Article Text


Differences in the management of shoulder pain between primary and secondary care in Europe: time for a consensus
  1. D G Kassimos1,
  2. G Panayi2
  1. 1Rheumatology Department, Dudley Group of Hospitals, Dudley, West Midlands, UK
  2. 2Rheumatology Department, Division of Medicine, GKT School of Medicine, Guy’s Hospital, London UK
  1. Correspondence to:
    Dr D G Kassimos
  1. D A W M van der Windt

Statistics from

We read with great interest the articles of Van der Windt and Bouter1 and Hay et al.2 There is no doubt that the study of Hay et al is well designed and has practical implications. They showed that physiotherapy or subacromial joint injection are equally effective for shoulder pain. This is new evidence as, so far, there has been little evidence to support the effectiveness of any common intervention for shoulder pain.3 However, the definition of “shoulder pain” illustrates the practical problem in diagnosis that general practitioners and hospital specialists face in routine clinical practice. We agree that the positive outcome for physiotherapy may reflect the increased contact time between physiotherapist and patient or the better understanding of the anatomical problem by the physiotherapist. The differences in management and in the effectiveness of physiotherapy by the British compared with the Dutch may also represent a cultural difference between the expectations and beliefs of patients in the two countries. It is likely that physiotherapy departments could be overloaded with referrals from primary care doctors if they are always the first next step in the pathway of managing shoulder problems. Hay et al did not carry out a cost-benefit analysis of the different treatments for shoulder pain (that is, injection v physiotherapy). A course of physiotherapy would cost around £200–320 (€284–454), whereas an injection would cost around £60 (€85).

There is a lack of consensus in the UK about the exact role of the general practitioner in the treatment of shoulder disease.4 A survey among rheumatologists and physiotherapists practising in the Southeast Thames Region of London (47 rheumatologists and 9 physiotherapists) showed that the management of adhesive capsulitis in secondary care varied widely. Nearly all the rheumatologists (98%) used intra-articular steroid injection, but the time, site, and frequency of injections were variable, with 72% believing that early injections are a priority. One of five rheumatologists (22%) believed that physiotherapy and mobilisation offered no benefit. Only a small number of rheumatologists (14%) believed physiotherapy to be the only means of treatment.5 Interestingly, 90% of physiotherapists working in secondary care wanted to see patients with a frozen shoulder as early as possible before or immediately after steroid injections. However their waiting time varied considerably (range of 3 days–3 months).

Similarly, across Europe treatment of shoulder pain varies considerably between primary and secondary care.6,7 Therefore we propose that European consensus guidelines on the management of the painful shoulder should be developed.8,9 This consensus may be weakened by the lack of an adequate evidence base. In addition, we would suggest a third and fourth arm to future studies—steroid injection with physiotherapy and a no intervention control group.


Author’s reply

Kassimos and Panayi deal with several important issues about the management of shoulder pain in their comments on the article by Hay et al1 and our leader.2 We agree that differences in the effect of treatment between the Netherlands and England may, at least partly, reflect differences in the organisation of care, as well as differences in expectations and beliefs between the two countries. We are also aware of the lack of consensus among general practitioners, physiotherapists, and rheumatologists about the management of shoulder pain. Between primary and secondary care, especially, the differences are large. This can partly be explained by the fact that the primary care doctor is confronted with an entirely different spectrum of disease than the specialist.3 Many patients in primary care present with signs and symptoms that are troublesome and cause worry, but are relatively benign and have a favourable prognosis. Patients referred to secondary care have been preselected by the nature and severity of symptoms, and have another prognosis, resulting in different treatment requirements.

The lack of consensus among health professionals, indeed, emphasises the need for multidisciplinary guidelines for the management of shoulder pain. Regardless of the quality of the evidence base, multidisciplinary guidelines will facilitate communication among health professionals and may optimise diagnosis and treatment of patients with shoulder pain. We suggest that the AGREE Instrument (Appraisal of Guidelines for Research and Evaluation)4 is used in the development of any guideline for shoulder pain. This instrument includes recommendations for the description of the scope and purpose of a guideline, stakeholder involvement, rigour of development, clarity and presentation, applicability, and editorial independence.

The development of a European guideline for shoulder pain will be quite an undertaking. The authors of the EULAR guideline for the management of knee osteoarthritis indicated that there was often discordance between research evidence and the opinion of experts.5 In this international guideline, variation across countries in healthcare delivery systems, access to health professionals, ways of funding, and attitudes towards the disease, all contributed to this discordance. The use of a Delphi system permitted consensus agreement on difficult issues, but still the applicability in individual countries may be limited. In the case of shoulder pain, it may be wise to start out with the development of national (multidisciplinary) guidelines. As yet, only a few European countries or professional organisations have developed such guidelines.

Finally, regarding the closing point by Kassimos and Panayi, we agree that there is a need for additional research comparing physiotherapy or corticosteroid injections with a no treatment control. It might be difficult or undesirable to carry out such a trial in patients with severe pain and limitations in daily activities, but controlled trials will certainly help to establish the effectiveness and cost effectiveness of physiotherapy and injections in patients with mild to moderate shoulder pain. Future trials may also evaluate the effectiveness of combined treatment (injections plus physiotherapy).


View Abstract


  • Dr D G Kassimos is on study leave from the Ministry of Defence of Greece.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.