Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow E-letters: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when E-letters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (9)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Frankel, B. S.
Right arrow Articles by Jackson, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Frankel, B. S.
Right arrow Articles by Jackson, D.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Family Practice Vol. 16, No. 3, 216-222
© Oxford University Press 1999

Guidelines for low back pain: changes in GP management

Ben SM Frankel, Jennifer Klaber Moffetta, Steven Keen and David Jacksona

Nuffield Institute for Health, University of Leeds and
a Institute of Rehabilitation, University of Hull, UK.

Ben SM Frankel, 19 Cheltenham Terrace, Heaton, Newcastle upon Tyne NE65HR, UK.

Frankel BSM, Klaber Moffett J, Keen S and Jackson D. Guidelines for low back pain: changes in GP management. Family Practice 1999; 16: 216–222.

Received 11 August 1998; Revised 17 December 1998; Accepted 28 January 1999.


   Abstract
Top
 Abstract
Introduction
 Methods
 Results
 Discussion
 Conclusions
 Appendix
 References
 
Background. Management guidelines are aimed at reducing inappropriate practice and im-proving efficiency; however, the effectiveness of many guidelines has yet to be confirmed. This study targets GPs' management of back pain and its relationship with the recent management guidelines.

Objectives. We aimed to investigate changes in GP management of low back pain, low back pain episode duration and time before consultation, over a 5-year period.

Method. A case series report of 574 patient notes was collected from a large practice in North Yorkshire (January 1992–March 1997). In addition, 713 referral notes from 26 practices across North Yorkshire were collected. The patients were potential subjects for a larger randomized controlled trial evaluating the effectiveness of an exercise programme. Both groups of data were analysed by identifying trends over time and using log linear regression.

Results. Recommendation of activity or exercise was found to be a trend increasing over time. In contrast, recommendation to rest was found to be a trend decreasing over time. Activity or exercise was more frequently recommended to younger patients. There was evidence that individuals are seeking a GP consultation more quickly.

Conclusions. The gap between GP practice and the recent guidelines appears to be reducing. However, many variations in practice still exist. Evidence of decreasing time before consultations may indicate that the level of self-care by low back pain patients is decreasing. Although management guidelines may have some effect, there is some way to go before back pain management is optimized.

Keywords. GP intervention, low back pain, physical exercise, rest.


   Introduction
Top
 Abstract
 Introduction
Methods
 Results
 Discussion
 Conclusions
 Appendix
 References
 
Management guidelines are aimed at reducing inappropriate practice and improving efficiency;1 however, the effectiveness of many guidelines has yet to be demonstrated. In response to the increasing economic and social costs of back pain to the National Health Service (NHS) and society, evidence-based guidelines for the management of back pain have been published.2 They recommend early resumption of physical activity. This contrasts with previous advice for patients to rest. Research on the effectiveness of guidelines for low back pain has, to date, used postal surveys. It has concentrated on availability of services3 and GPs' self-reported management of the condition, comparing results against guideline best practice.4,5 The results from these surveys are conflicting.

For example, Summerton and Fitzgerald, after completing a postal survey of 104 GPs (69% response rate) in Yorkshire regarding their management of back pain, concluded that the majority of GPs were managing back pain in accordance with guidelines.4

In contrast, Little et al.,5 who also conducted a cross-sectional postal questionnaire of 236 GPs (70% response rate) in one South West health district, compared GPs' reported management of back pain with the 1994 guidelines.6 They reported that many GPs did not routinely perform recommended examinations or give advice about exercise and everyday activities. Hence, they concluded that the management of back pain by GPs did not match guidelines. In addition, Underwood et al.3 have demonstrated that the availability of services for GPs to treat back pain fell short of guideline recommendations.

However, there may be a difference between what GPs say they do and what actually happens. One way to assess this is to analyse medical records. Thus, the objectives of this study were to investigate:

  1. trends in prescribing for back pain;
  2. back pain episode duration;
  3. the duration of back pain before consulting the GP; and
  4. to apply these results to current back pain guidelines.


   Methods
Top
 Abstract
 Introduction
 Methods
Results
 Discussion
 Conclusions
 Appendix
 References
 
This case series report was carried out as an adjunct to a larger randomized controlled trial evaluating the effectiveness of an exercise programme for back pain patients.7 All patients used for this study were referred as potentially eligible for the randomized controlled York back pain trial,7 except those who either failed to meet the inclusion criteria7 or refused (or were unavailable) to take part in the study. Here the ‘potentially eligible’ were any patients with low back pain of less than 6 months duration for the present episode and aged 18–60 years. Two different approaches were used to collect data from the above patients. These consisted of:

  1. Group 1, patients with back pain from one large fully computerized practice of 12 GPs; and
  2. Group 2, patients with back pain referred from 26 practices in the York area.

Group 1
Data was collected from usual GP records on a computerized practice in North Yorkshire between January 1992 and March 1997, in order to include a baseline period before the guidelines were published in 19946 and 1996.2 Data were then grouped into 9-monthly intervals. The data collected from the usual GP records included: consultation date, intervention given, consulting GP, repeat intervention (if applicable), episode duration and the time before consultation. Also, inclusion of patient number, patient age, patient gender and postcode enabled stratification regarding episodes of low back pain. Information from the first consultation of each episode contained all the interventions given. A new episode of low back pain was defined by a gap of a period of 1 month in both consultation and intervention; and the time before consultation was defined as the length of time a patient had low back pain prior to consulting their GP.

Criteria for group 1 excluded episodes of low back pain: (i) in pregnant women; (ii) involving litigation; (iii) involving non-mechanical or complex pathology; and (iv) which appeared in the notes to be chronic prior to January 1992, i.e. had been present for longer than 6 months.

All recorded GP interventions were coded and grouped into ten categories:

  1. Advice;
  2. Activity/exercise;
  3. Physiotherapy;
  4. Rest;
  5. Non-steroidal anti-inflammatory drugs (NSAIDs);
  6. Analgesics;
  7. Stop work;
  8. Manipulation;
  9. Nil; and
  10. Other.

Two main analyses of the data were carried out using Excel 5.0 and Minitab. First, using histograms based on percentages, trends in the use of interventions over time were identified. Once a trend had been identified visually, a line of best fit was drawn giving the slope of the trend and its related simple correlation coefficient r2 value. The closer an r2 value was to 1.0 the greater the correlation between the trendline and points on the graph. An r2 value greater than 0.8 was considered significant.

Secondly, log linear regression was performed to make sure that trends were not due to changes in the underlying number of patients. This provided a change in deviance across the different variables, (for example, different interventions), over time. The greater the change in deviance the greater the significance. The cut-off for statistical significance was set at the 5% level.

Group 2
Consultations from a further 26 practices in North Yorkshire were used to validate and compare the data from Group 1. The Group 2 data were collected directly from the referral forms (see Appendix) under the more limited headings: consultation date, name, age, date of back pain onset and treatment plan. The criteria for Group 2 included any ‘potentially eligible’ patient referred to the York back pain trial (see beginning of method section). The study period for Group 2 was between 1994 and 1996. Data were analysed as for Group 1.


   Results
Top
 Abstract
 Introduction
 Methods
 Results
Discussion
 Conclusions
 Appendix
 References
 
Group 1
The practice served a population of over 20 000 patients with 12 GP partners. Five hundred and seventy-four patients (300 female and 274 male) fitted the inclusion criteria for Group 1. In total, Group 1 analysed 1657 consultations and prescriptions from 30 GPs (locums included). Figure 1Go shows the percentage of recorded interventions over time (for all episodes) given by GPs at the North Yorkshire practice. One of the partners (Dr A) had a special interest in backs, was trained in manipulation and accounted for 49.5% of recorded consultations. Census data for the practice area showed that 82% of the population were in socio-economic classes 2 and 3, but because of the limitations of assessing a patient's social class from their postcode, any further analysis was not possible.



View larger version (23K):
[in this window]
[in a new window]
  		FIGURE 1 Changes in the percentage of recorded interventions given over time, from all episodes

 
The most frequent methods of management by all GPs were physiotherapy, analgesics, NSAIDs and manipulation. Figure 1Go demonstrates the changes in prescriptions and advice over time. This histogram clearly shows that recommendations to rest decreased over time, whilst recommendations for activity increased. Statistical analysis confirmed that the following trends were significant:

  1. increases in the recommendation of activity/ exercise showed a log estimate of the size of the change of 2.057 with a standard error (SE) of 0.072; and
  2. decreases in the recommendation of rest showed a log estimate of the size of the change of 1.635 with an SE of 0.07.

Both these recommendations 2 and 4, also showed a statistically significant change of deviance of 38.76 with nine degrees of freedom (P < 0.01).

The histograms and trendlines demonstrated no clear consistent differences in intervention frequency between males and females (change of deviance = 6.934; d.f. = 9, P > 0.05). However, one other intervention trend was apparent, (though not statistically significant): the interventions of Advice and Activity/exercise increased, in percentage terms, the younger the patient (change of deviance = 7.234; d.f. = 9, P > 0.05).

Differences and changes in methods of managing back pain were compared across the GPs in two different ways. The first was by including the doctor who specialized in musculoskeletal problems and spinal manipulation and then excluding him (Dr A) from the analysis. Table 1Go shows the changes in the average percentage of interventions between categories of doctor with corresponding trendline results. As expected, Dr A reported significantly more interventions than the other GPs, except for NSAIDs and Stop work. The most obvious difference was observed with intervention 8, Manipulation, which was Dr A's preferred method of management.


View this table:
[in this window]
[in a new window]
  		TABLE 1 Showing changes in the average percentage of interventions between each category, with some trendline results
 
Table 1Go also compares the average percentage of interventions at different stages of a patient's back pain history, the stages being:

  1. the patient's first consultation of their first episode of back pain;
  2. the patient's first episode of back pain; and
  3. the patient's subsequent (non-first) episodes of back pain.

Increasing trends from (a) to (c) were noted in interventions Stop work, Manipulation and Others; whilst decreasing trends were noted in Advice, Rest, Nil and NSAID interventions. NSAIDs were less likely to be given during subsequent episodes of back pain compared with first episodes; the converse was true for analgesics. These findings were confirmed by the r2 and slope values given in Table 1Go.

Table 2Go shows percentages of the average duration of time before a consultation and the average episode duration (ED), and compares them between and within the separate categories. There were no statistically significant changes in ED over time in any of the categories. However, all the first episode categories had a greater percentage of short EDs (1–15 days) compared with their respective subsequent episodes category. Some expected differences (not in Table 2Go) were worth noting, though not statistically significant. Male patients tended to have shorter EDs than women. Dr A's patients were associated with shorter EDs than those of his colleagues.


View this table:
[in this window]
[in a new window]
  		TABLE 2 Showing average percentages of: the duration of time before consultation (TBC) and the episode duration (ED), allowing comparisons between categories and within each category to be made
 
Analysis of the time before consultation revealed one significant trend. The percentage of patients with a time before consultation of between 1 and 7 days increased, whilst the percentage of patients with a time before consultation of between 15 and 45 days decreased (change of deviance = 12.83; d.f. = 3, P = 0.05). In other words, patients attended their GP earlier. Table 2Go also shows: (i) a greater percentage of patients with a short (1–7 days) time before consultation in the subsequent episode group compared with the first episode group; and (ii) a greater percentage of patients with a short time before consultation in those visiting Dr A compared with the other GPs.

Group 2
The data for group 2 were derived from 26 different GP practices in the York area and contained 713 interventions. The data were collected and analysed in the same manner as for Group 1. The interventions covered a period of 31 months; however, the 9-monthly intervals used did not coincide precisely with the intervals used for Group 1. Interventions of Activity/exercise, Rest, NSAIDs and Analgesics were the four most common interventions (see Fig. 2Go). There were two significant intervention trends: (i) the recommendation of activity/ exercise increased; and (ii) the intervention of NSAIDs decreased. Other trends were not statistically significant.



View larger version (16K):
[in this window]
[in a new window]
  		FIGURE 2 Changes in the percentage of recorded interventions given over time, from Group 2

 

   Discussion
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
Conclusions
 Appendix
 References
 
One of the results of this analysis suggests that GPs' tendency to recommend physical activity to back pain patients has increased. Also, over the same time period, patients were less often advised to rest. Both trends appeared in Groups 1 and 2 and appear to coincide with the introduction of the guidelines in 1994.2,6 The gap between guidelines and practice, therefore, does seem to be reducing, but a causal relationship cannot be assumed.

There are many possible reasons for change in management over this period. These include:

  1. the GPs may have read the management guidelines for low back pain;2,6 or
  2. they may have been aware of media discussion of the guidelines; or
  3. practice-specific factors, such as the presence of a back specialist, the change to fundholding and/or setting up of contracts with physiotherapists; or
  4. the involvement in the randomized controlled trial to evaluate an exercise programme for low back pain patients may also have been influential.

Other biases include the fact that Dr A treated far more patients with back pain than his colleagues (see Table 1Go), and Dr A used Manipulation together with NSAIDs almost 90% of the time, compared with an average use by other GPs of 37%. This practice was, therefore, in some respects atypical. However, the important findings of this study were reflected in all GPs involved.

Younger patients were more likely to be recommended activity or exercise. Although it was not apparent why this was so, perhaps younger patients may be perceived by GPs as more receptive to physical activity advice.

Table 1Go showed that prescriptions for NSAIDs were less likely to be given to patients with a recurring episode than to patients with first episodes of back pain; also, analgesics were prescribed more to patients with a recurring episode. Therefore, the intervention of NSAIDs was still found to be the common first line approach in spite of the recommendations to try simple analgesics. The common use of NSAIDs has cost as well as clinical implications. The evidence for their effectiveness is limited and they are associated with adverse effects, which are potentially serious with prolonged usage.8

Patients apparently visited their GP earlier, as evidenced by the increasing frequency of short times before consultation over time. This implies a low level of self-reliance on the part of the patient, and/or a high expectation of help from the GP.

But, how reliable were patient notes? The information recorded in the patients' notes, and therefore used in this study, relied on GPs routinely documenting their advice or intervention. The reliability of consistent documentation is dependent on memory and habit. It is uncertain whether the data are a true reflection of what happens. Further cross-checking of patient notes and analysis of more GPs' notes might have improved the validity of the findings. However, it is noteworthy that the two main findings (i.e. changes in rest and physical activity) were reproduced by both methods of data collection (i.e. Groups 1 and 2).

Little et al.5 critically appraised the CSAG6 report in the light of a survey they carried out. In so doing, they highlighted one of the fundamental problems of guidelines in general; i.e. that of generalizability, especially in general practice. This finding was criticized by Summerton,9 who considers that the precise setting is, in many situations, irrelevant. He9 does see a use for guidelines in auditing clinical practice, so that while interventions may occasionally vary from the guidelines, an overall picture of practice can be compared and criticized.


   Conclusions
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Conclusions
Appendix
 References
 
In general, management of back pain now seems more in line with the guidelines than it was 5 years ago. However, reasons for this trend are unclear and many variations in practice still exist. Further work in evaluating the effectiveness of GP practice and the current guidelines is required.


   Appendix
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Conclusions
 Appendix
References
 



   Acknowledgments
 
We wish to thank Mrs S Bell-Syer (Research Fellow at the Centre for Health Economics, University of York) and Prof. T Sheldon (Previously Director of NHS Centre for Reviews and Dissemination, University of York). Our thanks also go to Dr A Harris (GP), Mrs L Mays (Haxby Practice), Dr T Nash (Consultant in Pain Management) and Mr A Vale (Senior Medical Statistician).


   References
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Conclusions
 Appendix
 References
 
1 Grimshaw JM, Russell IT. Effect of clinical guidelines on medical practice: a systematic review of rigorous evaluations. Lancet 1993; 342: 1317–1321.[ISI][Medline]

2 Waddell G, Feder G, McIntosh A, Lewis M, Hutchinson A. Low Back Pain Evidence Review. London: Royal College of General Practitioners, 1996.

3 Underwood MR, Vickers MR, Barnett AG. Availability of services to treat patients with acute low back pain. Br J Gen Pract 1997; 47: 501–502.[ISI][Medline]

4 Summerton N, Fitzgerald J. ‘General practice management of acute back pain is evidence based’. Unpublished manuscript, 1996.

5 Little P, Smith L, Cantrell T, Chapman J, Langridge J, Pickering R. General practitioners' management of acute low back pain: a survey of reported practice compared with clinical guidelines. Br Med J 1996; 312: 485–488.[Abstract/Free Full Text]

6 Clinical Standards Advisory Group (CSAG). Back Pain. Report of a CSAG Committee on Back Pain. London: HMSO, 1994.

7 Klaber Moffett J, Togerson D, Bell-Syer S, Jackson D, Llewelyn Phillips H. A randomised controlled trial of exercise classes for back pain patients in the UK community setting. 25th Annual Meeting of the International Society of the Lumbar Spine, Brussels, June 1998.

8 Silverstein F, Graham D, Senior J et al. Patients with rheumatoid arthritis receiving NSAIDs drugs—a randomised double blind placebo controlled trial. Ann Int Med 1995; 123: 241–249.[Abstract/Free Full Text]

9 Summerton N. GPs' management of acute back pain. Br Med J 1996; 312: 1480.[Free Full Text]


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow E-letters: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when E-letters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (9)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Frankel, B. S.
Right arrow Articles by Jackson, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Frankel, B. S.
Right arrow Articles by Jackson, D.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?