Research Paper - (2007) Volume 15, Issue 1
Senior House Officer, Department of Medicine
Benjamin JL Kendrick BSc MRCS
Senior House Officer, Department of Orthopaedics
David M Williamson FRCS (TR & Orth)
Consultant, Department of Orthopaedics
Great Western Hospital, Swindon, UK
Received date: 5 July 2006; Accepted date: 8 August 2006
Introduction This study was designed to assess the impact of written guidelines on general practitioner (GP) referrals to an orthopaedic outpatient department. We looked at referrals of six common, yet diverse, orthopaedic conditions and planned a prospective audit to assess the number of referrals from GPs before and after the provision of referral guidelines. A secondary part of the study involved assessing the ‘usefulness’ of the guidelines.Methods All GP referral letters for new patients with the chosen conditions in a 13-week period were audited. Paper copies of referral guidelines were then distributed to all local GPs. After a period of four weeks for distribution, the process was repeated for a further 13 weeks. Each letter was analysed for its content of therapy or management already tried by the GP prior to referral, as suggestedin the guidelines. A feedback questionnaire was sent to GPs in the Swindon Primary Care Trust to assess the distribution and use of the guidelines.Results In total 471 referral letters were assessed, 304 before the provision of guidelines and 167 afterwards. The first 13-week period had 195 (64%) referrals that consisted of patients who either had not received the recommended management, or for whom this had not been mentioned in the referral letter. The second period had 103 (61%). There was no statistically significant difference (P = 0.49) and therefore little evidence that the implementation of guidelines had an effect on the management of patients prior to referral orthe consequent timing of seeking specialist opinion.Conclusion The provision of orthopaedic written referral guidelines to GPs does not affect the prereferral management of patients, or the overall number of referrals to the specialty. Further work with primary care trusts to develop a referral system that gives all the information required by the hospital specialist may help promote greater awareness of referral guidelines, and more strict criteriaon the timing of referral.
orthopaedics, referral guidelines
The effectiveness of communication between health-care professionals has long been a matter for dis-cussion in view of its great importance in correctly managing patients. With regard to the referral letter from a general practitioner (GP) to a hospital con-sultant, the information contained therein has a bearing on the allocation and prioritisation of the outpatient appointment. The letter should allow accurate assess-ment of the patient’s needs and clinical condition, to provide optimal management.
Without guidelines for referral, it is conceivable that referral letters may sometimes not provide the necessary information for such decisions to be made. This consequently could result in increased workload, unnecessary attendances and suboptimal treatment prior to the outpatient consultation.
With this in mind the orthopaedic consultants at the Great Western Hospital, Swindon, created a set of referral guidelines for the most common orthopaedic conditions relating to the upper limb, lower limb, back and children. The guidelines were designed to identify those patients who would benefit from a specialist opinion at the optimum point in their management. They were discussed and approved by a GP represen-tative of the Swindon Primary Care Trust (PCT). A selection of conditions were chosen for the study where guidance could have made an impact in advis-ing treatment or management in primary care prior to referral, thereby avoiding an outpatient consultation. Osteoarthritis cases were excluded from the study as it was felt that this common condition was encountered frequently and referred appropriately in the majority of instances. Table 1 shows the referral criteria for each complaint.
The recommendations contained within the guide-lines were based on the current evidence in the orthopaedic literature for managing the various con-ditions. Only treatment that was appropriate to pri-mary care in the NHS was advised. The aim was to give advice to GPs on the management of specific ortho-paedic conditions, in particular those conditions where orthopaedic surgical intervention has little to offer. It was also thought that this would have an impact on outpatient waiting lists, as patients would be dealt with more effectively by the GP, and unnecessary hospital appointments would be avoided. Thus we planned to distribute easy-to-follow referral guide-lines to the GPs in the catchment area of the Great Western Hospital, and to audit the difference in referral content and numbers before and after dissemination of the guidelines.
Copies of all referral letters from GPs to the ortho-paedic department were reviewed over a 13-week period. All GP referral letters for new patients, whether posted or sent electronically, were included. Letters referring patients for six specific orthopaedic com-plaints, namely anterior knee pain, back pain, carpal tunnel syndrome, in-toeing in children, sciatica and tennis elbow were selected. The content of each referral letter was assessed using the orthopaedic department guidelines (see Table 1), which were composed at the start of the study.
The referral guidelines, produced by the ortho-paedic consultants, were then distributed as hard copies to all GP practices in the hospital catchment area. After allowing four weeks to ensure distribution and awareness of the referral criteria, the referral letter audit process was repeated for a further 13-week period. GPs were unaware that a referral letter audit was being undertaken.
For each letter that did not meet the referral guide-line criteria, the reason was recorded and collated.
The number of letters for each condition was obtained for each 13-week period (see Table 2). These were divided into two groups, referral criteria positive and referral criteria negative letters. The differences were analysed using a Fisher exact test for statistical significance. The reasons for being deemed referral criteria negative were totalled for both periods (see Table 3), and the statistical significance assessed using a Chi-squared test.
Following completion of the study a questionnaire was emailed to 29 GPs (in Swindon PCT) after the audit period, to request feedback on the usefulness and relevance of the distributed guidelines. Questions related to the availability of the guidelines, the breadth of complaints covered, and the usefulness of manage-ment information. There was also the opportunity for free-text comments. The responses could be returned by either email or post.
This study evaluated 471 referral letters to one ortho-paedic department at a district general hospital during the study period. The results in Table 2 show the different number of criteria-positive and -negative referrals for each orthopaedic condition studied sep-arated for the two 13-week periods. The only condition showing a significant reduction in the percentage of criteria-negative referrals after the guidelines were distributed was tennis elbow (P = 0.038). However, if an adjustment is made for the number of cases this ceases to be significant. All other conditions showed no significant difference in the number of criteria-negative referrals after the guidelines were distributed. In the 13-week period before the guidelines were issued 64% of referral letters did not meet the criteria for referral, while the corresponding figure after the guidelines were distributed was 61% (P = 0.489).
The commonest reasons for a letter to be deemed as not meeting referral criteria included lack of physio-therapy treatment before referral, inadequate anal-gesia, or failure to administer a steroid injection. Table 3 shows the reasons and rates for criteria-negative referrals. As not all the evaluated conditions required each treatment modality, the absolute numbers were less important than the differences between pre- and post-guidelines (P = 0.26). Therefore there was no evidence to suggest that the implementation of guide-lines significantly affected the numbers of referrals meeting the criteria set out in these guidelines in terms of pre-hospital management or timing of referral.
The feedback questionnaire was completed and returned by 16 (55%) out of 29 GPs contacted. The information returned showed that 13 (81%) out of 16 actually received the guidelines. Ten (62.5%) out of 16 felt the guidelines were clear in their recommendations, one didn’t know, and five did not comment. GPs were asked to assess the guidelines for each individual condition, as being ‘no help’, ‘helpful’, or ‘very help-ful’. There were 80 assessments, of which 68 (85%) stated the guidelines were ‘helpful’, seven (9%) ‘very helpful’, and five (6%) ‘no help’. All five replies of ‘no help’ were for different conditions. Finally, on the question of whether the guidelines were used, three replied ‘often’, six replied ‘sometimes’ and seven GPs admitted to never using them when referring patients.
Orthopaedic outpatient clinics, like most other spe-cialties, have a lengthy waiting time for a new appoint-ment. At our trust in 2003/2004 24.2% of patients referred by GP letter waited longer than 13 weeks for their initial appointment.[1] Therefore a robust system of referral is needed to ensure those patients most in need are seen soonest, and those patients who can be managed in primary care do not take sought-after places. The appointments are allocated by the ortho-paedic consultants based on the content of the referral letter. Therefore the content of the letter is of para-mount importance. The orthopaedic referral guide-lines were designed with twin aims, firstly to ensure only relevant patients were referred and secondly to demonstrate the information required to enable ap-propriate appointment allocation. Our guidelines required a relevant history and examination in add-ition to basic management. It has been described in the literature how much the information contained within referral letters varies.[2] We found that the infor-mation included varied in its depth and relevance. The poorest section was current management, particularly which analgesic, if any, was being prescribed. It was not possible to distinguish whether a patient had not received a specific management option or whether the information was simply not included in the letter.
The GPs were not informed of the audit so as to avoid any confounding or ‘Hawthorne effect’.[9] We decided not to include letters in the study where it was specified that the referral was due to patient request, as we did not intend these guidelines to be a restriction on patient choice.
It has been described previously in other specialties that GP referral guidelines do not affect either the content of letters or the number of referrals.[3] How-ever, we could find no published data regarding orthopaedic referrals. Although orthopaedic surgery does not differ markedly from any other specialty it does have a wide range of simple, non-pharmacologi-cal, management options which can easily be utilised in primary care. Many orthopaedic conditions are self-limiting over a relatively short time span, with the mainstay of treatment being symptomatic control.
Because the orthopaedic department works closely with the physiotherapy department, it is susceptible to demands being made in response to elongated physio-therapy appointment waiting times. Consequently guidelines would have no impact on patients with complaints requiring physiotherapy, as GPs will tend to refer to the service with the shortest waiting time, even though physiotherapy may be the optimum treatment.
An unexplained effect of the guidelines was the total reduction in referrals following their distribution, but not a change in the proportion of letters that were criteria-negative. The second 13-week period occurred over the summer months when it is possible there was a seasonal reduction in referrals. We could not find any study in the literature to support this hypothesis.
The GP questionnaire showed that the distribution of guidelines via normal mail was flawed, with some GPs not receiving them. The questionnaire results also showed that although the guidelines were generally perceived as clear, available and helpful, they had little effect on referral patterns. This perhaps demonstrates that other factors come to bear on GP referrals such as patient demands, physiotherapy waiting times and lack of time or expertise for GPs to perform pro-cedures such as injections. It has also been recognised in experimental studies that passive approaches to disseminate information have limited effects.[4] Active learning, however, which is essentially interactive and participatory, has been shown to be more effective in producing change.[5]
The guidelines were composed based on the indi-vidual experiences of the consultant orthopaedic sur-geons. It is difficult to assess the quality of any guidelines, as there are no widely Accepted date: ‘gold standards’ with which to compare. Feedback from GPs did not include any comments on the content of the guidelines with regard to management requirements.
It has been proposed that GPs would benefit from feedback on cancer referrals, but this could be expanded to include all referrals.[6] Studies have also shown that GPs would welcome constructive feedback.[7] Com-ments from GPs regarding our guidelines included that referrals are driven by chronicity and severity of symptoms, and that if the guidelines were incorpor-ated into the online referral form they would be used.
Alternative methods of relieving the orthopaedic service include physiotherapy specialists and GP orthopaedic specialists to triage of referrals. These practitioners can also offer some treatments that GPs may not be able to provide, e.g. steroid injections.
For further study it would be useful to look at the number of new outpatient appointments in clinic whose attendance is deemed unnecessary when seen by a consultant. This could then be compared with the referral letter data. This may give more information as to whether treatment is not being instigated or is simply not documented. A previous study has shown that although a number of outpatient appointments are unnecessary, a significant proportion of those patients referred benefit from the appointment.[8] There-fore an idea of the information required by the patient could be obtained and made available to GPs, thus reducing referral numbers.
A criticism of the study is that it takes no account of the benefit patients may get from outpatient consul-tation, even if management is not altered. We appreci-ate that the study attempts to deal with referrals as ‘black or white’, but in the current climate in the NHS, with league table targets and increased patient choice, the emphasis is moving towards pure streamlined service provision.
The provision of written referral guidelines to GPs for common orthopaedic conditions does not affect the proportion of referrals meeting guideline criteria. A more effective form of guidance is required, possibly electronic or web based. This could include the opti-mum management for a specific condition being displayed on an electronic referral form, with treat-ments that have been tried prior to referral marked with tick boxes, and free text for further history, examination and any other information pertaining to prior management in the community. A comp-lementary strategy might include the provision of educational outreach visits, to actively discuss and encourage the use of the guidelines that have been created.
Thank you to Dr Gordon Taylor, Research and Devel-opment Support Unit, Wolfson Centre, Royal United Hospital, Bath, for his statistical help.
None.
