Results
GPs' constructions of 'routine' back pain work (interview findings preintervention)
There were no obvious differences in the views expressed by GPs who followed the recommendation of the subgrouping tool and those who did not. In the majority of cases GPs demonstrated a relatively high level of non-adherence to tool use. This finding is strongly borne out in the interviews with GPs who had not used the tool expressing similar views to those who had used the tool, citing the main themes (reported below) as the main reason for their behaviour.
The main focus of these first interviews was on GPs' current approach to back pain, and the issue of routinisation was central to their accounts. The GPs described non-specific low back pain (LBP) as a 'common' complaint that, when compared to the major chronic illnesses such as heart disease or diabetes, had a lower priority for them. GPs followed their own 'script' for all patients with LBP symptoms, allowing them to classify their complaint according to pain severity and follow a management plan which initially included information and exercise advice, followed by subsequent referral to a physiotherapist for those with persistent symptoms. Such a strategy could detract from exploring patients' unique account of their condition and experiences. Most GPs were confident in their ability to match their patients' symptoms with what they perceived to be the appropriate treatment pathway. By and large decision making seemed to be heavily 'scripted'.
GP13: 'Erm, generally I feel OK, in the sense that I am pretty clear about what I want to do with most of the patients. The vast majority, have probably got a self-limiting back problem, so chronic pains, and I have a good idea in terms of medication that I want to intervene with...'
GP14: 'I am confident that I know the right investigations to do and who to send them to for further advice, physios or orthopaedic and I am happy that I can pick out if there is a serious problem, but I am not specialised or I don't have any special interest in the back.'
Most GPs claimed that many patients with low back pain do not require a consultation with a GP, given that the problem is often reasonably self-limiting.
GP24: 'From my point of view I do feel I see a lot of people unnecessarily with simple lower back pain that doesn't necessarily require a GP appointment.'
GPs were reluctant to spend much time managing low back pain, reflected in the limited time they typically spent with patients during consultations, preferring to 'dispose' of patients [
27]. One strategy was to refer them to physiotherapy, occupational therapy or to other clinicians.
GP15: 'I think one of the difficulties I find is, if someone has a chronic back pain, I wish there were occupational physicians that we could say 'oh maybe you could go and see someone if it is affecting your work'.'
GPs also reassured patients that their symptoms were likely to be muscular and therefore self-limiting; a strategy used perhaps to minimise repeat consultations.
GP18: 'When they are in acute pain, when it's just happened, I think a lot of people are thinking, 'oh, I have broken my back', 'I've slipped a disc', or whatever and to be reassured at that point that it is more likely to be muscular and will settle, they are just glad to be reassured, but I always put in the proviso, but if it's not improving, or you get any red flags, to come back straight away.'
Others expressed conflict between their own and patients' expectations of treatment.
GP24: 'I think people are much, they much more want a sort of quick fix and they want a cure, rather than being prepared to take some responsibility, and follow the doctor's advice, and take exercise, lose weight, that sort of thing. So it's a bit frustrating because often you'll see people and they'll say 'Oh I've had back pain' and you'll see that they've DNA-ed (did not attend) the physio you know sometimes, or whatever, so probably for back pain I think that patients are less than committed often, and that's frustrating.'
GP30: '...people who have got chronic back pain want to have a diagnosis, and that's impossible to give most of them and they find it very difficult to accept that what you're trying to do is control and improve symptoms rather than cure them.'
GPs viewed patients' lack of understanding about the typical trajectory of back pain as a hindrance to patient recovery rather than an opportunity to learn about their experiences and how they relate to their perception of the problem and its secondary prevention.
GP7: 'I think the patients often perhaps don't appreciate the natural history [of back pain]... when we are just waiting for the natural history to evolve, isn't accepted very well, very favourably by the patients.'
Wider organisational and professional issues had a significant impact on back pain management, with GPs often reluctant to refer patients to physiotherapy services.
GP11: 'I am sure that physiotherapists are the most appropriate practitioners to treat back pain, so the only thing that stops me is use of resources... yes almost everybody with back pain will benefit from that, but we would run out of physiotherapy resources very quickly when we need it for other things.'
Likewise, the need for orthopaedic services to manage only the most complex low back pain cases resulted in a widespread reluctance by GPs to refer patients; referral in such circumstances could risk damaging professional credibility or even relations with colleagues in secondary care.
GP13: 'the orthopaedic people, who are just quite laid back about everything unless it is a barn door sort of [thing] standing up in front of you, they are just not really interested. So sometimes I think where I get a bit stuck with back pain is when the secondary care services and the registrars, or, they have got quite a high threshold for taking admissions... when you have the acute backs... in an ideal world, with some of those people you want an instant MRI scan to see what is going on and that's just not possible.'
After the first (observational) phase the GPs were introduced to the subgrouping for targeted treatment system and the six-item subgrouping tool through practice-based meetings with clinical opinion leaders and the study research team. Prior to offering the tool to all practices, it was piloted in one practice and found to be simple to use and acceptable, adding no more than 2 min to the consultation. The use of the tool was supported by follow up visits to the practices by informatics staff, to ensure the computer 'pop-up' screens were working and to provide regular feedback (every 2 months) via email and letter about number of patients with low back pain seen and for whom the tool had been used. Offers of additional guidance were made to practice managers and link GPs at each of the participating practices, in these feedback communications. The research team clearly discussed how they would monitor referral rates to physiotherapy and provide additional resource if and where needed. The tool provided the GPs with more systematic information about their patient's prognosis and gave a treatment recommendation based on each patient's subgroup classification, in the hope that patients at risk of poor outcome would receive earlier targeted treatment to help improve their outcome. Subsequent interviews were conducted approximately 12 months after the introduction of the new subgrouping for targeted treatment system in each GP practice.
Practical coherence (interview findings postintervention)
We use the term 'practical coherence' to denote the extent to which clinical routines and behaviours affected GPs' ability to use the subgrouping for targeted treatment system. By 'practical' we refer to those behaviours that enable a task to be managed efficiently, within the constraints of a busy clinic. The GPs perceived the subgrouping tool to have a lower priority relative to other tasks (for example, Quality and Outcomes Framework (QoF), national and local targets), and consequently claimed that time constraints and pressures intrinsic to a busy practice prohibited them from using the tool. Thus, the extent to which the tool could enhance existing clinical routines appeared to strongly influence tool use.
GP35: '...the problem is the time. It's definitely it's playing a major [part] on the issues you know. You're doing something a bit extra on top of what you normally do on a daily basis, which probably already a bit struggling with so timing probably is the most important thing you know.'
GP17: '...it's (the tool) not had very much visibility in the last year (laughing). When I'm talking about a 2 and a half million overspend, I'm afraid the IMPaCT Back study didn't have much of an impact. It might have a million pound public health resource but you know, but at my level it was I'm looking at overspend.'
The following respondent claimed that despite the simplicity of the decision tool, it was rarely used because the task of examining patients and discussing their problems took priority in the context of a busy working schedule.
GP35: 'You've got to examine, you've got to document everything you know, whether they've got any sort of red flags. These all take time... when you've got just about 10 minutes even the seconds is really important and probably that was something that I struggled [with] and if I couldn't really discuss anything in that particular time you know, it was just a timing issue but nothing else...'
A major finding from the interviews was that GPs felt the tool did not account for the complexity of decision making in low back pain. The following GP stressed the impact of patient demand on referral decisions, which is not taken into account by the subgrouping tool.
GP3: 'Whereas, with a patient it's more of, I don't know how to put it really, it's not like a bartering but it is a bit like that you know. They want physio, well sometimes they want physio, and you're thinking well you know, should they have it or not and they may only score one but if they really want physio, they going to get physio because we're here to meet the patient expectations and demand...'
The following GP claimed that medicine was an art, requiring awareness of the human elements of medical care, which the tool could not effectively capture.
GP3: 'It's difficult to say because you see, I'm one of these that thinks that there's an art to general practice and it's more a sort of conversation and a feeling between two people. Now, you can't put feelings into a questionnaire [reference to the tool] so what's right for one, is completely wrong for another and unless you had a questionnaire that was a thousand questions long, you're just not going to capture that, are you? ... It's not a tool that I would want to use within my consultation because it's not how I practice but it's not been obstructive to me.'
Others found that the tool conflicted with established patterns of care, encouraging what they saw as 'inappropriate' referrals to a physiotherapy service that was already struggling with waiting times.
GP10: 'The times I've used it, I've brought it out and said all right to the patient, '[X University] are doing a study let's look at this' you see. So you go through it and then it says 'referral to physio', and so that puts the idea into the patient's mind which you then can't go back on, and I have to say that the very few that I've done, were not people I would have immediately referred to physio, and our physio team can't cope at the best of times you know. They're about a month behind...'
Other GPs found it difficult to alter their routines to fit in with the decision tool, and often forgot or refused to use either the computer or paper-based tool during clinics. Even those GPs that asked for the paper-based tool appeared to struggle to remember to use it with patients.
GP11: 'My major difficulty was remembering to ask the questions because my practice is to enter my notes after the patient has left, so often I would then, you know three or four problems you put them all in, you give the back pain one and the (computer) template throws up and you think oh I've forgot to ask them.'
GPs expressed a general lack of familiarity with the subgrouping tool, claiming that they would have benefited from more frequent contact with the research team for feedback and guidance.
GP17: 'Certainly. I haven't accessed your tool for about a year to be honest, so my familiarity with it, which is part of the reason why last time we spoke, I spoke to someone, I said I would engage more but to be fair, it's probably a lack of familiarity at that point that just hindered the process.'
GP3: 'So that's not been an issue. It's not a time factor; it's remembering to do it if the computer's not prompting you.'
However, involvement in the research led some GPs to reflect on their current practice, which in itself appeared to have a positive effect on behaviour.
GP11: 'There were occasions when it made me rethink and I thought oh well, perhaps I should, because sometimes the process; none of us are consistent and sometimes you don't do the best job so, sometimes you've made a fairly quick decision and, you think you've got to move onto other things, other things are more urgent; more pressing so you haven't necessarily done the best job and the discipline of asking somebody those questions makes you rethink and you think actually perhaps I haven't done as good a job there as I would.'
Several GPs wanted more information about the physiotherapists' new role in the subgrouping for targeted treatment system.
GP13: 'But I wasn't I think initially when [X] may have presented, you know even before recruitment, I think they talked then about what they were looking at but I think it would have been useful as part of Y's and Z's [research team] presentation, to maybe have a physio there saying that this is you know, what we will offer as the intervention.'
GP13: 'Hum, yes in a roundabout way. So I know about the assessments that they [physiotherapists] do and also the psychological therapies that they're sort of putting in but not, I couldn't, if you said to me could you write an essay or a side of A4 of all the things that they do, no I couldn't...'
Relational coherence
Relational coherence refers to the impact of the subgrouping tool on interpersonal relationships between peers. For instance, the use of a new system for targeting and treating patients may affect referral decisions to certain services and directly impact on their workload, or the extent to which tool use is reinforced within an organisation through informal discussion, debate, and (dis)agreements between peers. Such activity may be a prerequisite for the introduction, wider acceptance and 'legitimation' of any new innovation in health care. A central issue affecting the acceptability of the subgrouping tool was the presence or absence of peer communication. Communication between colleagues served two aims: (1) raising awareness (often more important in larger organisations where dialogue between individuals may be less direct), and (2) individual or collective evaluation of a new way of working (integral to its acceptability). The GPs claimed that they generally did not discuss the decision tool or subgrouping for targeted treatment approach with colleagues.
GP11: 'The only time that we've really discussed that has been when you've been to see us.'
Interviewer: '...you ever got chance to have a real good chat with your colleagues as far as what you've been asked to do and how you felt about it?'
GP3: 'I mean we have so many things to discuss, but no, minimal I would say.'
GP10: 'I mean we perhaps talk for about 2 or 3 minutes after you've been to see us (laughing) but then it gets forgotten until the next time if I'm honest.'
Another GP was unclear about his role in the study, claiming that he would like to have received a formal 'contract' with the research team about his role and obligations, which would have improved communication with colleagues.
GP13: 'This is what you're signing up for. Erm, I would have been happy to do it. It wouldn't have been a problem but I would have been much clearer about my role and then I was getting newsletters or feedback from yourselves. I would have understood it was my role to disseminate the information and to discuss it at the, you know, the PLT [Practice learning time] or the lunchtime briefs that we have, and also to share the recruitment figures and to feedback to my partners where we were up to. I think that would have cascaded then.'
GPs discussed the tool together when concerns were raised about the impact that it had on their practice or on the referral rate to physiotherapy. The following GP claimed that discussion with peers resulted in a practice wide consensus that the tool was inappropriate as it recommended that too many patients be referred to physiotherapy.
GP9: 'Certainly in the earlier part of the study, I was referring anybody who came out with a score of 3 or greater but having sort of reviewed that as a peer group of doctors here and discussed whether we should be doing that, we actually came to the conclusion that there are a lot of unnecessary referrals being done purely because of the score 3 or more; ordinarily wouldn't need physiotherapy and probably would benefit relatively little from having it.'
GPs claimed that referral to physiotherapists was conditional on receiving regular feedback on patient progress and timely discharge.
GP13: 'I don't make assumptions about patients but I would [I think] feel more confident about referring patients who I think are going to be left with long-term sick really with an acute presentation so I have a lower threshold for referring patients to physiotherapy, as long as I knew that we were going to get the additional input. I think I'd be reluctant to refer patients with a low threshold who I felt were then going to be attached to physiotherapy, you know, for a long time.'
The danger of overwhelming physiotherapy services with patients presenting with underlying psychological health problems was mentioned by some GPs who felt that any decision instrument that encouraged referral of this nature raised concerns, and was less likely to receive support from GPs.
Interviewer: 'So what sort of clinician do you feel is the most appropriate to deal with patients that consult with low back pain that have the psychosocial issues?'
GP9: 'Traditionally it's been GPs because things like that tend to fall on GPs as sort of the gatekeeper to other services. I think there's no problem I would have with them seeing a physiotherapist as a first port of call but I think the physios would have overwhelming concerns about that because they really don't have the capacity to do that on the NHS.'
A similar view was raised by the following GPs, whose referral decisions were affected not only by the patient's physical complaint but also by the potential burden on physiotherapy services and consequent relations with physiotherapists [
10]. They obviously did not appear to have fully understood that additional physiotherapy resources were made available as part of the study, which perhaps raises questions about the communication process between the study team and the participating GPs. It is clear though that engagement in the new system required more than financial resources.
GP35: 'There was a constant struggle you know with the physio team because they were under a lot of pressure and you know, maybe as a GP you know, you were a bit under pressure even morally you know should I? Shouldn't I? You know, is it better or not, you know? What's the best basically to do? You know, this kind of struggle that you normally have and in sort of real life, you know sometimes the things that look really really nice on paper you know and in real life you know, sometimes it doesn't work unfortunately.'
GP10: '...I wasn't going to get my wrists slapped, because you do sometimes get your wrists slapped. There's one physio in particular that just bounces them. So you have to justify every referral really...we get messages saying what the wait is for physio and you know, please don't overload us and we're often being told that so I try and manage. I have to say, I try and manage the psychosocial side myself if I'm honest.'
Discussion
The task of integrating new ways of working in healthcare settings can be challenging [
28]. According to the NPT the implementation of a new approach is operationalised through four mechanisms: coherence (establishing meaning), cognitive participation (engagement of individuals); collective action (interaction with existing practices); and, reflexive monitoring (reflection and understanding of an intervention). The GPs in this study did not progress beyond the first stage of implementation, or coherence, the main focus of this paper. Coherence, or 'meaning', in practice may be difficult to achieve for several reasons. The work of healthcare professionals is often 'routinised' making it difficult to deviate from existing practices. In this study, low back pain was perceived by GPs as a common complaint, often considered to be a clinically 'uninteresting' condition, and deserving of less attention from clinicians in contrast to other health problems, which are perhaps more clinically challenging. GPs utilised well rehearsed management strategies for back pain which typically included the routine 'classification' of patients' symptoms into 'acute' or 'chronic' categories, leading to highly structured approaches; the first included predominantly advice about exercise and lifestyle, while the second involved referral to a physiotherapist for more intensive interventions. Neither typically led to a detailed exploration of the patients' experiences of their illness episode as means of guiding the decision-making task, even though in a different back pain study GPs claimed to personalise care [
29]. The implications are significant. For any strategy aimed at implementing a new way of working in health care, the importance of understanding clinical routines is clear. Although individual perspectives about the care of patients varied, it was evident that our GPs followed a 'script' that contributed to the 'standardised' approach taken. They appeared to follow the recommendations set out in the UK National Institute for Health and Clinical Excellence (NICE) Guideline for Low Back Pain (2009) in their daily interactions with patients, reinforcing the importance of activity, but providing only limited guidance about the long-term consequences of the condition [
21]. While asserting the need to address the psychosocial impact of back pain, the guidance offers limited practical recommendations. The consequence of this uncertainty may be important, perhaps reinforcing GPs' perception of back pain as a problem for which little can be done and for which patients should perhaps assume greater personal responsibility. GPs felt that their current approach to back pain appeared to work, as management often led to the effective 'disposal' of the patient [
27].
Despite a successful pilot at one practice, the GPs claimed that the subgrouping tool did not integrate well into their everyday management of patients. The organisational pressures affecting the care of patients, such as time constraints presented major obstacles to the use of the decision tool. The perceived motivation of patients to adhere to clinical advice has been shown to affect GPs' decisions to incorporate scientific evidence in practice, as reported previously in relation to smoking cessation [
5]; if patients fail to follow advice then GPs might be less likely to offer it routinely. In this study GPs claimed that clinical decisions frequently involved making subjective judgements about patients' likely motivation to act upon advice, where medical decision making was an 'art' often influenced by subjective considerations. In this regard the translation of evidence into practice is rarely straightforward, since it demands an awareness of patient-specific influences affecting outcomes. Other obstacles included GPs' apparent desire to avoid overloading physiotherapy services (even in the context of the service actively participating in the IMPaCT Back study and accessing additional support to do so), while adherence to the decision tool was perceived to have the opposite effect; increased referrals. Coherence in this context might therefore be defined in terms of 'role congruence' between existing healthcare routines and resource concerns and the extent to which they are 'disrupted' by an innovation [
30]. Clearly, a distinction exists between health professionals' willingness to participate in a research study and the adoption of new tools in routine clinical practice. Several GPs were unclear from the start if they were only required to recruit patients into the study via the tool or to also use the tool to improve their daily clinical practice. This raises important questions about recruitment of health professionals into implementation research, and the strategies required to maximise engagement; for instance, through discussion about the goals, benefits and potential limits of the study.
Peer communication also seemed to affect GPs' 'adoptive' behaviour. There is clearly some overlap between the themes reported here, though it is important to point out that there were different dimensions to the pressures prohibiting GPs from routinely using the tool. GPs perceptions about the economic consequences of their referrals to physiotherapy services were a factor affecting whether or not to adopt the tool. Occupational factors, including concerns about the displacement of labour to other groups of workers (in this case, physiotherapists) were another. The risk of undermining professional relationships with peers, for instance, as a consequence of over referral to physiotherapy, was a concern for some GPs. The same GPs incidentally challenged the suggestion that physiotherapists were the most appropriate professionals to manage patients with back pain, particularly those with more complex 'psychosocial' problems; claiming that in such cases patients would ideally require the attentions of a GP, though they expressed this view through a 'managerialist' discourse apparently concerned with the protection of physiotherapy services. This finding might be viewed differently, perhaps as an attempt to protect occupational work boundaries; GPs' reluctance to surrender control over patients could be a strategy aimed at defending their role and location in the status hierarchy. Thus, resistance to using the decision tool in practice may be a symptom of an underlying occupational desire to protect health care jurisdictions [
31]. A small number of GPs understood the enhanced role of the physiotherapists in the subgrouping for targeted treatment system and saw this as an opportunity to shift the care of back pain patients who exhibited significant obstacles to recovery to specially trained physiotherapists. The subgrouping for targeted treatment approach aimed to help GPs decide which patients were most likely to need onward referral. They were not always concerned with maintaining boundaries, but considered alleviation of their burden of work, especially with regard to potentially 'intractable' patients as a positive step.
Clinicians have historically been sceptical about the utilisation of formal, rationalistic measures to support the delivery of care, particularly if they perceived a threat to their clinical freedom. Pope [
32] claims, in relation to the routine adoption of evidence-based guidelines that 'By privileging technical knowledge that can be formulated and specified...EBM (evidence-based medicine) thus presents a significant threat to clinical judgement and ultimately control over medical work' [
32]. Resistance to healthcare interventions could be taken as a reflection of clinicians' desire to maintain control over their work [
30], with the intervention (which might include extended roles for other groups, such as physiotherapists) perceived as a direct threat to that autonomy. Equally, guidelines may enhance autonomy by presenting clinical decisions as scientific [
33]. Our GPs, however, rejected the standardised rules embodied by the subgrouping tool, questioning its ability to significantly empower them or enhance their practice in the management of back pain. Coherence in this regard could be viewed as the degree to which a healthcare intervention is perceived to be a threat to an occupation's influence over a sphere of work. Various mechanisms of professional closure may be deployed to aid this task and to safeguard control over services or labour.
Medicine's claim to autonomy, to set its own standards and control clinical performance without outside interference, is its central characteristic [
34]. Of course this principle can be compromised as primary care commissioners (such as Primary Care Trusts in England) are often able to exert a major influence over the context of clinical practice, such as through control over resources. Similarly, patient demand can affect clinical decision making and shift the balance of power away from doctors. Yet, for all these constraints, the cornerstone of the medical profession is its claim to clinical autonomy, and emphasising its importance lies at the heart of professional work despite the threats. Clinicians defend what may seem like irrational behaviour, in order to exercise 'discretionary power to cope with various patient, practice and workload pressures' [
35]. This position, far removed from the usual behavioural levers of coercion and economic incentive renders any change to their behaviour difficult to achieve [
35]. Consequently, the claim that discretionary power was necessary to manage the pressures of everyday clinical practice was implicitly made by our GPs in relation to the subgrouping tool. If the tool failed in its ability to at least assist the task of managing patients then GPs could refuse to use it, irrespective of its effectiveness. Our study shows that even when an innovation is developed by, and in collaboration with, those for whom it is intended, this does not guarantee its integration in practice [
19]. In the final analysis, it is difficult to conclude whether physician resistance, the subgrouping tool itself, or organisational factors had the greatest effect on the adoptive behaviour of clinicians. It would appear that a combination of all three had an important impact. However, underpinning these are the practical and interpersonal dimensions of routine general practice that may be the strongest predictors of behaviour change. In particular, the relevance and benefit of the new system ('coherence') to existing clinical work seemed to weigh heavily in the uptake of the tool.