Focused ethnography approach

This study used team-focused ethnographic methods. In a focused ethnography, rather than embedding a single researcher in a social setting for a lengthy period, more targeted data collection is used to explore the study topics. Using existing information from the literature and from what is known in clinical practice helps to determine the research question and, subsequently, to generate findings that are relevant and useful.¹³⁹

This project is a team-based ethnography as it involved three ethnographers working across a total of eight general practice sites in three regions (Scotland, Oxford and Bristol). Team-based ethnographies have become increasingly popular across disciplines.^140–142 This trend has been linked to funding bodies that encourage multidisciplinary team-based research, because of the wider range of expertise that will be brought to a project.^140,142 The use of a team-focused ethnography for this study has been explored further by Bikker et al.¹⁴³

Although ethnographic methods have been used in primary care for several notable studies,^144–146 they have usually required an intensive period of immersion (often by a sole researcher) in the study setting. The focused ethnographic approach lends itself well to health services research, because it is an efficient way to obtain an insightful understanding of concepts and processes within the fast-moving context of health-care policy development. This method was particularly suited to this study’s research questions as it allowed data to be collected on a predefined topic guided by the conceptual review work, while emphasising the importance of the context and the cultural landscape of the practices within a limited time frame.

Composition

The case study team consisted of three researchers with experience of ethnographic research who were employed to conduct the focused ethnographic fieldwork. Although having three researchers collecting data was time efficient, to minimise the costs, two of the researchers were employed solely for the duration of the data collection (12 months). The third researcher was employed for the duration of the project (27 months) and also took on the role of project manager. The ethnographic researchers were based in two universities in England and one in Scotland. They had different backgrounds, which included medical and social anthropology, nursing and mixed-methods research methodologies. The three ethnographers were managed by two co-applicants: Helen Atherton on a day-to-day basis with input from Sue Ziebland (see Table 1, Chapter 1). This subteam of five researchers made up the ‘focused ethnographic team’. The wider study team was also involved in providing a ‘check’ on the data the ethnographic team were collecting.

Communication

Steps were taken to ensure that the focused ethnographic team worked cohesively on the data collection and analysis. Over the period of the fieldwork and analysis, this involved:

• the focused ethnographic team setting out a strategy for teamworking during their initial meeting
• the three ethnographers having a weekly telephone meeting
• the three ethnographers and the day-to-day focused ethnographic team lead (HA) having a fortnightly telephone meeting
• the three ethnographers contacting the day-to-day lead (HA) via e-mail or telephone, when necessary
• the day-to-day lead meeting regularly with the senior focused ethnographic team lead (SZ)
• the entire focused ethnographic team attending three data collection and analysis workshops over the course of the project: two in Oxford and one in Edinburgh
• the ethnographers attending the project meetings (both face-to-face and telephone meetings) during the 12 months of fieldwork, allowing for updates to be made to the wider study team, and for the ethnographers to hear about the progress of all elements of the project.

Selection of sites

We recruited eight practices as planned. Two were in Scotland, three were in Oxfordshire and three were in Bristol. The scoping survey (see Chapter 3) was used to identify potential practices. Based on our survey data, we constructed a matrix of practices detailing various practice characteristics. Characteristics included experience of implementing different types of alternatives to face-to-face consultations; the size of population the practices serve; and location, in terms of whether it is urban or rural and deprived or affluent (based on deprivation scores). We then selected practices to approach, ensuring that we covered a range of demographic characteristics and applications of alternatives to the face-to-face consultation, for example ‘currently using’, ‘tried and rejected’ or ‘substantially modified their plans to use’ alternatives.

We sent invitation letters to practices, followed by telephone calls to the practice manager and/or GP to ensure that the invitation had been received. When practices were interested, a researcher visited to provide more information and discuss the implications of participation. Practices that declined to participate stated excessive workload and/or staff shortages as the reason.

Consent and observational work

The ethnographers observed practice staff in all areas of the practice, such as clinical areas, reception desks and administrative offices. Observations included consultations, both face to face and alternatives to face-to-face consultations.

In the first instance, the practice manager agreed to the practice participating in the research and informed all staff that the ethnographers would be situated in the practice for a number of weeks.

On commencing the research within the practice, informed consent for observation was sought by the ethnographer from all members of the practice. This involved providing all staff members with an information sheet about the study and providing a consent form. The information sheet outlined that the ethnographers may observe them at work, engage in informal conversation and take notes. They were given up to 1 week to read the sheet and complete the consent form before a further enquiry was made. Some staff members were not based in the practice and declined to consent, on the basis that we would not be observing them. A small number of staff members declined to consent. We ensured not to engage them in informal conversation or directly observe their work. We did not take notes on any work they were engaged in. In addition, those consenting were informed that, at any point during the research period, they could decline observation or conversation. Staff members engaging in interviews were given an additional information sheet and completed an additional consent form for participation in the interview.

Observations of consultations were at the discretion of the GP. Patients were invited to consent to their face-to-face consultation being observed, given an information sheet and asked to consent to the observation. For observation of telephone and e-mail consultations, the ethnographers were not party to patient-identifiable data and so consent was not required; however, the clinician was able to decline the observation if they felt that it was not appropriate. If the ethnographer observed the GP within their consulting room but outside a consultation, they took notes but did not make notes relating to a patient or their condition, being focused only on how the GP worked (e.g. when they responded to e-mail consultations or when they scheduled their telephone consultations). Data were collected using field notes in all cases, and these did not include patient-identifiable data.

Summary profile

In each practice, the ethnographers completed a summary profile for that practice. This was designed to capture the detail obtained by the ethnographers in the field in their own words. It included, among other details, the types of alternative consultations that are (or were) provided, how these are/were provided (e.g. timing, volume, staffing) and any parameters for the types of patients who are/were allowed or encouraged to use alternatives to face-to-face consultations, including variations between practitioners. Over the course of the data collection, the ethnographers added any new observations about the practice to the summary profile template for each of the eight sites. This allowed individual field notes to be transformed into a common format. As a result, comparing the observations between practices became more straightforward, even though the personal styles of completing field notes differed between the ethnographers. At the start of the data collection, the discussion about the research topics among the focused ethnographic team was more general, reflecting the exploratory nature of the study and the process of familiarisation with the field. However, as the data collection progressed, the format of the summary profile template evolved and became more focused on alternatives to the face-to-face consultation.

The ethnographers recorded their own field notes during observations. In more formal settings, such as practice meetings, minimal notes were made by ethnographers in real time, followed by more detailed field notes retrospectively.

Document collection

In each of the case study practices, we sought to review notes and minutes from practice meetings over the preceding 6-month period. Some practices were reluctant to give our researchers full access to these notes and minutes, and so the practice managers in each practice reviewed the minutes, identifying and extracting any relevant information.

In each practice, researchers requested that they be invited to observe practice meetings during which alternatives to the face-to-face consultation would be discussed. In the end, this was the case for only one meeting in one practice, where the researcher attended the meeting and kept notes.

Semistructured interviews

Each interview participant provided informed consent. Interviews were digitally recorded, using an encrypted recorder. The files were transcribed verbatim by a professional transcription service.

Staff interviews were usually conducted in the general practice (one GP was interviewed at home). Additional interviews relating to video consultation at non-case study sites were conducted over the telephone. Information was collected on sex, age, ethnicity, role in the practice, length of time in the role, length of time since qualifying and the type of alternative to the face-to-face consultation used.

Patient and carer interviews were conducted at a location of their choosing, usually their home (or the general practice). Information was collected on sex, age, ethnicity, whether or not they were a carer, whether or not they had a long-term condition or disability, level of education and occupation.

The initial design of the topic guides was based on the findings of the conceptual review and templates used by focused ethnographic team members in previous studies. The topic guides were then revised iteratively among the ethnographic team once the ethnographers were working in the field, and a final topic guide was agreed on for both staff and patient interviews. The staff topic guide differed slightly according to the staff member being interviewed; for example, GPs were asked specifically about how the alternative to the face-to-face consultation affected their relationship with patients (see Appendix 5 for the topic guide).

Case study site staff

Once they had been conducting observations in the practice for a few weeks, the ethnographers identified relevant staff members to participate in the interviews. The ethnographers had made contact with potential interviewees during non-participant observation, and provided them with the information sheet relating to the interview. If they were willing to participate, an interview was arranged at a time that was convenient for the staff member.

In each practice, administrative staff (including receptionists), GPs and nurses were recruited to be interviewed. When relevant, other members of the practice team were also interviewed, for example, a rural health worker in one practice, a patient manager and an IT manager in another. The team of ethnographers remained in close contact throughout the process, to ensure that the range of staff members being interviewed was suitably varied in regard to factors such as role in the practice, knowledge or involvement in alternatives to the face-to-face consultation.

At the protocol stage, we intended to interview allied health professionals working in general practice, such as phlebotomists and community-based pharmacists. However, it became apparent during observations that, where allied health professionals were working in the practices, they had little to no involvement in the introduction or use of alternatives to face-to-face consultations. Recognising the importance of the wider study team members within general practice, each ethnographer engaged in informal conversations with these staff members to ensure that their perspectives were covered. These conversations were recorded in the researcher’s field notes and subsequently referred to in the structured summary profile.

Users of video consultation

As described in Chapter 3, we recruited an additional four participants from outside the case study sites. These were general practice staff members who were using, were about to use or employed a system that used video consultations.

• We approached practices that had been funded by the GP Access Fund¹⁴ to employ video consultations.
• We approached practices that had publicised their use of video consultation in news articles or reports.
• We posted a message on Twitter:

  Has your practice offered patients video/skype consults? We are looking for GPs to do a 30m phone interview for AltCon study.

  Helen Atherton has 865 Twitter followers and tweets about digital health in primary care.
• We used personal contacts (HA and CS) obtained via related research studies.

Two participants were recruited via Twitter, and a further two were recruited via personal contacts (HA). We were unable to contact all of the GP Access Fund¹⁴ practices and, in the case of those we did contact, we could not identify individuals who were using video consultation. We received no reply from the practices that had publicised their use of video consultation via news articles or reports.

Patients/carers

The aim was to interview patients with different characteristics in relation to age, sex, ethnicity, disability, frequency of attendance and whether or not they had long-term health conditions. All patients invited to participate in interviews had experience of using an alternative to the face-to-face consultation within the practice.

Initially, patients were identified opportunistically, based on those who had engaged in contact with the practice via an alternative to the face-to-face consultation. In subsequent interviews, patients were purposively sampled to ensure that participants with the range of characteristics listed above were included. Practice staff and GPs helped to identify patients and carers and provided potential participants with a study information pack, either via post or in person when attending the practice. This pack included information about the study, an invitation to take part and a reply slip, which they could return via prepaid post. The researcher then called to arrange a convenient time for the interview.

Hard-to-reach and disadvantaged groups

In using a purposive sampling technique, we specifically included people who were identified in the protocol as being in hard-to-reach groups with regard to accessing general practice. Examples included parents/carers of people with complex needs, young men, the vulnerably housed and minority ethnic groups.

The other groups of interest were those that might be disadvantaged by limited provision of alternatives to the face-to-face consultation. These groups were described in the protocol, and further relevant groups were identified in the conceptual review (see Chapter 2). These included patients with mental health conditions, patients living in rural areas, patients with restricted mobility, patients with hearing loss, patients at a great distance from the practice (e.g. working away) and patients with low health literacy and/or low computer literacy.

The aim was to look at the range of problems and issues for these groups, rather than making statements about specific population subgroups. To make it easier for people in these groups to participate, there was flexibility about timings and locations for interviews, with telephone interviews offered when appropriate. Participants in hard-to-reach and disadvantaged groups were identified by the practices so that they could be invited to interview.

Coding

The coding frame for analysis of the ethnographic data was devised by the focused ethnographic team at a face-to-face meeting early in the data collection period. Each ethnographic team member had read a series of field notes and transcripts, and contributed to devising a coding structure for the staff data and another different coding structure for the patient data, which comprised interview data only. The three ethnographers read and coded interview transcripts and field notes [using the comment facility in Microsoft Word 2013 (Microsoft Corporation, Redmond, WA, USA)]. Once the coding frame was in place, the field notes were coded and each ethnographer condensed their findings into a summary profile. The day-to-day lead (HA) read the coding of the field notes into the summary profile and read every transcript, checking the coding to ensure reliability and comparability, and adding or making adjustments where relevant. The transcripts and summary profiles were then entered into NVivo software, which allowed thematic reports to be generated.

Analysis

A series of NVivo reports were generated to gather related sections of the data together. At this point, two of the ethnographers had completed their contracts. The day-to-day lead of the focused ethnographic team (HA), senior team lead (SZ) and remaining ethnographer (HB) read all of the reports. They applied the OSOP method⁹⁰ to identify the line of argument in each thematic report and identify outliers or negative cases. The data from the staff and patient interviews and the field notes were integrated, and a condensed summary was produced for each thematic code.

At this point, the wider study team became involved and were paired with the remaining members of the ethnographic team to discuss the interpretation of the data (HA and BM, HB and CS, SZ and JC). The core messages were presented at a wider team meeting and the analysis refined through discussion among all the team members.

In addition, we held a stakeholder workshop to present and discuss the initial findings and their application. Attendees included academics, policy-makers and health-care professionals. The responses from the workshop delegates were considered during the final stages of the data synthesis. See Chapter 7 for further information on the stakeholder workshop.

Staff interview participants

We interviewed 45 members of staff from the case study sites, and a further four from other practices. In the case study sites, we interviewed 19 GPs, eight practice managers, two deputy practice managers, one practice co-ordinator, two senior practice nurses, three practice nurses and one nurse practitioner, one rural health worker, four senior receptionists, one receptionist, one patient service manager, one practice administrator and an IT manager. The four participants from practices outside the case study sites were three GPs and one practice manager. Various different types of alternatives to the face-to-face consultation were used by staff in the case study sites. All GPs and nurses were using telephone consultations.

The participants from outside the case study sites were interviewed about video consultation; one had used video consultation with patients, two had limited experience of using video consultation with patients and one practice manager was in the process of setting up a video-consultation service.

Of the 48 staff interviewed, 33 were female. The age of participants ranged from 31 to 68 years. The majority of participants were white British. For clinical staff, the length of time since qualifying varied from 7 years to 40 years. Staff had been in their current role for varying amounts of time, from 10 months to 31 years. Of the GPs, 16 were partners and six were salaried.

Full details about staff interview participant characteristics can be found in Appendix 7.

Patient and carer interview participants

We interviewed 39 patient and carer participants. All participants were using or had used an alternative to the face-to-face consultation with a health-care professional at their practice. We collected information on their ethnicity, age, sex, current health conditions and whether or not they were carers, alongside information about their level of education and current employment status. Full details about patient interview participant characteristics can be found in Appendix 8.

Hard-to-reach and disadvantaged groups

The sample included a wide range of participants who may find it difficult to engage with general practice settings for varying reasons. These included a parent looking after a disabled child, an asylum seeker who was vulnerably housed, two men aged < 30 years (who typically do not engage in health care) and three participants in minority ethnic groups with English as a second language.

The sample also included patients who may find themselves disadvantaged by the current limited provision of alternatives to the face-to-face consultation, namely six patients with mental health conditions, four patients living in a very rural area, nine patients with restricted mobility and two patients with hearing loss.

Patient and public involvement and young people

The age range of our patient/carer participant sample was wide, but it did not include any participants younger than 24 years (range 24–91 years). Young persons below the age of 18 years were not eligible for inclusion in the study. Being mindful of the potential importance of alternatives to the face-to-face consultation in young people, we sought to obtain the views of young adults to supplement the study. We did this by organising a PPI event for young people. We were able to do this via ‘Bristol Young Health Watch’, a group of young people aged between 16 and 19 years who work alongside the Bristol CCG. The study PPI lead (AG) met with the group during one of their regular existing meetings. Group members were asked to comment on a number of issues related to alternatives to face-to-face consultations in primary care. These comments were recorded on flip charts during the meeting.

Although the people attending the workshop felt comfortable with potentially using new technology to access GP services, they reported similar concerns to other patient groups about issues of confidentiality, being able to exert choice over when to use alternatives and seeing these as a supplement rather than an alternative to face-to-face consultations. The content of the discussion was used to provide context for the findings obtained via interviews and observation.