasted between 20 minutes and an hour. Following each interview, the researcher made field notes about the engagement of participants, any key points that had arisen, and whether there had been any technical issues, such as problems with Wi-Fi.
Data analysis
The interview audio-recordings were transcribed verbatim by a transcribing company approved by King’s College London (KCL) and were quality assured by a researcher. The interviews for each participant group were assigned to one researcher (staff interviews – SS, resident interviews – AD and family interviews – JF). We adopted an inductive orientation to thematic analysis – analysis was located within, and coding and theme development were driven by the data content.28 At the beginning of the analysis process, a sample of staff, residents and family/friend transcripts were each read and coded independently by five researchers (JF, AD, SS, RH, SH). The researchers met and discussed their coding and, as a team, compiled a specific coding index for each participant group. The themes identified in the rapid review (phase 1) were used as deductively derived main themes in the coding indexes, and subthemes and any additional themes were inductively derived from the transcripts. AD and other team members began developing a coding framework for resident interviews based on the themes from the review and initial readings of the transcripts. However, it quickly became apparent that the elliptical nature in which many residents spoke in response to questions meant that using a framework was a blunt and, therefore, not particularly useful way of analysing this data set. AD read and reread resident transcripts and generated key themes of resident discussion (loosely described as ‘codes’) and compared how these themes were expressed across the interview data set. Researchers each analysed transcripts from their assigned participant group, but SS, AD and JF met regularly to discuss and compare their findings and modify their indexes accordingly. JF also read and analysed a subsection of transcripts coded by SS and AD for quality assurance.
