Aims and objectives of the qualitative exploration
The qualitative exploration was conducted to investigate in-depth, and from the perspective of practice staff, patients/carers and other stakeholders, the potential benefits and challenges associated with different models of DF. The study was informed by the research undertaken in earlier phases of the Di-Facto study, reported in previous chapters of this report.
The qualitative exploration comprised two elements:
- Focused ethnographic case studies in general practices:
Focused ethnography was used to explore in-depth the use of DF in general practices. General practitioner practice case study sites were selected and through non-participant observation, semistructured interviews of staff and patients and secondary analysis of documentation, the researchers sought to build up a detailed picture of DF.
- Interview study with key stakeholders:
In addition to the in-depth understanding of experiences of DF within individual practices, we sought to locate this study within a broader context and understanding what influences DF in primary care as a setting. We conducted semistructured interviews with stakeholders who may have critical oversight at a level beyond individual practices, for example, policy-makers, commissioners and third-sector organisations.
COVID pandemic
The study was designed prior to the COVID pandemic. COVID led to significant changes in the provision of primary care with an increase in remote consultations and use of digital services. It was necessary to adjust our approach in light of both the changes in primary care due to COVID and national COVID guidance. Fieldwork was conducted between September 2021 and July 2022. During this time, on 8 December 2021, England moved to ‘Plan B’ for Omicron which affected the fieldwork and recruitment, and the ‘living with COVID’ guidance was published on 24 February 2022.173 The case studies focused on DF reported at the time of the research and within the context of the COVID pandemic, the research explored how general practice made changes to service delivery in light of COVID.
Rationale for the methodological approach
Rationale for the methodological approach
The study required an approach which would enable the research team to explore in-depth how DF occurred within general practice. As noted by Reeves et al.174 ‘the central aim of ethnography is to provide rich, holistic insights into people’s views and actions, as well as the nature (that is, sights, sounds) of the location they inhabit, through the collection of detailed observations and interviews’. A focused ethnographic approach was selected. Unlike traditional ethnographic approaches, focused ethnography enables shorter and more focused periods of ethnography and focus on a predetermined topic. The approach had also been successfully used by researchers on this project (HA) in previous work.175 Interviews were used to explore the views of stakeholders on DF and complement the general practice-based research.
Qualitative exploration team
Team composition
The core qualitative exploration team comprised nine members (see Contributions of authors).
Communication
Throughout the study the qualitative exploration team met monthly to ensure communication between team members. In early 2022 two whole-team face-to-face meetings occurred to discuss and develop analysis plans. In addition, the case study team met fortnightly during the data collection and analysis phase to ensure consistency of approach and discuss any issues arising. Researchers conducting the focused ethnography met weekly during the data collection phase. Stakeholder interview study researchers and mental health interview study researchers also met regularly during the relevant parts of the study.
Focused ethnographic case studies
The focused ethnographic case studies involved a researcher spending 6–8 weeks conducting fieldwork in each practice. The intensity of attendance varied according to the extent and types of DF activities within the practice, the size of the practice and COVID restrictions in place. We sought to understand the practice context (i.e. patient demographics, location, ways of working), how and why the model of facilitation in use might lead to benefits, the key processes involved in DF, and identify outcomes of importance to patients/carers and staff. We also investigated factors that may act as barriers and facilitators both to the practice being able to offer DF and to patients or carers using online services.
Case study recruitment
Sampling of general practices
Our aim was to include a varied sample of eight practices that would provide findings that are relevant across a range of primary care settings. Sampling of general practices focused on seeking variation according to two factors.
Experience of and/or delivery of digital facilitation using different approaches
As outlined in the protocol we aimed for variation according to the types and level of DF in use within practices. Individual practices were identified via the practice survey responses (see Practice survey) and where needed by obtaining information from practices directly about the types of DF in use.
Practice characteristics considered to inform sampling (illustrated in Figure 14)
FIGURE 14 Illustration of sampling considerations for the selection of case study sites. (Figure not reproduced here; see the original article.)
We selected practices for variation according to range of characteristics:
- Practice setting (rural/semi-rural/urban) appears to be systematically associated with variation in access to general practice, for example, older people in rural areas have poorer access to GP services176 and face digital exclusion due to poor availability of broadband internet.177
- A range of IMD scores at practice level, as socioeconomic status appears to mediate health disparities via reduced health literacy.167
- Number of patients aged ≥ 65 years in the general practice, as although levels of use are increasing, older adults (≥ 75) are less likely to use the internet, and the 16–34 age group are the highest users of the internet.
- List size of the practice [from a range across small (< 6000) to large (> 12,000) registered patients].167
Recruitment of practices
The results of the practice survey conducted in WP2 (see Practice survey) were used to identify general practices for potential participation. The sampling considerations indicated in Figure 14 were used to identify a range of potential sites via a sampling matrix. Where this was not possible, the researchers worked with local CRNs to support the initial approach to practices. Where this was not possible the researchers contacted general practices directly. Some CRNs also approached or provided contact details for practices that had not completed the survey but matched the matrix criteria. An eighth practice was recruited to ensure greater diversity in the study sample.
Practices were sent an introductory e-mail which was followed up with a phone call or e-mail up to two times with a week between each contact. In the case of no response, it was assumed that the practice did not wish to take part and they were not contacted again.
Interested practices were given written information about the study to share with staff members. For practices agreeing to participate, an online site visit was arranged where the research team, and where appropriate the CRN, explained the study and answered any questions. Sampling was continuously reviewed and further recruitment was tailored to ensure the intended range and variation in participating practices.
Data collection and management
Data collection approaches included in each case study site are given below.
Non-participant observation
Observations focused on DF occurring within the general practice, specifically, what was involved in the process of DF and how practice staff and patients/carers interact with different types of DF. We did not directly observe consultations with patients but did ask about DF in clinical consultations during interviews with both clinicians and patients. Observations and informal conversations were recorded in extensive handwritten fieldnotes, without use of identifiers or names. Where possible, fieldnotes were made at the time of observation but if not, were written up as soon as possible afterwards. Each participating practice was allocated a letter (i.e. Practice A, Practice B, etc.) so that any information stored about the practice was de-identified.
Posters were put on noticeboards in both public and staff-only areas of the practice, and, where possible, on the practice website. The poster included a photograph of the researcher and the researcher’s contact details, along with an explanation of the study and the dates that the researcher was to be in the practice. Staff members and patients had the right to decline to be observed; however, this did not occur.
The researchers were sensitive to situations where their presence could cause any issue. Researchers were trained in observation techniques and had the support of an experienced qualitative researcher (HA).
Attending practice meetings
Researchers explored whether there were any relevant practice-level meetings to attend. These meetings would include those where DF or related issues were discussed. When attending such meetings, the researcher ensured that all participants were aware of who they were and why they were there. Fieldnotes were maintained both during and after meetings and, with permission, relevant documentation collated.
Collection of relevant data and documentation
Researchers collected documentation pertaining to digital services or facilitation within the practice. Staff were asked to provide these with any identifiable information removed. The researcher discussed with the practice staff member the relevance of the documents and why they were important in relation to DF.
Semistructured interviews
Semistructured interviews were conducted with patients/carers and practice staff within case study practices.
Sampling and recruitment of practice staff
We sought to interview three to four staff at each practice during the fieldwork, based on their role and involvement in DF activities, and invited to participate in an interview but were free to decline without providing a reason.
Sampling and recruitment of patients/carers
We sought to recruit three to four patient/carer participants per practice, sampling to include a mix of characteristics in terms of age, ethnicity and experience (or not) of DF. Patients were identified by practice staff who also checked the suitability of those selected (aged 18 years and over) to ensure that they had capacity to consent and that there was no other reason why they should not be approached. The practice sent patients an invitation (by post, e-mail or given in person) to participate along with an information sheet about the study. Recipients were invited to reply to the study team (using an enclosed reply slip or e-mailing the researcher directly via contact information on the reply slip) to indicate that they were willing to participate in the study.
Interview procedure
Topic guides for staff and patient/carer interviews were informed by the review and practice survey. The topic guide for staff covered the drivers for facilitation in the practice, the type of facilitation in use (resources, processes, valued outputs), the perceived success of the facilitation model and barriers or facilitators to implementation (see additional files www.fundingawards.nihr.ac.uk/award/NIHR128268; accessed 3 April 2024). The topic guide for patients/carers explored their use of online services outside of health, experiences of DF with the practice and barriers and facilitators to using online GP services (see additional files www.fundingawards.nihr.ac.uk/award/NIHR128268; accessed 3 April 2024).
Interviews were conducted either in person, by online video conferencing or telephone according to participant preference. With permission, all interviews were audio-recorded with the use of encrypted recorders. Consent was obtained prior to all interviews. Interviews were transcribed verbatim by a professional transcription service. Patient/carer participants were given a £10 shopping voucher to thank them for their time.
Data management
Data files were stored on secure servers and only accessible to the study team after being anonymised. Recordings were stored until transcribed and checked then deleted from servers.
Vulnerable groups
In the context of digital inequalities, individuals from older age groups, those in minority ethnic groups, those in lower socioeconomic groups, those in poorer health and individuals in rural settings are recognised as vulnerable groups.16–19 Throughout the qualitative exploration, researchers sought to include participants from these groups. This approach was embedded throughout the study and something researchers were conscious of in the design, conduct and writing up the research. For example, in the focused ethnographic work the research team sought to include a range of general practices from diverse areas to include vulnerable groups.
Stakeholder interview study
Figure 15 indicates the design of the stakeholder interview approach.
FIGURE 15 Design of stakeholder interview approach. (Figure not reproduced here; see the original article.)
Stakeholder sampling and recruitment
This element sought to recruit 12–20 stakeholders relevant to the provision or impact of DF beyond the level of individual practices. This included stakeholders who may have critical oversight of DF in primary care or those representing patient groups at local, regional and national levels, including those from CCGs and ICSs, NHS England and third-sector organisations.
Identification of stakeholders was informed initially by integrating policy reviews, knowledge of the health system and patient and professional bodies, networks of the wider research team and internet searches. Snowball sampling was then used to identify further stakeholders and to respond to emerging findings from the study, including recruitment of representatives of organisations which focused on DF or digital primary care for potentially vulnerable populations.
Interview procedure
Individuals were invited for interview by e-mail and sent an information leaflet about the study. Recipients were asked to reply to the e-mail to indicate if they were willing to participate. The topic guide (see additional files www.fundingawards.nihr.ac.uk/award/NIHR128268; accessed 3 April 2024) explored the key drivers of DF, how stakeholders thought DF worked and the intended consequences of its application, both positive and negative. Interview procedures were followed as outlined for staff interviews.
The experience of patients living with mental health conditions
The project was extended to include an additional focus on patients with an experience of a mental health condition. The aim was to explore within case study practices how patients living with mental health conditions experience DF.
Practice sampling
This element of the study aimed to include four of the case study practices, seeking practices from each geographical area (Southwest, East of England, West Midlands, East of England and the North West).
Patient sampling and recruitment
Interviews with four to five patients with mental health conditions were sought per practice. We focused on patients with anxiety, depression, or severe enduring mental illness, identified from diagnoses or from recent prescriptions recorded in the patient record.
Searches of electronic records were run at participating practices to identify potential participants. Practice staff checked the suitability of patients, including capacity to consent, before sending them a postal invitation along with an information sheet about the study.
Patients were informed that they had been invited to take part because they had been identified by the practice as living with a mental health condition. Invitees were reassured that the invitation had come directly from the general practice therefore researchers were not aware of their details and health condition and would not be unless they chose to take part in an interview.
Interested participants responded by returning a reply slip to the research team.
Interview procedure
The topic guide was an adapted version of that used for the main study. Changes were made in conjunction with members of the PAG with specific lived experience of mental health conditions, set up to support this element of the study. The topic guide was adapted to focus on the specific experience of this group of patients (see additional files www.fundingawards.nihr.ac.uk/award/NIHR128268; accessed 3 April 2024). Interviews were handled as in the main study (see above).
Data analysis
Data from the focused ethnographic case studies and stakeholder interviews were analysed together using reflexive thematic analysis.178 Analysis of the mental health interviews was conducted later using the same coding framework.
Reflexive thematic analysis178 was applied through the following steps:
- Reading transcripts and developing the coding frames; transcripts were read, anonymised and accuracy checked by the researcher who conducted the interviewer. A selection of transcripts were read by the wider qualitative team. Coding frames were developed in a series of team meetings. A patient coding frame, a practice and staff coding frame (to include fieldnotes and interview data) and a stakeholder coding frame were developed. Analysis of relevant data and documentation obtained from each case study site was used to provide context.
- Agreeing a final coding frame at an analysis meeting; we held an in-person meeting to finalise the coding frames and ensure that they captured data across all sources and of all types.
- Gathering-related sections of transcripts, fieldnotes and documents under thematic codes; researchers coded the data using Microsoft Word templates for each code.
- Applying thematic analysis to each line of argument in the text, looking for outliers and negative accounts; we used the ‘one sheet of paper’179 method to support the collation of the coded data into initial themes. The one sheet of paper approach is iterative and involves reading through the data under each code and creating a summary of the different issues. The issues for each code were then grouped into initial broad themes, or axial codes. This approach allows all of the issues raised by the data to be explained and not just those that are most commonly appearing. These axial codes were then compared across the one sheet of paper summaries and themes were derived and summarised.
The summaries of each theme were discussed refined at team meetings until the final themes were agreed.
- Finally, sharing findings with the wider research team to finalise interpretation.
Patient and public involvement and engagement
Box 5 provides information about the PAG and how PPIE influenced the qualitative explorations.