Applying GRADE to a network meta-analysis of antidepressants led to more conservative conclusions

Abstract

ObjectiveTo explore the impact of applying the Grading of Recommendations and Assessment, Development, and Evaluation (GRADE) approach to assess the certainty of the evidence in a published network meta-analysis (NMA) of antidepressant therapies. Study design and settingsWe applied the GRADE approach to rate the certainty of the evidence for two outcomes, efficacy and acceptability, in each of the 66 paired comparisons within a previously published NMA assessing the relative efficacy and acceptability of 12 new-generation antidepressants. ResultsFor the outcome of efficacy, of the 25 comparisons in which the 95% CrI of OR excluded 1, 18 had certainty of evidence rated high or moderate. For the outcome of acceptability, of the 13 comparisons whose 95% CrI excluded 1, 10 had certainty of evidence rated high or moderate. Of the 11 comparisons involving sertraline, the antidepressants that the authors of the NMA suggested to be best, only 3 demonstrated it to be more effective and only 3 showed better tolerance, based on a 95% CrI excluding 1 and a high or moderate rating of certainty. ConclusionsIn this example, application of GRADE highlighted varying evidence certainty, led to more conservative conclusions, and potentially avoided unwarranted strong inferences based on low certainty evidence.