Aims

The research objectives for the qualitative aspects of the study were:

1. Is the CMGC acceptable to healthcare staff?
2. How might the CMGC become embedded into practice?
3. In what ways does NPT help to understand the barriers and facilitators to implementing the CMGC?

Together these study aims looked to identify issues associated with how the CMGC was implemented, primarily from the perspectives of the healthcare staff working with it. The application of NPT also afforded opportunity to assess which elements of the intervention, if any, were deemed particularly problematic for staff and how these may be best resolved.

Design

The overarching study was a single arm non-randomised prospective observational study investigating the CMGC intervention outlined in Fig 1 [12]. For the qualitative elements, an ethnographic approach [20] incorporating observations and interviews was adopted exploring how healthcare staff responded to the introduction of the CMGC pathway and how the context of the implementation affected their responses.

The CMGC intervention

Key features of the CMGC intervention were additional patient appointments built in to identify those with less than 15% reduction in IOP, deemed as non-responsiveness to treatment, when compared to their baseline IOP measurement [12]. The staff followed a prescribed algorithm which detailed how to proceed at each of the two clinic visits over four weeks [12]. After being provided information on the study and the new care pathway, consented patients arrived into clinic for baseline IOP measurement before having an IOP-reducing eye drop instilled on their behalf by a member of staff. Four hours’ later, the patient arrived back for another appointment where their IOP was re-measured and assessed for any noticeable effect. Patients were informed of their results on each occasion. Another set of appointments was then arranged four weeks’ later where this process would be repeated; a further re-measurement of IOP, followed by the instillation of an eye drop by a staff member. Again, four hours’ later, the effect of this drop on the IOP was assessed, establishing whether patients responded to eye drops and/or needed support with adherence before making a decision on the ongoing treatment plan.

Sample/Participants

After being provided information on the study, participants were purposively enrolled into the CMGC from four routine glaucoma clinics (see Table 2 for pseudonymised site characteristics).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Eye clinic site characteristics.

https://doi.org/10.1371/journal.pone.0255564.t002

Staff were purposively recruited to assist with implementing the CMGC intervention, carrying out its procedures and offering their views on any barriers or facilitators to a wider implementation. Inclusion criteria for these participants were: being employed by one of the research sites; and being either a doctor, nurse, orthoptist or optometrist. Staff invited to work within the intervention were selected based on their involvement with existing services and in conjunction with the participating health boards.

Data collection

This article reports two different sources of qualitative data. In assessing the extent to which the CMGC pathway was acceptable to participants, we carried out observations, as well as semi-structured interviews with the healthcare staff responsible for implementing the intervention. All data collection took place between June 2018 and March 2019.

Ethical considerations

The study received ethical approval from West Midlands–Black Country NHS Research Ethics Committee on 16 November 2017, IRAS Project ID: 232242. All participants were given information sheets about the study prior to gaining their written informed consent and all practices followed the guidelines of the Declaration of Helsinki [22].

Data analysis

The analytical process followed the principles of framework analysis, commencing with a period of familiarisation where transcripts, field notes and recordings were revisited by the team to identify consistently emerging themes [23]. Both sets of qualitative data were analysed together to allow cross-comparison between them. With these preliminary themes established, the next phase of analysis involved the development of a framework to make collective sense of them. Many of the study’s a priori concerns relating to the normalisation of the intervention factored into this process through the core constructs and subcomponents of NPT, although analysis was also sensitive to localised themes emerging from the data.

Rigour

Data trustworthiness was enhanced by the collection and analysis of multiple data sources: longitudinal organisational observations; clinical consultation observations; and staff and patient interviews [24]. Furthermore, the input of the experienced research team during coding to the developed framework informed the indexing of datasets to applicable themes. This was done with the assistance of QSR International NVivo software to organise data, with regular check-ins between researchers and senior investigators to ensure agreement on where data fitted within the developing themes [25]. Once coding was complete, the final thematic framework of data was mapped and interpreted collectively by the research team, establishing the key themes relating to the implementation and how these interrelated [23].

Staffing for CMGC

The research sites each had varying staff numbers, job role, clinic timings and follow-up strategy (Table 2). These had potential to indirectly influence key aspects of how the CMGC was perceived by staff implementing it. Each research site was given autonomy when selecting staff to be involved with the implementation of CMGC, although each clinic required a consultant ophthalmologist to act as Principal Investigator. Based on each site’s pre-existing staff composition the number of staff allocated to the CMGC varied. In three of the clinics, the number of CMGC staff within each site was not viewed problematically in clinical observations, with capacity rarely provoking delays to appointments. However, the limited number of staff enrolled in the CMGC from Oak Clinic, alongside broader departmental capacity issues led to CMGC recruitment being halted and ultimately abandoned as key staff resigned or had long periods of leave:

“As OH302 (the assigned CMGC clinician) leaves, there is little clarity on how the CMGC will be hosted in Oak Clinic after that point”

Week-by-Week Observation 180716; Oak Clinic

Efforts to identify alternative CMGC staff to facilitate the intervention for this site were unsuccessful. This linked to the NPT construct of cognitive participation, and particularly the subcomponents initiation and enrolment. Both are concerned with the correct staff being in place to build and maintain the set of new practices around the intervention. Within the context of Oak Clinic, this was evidently not the case; at the point of initiation greater staff cover would have ideally been enrolled to mitigate the issues of any departures once the intervention was active.

CMGC staff and job specialism

Each of the CMGC clinics offered a different dynamic of staff. For most clinics, the CMGC was staffed by those with specific knowledge of existing glaucoma care pathways. However, Birch Clinic provided a blend of glaucoma specialists working alongside general ophthalmic nurses, with the latter predominantly providing patient-facing care.

In terms of the CMGC procedures, these staff were all comfortable performing procedures such as IOP measurement and eye drop installation. That said, based on a less extensive understanding of glaucoma, certain facets of consultations left them feeling illegitimate to perform their roles: “If the patient asked me anything technical, I would have struggled” BH303; “Glaucoma is not a part of my domain” BH302. These staff felt that they had insufficient glaucoma-specialist knowledge and should it be required from them, they would feel inadequate to give it. This theme aligns closely with the NPT subcomponents of legitimation (coherence) and skill set workability (collective action). The clinicians, while legitimately placed in some respects based on their broader ophthalmic knowledge and technical capabilities, felt that in practice discussing the condition in depth with patients was not consistent with their natural skill set.

When this is compared to Cedar Clinic or Maple Clinic where the patient-facing CMGC staff were glaucoma-specific ophthalmic nurses or orthoptists, these themes appeared less of an issue. Staff who had greater tacit knowledge of the treated condition, e.g. consultants, glaucoma nurses, as well as the practical skills to perform key clinical procedures, reported more comfort with the CMGC: “That’s part of my standard care…that wasn’t any different than I do normally” CH302; “The clinical stuff, the clinical side of actually doing it is fine” MH306.

Beyond this, those staff with more formal knowledge of glaucoma were also observed to have a more instinctive understanding of the benefits of the CMGC, offering perspectives on the future use of the intervention as well as the usefulness of the data: “If I knew on day one who was going to work on drops and who wasn’t that would be so beneficial” OH301; “If the patient responds to the dose given…and that dose is confirmed as being effective on intraocular pressure at the four hour mark…that is a premise on which you can build the whole of the rest of your care” CH301. Tying to the NPT construct of coherence, those staff with less tacit knowledge of the glaucoma care pathways and the condition itself, appeared less able to perform sense-making work. This extended to differentiating between the CMGC and standard care, as well as ultimately internalising the intervention within their day-to-day duties: “I mean this was for a given period of time, so I was quite happy to see those patients ad hoc” BH302. By the end of the study period, those staff with less glaucoma expertise did not feel their involvement with the intervention would be maintained. As such, while performing the duties within the confines of the research was acceptable, issues of internalisation, skill set workability and legitimation all had potential to negatively influence long-term staff response to the CMGC.

Pre-existing clinical structures

One of the more commonly reported issues was a perceived rigidity in the CMGC’s arrangement of appointments. This included prescribed four-hour follow-up appointments for IOP re-measurement following the instillation of an eye drop, as well as four-week follow-up appointments to monitor patient efficacy and adherence with instilling drops themselves. For each clinic, these requirements prompted varying levels of operational reorganisation so that the CMGC pathway could be integrated. By example, Maple Clinic hosted initial CMGC appointments on an ad hoc basis during administrative sessions on a weekday morning when clinicians were free. However, four hours’ later these same clinicians were then hosting an afternoon clinic with limited capacity for any CMGC appointments or IOP re-measurement: “The trouble came when we were, after the four hours…trying to identify somebody to do the measurements” MH306. Likewise, Oak Clinic encountered similar issues with the two designated CMGC clinicians available during a morning clinic but either off site or in surgical theatre during the afternoon session:

“Patient arrived at 14:25 for return appointment at 14:30. OH301 was in theatre with a trauma patient…OH301 arrived at 14:55 based on complications in theatre, apologising to patient for the delay”

Clinical Consultation 181001; Oak Clinic

While Oak Clinic and Maple Clinic were able to host CMGC appointments, this often required staff fitting them in alongside other clinical commitments, or delays to appointments based on lack of staff availability. However, within Birch Clinic and Cedar Clinic, where the same staff were available from morning to the afternoon, this problem was much less noticeable. This highlights the importance of NPT subcomponent contextual integration for the CMGC, with structural variations in timetabling and staffing providing examples where the CMGC operated relatively smoothly alongside those where it was significantly more problematic.

Recent debates over contextual integration within NPT, and the plasticity and elasticity of an intervention and setting, suggest that flexibility within both ease an implementation’s progress [5]. Regarding the CMGC, some participants reported that both the prescribed pathway, as well as restrictions within their respective health boards, made future implementations difficult. The clinics encountering issues with four-hour follow-up appointments also reported issues relating to the prescribed four-week follow-up appointments:

“Most places are working at least six weeks ahead. So I know that my clinics for the next six weeks are already booked. So therefore, if we identify somebody today who needs to go onto treatment, when do I bring them back?”

MH306

Administrative procedures in each site were reported to restrict the ease of embedding the CMGC within standard care in the long term. This quoted clinician from Maple Clinic highlighted an inelasticity within the appointments system, based on existing processes guiding appointments towards six rather than four-week follow-ups. Other sites also reiterated this point: “That’s the same across Wales, you’ll find that most patients that you want to see will appear back at six weekly intervals, but if you put in for seven weeks you won’t see them for months” OH301. Limitations within an intervention’s intended setting would ideally have been identified during earlier stages of design and planning, suggesting that stakeholders with more detailed oversight of such administrative issues may have been overlooked during this process. This inclusive co-production of an intervention’s components with staff localised in each setting links back to the NPT subcomponents of enrolment and initiation. While every effort was made to identify organisational and technical issues in advance of implementation, the issues of follow-up periods was not highlighted at this stage meaning there was a lost opportunity for greater plasticity within the CMGC.

Beyond this, though, these findings in relation to pre-existing clinical contexts, and their perceived elasticity, are pivotally important to how an intervention may be perceived by staff. When staff felt empowered to enact organisational changes, it was noticeable that their local setting was perceived to have much greater elasticity: “It’s a capacity variance requirement rather than a huge amount of additional activity” CH301; “Yes, the system’s a bit clunky…it’s getting better as time goes on, is what I think I would say, it’s not as bad as it was” BH301. These were both more senior members of staff for whom the systems and setting around them did not appear to be significant barriers. This was a notable contrast from those staff who perceived the setting and, consequently the CMGC itself, as being inflexible: “I think it is far too interventionist with a complete blanket of people too soon” MH307. In terms of NPT, this has implications for how plasticity and elasticity are understood in relation to contextually integrating interventions into healthcare settings.

Study limitations

Recruitment to the qualitative staff interviews, while achieving the study targets, would ideally have been increased to discuss emergent issues with a wider pool of participants. The issues experienced with contextual integration may have proved more problematic if a greater range of local contexts were introduced. That said, the variability highlighted within the sampled four clinics did provide insight into many of the broad issues that may be uncovered within any given local context. In turn, this enabled iterations of the CMGC that may prove more workable in other glaucoma clinics.

Regarding the use of NPT as a framework to understand the implementation of the CMGC, it was clear that most core constructs and subcomponents were applicable to the study data. For example, the relevance of certain subcomponents for the CMGC, such as contextual integration, are clearly demonstrable in the outlined findings and have contributed to a remodelling of the intervention. However, many of the subcomponents, particularly skill set workability and legitimation appear to overlap leading to potential confusion in data analysis, despite them emerging at different points in the implementation timeline [17]. Additionally, issues relating to staffing of the CMGC and its uptake, while loosely pointing towards themes of leadership and resource constraints, may have been better explained by alternative implementation frameworks where more overt solutions are offered. These could include the use of integrated knowledge translation approaches where more staff are likely to have been closely involved prior to initiation of the intervention [36], as well as the use of techniques such as change champions [37].