Systematic Reviews to Answer Health Care Questions

12 CHAPTER

Assessing and Rating the Strength of the Body of Evidence

Heidi D. Nelson

■ ■ INTRODUCTION The goal of a systematic review is to provide the best evidence available to make health-care decisions. 1 The final step in the synthesis of a systematic review is determining how strong the best evidence is. This involves evaluating the strength or quality of the body of evidence for specific research questions and outcomes. Previous chapters outlined the issues to consider and steps required to select and evaluate stud ies addressing the research questions posed in the systematic review, including both qualitative and quantitative analysis. These steps result in an understanding of the quality and applicability of individual studies and synthesis of their results across studies. However, at this point, a systematic approach to understanding the strength of the body of evidence for a given outcome or question is needed to provide an assessment of the reliability of the findings. It is important to distinguish the difference between assessing the strength of the evidence itself versus determining how well the evidence supports a specific recommendation or guide line. These are separate processes. For example, the strength of the evidence may be high for a given intervention and outcome, but the strength of the recommendation made by a guideline development group may be low because of other factors, such as patient preferences, adverse effects, or relevance to current practice. This chapter describes how to evaluate the strength of a body of evidence, including the development of current methods; definitions of the domains or characteristics of evidence and how they are evaluated; criteria unique to observational studies; and how to make final assessments of the strength of evidence. Accepted standards for rating the strength of a body of evidence guide these methods (Table 12.1). 2–8 ■ ■ DEVELOPMENT Systems to provide structure and hierarchy to bodies of evidence have been developed by many groups for multiple purposes. In general, they were intended to help users quickly and simply understand the relevant research to make evidence-based health care, guideline, or policy deci sions. Traditionally, members of guideline groups were clinical rather than research methods experts. Groups with greater expertise in methodology created systems that were more flexible and detailed, but also more complicated and difficult to understand. As a result, more than 50 types of grading systems have been used. 9,10 Dissatisfaction with the methodology, variation, and number of competing systems led a group of evidence-based guideline developers, researchers, and methodologists to create a new, more universal, and transparent approach to rating the strength of a body of evidence, Grades of Recommendation Assessment , Development , and Evaluation ( GRADE ). 11 Although these efforts focused on a system primarily for guideline development groups, the essential concepts are relevant to systematic reviews used for other purposes. Key principles of GRADE include

Copyright © 2024 Wolters Kluwer, Inc. Unauthorized reproduction of the content is prohibited.

183