General Practice

Effect of UK national guidelines on services to treat patients with acute low back pain: follow up questionnaire survey

BMJ 1999; 318 doi: (Published 03 April 1999) Cite this as: BMJ 1999;318:919
  1. A G Barnett, statistician,
  2. M R Underwood, clinical scientist (M.Underwood{at},
  3. M R Vickers, senior scientist
  1. MRC Epidemiology and Medical Care Unit, Wolfson Institute of Preventive Medicine, St Bartholomew's and the Royal London School of Medicine and Dentistry, Queen Mary and Westfield College, London EC1M 6BQ
  1. Correspondence to: Dr Underwood
  • Accepted 30 October 1998

In 1994 the UK Clinical Standards Advisory Group recommended eight treatment standards for back pain.1 In 1995 availability of these services to general practitioners was generally poor.2 We conducted another survey two years later to assess change in availability.

Subjects, methods, and results>

For the 1995 survey we approached a random sample of 342 practices (out of 870) in the Medical Research Council's General Practice Research Framework (a UK-wide network of general practices that participate in research). Of the 307 practices that replied, 290 were still framework members in 1997 and were sent a questionnaire identical with that used in 1995 (study panel). Completing a questionnaire twice can affect responses (panel conditioning).3 To assess this effect we randomly selected a second sample from the members of the framework.

Both surveys asked the practices whether eight specified services (see table) had been routinely available to their patients during that financial year (1994-5 and 1996-7) and whether they would refer patients to them if they were available.

Availability of recommended back pain services reported by general practices in initial study (1994-5) and in follow up study (1996-7). Values are proportions (percentages) of practices giving valid responses to each question unless stated otherwise

View this table:

Responses obtained in 1997 from practices that had also replied in 1995 were considered equivalent to those approached for the first time in 1997 if the limits of the 95% confidence interval for the difference in proportion of positive replies were no greater than 10%. For equivalence with 80% power at the 95% confidence level,4 on the basis of the service “physical therapy before six weeks for patients off work,” 211 responses were needed from the second sample. To allow for non-response, we approached 232 practices.

Paired responses from the 1995 and 1997 surveys were compared by using McNemar's test. A logistic regression model including region, list size, and panel membership, as appropriate, was used to assess the effect of fundholding status on the availability of services.

The response rates were 87% (251/290) for the study panel and 85% (198/232) for the second sample. Members of the study panel were representative in terms of region and deprivation score, but larger practices were overrepresented. Response probability was unaffected by region, list size, panel membership, and practice deprivation score. The vast majority of general practices reported that they would use the recommended services if available (ranging from 88% (urgent referral to a physical therapist) to 99% (emergency referral for possible cauda equina compression)).

For three services the study panel reported a significantly better service in 1996-7 than in 1994-5 (table). Changes in practices' fundholding status did not explain this. For two of these improved services panel conditioning may have occurred (table). Only for the service “assessment by multidisciplinary team” were responses unaffected by panel conditioning, suggesting a genuinely improved service. Fundholders reported that three services were more available: urgent and routine referrals for physical therapy (P=0.01 and P<0.0001 respectively) and physical therapy before six weeks for patients off work (P<0.0001).


The national guidelines have not had a clear effect on the reported availability of back pain treatments. In contrast, national guidelines on asthma have changed general practitioners' behaviour.5 Commercial support for the dissemination and implementation of the asthma guidelines—plus the fact that prescribing recommended drugs is easier than developing services that require additional health authority resources—may explain this difference.

Possible explanations for differences between the study panel and the second sample include: (a) the behaviour of practices that complete two questionnaires on the same subject is not typical of all general practices and (b) completing the first questionnaire increased the practices' awareness of the recommendations, leading them to use existing services more efficiently or possibly to influence purchasing decisions.

We did not assess general practitioners' knowledge or application of the advisory group's guidelines. The practices expressed willingness to use the recommended services, and fundholders' ability to provide more physical therapy suggests that efforts to improve access to these services should be focused on the health authorities and primary care groups.


A poster based on this work was presented at a conference held by the World Organisation of National Colleges, Academies and Academic Associations of General Practitioners/Family Physicians in Dublin in June 1998.

Contributors: AGB carried out the study and performed the data analysis. MRU had the original idea for the study, developed the questionnaire, and wrote the first draft, subsequently incorporating the ideas of AGB and MRV. MRV supervised all stages of the work. All the authors are guarantors for the study.


  • Funding The Medical Research Council and the National Back Pain Association.

  • Competing interests None declared.


View Abstract