Box 3.1: Approach to rating the quality of evidence and strength of recommendations using the GRADE system
Consolidated ARV guidelines, June 2013
Since 2008, WHO has followed the GRADE system. GRADE separates the rating of the quality of evidence from the rating of the strength of the recommendation.
The quality of evidence is defined as the confidence that the reported estimates of effect are adequate to support a specific recommendation. The GRADE system classifies the quality of evidence as high, moderate, low and very low (Table 3.1) (4–10). Randomized controlled trials are initially rated as high-quality evidence but may be downgraded for several reasons, including the risk of bias, inconsistency of results across studies, indirectness of evidence, imprecision and publication bias.
Observational studies are initially rated as low-quality evidence but may be upgraded if the magnitude of the treatment effect is very large, if multiple studies show the same effect, if evidence indicates a dose–response relationship or if all plausible biases would underestimate the effect (10). The higher the quality of evidence, the more likely a strong recommendation can be made.
The strength of a recommendation reflects the extent to which the Guideline Development Group was confident that the desirable effects of following a recommendation outweigh the potential undesirable effects. The strength is influenced by the following factors: the quality of the evidence, the balance of benefits and harms, values and preferences, resource use and the feasibility of the intervention (Table 3.2).
The GRADE system classifies the strength of a recommendation in two ways: “strong” and “conditional” (11). A strong recommendation is one for which the Guideline Development Group was confident that the desirable effects of adhering to the recommendation outweigh the undesirable effects. A conditional recommendation is one for which the Guideline Development Group concluded that the desirable effects of adhering to the recommendation probably outweigh the undesirable effects but the Guideline Development Group is not confident about these trade-offs. Table 3.3 summarizes the implications of a strong or conditional recommendation for individuals, clinicians and policy-makers.
The reasons for making a conditional recommendation include the absence of highquality evidence; imprecision in outcome estimates; variability in the values and preferences of individuals regarding the outcomes of interventions; small benefits; applicability in all settings versus specific settings; and benefits that may not be worth the costs (including the costs of implementing the recommendation).