One of the most common arguments for the GP-led model of primary care for musculoskeletal pain problems centers on the importance of the GP in identifying serious pathology or so-called 'red flag' indicators of possible serious pathology 12 and ensuring urgent referral for those cases. Patients and GPs may be concerned about changing the model of care for musculoskeletal problems given the argument that the doctor is best placed to identify serious causes of musculoskeletal pain such as tumor, fracture, or infection by paying close attention to these red flags - clinical signs that supposedly indicate serious pathology requiring further diagnostic investigations or immediate medical attention. Fears about missed pathology have led some to question whether other health professionals, such as physiotherapists, have sufficient knowledge of diagnostic strategies 1314. Understandably, many patients are concerned that their musculoskeletal problem may signify a serious or progressive disease that, if treated early, can be cured. Research, however, suggests that serious disease is rarely the case in common musculoskeletal presentations. For example, in the most common musculoskeletal presentation, low back pain, the frequency of diagnoses of serious pathology is very low in patients presenting in primary care. An inception cohort of 1,172 consecutive patients receiving primary care for acute low back pain in Sydney, Australia, demonstrated a very low prevalence of serious pathology, namely 11 cases (0.9%), eight of which were vertebral fracture 15. However, most patients in the Australian series (80.4%) had at least one red flag, indicating that, when used in isolation, they have little diagnostic value in the primary-care setting 16. Evidence suggests that concentration on differential diagnosis and red flags may even divert the GP from evidence-based practice and contribute to unnecessary investigations, over-medicalization, and increased disability and costs 17. There is evidence that some relatively uncommon musculoskeletal conditions (notably rheumatoid arthritis, gout, and polymyalgia) benefit from early diagnosis and treatment, but there is also evidence of substantial variation in the quality of early GP diagnosis and treatment (for example, 1819) and this variation may be related to lack of evidence about diagnostic utility of early symptoms and signs. It seems reasonable to ask for evidence about the effectiveness of musculoskeletal therapists in referring patients who may have such early conditions before assuming that doctors do it better and that all musculoskeletal patients must therefore be seen first by a GP. Some red flags (such as weight loss) are common between musculoskeletal conditions and some are condition- or site-specific, highlighting the need for primary-care professionals to have adequate training in differential diagnosis and in spotting unusual presentations of patients. There is no evidence that GPs are better than other well-trained health-care professionals at spotting these rare cases. Rather, the evidence to date suggests that there is no difference in the accuracy of diagnoses reached by GPs and other professionals for musculoskeletal disorders 20. This contrasts with GP management of conditions such as angina or diabetes, in which toolboxes of diagnostic and practical management skills have been acquired throughout training and in which GP interventions clearly make a difference.

A second argument focuses on the GP's role in the care of patients with multiple chronic conditions or multimorbidity in primary care. Multimorbidity is common in the population (58%) and in people with back pain 21; indeed, most consultations in primary care involve people with multimorbidity (78%) 22. These patients are likely to be more complex to assess and treat and are likely to proceed to poorer clinical outcomes over time. For example, the combination of chronic musculoskeletal pain and depression is associated with clinical outcomes that are worse than those of either condition alone 23. Some argue that musculoskeletal practitioners such as physiotherapists or chiropractors do not have the breadth of knowledge across common multimorbidities to identify and manage these patients well. It is certainly true that the patient requiring medical management of a range of chronic conditions such as diabetes or coronary heart disease, both of which occur more frequently in persons with chronic musculoskeletal pain, expects and requires the attention of a medical practitioner. However, it is also true that, ironically for the patient who has multiple health problems and who places priority on their musculoskeletal problem, there is evidence that the GP will tend to place priority on the other health conditions rather than the musculoskeletal problem 24. Furthermore, many of the general principles of long-term management of chronic health problems, such as education, support for self-care, and enhancing the individual's functional ability and quality of life despite disability, represent central tenets of physiotherapy, for example, and apply outside the field of musculoskeletal pain as well 25.

For up to 100 years in many countries such as the UK, Denmark, Norway, and Sweden 26, the GP has had a pivotal role in the referral of patients to medical specialists and many other health-care professionals. The core values of general practice include comprehensiveness of care, a focus on the person with the disease and their psychosocial context, continuity of care, and the doctor-patient relationship over time, and these encourage the GP to take responsibility for the whole patient, irrespective of the specific health problem 27. It has been argued that most patients in such 'gate-keeping' systems of health care value having one point of initial contact with a health professional they know and trust when experiencing significant health concerns and that, if allowed to choose their primary care through direct access to specialists, patients often do not go to the right specialist, because they do not have the ability or confidence to select appropriate care 28. Although the primary reason for introducing this principle of referral, or the 'gate-keeping role', was the protection of the income of GPs 26, some believe that it has proved to be a sensible and important way of regulating and co-ordinating primary and secondary health care 26, ensuring the cost-effectiveness of health services such as the National Health Service (NHS) in the UK 29 and the cost-effective delivery of health services more broadly 30. The move to 'single issue' services such as those for diabetes, depression, or musculoskeletal conditions can be thought of as representing a 'cherry picking' approach to health care 27 that may not, in the longer term, lead to improvements in population health and risk diversification but rather risks increased diversification and fragmentation of primary care.

Concerns about the long waiting times for consultant appointments following GP referral and claims that many referrals by GPs to specialists were inappropriate or unnecessary fuelled the call for, and development of, GPs with special interests (GPSIs). The emergence of GPSIs in some countries such as the UK and Denmark offers the potential for more care to be provided closer to home 31 and for referrals to hospital consultants to be reduced. In 2004, there were approximately 1,300 GPSIs in the UK across a wide range of health conditions; by 2011, there were in the region of 3,000 to 4,500 GPSIs. Each of the 152 Primary Care Trusts in England, for example, has approximately 20 to 30 GPSIs across clinical fields such as dermatology and gynecology and, less commonly, rheumatology, pain, and musculoskeletal. There is great variation across the UK in terms of the GPSI role, job specification, qualifications, and governance arrangements. Any one Primary Care Trust may have only two or three GPSIs in musculoskeletal pain, and, although GPSIs clearly have the potential to enhance the primary-care management of patients with musculoskeletal conditions, their small numbers relative to patient demand means that GPSIs alone cannot provide a comprehensive solution in the UK. Similarly, in Denmark, the Society for Musculoskeletal Medicine lists fewer than 100 out of a total of around 3,500 GPs who are certified in musculoskeletal medicine.

One challenge for GPs is that the management of many musculoskeletal conditions in primary care is about symptoms and function, movement and rehabilitation, activity, and positive attitudes rather than the traditional medical model of diagnosis and medical treatment. Previous studies confirm that patients with common musculoskeletal conditions such as osteoarthritis or back pain report GPs 'not taking their complaint seriously' 33 and therefore patients are left with the message that 'there is nothing to be done'. It seems sensible, therefore, to consider whether professional groups who actively embrace evidence-based care of patients with these conditions and who are actively engaged in leading training and research in these fields should be 'keeping the gate' for people who seek care for musculoskeletal problems.

Current GP care for musculoskeletal conditions is variable, consisting mostly of medication, brief advice 1734, and relatively poor information about prognosis 35. A recent Australian survey showed that usual care for patients with back pain does not, on average, match care endorsed by international, evidence-based guidelines 17 and that GPs favor expensive management strategies, including medication and imaging, over simpler and universally recommended treatments. Pathways of care for patients presenting with musculoskeletal pain are often chaotic 36, and GPs manage patients themselves or refer them to any one or more of several providers and agencies, ranging from interface services, telephone triage services, physiotherapy services, complementary and alternative medicine (CAM) practitioners, podiatry services, and pain management services to traditional hospital orthopedic and rheumatology services. Furthermore, there is evidence that many GPs lack an understanding of what musculoskeletal professionals such as physiotherapists or chiropractors can offer their patients 37 or the value of treatments (such as exercise) offered by these professionals 38. Clinical guidelines for the management of common musculoskeletal problems such as back pain 5 and joint pain attributed to osteoarthritis 3940 recommend pharmacological treatments, physical treatments such as exercise, and (for those not responding well to these treatment options) considerations of psychological treatments or surgery or both. At present, patient care is fragmented, GPs mostly offer advice and medication, and onward referral to other professional groups determines access to additional treatments. Recent initiatives to widen medication-prescribing rights to allied health professionals, including physiotherapists 41, and evidence that such professionals can be up-skilled to deliver clinically effective and psychologically informed interventions using principles of cognitive-behavioral therapy 4243 highlight the possibilities, and benefits, of greater engagement of these professionals in the provision of best primary care. These developments may relieve a currently unnecessary burden on GPs who are asked simply to write a prescription for pain medication recommended by the physiotherapist.

In addition, there is evidence that early intervention in general, and early treatment by physiotherapists in particular, for common musculoskeletal problems such as low back pain can reduce the amount of time people are off sick and can help to prevent acute problems from becoming chronic 5444546. Although early intensive treatment is not always beneficial for recovery in musculoskeletal injuries 47, services that ensure timely access to care for musculoskeletal pain contribute importantly to patient experience and satisfaction 48. Furthermore, there is evidence that patients going directly to see musculoskeletal practitioners such as physiotherapists are not at risk of having their serious medical conditions overlooked 49 and that experienced physiotherapists have the same level of knowledge as orthopedic specialists 5051, demonstrating good clinical diagnostic accuracy 52 and management decisions 53. There is also evidence to suggest that models of care led by musculoskeletal professionals such as physiotherapists lead to fewer prescriptions and investigations, decreased need for expensive and invasive treatments 5455, and fewer consultations back in the health-care system 56.

The best evidence from clinical trials indicates that primary-care treatments can achieve modest but definite improvements for patients with back pain. Encouraging people to stay active and at work, helping patients adjust their beliefs and expectations to realistic but achievable goals, and offering simple analgesia and a range of interventions such as exercise, manual therapy, and acupuncture as well as support for rehabilitation to the workplace should result in less suffering, disability, and missed work. These core treatments are consistently recommended in national and international guidelines 557 and are those that musculoskeletal practitioners such as physiotherapists, chiropractors, and osteopaths are specifically equipped to deliver. Many high-quality trials of different interventions test treatments versus a control group of usual primary care initiated by the GP; in most of these trials, the treatments are shown to be superior to such usual primary care 58, although the size of the average differences between groups is generally small to moderate. Cost-effectiveness data from randomized clinical trials indicate that primary care for patients with low back pain is not cost-effective unless it also involves one or more added components such as exercise, spinal manipulation, or behavioral counseling 59.

Clearly, it is important that the overall effectiveness and potential challenges of a front-line service run by primary-care musculoskeletal specialists be evaluated critically and that appropriate training, education, and development of practitioners to provide such a service be supported and maintained. Ferguson and colleagues 60 have highlighted the need for ongoing education of physiotherapists in the systematic recording of red flags. The evidence for specific interventions offered by musculoskeletal practitioners such as exercise and manual therapy varies from trial to trial; although the results of these trials may be attributed, at least in part, to the heterogeneous populations of patients recruited 58, the varying effectiveness of practitioners and their interventions needs to be considered also. The performance of such practitioners needs to be critically reviewed; there is evidence that many physiotherapists in the US, for example, may not be delivering guideline-based care 61 and that their counterparts in the UK may be working, as do many GPs, within more of a traditional biomedical model rather than an active behavioral model of rehabilitation 34. However, our view is that the evidence to date provides good reason to suppose that a model in which front-line primary care is provided by a range of musculoskeletal professionals such as physiotherapists and chiropractors and in which GPs serve as a route of second-line referral will be just as safe and effective as the current GP-led model and provides sufficient reason to suppose that it could provide more appropriate, efficient, and effective care for most primary-care consulters with musculoskeletal problems. These suppositions, of course, need to be tested to estimate the costs and benefits of making a switch from the current model; the evidence to date supports the rationale for large-scale prospective evaluations of such service development and change, and pilot studies have shown that musculoskeletal practitioners such as chiropractors can, and are ready to, be included in national quality development systems 62.

Greater freedom of choice in addition to improved and faster access to musculoskeletal care are further advantages for patients offered an open service to musculoskeletal practitioners 1363. This is important because the expected increased burden of musculoskeletal pain over the next 50 years means that current models of care need to be re-evaluated in order to provide musculoskeletal services that meet rising patient demand. Musculoskeletal pain has not yet been a national health-care priority in most countries, but, given the aging population, the burden to society will continue to rise in the future 64.

Previous changes to musculoskeletal services have already devolved much care from hospital and specialist centers to primary care (for example, in the UK through the introduction of Clinical Assessment and Treatment (CAT) services 4 at the interface between primary and secondary care). Many of these CAT services employ health professionals such as physiotherapists to carry out initial patient triage and place the patient in the most appropriate pathway of care 6566. In Sweden, many orthopedic departments now use physiotherapists as front-line diagnosticians in triaging patients with osteoarthritis. In both countries, this has resulted in dramatic reductions in waiting lists for patients waiting to see rheumatologists 65 and surgeons 67 as well as good patient outcomes over time 66. Patient choice suggests that this could be achieved in primary care. In Denmark, at least one third of patients with back pain now choose to see a chiropractor as their entry into the health-care system 6; in the US, more than half of people who had suffered from back or neck pain during the past year had consulted an alternative health-care practitioner, most commonly a chiropractor or massage therapist, whereas only one third had been seen by a conventional provider 9.

There is growing evidence in favor of changing the GP-led model of care. Self-referral to physiotherapists is well established in countries such as Australia, New Zealand, Canada, most states of the US, The Netherlands, and Scotland and in some services in England 686970. The Netherlands has operated direct access to physiotherapists since 2006, and an evaluation shows that this is particularly popular with younger patients, those with higher levels of education, those with the most common musculoskeletal complaints of back and neck pain, and those with recurrent pain problems 13. In addition, data show increasing proportions of patients choosing to directly access physiotherapists, from 22% in the year of introduction of self-referral to 43% over the course of a 4-year period 11. Direct access and freedom of choice about first-contact care for musculoskeletal problems thus clearly satisfy a need among patients. Evaluations of self-referral to physiotherapy have shown high patient satisfaction and have shown that GPs and physiotherapists strongly support having physiotherapists work at the first point of contact 1371 for musculoskeletal problems. It also appears, from non-randomized studies, that self-referral to physiotherapy can be cost-effective 55. Data on self-referral to physiotherapy in Scotland indicated that the average cost of an episode of care was £66.31 compared with £88.99 for a GP referral, suggesting an estimated cost benefit to NHS Scotland of £2 million per year 55. Recent self-referral pilots across six Primary Care Trusts in England highlighted a reduction in the number of associated NHS costs, particularly for investigations and prescribing, and 75% of patients who self-referred did not require a prescription for medicines. Self-referral to physiotherapy did not lead to an increase in demand for services and led to reduced work absence among patients 72. Evaluations of the introduction of self-referral to physiotherapy in The Netherlands showed that self-referring patients were treated in fewer treatment sessions (average of 2.3 fewer treatment sessions) and that overall there was no increase in the number of patients visiting a physiotherapist in comparison with the year before the introduction of self-referral 13. In regard to chiropractors, patients have traditionally self-referred to chiropractors, and recent evidence suggests that chiropractors in the UK already view themselves as primary-contact practitioners within the musculoskeletal field 73. Patient satisfaction for chiropractic treatment is high 74, and there is evidence that back pain patients treated by chiropractors incur lower costs (due mostly to less advanced imaging) than patients treated by GPs 75. There is, however, also evidence suggesting that the characteristics of patients seeking care may not be comparable to those of patients seeking care from GPs 65471.

In many countries, health-care professionals such as physiotherapists, chiropractors, osteopaths, and exercise therapists are in important positions to provide support for active self-management and positive treatment options - especially interventions related to exercise and prevention of future episodes - for patients with musculoskeletal conditions. Furthermore, there is emerging evidence that prevention and treatment of musculoskeletal problems in the aging population can be tied to prevention and treatment of other public health problems through the promotion of an active lifestyle and targeted exercise 76, an area in which primary-care performance is suboptimal 77.

Alternative models of care led by other health professionals are already well established in private practice and in many health services, including those in Australia, Denmark, and The Netherlands and some in the UK. Such services provide the opportunity for increased professional responsibility and challenge for musculoskeletal practitioners as they make their own decisions autonomously and in direct partnership with their patients. These professionals are interested and well educated in the diagnosis and management of musculoskeletal conditions and their care is associated with better clinical outcomes 45 and greater patient satisfaction 67478 and cost-effectiveness 557579 in comparison with GP care alone 59. If patients first see these musculoskeletal practitioners, the majority are unlikely to require the input of GPs or secondary-care specialists 45 and those who do may be more likely to benefit from those consultations.

It is important to note again the lack of randomized controlled trials to assess the effectiveness of these new models of care. There are clear parallels with other services, however. Examples include oral health and dentistry and eye health and optometry, which provide well-established models of first-contact care for patients and from which patients with suspected serious or other pathology are then referred to their GP. Dentists are even licensed to perform surgery and have limited prescription rights. The potential benefits for musculoskeletal practitioners such as physiotherapists and chiropractors are considerable, but further evidence, preferably in the form of controlled clinical trial evidence about clinical and cost outcomes, is needed.

Changing the care pathways for patients with common musculoskeletal problems is expected to be met with resistance at the organizational level but is likely to be a relief to many GPs 80. Direct access to musculoskeletal specialists may reduce the workload of GPs. For example, in the year of introduction of direct access to physiotherapy in The Netherlands, more than one fifth of all patients seen by physical therapists came via direct access and these were not a new group of patients 13 but were those with recurrent musculoskeletal problems who normally would have consulted their GP.

Traditionally, GPs receive little training in common musculoskeletal problems in undergraduate medical school, during medical internship, and in post-graduate education 81 and often have limited knowledge about the suite of non-pharmacological treatments available to patients. Surveys and interviews indicate a lack of confidence in examining and providing treatment to patients with back pain, and many GPs feel ill equipped, either relying on pharmacological management or subsequently referring patients to doctors with special qualifications or to physiotherapists, chiropractors, or osteopaths 61780. This limited knowledge base contrasts starkly with that of musculoskeletal professionals such as physiotherapists, chiropractors, and osteopaths. Research and academic developments in the field of musculoskeletal pain are led by a range of health professionals, from physiotherapists and chiropractors to rheumatologists and psychologists, and musculoskeletal pain is a relatively neglected academic area for general practice despite its substantial impact on the workload of GPs. This means that much of the knowledge being generated about the assessment and management of musculoskeletal pain is seen as core business by those within physiotherapy and other musculoskeletal professional groups but rather less by the professional group (GPs) that most often provides front-line care for these patients.