Discussion
A review of referrals to the Neck and Back Outpatient Clinic at Oslo University Hospital shows that many referrals lack information. Based on the data, just over half of the patients had been clinically examined, while more than 90 % of the referrals included an MRI report. Patients were more likely to be assessed by the specialist health service if the referral included their medical history, findings from a clinical examination, and a status of employed.
We found that the information was incomplete in many referrals, which is consistent with a report from NAO Norway (2) . The proportion of referrals with information on clinical examination in our study is somewhat higher than in a study of patients referred to chiropractors (12) . In a survey among GPs, 80 % stated that they examined patients with neck and back problems (13) . The discrepancy may be due to a lack of documentation of the examination in the referral, overreporting, or regional differences. Another possible explanation is that primary care clinicians who examine patients are less likely to refer the patient to the neck and back clinic.
NAO Norway's report shows that 40 % of GPs believe they could treat more patients with musculoskeletal pain if they had more time. This is something the authorities should take seriously as musculoskeletal issues are the primary cause of morbidity and lost work capacity (14) . Recent research supports the idea that GPs can teach patients to better manage back pain, and can reduce the use of MRI and sick leave (15) . In addition, Abrahamsen et al. have shown that a conversational tool can help reduce symptoms in patients with medically unexplained symptoms, including back and neck pain (16) .
Clinical guidelines in most countries recommend that in cases of back pain, healthcare personnel should take a patient's medical history and perform a clinical examination, including a neurological assessment, whether nerve root involvement is suspected or not (5, 17, 18) . Our study suggests that most referrals contain adequate information about medical history, but often lack descriptions of clinical examination, medication use, previous treatment and employment status.
Current guidelines recommend limited use of MRI for neck and back pain, but consistent with previous studies, we found that a large proportion (91 %) of referred patients had had an MRI scan (19) . A Japanese study showed that only one in five patients with significant radiological stenosis had clinical symptoms (20) . MRI findings increase with age; over 30 % of 20-year-olds have disc changes compared to almost all 80-year-olds (21) . It is therefore concerning that referrals contain more information about imaging performed than thorough medical histories and clinical findings. This information is necessary for interpreting and conveying MRI results. Patients frequently arrive at our department in an anxious state about MRI findings that are, in fact, normal age-related changes. This has caused unnecessary worry for patients and can have a negative impact on the prognosis. Consequently, it is perhaps not surprising that two-thirds of patients believe MRI findings are more important than clinical examination findings, and that nearly half would opt for neck surgery based solely on MRI findings, even if they have few or no symptoms (22) .
In the prioritisation guide for hospital doctors, there is considerable scope for individual judgement regarding which patients should be granted the right to health care. Both the quality of the referral and differences in the interpretation of the referral text can lead to disagreement. Disagreement was greatest for adequate medical history (6 out of 52 referrals). Since the same person classified the medical history and decided whether the referral should be accepted, it is not surprising that this variable has strong predictive power. It is also uncertain whether these results can be generalised to other outpatient clinics. The aim is to ensure consistent assessment of all referrals, with as little subjectivity as possible. We will therefore review this process with a view to improving consistency in assessments and devising a guide that GPs can access on our website. We will provide information about this via the collaborative forum between primary care and the specialist health service.
In the analysis of factors that predicted whether a referral was accepted or rejected, several categories lacked information, which introduced uncertainty into the analyses. It may seem unreasonable that the use of strong pain medications (opioids) was associated with referral rejection, but the Neck and Back Outpatient Clinic does not have a tapering programme for such medications, and such referrals are instead forwarded to the pain clinic. Age should not impact on the assessment but we found that it does play a role in whether patients are accepted by the outpatient clinic. One possible reason is that the oldest age group may have previously had their condition assessed at our clinic, while for younger patients, we are more likely to accept referrals that do not meet the criteria.
A methodological challenge in the study is that the same person recorded the information and decided whether the referral should be rejected or accepted, which may have influenced the assessment in a way that differs from normal practice. Nevertheless, the rejection rate was about 30 %, which is consistent with the average at our department both before and after the study.
Conclusion
Conclusion
We concluded that 43 % of the referrals to a multidisciplinary back outpatient clinic lacked information about clinical examination, and that an MRI scan was often requested despite this. The referrals also frequently lacked details about medication use, previous treatment and employment status. Identifying factors that increase the likelihood of a referral being accepted could promote a more standardised referral process. This will facilitate the work of both the referring clinician and the specialist.