Challenges and visions for managing pain-related insomnia in primary care using the hybrid CBT approach: a small-scale qualitative interview study with GPs, nurses, and practice managers

Participating practices were geographically spread across Coventry and Warwickshire identified by the Local Comprehensive Research Network to ensure different demographics, practice settings, and experience with research participation. Practices had between 5000 and 8000 registered patients, with between two and six GPs. Practices’ demographic composition was widely varied (respective samples 2.1, 3.7 and 25.8% in each were of ethnic minority backgrounds). These practices also represented a wide spread of socio-economic backgrounds with respective scores of 8, 4 and 1 on the 1-10 Index of Multiple Deprivation (IMD), with lower scores indicating higher levels of deprivation [59, 60].

The IMD was used for two reasons; first, to gain a robust sense of relative deprivation of the populations served by our participating practices. Second, as it utilises seven domains known to impact deprivation levels (Income; Employment; Education; Skills and Training; Health and Disability; Crime; Barriers to Housing Services; Living Environment), the resultant data gives a more rounded understanding of potential complexities and barriers (or lack thereof) to a range of questions regarding access to healthcare and relative needs. Previous research has linked indicators of deprivation, such as lower social economic status and other associated factors with poorer sleep and pain outcomes, emphasising the reach of these issues in each of our practice’s served populations [61, 62].

The sample comprises nine primary care service providers (one GP, one nurse, and one practice manager from each participating practice). Three separate group interviews by practice were conducted. One practice manager opted to sit in as an observer of the discussion instead. Participants were purposively recruited via email through the linked feasibility trial. Inclusion of medical, caring and management perspectives was to gain a more thorough understanding of practices’ day-to-day operations.

All group interviews were conducted on-site in each practice, ranging from 30 to 60 min in duration. Participants were interviewed once post-trial, within 3 months of recruitment completion (May and June 2016, February 2017). A semi-structured interview schedule with 10 seed questions was used to anchor discussion, whilst allowing for organic interactions, i.e., interactions between the interviewer and participants, as well as among participants allow for spontaneity and natural tangents in conversation (Table 1). Digression from the schedule was also allowed, e.g., for the interviewer to follow up on a response with prompts or questions for clarification where appropriate or for the interviewees to discuss related topics. The seed questions were devised during the planning stage of the linked feasibility trial [56] through multiple consensus meetings between research team members to aid with overall process evaluation.

One interviewer conducted each of the group interviews (VN) and made field notes during and immediately after interviews. These did not form part of data analysis but helped keep track of key discussion points and areas for further exploration. Interviews were tape-recorded and transcribed verbatim for later analysis (VN, CM, VEJC). VN was a Research Fellow on the study’s team with extensive experience in mixed-method methods approaches to health service research and working in the National Health Service (NHS) as an allied healthcare professional. Data transcription was supported by the project co-ordinator (CM) and a psychology PhD researcher affiliated with the study (VEJC), who also observed one of the group interviews. These three researchers did not know the interviewees nor have established links with the practices prior to the study.

The study protocol was reviewed and granted favourable ethical opinion by the Solihull NHS Research Ethics Committee (REC NRES reference number: 14/WM/1053). Written informed consent was obtained from each participant prior to group interview commencement. Assurances of confidentiality and omissions of potentially identifiable data were made alongside the option for participants to decline any or all the data usage in subsequent analysis.

Three group interview transcripts formed the data corpus. Thematic analysis was employed, guided by six key processes as set out by Braun and Clarke (2006) [58]. Findings are presented with exemplar quotes that substantiate interpretation. Exemplar quotes were chosen via an iterative process between VEJC, CM, and NT during the latter stages of analysis. The process was then repeated with the wider study team during report production to certify that the chosen quotes best corroborated the thematic narrative. Any quotes which team members felt did not fully reflect or support the narrative were discussed and rectified. All quotes are accompanied by an anonymised practice identifier, such that the extent to which each practice is represented in the results is transparent.

Two researchers (CM and VEJC) performed the analysis independently using NVivo, as a means to further establish rigour and credibility, and allow for unanticipated independent insights [63, 64]. Initial codes and themes were discussed between them to resolve any differences in opinion. The agreed codes and themes were then discussed and re-organised with senior team members, who served as content-experts to contextualise generated themes in relation to previous research and clinical practice experience (DE, VN & NT). Themes were derived from the data within parameters discussed in the interview schedule. Themes that ‘survived’ this stage of analysis were then sent to all group interview participants for comments, as a further step to enhance credibility and authenticity of the analysis [63]. As an additional measure for credibility, follow-up meetings (conducted November 2018) with NT and VEJC were offered to all three practices as an opportunity to discuss and review initial findings. Two face-to-face meetings took place, each was approximately one-hour long. One practice could not commit to a follow-up meeting due to key personnel change and heavy workloads, however, did provide email confirmation that they were satisfied with initial analysis and direction.

Implementing an integrated service in primary care will incur some challenges, namely that the appropriate infrastructure, is not yet in place to deliver such treatment for this patient group [39, 71]. Infrastructure is manifold in describing the sum of physical, technical and organisational components needed in the context of healthcare delivery [72]. This involves complex, highly integrated areas such as staff structure and training, framework situation and physical location, and treatment pathways/access points [73] when considering developing existing or new services. This is inclusive of a lack of a fully empirically validated protocol specific to pain patients presenting with pain-related insomnia [70] as well as a shortage of trained therapists to ensure appropriate delivery and follow-up [48, 50]. A previous qualitative systematic review synthesised these issues in relation to differing beliefs and expectations of both patients and GPs [74]. Specifically, in the context of beliefs about pain, treatment expectations, trust, and patient education. Findings also advocate a shift away from specialist services to treat pain in intensive, short time periods to community services that would provide ongoing holistic patient assistance including managing comorbidities [75]. However, there are numerous common barriers that contribute to this evidence-practice gap, inclusive of primary care staff’s misconceptions of addressing and implementing recommended treatment guidelines for best practice [76].

Arguably, these issues around infrastructure and implementation of best practice are not new nor specific to the management of pain-related insomnia. Given that comorbidity in chronic conditions and simultaneous presentation tends to worsen prognoses, intervention and management at the primary care level may lessen some of these associated challenges [77]. This may address previously stated undesirable outcomes associated with current pharmacological provision as ‘go-to’ treatment options for both chronic pain and insomnia management [29‐38]. Moreover, there is some evidence demonstrating cost-effectiveness from successful collaborative care models in depression to promote biopsychosocial care models in primary care for chronic pain and insomnia [78]. Further integration and education of psychological care and PMPs in practices could also meet patients’ described needs and boost clinical adherence [74, 76].

Linked to infrastructure issues is information discrepancy surrounding how the pain-sleep relationship is conceptualised. This was reflected in differences between practices regarding how pain-related insomnia was described and treated whilst utilising secondary management models [79]. This may stem from clinical education around pain during medical training and post-qualification. Pain education is fragmented, falling short of what might be expected given the prevalence and burden of chronic pain across Europe [5]. Of 28 UK medical schools surveyed, only 11 had a dedicated pain medicine module, of which 4 offered it as a compulsory element [80]. None of which indicated addressing pain-related insomnia. However, evaluative research with the same dataset identified potential facilitators including reframing clinical management using a biopsychosocial framework [81]. Such was reflected in our participants’ considerations for improvements and ideal treatment scenarios. Since research has demonstrated the existence of a complex, bi-directional relationship which informs pain-related insomnia [17, 18, 20, 21, 25, 66, 82‐84] further education and information exposure should focus on its clinical application and treatment in line with recommendations of integrated, multidisciplinary service provisions [85].

Even with differences between our participants about where such a service should situate in a treatment framework, there was clear support regarding psychological input and allied health professionals to run and manage it. Previous evaluation of nurse administered CBT-I for persistent pain has shown promising results through using relevant professionals to conduct effective delivery and specialist training of other relevant staff [86] as have other self-management programs for other chronic conditions such as arthritis, diabetes and asthma [87]. This would certainly fit with pain management recommendations advocating multidisciplinary, biopsychosocial approaches to chronic pain and insomnia management [88, 89]. However, according to the most recent National Pain Audit [90] there is an overall lack of multidisciplinary provision, even though 64% of NHS England services describe themselves as such, only 40% meet criteria (inclusion of a psychologist, physiotherapist and physician). Moreover, the Midlands (our practices’ geographical situation) are relatively poorly served with the least access to such provisions. Guidance on future planning of pain management service strongly encourage psychologically informed, multidisciplinary treatment of chronic pain to be available at the primary care level [91]. Thus, mirroring views expressed by our participants and policy guidance could be the basis for productive discussion on how to further research such provision in primary care.

By boosting psychological support in primary care through multidisciplinary avenues, there is potential to make sure that service provision is not solely GP dependent [1, 92, 93] and meet this patient group’s treatment expectations [94]. Interestingly, the clinical and cost effectiveness of sleep management in pan management programmes has been noted as a key issue to be addressed in future research as per the 2020 NICE draft guidelines [95]. Another option may be to augment treatment pathways and redistribute resources in PMPs or IAPT to meet described needs and provide a more cohesive, integrated service [96]. However, participants highlighted concerns that in their current forms these services are not adequately resourced to deal with the complex needs of pain-related insomnia. Indeed, there is review evidence that implementing clinical networks can improve overall healthcare delivery for complex conditions, thus bring together some currently disparate treatment avenues and build rationale for more specialist interventions to slot into some currently available services [96]. However, these issues warrant more definitive discussion between researchers, practitioners, and CCGs in the design of future trials or service development [91].

Recently, there has been some considerable progress made in terms of prescribing and delivering psychological interventions for insomnia using digital platforms. CBT-I is a manualised psychological intervention with a rich evidence base, as such is well placed as a scalable solution to bridge gaps in feasible delivery and access for people presenting with sleep complaints [97]. After rigorous testing and evaluation “Somryst” (previously known as Shut-i) has been given Food and Drug Administration (FDA) approval in the USA as a prescribed treatment for insomnia [98]. Similarly, in a UK/NHS context “Sleepio” has accrued an extensive evidence base supporting its acceptability, feasibility, and efficacy across a range of populations presenting with sleep complaints [97, 99‐102]. Interestingly, there is a trial protocol to explore the efficacy of “Sleepio” for people with lower back pain which could prove fruitful and leverage the case made for addressing sleep complaints as a core component of PMPs [103]. Given the success of using digital platforms to disseminate CBT-I, it is possible that providing CBT-P and hybrid CBT through digital platforms could follow similar trajectories. The potential utility and cost-effectiveness of prescribing and delivering hybrid CBT using online platforms could help mitigate some of our participants’ mentioned barriers to treatment access. Central to these are related to an observed lack of trained therapists trained to understand and deal with the often-complex presentation of pain-related insomnia, as well as barriers to travel for older people, those on fixed incomes, frail, or with caring responsibilities.

As the purpose of these interviews was to inform the perceived feasibility of setting up a Hybrid CBT for pain and insomnia RCT [56], gaining insight from managerial, allied, and medical perspectives was deemed suitable. It is however acknowledged that there were possible demand characteristics on our clinical partners that might bias their views. Whilst there is no way to formally assess this, taking account of this possibility is important in the overall interpretation of these findings. Our sample size was small but in line with what was practically possible for a small-scale feasibility study [56]. The principle of ‘information power over data saturation’ has been applied here [104], whereby aim, specificity, theory, dialogue, and analysis are used to reflexively decide appropriate sample size [104, 105]. Information power suggests that the more information a given sample holds relevant to the study, then a lower number of participants is needed to develop new knowledge. Data saturation, in contrast is a process of constant comparison where every new observation is compared with previous analysis until no new information can be gained [106].

Differences in current practice between our practices may reflect differences in the operational characteristics as they were purposively sampled for the linked feasibility trial to consider geographical location, register size and population served, and experience in research participation. However, it is unclear exactly how these operational characteristics are shaping these practices’ differences in terms of chronic pain and insomnia treatments and would be an interesting focus for further research. Appropriate measures were taken to ensure rigour and transparency. Moreover, the interview schedule did not allow the examination of how GPs treat individual case presentations of chronic pain and/or pain-related insomnia. Thus, we could not delve into the nuances of current practice at the individual level. As a final point, because discussion was based on general recollection, there is always the possibility that these may not be wholly accurate representations of current practice.