How to Read the New Recommendation Statement: Methods Update from the U.S. Preventive Services Task Force

The U.S. Preventive Services Task Force (USPSTF) makes recommendations for primary care clinicians and practices about preventive services for asymptomatic patients. Each recommendation is based on a careful review and synthesis of the evidence and is released with an accompanying summary of the evidence reviewed, usually in a journal publication. All recommendations and complete evidence reviews are available on the Agency for Healthcare Research and Quality (AHRQ) Web site at www.preventiveservices.ahrq.gov. The USPSTF last described its methods in 2001 (1). Since then, it has worked to refine its methods of evidence review and assessment and to create more usable documents in response to clinicians' needs. These changes have resulted in a revised grading system, as well as a new format and new language for the recommendation statement. Here, we focus on the changes to and the new look of the USPSTF recommendation statement. Discussions of other aspects of the methodological developments will unfold as a series of papers progresses. Another paper in this issue (2) describes the processes whereby the USPSTF develops and communicates its recommendations. Future papers in this series will include a discussion of how to approach the consideration of a clinical preventive service when evidence is insufficient to make a recommendation for or against its use and an explanation of the process by which the USPSTF evaluates evidence and determines the certainty and magnitude of net benefit of a clinical preventive service. Why Change the Recommendation Statement Now? The medical literature has seen an explosion in the number of systematic reviews published in the past 10 years, both from groups using specific evidence-based methods (for example, the Cochrane Collaboration) and from other independent institutions. This change in the field of evidence assessment and synthesis, and the changes described in the following paragraphs, have made it advisable for the USPSTF to update its methods for the development of its recommendations. The advancing methodology of systematic reviews draws attention to the fact that there may be important evidence from many types of studies. Although the well-conducted randomized, controlled trial often provides uniquely useful evidence (3), evidence from other types of studies is also critically important for making evidence-based recommendations. An important development in the field of making recommendations from systematic reviews is reflected in the work of the GRADE (Grades of Recommendation Assessment, Development, and Evaluation) working group. This group comprises experts from around the world and is working to develop standard processes and language for assessing bodies of evidence and making recommendations on the basis of the evidence. The approaches of the GRADE working group and the USPSTF have many elements in common. Both place separate attention on assessing the evidence and making a recommendation on the basis of the evidence. The GRADE approach assigns evidence quality at 1 of 4 levels: very low, low, moderate, and high, on the basis of specific criteria. The USPSTF assigns evidence certainty at 1 of 3 levels: high, moderate, and low, on the basis of 6 critical appraisal questions. The GRADE criteria are similar to the USPSTF's 6 questions. The recommendation phase for both GRADE (46) and the USPSTF rely on a judgment of net benefits (benefits minus harms), including whether net benefits are positive, negative, or uncertain. The GRADE process more directly includes costs than the USPSTF approach, although the USPSTF does consider the time and effort of patients and providers. The GRADE working group is developing a system that will apply to many areas, including public health, diagnostic, treatment, and prevention issues, whereas the USPSTF is more narrowly focused on prevention. A full description of the steps in the production of recommendations in the GRADE framework is not yet available, because several considerations in the GRADE recommendation phase are still under development. The USPSTF looks forward to an ongoing dialogue with the GRADE working group, with the hope of coming to consensus on a standard process and language to minimize confusion and maximize communication. In 2004, AHRQ conducted focus groups with 23 community-based and academic primary care physicians in Washington, DC (2 groups), and San Diego, California (1 group), to assess the extent to which current USPSTF recommendations and products are understandable and useful to them. Focus group respondents suggested improvements in the format and dissemination of the Task Force's recommendations. They reported an interest in being able to choose the level of detail they accessed in recommendations and the form in which they accessed them (for example, in print or on a Web site). Using this first set of focus group findings, the Task Force pretested 3 possible new formats in 2005 in 4 focus groups held in Baltimore, Maryland. Participants provided feedback about how formatting could highlight key information. Further refinements to the draft new recommendation statement were reviewed in 2006 with 4 focus groups of practicing primary care clinicians in Baltimore and the metropolitan Washington, DC, area, during which participants offered the consistent message that busy practicing clinicians require efficient tools that are clear and concise, use simple language, and have a clear format. Clinicians want to be able to scan written documents quickly, identify the relevant patient population, and see what actions are recommended. The New Recommendation Statement Recommendation statements now comprise 9 major sections (Table). Table. Contents of the U.S. Preventive Services Task Force Recommendation Statement* Preamble The preamble stresses that although evidence is the primary basis for USPSTF recommendations and statements about preventive services, the decisions made by clinicians for individual patients include other important considerations, such as the patient's clinical state and circumstances and personal preferences, factors that are important to consider when implementing any USPSTF recommendation (7). Likewise, the preamble states that policy decisions should consider local resources, constraints, expertise, and priorities. In addition, decisions about the screening and treatment of individuals and policy decisions should include a clear understanding of the evidence, which the USPSTF seeks to provide. Summary of Recommendation and Evidence The second part of the recommendation statement is the Summary of Recommendation and Evidence (for an example, see the recommendation statement that also appears in this issue [8]). This statement describes the recommendation and includes the letter grade. This is the bottom line of the USPSTF's statement. The USPSTF will continue to assign a letter grade to signify its assessment of the level of its recommendation. The grade will be based, as before, on the USPSTF's assessment and synthesis of the overall evidence and the magnitude of net benefit (benefits minus harms). The evidence will no longer receive an overall assessment of good, fair, or poor; rather, the product of the evidence assessment and synthesis by the USPSTF will be expressed as levels of certainty. This change in terminology is intended to add precision to the description of the recommendation- making process and does not indicate a change in the process of evaluating the evidence. In brief, certainty represents the USPSTF's judgment about the overall evidence of net benefit. The Task Force's recommendation letter grades are explained in Table 1 on page 132. While the USPSTF continues to use the same letter grades as it used in the past, some of the wording has changed. The description of an A recommendation no longer contains the word strongly; therefore, the A and B recommendation language is now the same. The USPSTF intentionally wanted to emphasize the importance of offering interventions with A and B recommendations, rather than distinguishing them on the basis of the certainty and magnitude of net benefit. The wording of the grade C recommendation represents perhaps the most important change in tone. The previous grade C recommendation read: The USPSTF makes no recommendation for or against routine provision of the service. The USPSTF found at least fair evidence that [the service] can improve health outcomes but concludes that the balance of benefits and harms is too close to justify a general recommendation. The new version will read: The USPSTF recommends against routinely providing X service for Y population. There may be considerations that support providing the service in an individual patient. The concept of the close balance of benefits and harms from the previous version (the italicized sentence in the preceding paragraph) is now captured by a new summary statement in the rationale section of the new recommendation statement: There is at least moderate certainty that the net benefit is small. This change is meant to indicate that although there is evidence of a small net benefit, the USPSTF has judged that this net benefit is too small to justify routine implementation of the service in the target population. When the USPSTF cannot estimate the magnitude of benefits or harms with any certainty, it assigns a grade of I to indicate that there is insufficient evidence to support a recommendation for or against provision of the service. In the new format, this grade will be associated with a statement, not a recommendation, because the USPSTF is not issuing a recommendation for the use or nonuse of the particular service. The USPSTF is aware of the conundrum faced by clinicians who must decide whether to offer a service in the face of insufficient evidence. If such services are used, clinicians and patients should understand that there is uncertainty about expected benefits and har