For healthcare professionals only

Primary Care

Early psychosocial interventions for low back pain in primary care

BMJ 2005; 331 doi: (Published 07 July 2005) Cite this as: BMJ 2005;331:88
  1. Chris J Main, professor of clinical and occupational rehabilitation1 (cjmain{at}
  1. 1 University of Manchester, Calderbank Research Unit, Manchester M20 3JQ

    The paper by Jellema and colleagues is the first to attempt systematically to incorporate psychosocial management in the treatment of back pain by general practitioners in a primary care setting.1 The success of pain management programmes in “reversing” the impact of chronic incapacity in a proportion of patients with apparently intractable chronic pain has led to an interest in secondary prevention,2 3 as the sort of factors that seemed to respond to cognitive behaviour therapy had also been shown to be early risk factors for chronicity.4 A specific focus by general practitioners on psychosocial obstacles to recovery has recently been recommended,5 but no controlled evaluation of this type of intervention has previously been carried out. Jellema and colleagues, in their well designed and controlled study, seem to have shown that such an approach, compared with “usual treatment,” is no more effective. This seems puzzling at first sight and certainly merits some consideration.

    It seems unlikely that the focus of the 20 minute intervention was inappropriate. The study seems to be well designed, with careful assessment and evaluation of outcomes based on an intention to treat analysis. Why then was no superiority shown for this new approach over usual treatment? Several possible explanations exist.

    Firstly, the purpose of the intervention seems to be appropriate and is clearly stated, but as the authors point out they could not within the constraints of the study determine whether the doctors had in fact carried out the intervention as required. Thus, crucially, we do not know whether the psychosocial risk factors were successfully identified or competently managed, as we have no data on the impact of the minimal intervention strategy on the patients' beliefs, emotional responses, or pain behaviour. Secondly, there is no reason to postulate a difference between the intervention and usual treatment for patients who did not show psychosocial “risk factors.” In terms of natural history, we know that a significant proportion of patients will improve in any event. The study was not powered to detect subgroup differences, and the use of median splits in the subgroup analyses, although pragmatic, may not reflect the level of severity at which the psychological factors become obstacles to recovery. Thirdly, we know that self reported disability is multiply determined. Although it might be argued that if no differences can be shown in outcome measures at a year further investigations are unwarranted, the data clearly indicate that most of the therapeutic chance has occurred by six or certainly 12 weeks, so presumably other factors thereafter contribute to longer term outcome.

    The findings of the study are, however, important. The study found no evidence of an effect, but it did not show that the psychosocial intervention was ineffective. In considering longer term outcomes from psychosocial interventions, and in the prevention of unnecessary disability, research now needs to focus on the nature and immediate effect of the interventions (including perhaps the contextual determinants of adherence), thereby providing a platform for prospective controlled studies into more clearly targeted interventions with a clearer focus on behavioural change and its determinants.


    • Competing interests None declared.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    View Abstract