GRADE good practice statements: a time to say “good-bye”? A new typology for normative statements on interventions

Abstract

BackgroundClinical and public health guidelines include a variety of types of normative statements concerning interventions. “Recommendations” are usually the central focus, and are based on one or more systematic reviews of research evidence. Guidelines may include other types of normative statements, however, including Grading of Recommendations, Assessment, Develoment, and Evaluation (GRADE) good (or best) practice statements (GPS), which represent recommendations that guideline panels feel are important but are not appropriate for formal ratings of quality of evidence because it is sufficiently obvious that desirable effects outweigh undesirable effects. These normative statements are typically supported by a great deal of high-certainty, indirect evidence, which the authors feel would be a waste of time to examine. There are a number of conceptual and methodological issues with GRADE GPS, however, and these are manifested in guidelines, including both inappropriate overuse and underuse, and unclear interpretation and impact among end-users. This situation has arisen in part from lack of clarity in, and misunderstandings of, GRADE guidance, the lumping of many different types of normative statements under one label (“GPS”), from limitations in GRADE's approach to linked bodies of evidence, and because the appropriate basis for many normative statements about interventions is not reviews of research evidence. A new typology is needed for normative statements on interventions and policies that are not optimally based on reviews of research evidence. Proposed TypologyThis proposed typology differentiates normative statements about interventions by the type or nature of the most appropriate basis for the statement. The typology encompasses the range of statements encompassed by GPS, but provides a more nuanced categorization designed to assist both guideline developers and end-users. This typology encompasses two main types of normative statements about interventions (including policies): (1) statements that indicate when to use (or not) an intervention, which intervention to use, and if, when and how to use it and (2) the principles, practices, or norms that inform or underpin such interventions. These correspond to normative statements based on empirical evidence, and those based on human rights, ethics, or norms, respectively. Normative statements based on empirical evidence include: (1) recommendations based on systematic reviews of human or animal evidence on effectiveness and harms, including linked bodies of evidence; (2) normative statements based on scientific fundamentals (eg, physical/biological/chemical properties, theories, laws, or principles); and (3) implementation guidance based most commonly on experiential evidence such as case studies. Normative statements based on human rights, ethics, or norms include: (1) guiding principles, based on human rights standards and conventions and/or ethics principles; and (2) practice norms and standards, based on clinical and public health norms and/or professional standards. ConclusionThere are conceptual and methodological problems with GRADE GPS, leading to their misapplication, with overuse and underuse. This paper presents a proposal for a new typology for normative statements on interventions, according to the basis for the statement. This typology encompasses and replaces GPS, providing a more nuanced set of statements. Testing of this proposed approach is needed among both guideline developers and end-users.