| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
Basel, Switzerland
Dr. Kunz is affiliated with the Basler Institute for Clinical Epidemiology.
Correspondence to: Regina Kunz, MD, MSc (Epi), PhD, Basler Institute for Clinical Epidemiology, University Hospital Basel, Hebelstrasse 19, CH- 4031 Basel, Switzerland; e-mail: RKunz{at}uhbs.ch
Since the introduction of systematically developed clinical practice guidelines > 20 years ago,1 guidelines have become a broadly accepted tool to support health-care professionals in their decision-making process. The usefulness of practice guidelines has been underlined by evidence that they can initiate changes in the way doctors practice medicine and promote improvements in patient care.23
Over the years, many guideline programs have developed a structured methodology following "evidence-based" principles that are complemented by consensus processes. However, others are still in search of guidance. It is encouraging to observe a growing understanding of the assessment of the evidence, but how to run the consensus process and come up with a transparent recommendation seems to be less clear. While the process of appraising evidence and formulating a recommendation has been frequently described in principle45 (Fig 1 ), a specific, easy-to-use explanation of how to apply those principles in practice has so far been missing.
|
High-quality evidence-based practice guidelines require an enormous amount of financial and human resources to identify, appraise, and summarize all relevant evidence, and to come up with useful and applicable recommendations. Nevertheless, national and regional agencies, professional bodies, or scientific organizations, even individual hospitals, produce their own guidelines, all of which are based on the same body of evidence. For that reason, > 400 clinical practice guidelines for diabetes and diabetes-related conditions are listed on the Web site of the National Guideline Clearinghouse of the Agency for Healthcare Research and Quality.6 Duplication and redundancy of work are enormous from a global perspective, and a number of initiatives have come about in search of collaborations.7 One of them, the Guidelines International Network, which was founded in 2003 and now has > 60 guideline organizations from 31 countries, promotes the idea of sharing and adapting high-quality guidelines locally, regionally, or transnationally. Experience, however, shows that sharing might not be so straightforward, given the lack of general agreement of what constitutes a good evidence profile and the "black box" between the evidence and the recommendation. And last, but not least, each organization has developed its own cipher to communicate its recommendations to physicians and patients, which has become a modern Tower of Babel.8
In this issue of CHEST (see page 174), Guyatt and members of the American College of Chest Physicians (ACCP) Task Force910 describe a new and comprehensive guide for the development of recommendations that they applied in their most recent update of guidelines on antithrombotic therapy.11 Basically, the process comprises the following three steps: a structured workup of the evidence on quality, effect, and uncertainty; informed judgments on the clinical relevance of findings, and the balance of benefits, risks, burdens, and costs; and providing simple and meaningful recommendations. While at first glance this process does not sound like something new, a closer look reveals that the novelties are in the details and in one overriding rule: be explicit and transparent in all your decisions. Whoever has sat on a guideline panel will recognize the inherent challenges of the request for explicitness and transparency. All members of the guideline panel have to participate in joint reasoning and mutual agreement. But the predefined criteria laid out in the article will guide any panel through the assessments and will help to formulate the following simple recommendation: do or do not do something, with a strong or weak emphasis. These are the fundamental novelties of the articles.
Hierarchy of Design
Assessment of the evidence reveals more flexibility without loosening methodological rigor. The starting point is the familiar hierarchy based on the following design: randomized designs on top; and observational studies on bottom. However, obvious weaknesses in the planning, performance, or analysis of randomized trials can bring the strength of the evidence down, as do inconsistent results, sparse data, or indirect comparisons, while certain features of observational studies, such as a strong association, can improve the strength of the research. Supporters of observational studies will not be the only participants to applaud this additional flexibility.
Balancing Benefit, Risks, Burden, and Cost
Not all study end points are of similar importance to patients and their clinicians; patients might value the same end point differently. Those important distinctions need to be captured and integrated into a more global judgment. How should a trial with a large effect of moderate importance be valued? How should high-quality observational research with a strong association be judged compared to research from a randomized trial with only 30 patients? How should a procedure with a small effect and high demands on the health-care budget be looked at? Balancing such aspects of research is a true challenge, and most guideline groups pull out and present their decisions as a black box. Expert panels can do better on this core guideline activity. In their guidance, Guyatt et al910 provide criteria for an explicit and structured decision-making process, which is complemented by multiple examples from the antithrombotic guideline that capture typical situations in which judgment was required. Sound groundwork in the assessment process of step 1 and step 2 then pays off.
Keep It Simple
The decision of the task force to limit the recommendations to two options ("strong" and "weak"), and to provide a translation into free text fits the objective of simple and meaningful recommendations. Is there a clear benefit when balanced against the risks and burdens? Is there a constant benefit across a wide spectrum of prognostic factors and baseline risks? Is there an overwhelming agreement among informed patients in favor of the intervention? If the answers are "yes," a strong recommendation can be issued; otherwise, the recommendation would qualify as weak. Using the terms strong and weak makes the concept amenable to panelists and users alike, and has been proven to be very helpful in specific situations.
Considering Cost?
Guidelines always have an impact on health-care resources, and this needs to be considered when issuing recommendations. Nevertheless, many guideline panels lack familiarity with economic assessments and prefer to leave this area to other professional bodies. Unfortunately, content expertise and resource allocation become disconnected. The ACCP Task Force developed a handful of easy-to-follow key issues to encourage panels with little economic expertise to actively address those issues. These criteria follow the same pragmatic approach as the methodology criteria in the first article by Guyatt et al.910 There is a learning curve to overcome, and the advice to seek support by a health economist for those economic discussions seems very sensible and a good investment in the future where decisions integrating economic issues will be even more in demand.
The criteria of the ACCP Task Force for a modern and informative grading system of recommendations approximate the ones established by the Grading of Recommendations Assessment Development and Evaluation (GRADE) Working Group.12 It is reassuring to see that their criteria fit the needs of the medical community and that their guidelines work in practice. The next step is to demonstrate that this methodology facilitates the development of recommendations in a global world where the guidelines of one country will be read and considered by many other countries. Several major urology societies from North America, Europe, and Asia, leaders in evidence-based guidelines themselves, are the first to take up this challenge. This could be a further contribution to reducing the redundancy and duplication of guidelines. The Guideline International Network and other organizations currently explore the potential of the methodology for their own objective of sharing guidelines. Furthermore, this approach can be easily applied to the following related areas in which recommendations are issued: the technology assessment movement, another ongoing activity in most health-care systems, is in need of more transparency with regard to evidence, values, and judgment, and faces similar problems of excessive duplication worldwide. A slightly modified approach could produce, on the one hand, consistent evidence profiles to be shared, and on the other hand, transparent judgments that will always be influenced by national needs and priorities.
What then are the limitations? Methodology is a tool with all its imminent strengths and limitations. Basic knowledge in critical appraisal is mandatory. A learning phase for the whole panel is highly recommended. Although the methodology has been applied and tested for many topics, there might be content areas in which other criteria would better catch the core problems. Sensible judgement is required when applying the method, as is required from all tools.
In 1986, the ACCP Task Force was the first to provide recommendations based on the level of evidence.13 Twenty years later, this same group is the first to incorporate the most recent methodological developments into their guidelines. I hope many others will follow this track.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
