Professor A Dickson FRCR, Addenbrooks, Cambridge
The purpose of vignettes is to aid the advisory panels of the HTA programme in prioritising topics for possible commissioning. Vignettes are a short briefing on the requirement for research in a particular topic area. They are researched and written in a very short time (typically no more than 2-4 days) on the basis of information from a quick search of Medline, the National Research Register and the Cochrane Library, consultation with a small number of experts in the field and occasionally other information.
Research Question
What is the cost-effectiveness of routine referral for x-ray for patients presenting to GPs with low back pain?
Why is research required?
No systematic reviews or clinical trials have been published on x-ray referral as opposed to no referral for patients with low back pain. Expert opinion suggests that since 1995/1996 routine referrals for x-rays are no longer common practice.
Who are the patients?
Low back pain is common. In 2 separate UK surveys (published in 1992 and 1994) the one year adult population prevalence of low back pain lasting for more than 1 day was 38%. Peak prevalence was in the 45-59 year age group. The proportion of people who developed a new episode of low back pain in a year was 31%.
In 1991-1992, approximately 7.2% of adults in the UK consulted their GP at least once within 12 months because of low back pain. About two-thirds of these consultations were for new episodes (first episodes or recurrences) and one third for more long term problems.
The majority of identifiable causes of low back pain are mechanical or degenerative (arising from the muscles and ligaments, joints or discs). Other causes include inflammation including conditions such as ankylosing spondylitis, infections (such as bacterial osteomyelitis or tuberculous osteomyelitis), neoplasms (such as multiple myeloma or lymphoma) or bone disease (such as osteoporosis or osteomalacia). However, up to 85% of patients with low back pain cannot be given a definitive diagnosis because of poor associations between symptoms, signs, imaging results and pathological findings.
What is the technology?
Patients are referred for routine lumbar spine x-rays to: (1) investigate the possibility of a serious but rare diagnosis, such as malignancy, fracture or infection, (2) to exclude congenital disorders in patients aged under 25 years, (3) to explain possible mechanical factors to the patient or (4) to assist physical therapists in their management.
There is concern about the number of unnecessary x-rays. Three standard views of the lumbosacral spine involve about 120 times the radiation dose of a chest x-ray. The yield of positive findings is very low. It has been estimated that x-rays of the lumbar spine cause 19 radiation deaths each year in the UK.
Recent Royal College of General Practitioners guidelines have noted that acute back pain is usually due to conditions which cannot be diagnosed on plain x-ray. Normal plain x-rays may be falsely reassuring.
The management of simple backache includes: Drug therapy with paracetamol (paracetamol-weak opioid compounds or NSAIDs are second line treatment), advice on staying active and manipulative treatment for patients who need additional help with pain relief. If patients do not return to ordinary activities and work by 6 weeks they are referred for reactivation/rehabilitation.
Current and projected use
In 1994 it was estimated that about 15 to 20% of general practice attenders with back pain were referred for a lumbar spine x-ray with wide variation between practices. No studies assessing GP x-ray referral have been published since this date.
Expert opinion suggests that since 1995/1996 routine referrals for x-rays are no longer common practice. The change in practice came with the release of CSAG recommendations (published in 1995) and Royal College of General Practice and Radiologist's guideline recommendations on x-ray referral for low back pain.
In Oxford, for example, plain x-ray films have not been used in the initial investigation of patients with low back pain since early 1996. A back pain triage service co-ordinated by the senior physiotherapist is used. Patients are referred from GPs to this service. Magnetic Resonance Imaging (MRI) is used for initial investigation, but only about one third of patients receive this test. MRI is indicated for urgent symptoms and when patients have undergone an appropriate period of conservative care. X-rays are only performed in patients with suspected osteoarthritis or metatastic disease.
In Liverpool a triage system has also been adopted. Patients referred by GPs are allocated to physiotherapists or consultants by computer and are not routinely referred for x-ray.
Cost
The cost of plain radiography (one or more films) is between about £10 and £30. The estimated annual cost to the NHS of back pain was approximately £480 million in 1993. This cost includes that of 1.5 million x-rays, at a cost of up to £45 million. The lost production costs were approximately £3.8 million and DSS benefits £1.4 billion.
Quantity and quality of the research so far
No systematic reviews or clinical trials have been published on x-ray referral as opposed to no referral for patients with low back pain.
Royal College of Radiologist's evidence based guidelines (published in April 1998) on making the best use of a department of clinical radiology made recommendations on investigations required for low back pain. X-rays were not recommended routinely for chronic pain with no pointers to infection or neoplasm. This recommendation was based on band C evidence namely: Other evidence where the advice relies on expert opinion and has endorsement of respected authorities. X-rays were also not recommended routinely for acute back pain and this was also based on band C evidence.
Royal College of General Practitioners evidence based guidelines on acute low back pain (published in April 1998) recommended that x-rays of the lumbar spine should not be routinely performed. The report noted that acute back pain is usually due to conditions that cannot be diagnosed by plain film radiography and that pain correlates poorly with the severity of degenerative change found on radiology.
The exceptions to this recommendation were symptoms getting worse or not resolving, neurological signs or a history of trauma. This recommendation was based on consensus and not directly linked to evidence. These recommendations reflect the conclusions of the Clinical Standards Advisory Group (CSAG) report published in 1995.
A US preliminary study of MRI and plain radiography has been published (in 1997). This study was performed to demonstrate the feasibility of an RCT of MRI vs plain radiography for low back pain. Sixty two patients with low back pain were randomly selected to undergo either rapid MRI or plain radiography. After 3 months, there were no statistically significant differences between the 2 groups in Roland score. MRI provided more useful information to clinicians and resulted in greater patient reassurance. A large US multi-centred RCT has been funded to evaluate this research question.
Three RCTs funded by the HTA Programme are underway, namely:
• 93/17/11 RCT to determine whether the outcomes of patients with low back pain are influenced by GPs referral for plain radiology.
• 93/17/13 RCT of the effectiveness, cost-effectiveness and cost benefit of routine referral lumbar spine x-ray in patients with low back pain.
• 93/17/43 Does early imaging influence management and improve outcomes in patients with low back pain?
What is the potential effectiveness of the technology?
The best case scenario is that routine referral for x-ray provides information that aids the management of patients with low back pain. The worst case scenario is that resources are inappropriately used and deaths result through radiation exposure.
Information from:
Professor A Dixon, Professor of Radiology, Addenbrookes Hospital, Cambridge
Dr D Wilson, Consultant Radiologist, Nuffield Orthopaedic Centre, Oxford
Mr R Bartley, Superintendent Physiotherapist, Nuffield Orthopaedic Centre, Oxford
NCCHTA 1998 - Panel Name