ReacciónThe formulation of WHO recommendations is based on the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach. Once evidence has been synthesized through a systematic review approach, evidence reviewers use the GRADE methodology to categorize the quality of the evidence into four levels: high, moderate, low or very low (see Table A1.1). The starting point for rating the quality of evidence is always the study design, whereby evidence from RCTs is rated as high quality, while evidence from non-randomized or observational studies is rated as low quality. This value is then adjusted based on additional considerations. Five factors may lower the quality of evidence, namely: limitations in study design and execution, indirectness, imprecision, inconsistency and publication bias. Three factors may increase the quality of evidence from observational studies: dose-response gradient, direction of plausible bias and magnitude of the effect (1).
Table A1.1 Quality of evidence in GRADE (1)
A recommendation may be strong or conditional (see Table A1.2), reflecting the degree to which the GDG is confident in the balance between desirable and undesirable consequences of implementing the recommendation. The strength of a recommendation is primarily determined by four main factors, namely: the confidence in the estimates of effect of the evidence (that is, the quality of the evidence as assessed through GRADE); the values and preferences related to the outcomes of an intervention or exposure; the balance of benefits and harms; and resource implications. Other considerations that may affect the strength of a recommendation include priority of the problem, equity and human rights, acceptability and feasibility.
Table A1.2 Interpretation of strong and conditional recommendations for an intervention (1)
Annex 1 References
1. WHO handbook for guideline development – 2nd ed. Geneva: World Health Organization; 2014 (https://iris.who.int/handle/10665/145714, accessed 17 October 2023).
