Article Text

Download PDFPDF

Pain management
  1. Peter Leman
  1. St Thomas's Hospital, Lambeth Palace Road, London SE1 7EH, UK (Peter.Leman{at}
    1. Anne-Maree Kelly
    1. Department of Emergency Medicine, Wstern Hospital, Private Bag, Footscray, Vic 3011, Australia (anne-maree.kelly{at}

      Statistics from

      Editor,—I read with interest the paper by Kelly1 on a “process approach” to pain management that outlined some change strategies for a pain protocol implementation. It was unfortunate that the journal did not include a copy of the tritrated IV narcotic policy for readers to assess separately. Novel approaches to improving pain management are always welcome and the author clearly illustrates the failure in emergency departments 10 years ago to provide adequate analgesia. It is interesting to note that despite the stated successful implementation of the policy, 10% (5 of 50 who received narcotic analgesia in 1997) of patients still had at least one dose of analgesia intramuscularly.

      Many of the processes described in this paper are based on subjective assessment rather than any more robust analyses. The rationale for using a nursing led process seems to have been justified because “it was felt” that they would provide more formal review and assessment, rather than any evidence that the emergency department doctors were unable to do so. In addition the conclusions that practice has changed permanently seems to be attributed to the “example of senior staff” and that the policy is now “everyday practice” with no supporting evidence.

      The author has used a χ2 test to illustrate that over time the outcome of the process has changed—that is, more patients now receive intravenous rather than intramuscular narcotics. As there is no temporally related control group the obvious bias of temporality has been ignored. It may well be that clinical practice has changed in the study emergency department and other departments over time, this secular trend is not necessarily related to the implementation of a local pain policy. Thus, the author's conclusion that a “major and sustained change to analgesia ordering” is attributable to the described process approach lacks validity.

      Bias in the matching of subjects has not been fully resolved. While the author states that the two groups are comparable for age and sex no supportive data are provided. The author states that the reason only one patient had a fracture of the tibia in 1993 and 21 in 1997 is “attributable to chance”. Simple analysis of difference of proportions would show that the probability of such an event occurring is very unlikely (p<0.0001, standard normal deviate −6.03). Although a χ2 test on table 1 supports the author (p=0.0001, df = 3, χ2 =20.88) for no difference in the overall fracture type, between the study periods.

      The author has described some important aspects of departmental change management in relation to analgesia policy. However, the author has failed to prove that the implementation of such a policy has influenced the outcome of this process.


      Author's reply

      Editor,—I thank Dr Leman for his thoughtful letter but am disappointed that he seems to have missed the important message of this paper—that pain management can be improved by innovative process change.

      Dr Leman makes several points that I will answer in turn. The process of developing a pain management policy required the emergency department (ED) team to take an honest look at our work practices and environment. It was the team's assessment that, in our ED, members of the nursing staff had more regular contact with patients as part of scheduled observation that forms part of the nursing process. Doctors, on the other hand, had less regular contact and were often occupied with other duties. While it would have been possible to have doctors perform the review and augmentation role, it would have meant a major change in work practice and thus was less likely to be successful. The issue is not one of who performs which steps in the pain management process, rather that all steps are performed consistently in a way that fits well with established work practices. Different departments may well adopt different strategies to achieve this end.

      The question of a control group for comparison was carefully considered at the time the process change was being developed. We had considered investigating time to analgesia between a group treated by the protocol and one that was not, but this was considered unethical in light of our knowledge that previous practices were ineffective. I agree that there may well have been gradual change in analgesia practice between the time periods studied, however the magnitude of change shown in this study is large and is as impressive for patients treated for other painful conditions, such as renal colic.1

      The question of bias in the matching of subjects is well made. The study aimed to compare two groups with long bone fractures and it was this larger group rather than specific fracture subgroups that was sampled, giving a reasonable match for overall fracture type between the periods as Dr Leman agrees. Only on subgroup analysis was the mismatch for tibial fracture identified—a chance finding appropriately acknowledged.

      That some patients were treated outside the protocol is almost inevitable! For this fracture group the rate of non-compliance (that is, giving intramuscular analgesia) was 5% of patients (8% of those receiving analgesia) and for a renal colic group, it was 3% of patients treated.1 This rate is low compared with other studies that have investigated adherence to protocols in acute medicine.2, 3

      My aim in reporting the Western Hospital experience in developing a new process for managing pain was to demonstrate that a commitment to improving patient care, an open and honest appraisal of the barriers and a flexible approach to solutions can result in innovative and effective treatment strategies. That said, the specific solutions may well be (appropriately) different in different environments reflecting different staff mix or work practices.

      Copies of the protocol are available on request from the correspondence address given in the paper.


      Request Permissions

      If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.