Discussion
In this article we have presented the perspectives of GPs and patients on the acceptability, and anticipated barriers and facilitators to the use stratified care for the five most common musculoskeletal conditions, with both similarities and differences in views reported across identified themes, which were analysed in relation to the TDF.
A key finding within the first theme – acceptability of clinical decision-making guided by stratified care – related to how the approach is perceived to impact upon the theoretical domains of
skills,
knowledge and
professional role and identity, and how this in turn may inform the GP’s
decision processes. A salient concern of some GPs was that the approach could undermine the GP’s professional identity by leading to reduced clinical autonomy. This concern is consistent with literature on clinicians’ desire to maintain independence from external influence on their clinical work (e.g. [
24,
25]). As Powell and Davies ([
26], p.808) contend, ‘a key facet of professional identity for doctors is the desire to practise as autonomous individuals who retain personal control over how they define, sequence and evaluate their work’. In our data, the desire for clinical autonomy was commonly discussed in generalised terms, not specifically relating to the anticipated impact of stratified care, appearing therefore to indicate a general resistance by some GPs to the rise of what Flynn [
27] refers to as the ‘prioritisation of codified knowledge’ – that is, clinician knowledge becoming increasingly systematised, with tacit, experience-based knowledge gradually being undermined. Whilst stratified care is not intended to act to ration access to healthcare, it was clear that some GPs homogenise all such clinical tools in terms of their perceived negative effect on clinical autonomy. Concern over the reduced role of clinical experience was also reflected in the patient data, perhaps supporting Lupton’s [
28] argument that despite a contemporary trend that sees clinicians now more open to question, an enduring respect for the GP’s clinical judgement remains strong amongst many patients.
In contrast, the contributions of some GPs reflected a different manifestation in theoretical domains – they perceived their existing knowledge and skills in the assessment and management of musculoskeletal pain patients as potentially enhanced by stratified care, suggesting a positive impact of codified knowledge. Therefore, whilst much of the existing research literature focuses on clinicians striving for autonomy, this finding aligns with recent studies reporting both positive and negative clinician views towards clinical decision-aids [
29], including a study of the acceptability of stratified care for low back pain in Germany [
30]. Such findings could indicate a shift on the part of some clinicians towards becoming increasingly accustomed to, and accepting of, the use of clinical tools to support decision-making; a view also reflected by some patients who perceived stratified care as an acceptable addition to their usual treatment, showing recognition of its potential added value in informing treatment decisions.
Importantly, however, the reported views of those GPs who perceived stratified care as an acceptable addition to practice did not suggest the abandonment of autonomy, but autonomy being exercised alongside, or within the frame of, the use of stratified care; helping rather than replacing good clinical judgement. This lends support to Evetts’s [
31] proposition that ‘discretion’ may be a more relevant concept in medical decision-making, a concept also identified by Armstrong and Hilton [
29] in relation to use of a urological diagnostic tool in secondary care. In the present study, it appears that whilst seeing the value in stratified care, the GPs wish to maintain the discretion to use their own skill and judgement where they feel appropriate, and in doing so to maintain a balance between ‘codified’ knowledge and ‘tacit’ knowledge.
It appears, then, that for GPs to feel that stratified care might be a useful addition to practice, it must be seen by them to enhance their existing skills and knowledge, i.e. the codified, systematised knowledge/information from the tool and recommended matched treatments needs to provide additional clinical knowledge. However, they must not feel they are losing a salient part of their professional identity: the freedom to exercise discretion, to ‘over-rule’ the tool and choose different treatment options for a patient when felt necessary. How patients perceive stratified care to affect GP discretion and decision-making will clearly have implications for how, they too, respond to its use in practice. Whilst it is intended that, in the future RCT testing stratified care, GPs will follow the matched treatment recommendations, there will be several treatment options for each patient subgroup, allowing GPs and patients to reach a shared decision. It is this balance between direction and discretion that must be emphasised to both GPs and patients to promote the use of stratified care in the RCT.
Another key finding was the extent to which stratified care could foster either agreement and increased patient engagement in treatment decisions, or conversely, disagreement and conflict, negatively impacting upon the therapeutic relationship. In relation to this theme, the theoretical domain of professional role and identity was again identified as salient, and specifically how this related to the perceived goals of the consultation.
Concerns expressed about the use of the prognostic tool impeding upon GP-patient rapport-building appeared to reflect some GPs’ aversion to the artificiality of asking patients questions from a pre-defined list, which they saw as an unwelcome departure from the their usual way of talking to patients. Another identified barrier was the concern expressed by both GPs and patients that the matched treatment options could lead to restricted patient choice. For GPs, these concerns drew on a consumerist view of contemporary healthcare in which patients are more inclined to impose their own views and expect greater treatment choice; which may lead patients to reject the recommended matched treatments. This correlates with the findings of Lupton [
32] who reported Australian GPs’ views that patients have become ‘more assertive and knowledgeable, and more willing to challenge doctors’ ([
32], p 486). Whilst we found evidence of patients also expressing concerns about potential conflict, these do not fit with the consumerist notion of patients vying for greater choice, but instead portrayed GPs exercising their authority to make the final decision. This aligns more to Pilnick and Dingwall’s [
33] argument that the notion of patients as consumers is over-emphasised, and in fact GP-patient power asymmetries remain inherent in consultations. For some patients, this asymmetry was seen to be problematic if it meant not receiving desired treatments. Whilst views of the contemporary clinician-patient relationship differed between GPs and patients, the importance placed on constructive dialogue and discussion, and concern about patient choice being restricted, were common, perhaps signalling similar beliefs about the primary goals of the consultation. Any threat to this relationship was seen as having the potential to undermine key elements of both the GP’s and patient’s respective identities.
However, for those GPs and patients who reported positive beliefs about the consequences of using stratified care, it was felt this new way of working could facilitate discussion and patient engagement, with GP and patient participating in constructive dialogue that can be seen to align to the principles of a patient-centred approach based on mutual decisions around management [
34]. From this perspective stratified care was represented as being an acceptable addition to the consultation.
Thus, stratified care being seen as a useful addition to practice depends upon GPs and patients seeing it as enhancing their relationship, as opposed to fragmenting it either through GPs perceiving that patients will disagree with recommended treatments, or through patients’ concerns about restricted access to desired treatments.
Another key finding related to the reported predominance of a primarily biomedical approach in GP consultations. For some GPs, a focus first on biomedical priorities reflects the organisational set-up of general practice, and beyond that, generalised views about the role of medicine. Western medicine has always prioritised biomedical thinking – reflecting Mishler’s [
35] famous dichotomy between the ‘voice of medicine’ and the ‘voice of the lifeworld’. The former centres on diagnosis, which is commonly afforded a privileged status as being what Jutel and Nettleton ([
5], p.739) term ‘the foundation of medical authority’. This privileged status was reflected in the data; with evidence suggesting some GPs perceive their primary professional role being as a diagnostician; a role supported by some patients’ apparent desire for a ‘quick fix’ through scans and surgical intervention.
Another factor driving a biomedical focus was a sense of fear amongst some GPs, with regard to overlooking serious, albeit rare, pathologies. Our data suggest that such fear may be driven in part by contemporary changes in general practice, whereby GPs face greater scrutiny and accountability, which Nettleton et al. ([
25], p.333), drawing on the earlier work of Power [
36], refer to as the ‘contemporary thrust of audit and regulation’. More than this, the data also suggest a professional fear of lasting damage to the therapeutic relationship and a personal fear of reputational damage rooted in a lack of confidence in the assessment and management of patients with musculoskeletal conditions. Such multi-level fear may result in an over-preoccupation with missing serious pathologies such as cancer. This finding reflects literature on how personal fear of patient complaints and risk of litigation has led clinicians to adopt a more cautious approach in their way of practising [
25,
29].
In relation to the TDF, these findings were represented through a proposed relationship in which the domains of environmental context and resources, professional role and identity, beliefs about capabilities, and emotions: fear could all potentially influence GPs’ and patients’ perceived goals of the consultation as centring on biomedical priorities. Whilst stratified care is intended to supplement examination and diagnosis, strict adherence to a biomedical approach─ whether through factors related to the environmental context or individual determinants─ could present a barrier to its use if GPs are less receptive to an approach based on prognostic factors.
Despite the emphasis in the data on the role of the GP as a diagnostician, there was recognition that making a definitive diagnosis is often not achievable; this may be especially the case for many patients with musculoskeletal conditions, which can often be non-specific in nature [
37]. This may have led some GPs to see the added benefit of the stratified care approach in allowing them to more confidently assess the patient’s likely prognosis. For GPs and patients to perceive stratified care as an acceptable addition to usual care in future, it will be important to highlight how it can add to the existing biomedical, diagnostic approaches, as well as emphasise its potential to improve GPs’ confidence by providing prognostic information in the face of diagnostic uncertainty.
The practical issue of embedding clinical tools to support decision-making within the short timeframe of the GP consultation, coupled with the concern that tools will not fit ‘naturally’, are frequently cited in the literature [
10,
12,
13], and were again apparent in our data. As with the other identified themes, the relationship between the theoretical domains of
environmental context and resources, and
goals was identified as salient. For some GP participants, the impact of clinical tools in general can alter the structure of the consultation, impede upon their usual consulting style, and distract from the goal of assessing the patient; a view which resulted in
pessimism about the use of stratified care. Again, however, the variation observed in the data suggests an increasing acceptance of the use of clinical tools on the part of at least some GPs, who indicated that their common use in general practice has led to them appearing more naturalised, rather than intrusions on the consultation.
Despite these contrasting perceptions, all the GP participants stressed that above all the process of completing the prognostic tool and accessing the recommended matched treatment options must not be time-consuming, a view reflected in the patient data. These views can again be seen to reflect the contemporary nature of general practice ─ the increased pressure on GPs to incorporate ever more protocols and measures into a limited timeframe – 10 min in the UK context – in the face of greater scrutiny and accountability, means that the GP’s time is a precious commodity. However, participants emphasised that the key to the use of stratified care was that it be seen to be clinically useful, and if this is the case then GPs would make time for it regardless of the time constraints.
Another factor cited as crucial for stratified care to be seen as an acceptable and useful addition to practice was that recommended matched treatment options involving onward referral to secondary care must be aligned with available services. GPs highlighted the futility of treatment options matched to the patient’s prognostic risk that are unavailable in their locality, or involve very lengthy waiting times. GPs and patients seeing stratified care as a useful addition to practice in future will therefore be contingent both upon recognition of its potential for added value, as well as whether it can be incorporated seamlessly into the existing structure of the consultation and aligned with local service provision.
Strengths and limitations
A limitation of this study is that views were sought prior to the final development of the prognostic tool and matched treatments; discussions in focus groups/ interviews were therefore limited to the general principles of stratified care as opposed to the specifics of the intervention. However, this did allow for the findings to be incorporated into the refinement of the intervention, thus representing an advantage of the timing of the study. An additional limitation is that exploring views of the anticipated impact of the approach on clinical practice rather than participants’ actual experiences can be problematic, as this involves speculating about hypothetical barriers and facilitators to its use. Observational research to be carried out during the RCT will generate further insights into how stratified care is actually used and experienced in general practice.
Though attempts were made during focus groups/ interviews to encourage participants to discuss issues that they saw to be salient, the influence of the researchers’ contributions on participants’ discussions must be acknowledged; giving due consideration to their subjective views/beliefs. In particular, perhaps, the fact that the researchers in this study are closely involved with the development of the stratified care intervention, could have had the potential to influence the way in which participants’ views were elicited. However, the role of the researchers’ own subjectivities in the research process need not represent a limitation, but can be seen as an inevitable and integral part of that process [
38]. In any case, the researchers made a conscious effort where possible not to impose their own priorities on the data-collection, and the variation in views observed suggests that participants were not led into adopting a particular stance in line with that of the researchers.
A strength of the study is the parallel investigation of both GPs’ and patients’ views. This is uncommon in the behaviour change literature using the TDF, and these dual perspectives add robustness to the conclusions drawn. The critical approach to using the TDF – in which relationships were identified between theoretical domains within themes developed through an initial thematic analysis – also represents a strength of the study.
Acknowledgements
The authors would like to thank the study participants, the wider Stratified Primary Care for Musculoskeletal Pain programme grant team and Dr Tom Sanders for his contribution to early data-collection.