FORT LAUDERDALE, Fla--Unlike the piecemeal development of clinical practice guidelines in the United States--by individual institutions, networks, or managed care plans--in Canada, guidelines development is a provincial effort, with the resulting product applying to all oncologists in the province, Mark Levine, MD, said at the first annual conference of the National Comprehensive Cancer Network (NCCN).
FORT LAUDERDALE, Fla--Unlike the piecemeal development of clinicalpractice guidelines in the United States--by individual institutions,networks, or managed care plans--in Canada, guidelines developmentis a provincial effort, with the resulting product applying toall oncologists in the province, Mark Levine, MD, said at thefirst annual conference of the National Comprehensive Cancer Network(NCCN).
In the province of Ontario, with a population of 10 million, theprocess has been ongoing for 4 years, including an initial pilotprogram, said Dr. Levine, director of the Comprehensive CancerCenter, Hamilton, Ontario, and a member of the steering committeeof the Ontario Practice Guidelines Initiative.
An 8-member coordinating committee oversees the entire process,while disease site groups do the work of writing specific guidelines.The cancer site groups each include, among others, a medical oncologist,radiation oncologist, surgeon, nurse, and radiation therapist,and they receive assistance from a methodologist and librarianto help with literature searches. Dr. Levine stressed that thedisease site groups must include practicing physicians. "Ifit doesn't come up from the grass roots, then it's imposed fromon top, and it isn't going to work."
A framework, known as the practice guidelines cycle, was developedto aid the cancer site groups in writing the guidelines by outliningspecific steps to follow.
The group first selects and frames the clinical problem. Topicsmay be prioritized by different means, Dr. Levine said, includingprevalence of a condition, burden of illness, potential for significanthealth benefit, relevance to local practice patterns, degree ofvariation in practices, and availability of high-quality evidence.
The site group next selects the main outcomes, reviews the literature,and, after consensus-building discussions, generates evidence-basedrecommendations, using levels of evidence to indicate the panel'sconfidence in specific recommendations. Dr. Levine said that thepanel must "write down its logic and make explicit its rationalefor decisions."
A particular problem faced by oncol-ogy guideline site groupsis that standard practices are often not supported by good level1 or 2 evidence. To get around this, the groups recommend changesin conventional practices, when the practice is unsupported byrandomized trials, only when there is level 1 or 2 evidence tosupport the change.
One lesson learned in the pilot project was not to "bringeconomics in at the beginning. Once you get economics confusedwith critical review of the evidence, you don't get past squareone," Dr. Levine said. The framework provided by the practiceguidelines cycle attempts to dissociate the two, he said.
He noted that the guidelines will not necessarily represent thepolicy of the Ontario Cancer Treatment and Research Foundation,which oversees cancer care in the province. The final policy decisions,he said, are based not only on the guidelines but also on considerationsof cost and available resources.
Once the site groups come up with evidence-based recommendations,community feedback is obtained by mailing the recommendationsand a questionnaire to qualified community physicians. These physiciansare identified by the directors of Ontario's eight regional cancercenters and members of the site group.
Dr. Levine said that the recommendations for surgical managementof early stage breast cancer, for example, were sent along witha questionnaire to 188 practitioners involved with breast canceracross the province.
Almost all respondents (96%) agreed that the recommendation wasrelevant, and 84% gave it outright approval. Another 8% or 9%gave provisional approval with suggestions for improvement, while1% to 2% were opposed to any recommendations that changed theirpractice.
Incorporating suggestions from the community, the panel then modifiesthe evidence-based recommendations and transforms them into formalpractice guidelines. The guidelines undergo an independent review,and the final product is then circulated to all regional cancercenters and community oncology clinics in the province.
Each guideline is published in "reprint form," witha structured abstract on the front page. This document also reviewsthe data from the literature, discusses the consensus buildingprocess and any ongoing trials that might affect the next generationof the guideline, and reviews the evidence-based recommendationsand the feedback from practitioners.
In addition, Dr. Levine said, a World Wide Web site is being developedto disseminate the guidelines, which also can be downloaded fromthe computer network that links the regional cancer centers. Finally,the Canadian Journal of Oncology will publish the guidelines.
Evaluation of how the guidelines are being used, including outcomesdata, is also underway, he said. For example, a guideline existsfor the use of paclitaxel (Taxol) in patients with metastaticbreast cancer who fail anthracycline therapy. Whenever Taxol isused for this indication, he said, physicians must fill out aform providing information on the patient at baseline and afterthree treatment cycles if responding, including data on responseand ECOG performance status.
In a preliminary evaluation, over a 6-month period, 418 patientsreceived Taxol per this guideline, and compliance with reportinghas been good, he said. Only six forms have missing data, andforms have not yet been received on about 100 patients (some ofwhom have not yet completed all cycles).
Dr. Levine noted that the recommended Taxol dose range in theguideline is from 135 to 175 mg/m². A few patients receiveda lower dose, and a few got a higher dose. The hospitals wherethe higher dose was used were reimbursed only up to 175 mg/m².The first assessment on 262 patients after 3 cycles showed 2 completeand 42 partial responses, for a response rate of less than 20%.
"This is what you would expect based on the literature (phaseII trials with highly selected patients)," Dr. Levine said,"but this is the real world, and in fact the response rateis probably much less than 20% because we are waiting for dataon 100 patients, and I suspect that many of these terminated treatmentearly."
Of the patients who had CR, PR, or stable disease, 60 had no changein performance status, 25 improved, 8 worsened, and data are missingon 5. "So a small number in terms of performance status feltbetter," he said.
These outcomes data, along with new information from clinicaltrials, will be used to help the cancer site groups revise andupdate the guidelines.