Back to presentation overview

Title: Patient treatment preference influences outcome of low back pain: a randomised controlled trial of a community-based programme of exercise, education and cognitive behavioural therapy.

Authors and affiliation: Gareth T Jones1, Ruth E Johnson2, Nicola J Wiles3, Carol Chaddock4, Richard G Potter4, Chris Roberts2, Deborah PM Symmons2, Paul J Watson5, David J Torgerson6 and Gary J Macfarlane1

1 Aberdeen Pain Research Collaboration, Epidemiology Group, Department Public Health, University of Aberdeen, UK.
2 Division of Epidemiology and Health Sciences, The University of Manchester, UK.
3 Academic Unit of Psychiatry, Department of Community Based Medicine, University of Bristol, UK.
4 Eastern Cheshire Chronic Pain Management Service, Eastern Cheshire Primary Care Trust, Macclesfield, UK.  5 University Department of Anaesthesia, Critical Care and Pain Management, University of Leicester, UK.
6 York Trials Unit, Department of Health Sciences, University of York, UK.

Email address: gareth.jones@abdn.ac.uk

Introduction: Each year, in the UK, 7% of the adult population consult their family physician with low back pain (LBP) and recent evidence suggests that half of patients will still experience pain and disability 3 months later.  LBP is multifactorial in aetiology and includes both physical and psychosocial elements.

Objective: Firstly, to determine whether a treatment programme for persons with sub-acute LBP, based on exercise, education and cognitive behavioural therapy, would result in a reduction in pain and disability over 12 months, compared to usual care supplemented with educational material.  And secondly, to determine whether a priori patient preference for type of treatment would influence outcome.

Methods: Patients, aged 18-65yrs, in Northwest England, consulting their family physician with LBP between Jan 2002 and July 2003.  Patients were followed up at 3 months post-consultation.  At this point, persons reporting pain of >20mm on a 100mm visual analogue scale (VAS) and disability of >5 on the Roland and Morris Disability Questionnaire (RMDQ) were randomised into one of two groups:

• Control Group: usual care, supplemented with written and audio educational material.

• Intervention Group: a 6-week community-based programme of exercise and education, delivered using a cognitive behavioural therapy approach, and the same educational material as the control group.

Participants were sent a questionnaire 12-months post-intervention in which pain and disability were assessed.  The effect of the intervention was evaluated using analysis of covariance, adjusting for age, sex, LBP history and baseline pain, disability and psychological distress.  The trial had 90% power to detect a clinically significant difference of 12mm in pain score and 3 units in the RMDQ.

Results: Of 2068 patients invited to participate, 1152 completed a 3-month post-consultation questionnaire.  448 (39%) reported disabling LBP, of whom 280 underwent a pre-trial assessment and 234 were randomised, 116 to the intervention.  Follow-up was high (84%) and both groups experienced a reduction in pain and disability over the 12-month follow-up.  However, compared to the control group, persons who received the intervention experienced no significant additional benefit in terms of reduced pain (-3.6mm; 95%CI: -8.5 to 1.2) and disability (-0.6; -1.6 to 0.4).  A significant interaction was observed between patient preference and treatment effect.  Prior to randomisation, 134 participants stated a preference for one particular treatment (114 for the intervention, 20 for the control).  Those who expressed a preference for, and received, the intervention experienced a clinically important reduction in pain (-11.7mm; -26.2 to 2.7).  Whereas, those who received the intervention, but had had a prior preference for the control, experienced an increase in pain at 12-month follow-up (+15.0mm; 13.6 to 43.5).  A similar finding was observed for disability.

Conclusion: Our results add to accumulating evidence that current interventions for LBP produce, at best, only a moderate additional benefit over usual care.  Secondly, these findings provide evidence that, in seeking to optimise treatment efficacy, the area of patient preference warrants further study.