Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Early radiography in acute lower back pain

Three Part Question

[In patients presenting with simple lower back pain] is [routine plain radiography of the lumbar spine better than directed radiography] at [providing reassurance to patients and therefore a more rapid recovery]

Clinical Scenario

A 32 year old male presents to the primary care emergency centre with a 2 day history of backache. There is no history of injury or illness and he has no "red-flag" symptoms to suggest a serious underlying cause. You reassure him and advice analgesia and mobilisation. He is very anxious and thinks an X-ray would reassure him that nothing is wrong and asks you to perform one. You disagree as you believe that clinically he has a low risk presentation and that radiography is not indicated. He eventually leaves, very unhappy with your care and stating that he will get his usual GP to refer him in for an X-ray anyway. Reluctantly, you reflect on the consultation and wonder if he migh be reassured by an X-ray and therefore might he be able to mobilise better and therefore get better faster.

Search Strategy

OVID Medline on the www. via Athens interface. 1966-March 2005.
[ or exp Back Pain/] and [exp RADIOGRAPHY/ or] limit to "prognosis (sensitivity) and (humans and english language and abstracts)

Search Outcome

Ovid Medline - 317 papers of which 2 were relevant to the clinical outcome.
An additional paper was found by examination of the references.

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Kerry S
94 GP practices in SE England. 659 patients included. 506 observed under routine care. 153 randomised to either immediate X-ray or observation. Followed up at 6 months and 1 year. Participants had Roland disability scale, SF-36 and Hospital Anxiety and Depression Scale assessments.PRCT with separate non-randomised observational armFollow up in 506 in observation arm427 completed initial assessment, 352 completed at 6 months, 317 competed at one year. GP records reviewed on 480 at one year.This is the final paper from an original Health Technology Assessment published in 2000. There is poor follow up of patients, particularly in the observational group. In the RCT group there was a difference at baseline due to chronicity of symptoms which may have influenced the result. No assessment of pathology found, or missed!
Follow up in 153 trial patients153 completed initial assessment, 126 completed at 6 months, 108 competed at one year. GP records reviewed on 140 at one year.
Baseline characteristicsNo difference in RCT group. In the observational group X-ray more likely in older population and for those with back pain more than 8 weeks
Outcome at 6 weeks in RCT groupNo difference in physical function, pain or disability. Better mental health (7 points in mental health score, 8 points on SF36 vitality score) in patients who had X-ray
Outcome at 6 weeks in observational groupPatients referred for X-ray had more pain and more disability though not if adjusted for chronicity. X-ray's patients had lower adjusted depression scores.
Outcome at one year in RCT group.Only difference was small benefit in mental health score (7 points) in X-ray group.
Outcome at one year in observational groupPatients referred for X-ray had more pain and more disability though not if adjusted for chronicity. X-ray's patients had lower adjusted depression scores.
Patient expectation, satisfaction, repeat consultation and referrals.No difference in RCT group. In observational group, those X-ray'd were more satisfied, more likely to reconsult and more likely to be referred for physio or top another health professional.
Kendrick D
421 patients with back pain presenting to the GP at 73 practices in Midlands, UK. Patients had to have had pain for more than 6 weeks and on the day of randomisation. No "red-flag" symptoms and aged 20-55. Randomised to either immediate radiography or later imaging at the discretion of the GP. Participants were assessed using the Roland score (the primary outcome measure), VAS, EuroQol, patient satisfaction, duration of pain, duration of sick leave, use of other services and sck leave. Assessments made at 3 and 9 months.PRCTPatient follow up394/421 completed the trial.These patients had medium term back pain. As most back pain episodes resolve spontaneously they are a select group.
3 month assessment - pain74% for X-ray vs. 65% for control p=0.04
3 month assessment - Roland disability score4 for X-ray vs 3 for control p=0.05 (control better)
3 month assessment - repeat visit53% for X-ray vs. 30% for control p<0.01
3 month assessment. selected factors showing no statistical differencetime off work, Median Euroqual score, Satisfaction, use of meds, physiotherapy, osteopathy, chiropractic, acupuncture.
9 month assessment - pain65% for X-ray vs. 57% for control p=NS
9 month assessment - Roland disability score3 for X-ray vs. 2 for control p=0.06
9 month assessment - selected factors showing no statistical differencevisited doctor, time off work, Median Euroqual score, Satisfaction, use of meds, physiotherapy, osteopathy, chiropractic, acupuncture.
9 month assessment - satisfactionbetter for X-ray group 21 vs. 19 points. (RR-2.69). p<0.01
Serious pathology foundNo life threatening lesions found in any patients.
Miller P
Same group as Kendrick paper. This is the economic analysis of the same study.PRCT withData collectionProspective collection of 8 direct (radiography, inpatient admission, outpatient attendance, gp visits, physical therapies, medications, special equipment purchase) and 5 indirect cost variables (practical help, extra expenses incurred, social security paymenyts, loss of earnings, loss of productivity for the employer). Valuation of radiography was defined by how much the patient would pay for it if they had to.Like many economic analyses this is difficult to do and to interpret. The original figures suggest little benefit to radiography but the hypothetical analyses (like a sensitivity analysis) show that the results could be significantly skewed based on some of the assumptions made in the paper.
Economic calculationNet economic impact = change in direct costs + change in indirect costs + value of reassurance from radiography - perceived value of risk of radiation
Clinical outcomesAs detailed in the Kendrick paper. Satisfaction did differ at 9 months with the X-ray group having greater satisfaction.
Cost-benefit analysisAssuming reassurance from radiography=30 and risk=43 then overall cost is 115 per patient
Cost-effectiveness analysisHeavily in favour of control group.
Boot strap analysis of cost effectiveness dataA hypothetical analysis of the data based on a higher value of radiography to the patient suggests that if satisfaction is considered at a higher level then X-ray may be of benefit to patient satisfaction.


This is a common problem in primary care and emergency medicine. The patient wants an X-ray, but you don't and you perceive that the evidence for it being useful in defining pathology is in your favour. These papers suggest that from a pathological point of view you are right. X-ray is unlikely to find significant pathology (without clinical indicators - the so called "red-flags"), the patient will probably have less pain and will get better quicker. However, and it is a big however, they may well be less satisfied with the experience. The choice is yours: Good medicine (don't X-ray) or popular medicine (X-ray). With respect to the three part question there is no data on patients with acute back pain. This is probably because of the overwhelming evidence that acute back pain has a very favourable prognosis. The patients in these trials had medium term (sub-acute) back pain and are therefore not representative.

Clinical Bottom Line

There is no published evidence to indicate if X-ray would have an effect in the acute setting. However, in the sub-acute setting there appears to a worsening of prognosis but an increase in satisfaction in those patients given X-rays. We therefore infer that there is little to be gained from routine X-ray in patients with acute back pain and no "red-flag" symptoms or signs.

Level of Evidence

Level 2 - Studies considered were neither 1 or 3.


  1. Kerry S, Hilton S, Dundas D, Rink E, Oakeshott P. Radiography for low back pain: a randomised controlled trial and observational study in primary care. Br J Gen Practice 2002;52:469-474
  2. Kendrick D, Fielding K, Bentley E, Kerslake R, Miller P, Pringle M. Radiography of the lumbar spine in primary care patients with low back pain: randomised controlled trial. BMJ 2001;322:400-405
  3. Miller P, Kendrick D, Bentley E, Fielding K. Cost-effectiveness of lumbar spine radiography in primary care patient with low back pain. Spine 2001;27:2291-2297