Question

summary of three at a table​

Answer 1

Summary of the overall strength of three example recommendations

Question/StepR1R2R3PhrasingUse two or more members of the review team, working independently, to screen and select studiesUse the Peto o or the Mantel-Haenszel method rather than inverse variance weighting for meta-analysis of rare eventsUse random effects models for diagnostic test meta-analysisIs the recommendation testable?YesYesNo*Adequacy of evidentiary basis or scientific rigor (testable statement) / face validity (nontestable statement)See

Table 9See

Table 8A random rather than equal (“fixed”) effect model is more plausible. Choosing models based on data is a needless compromise if a plausible generative story can be posed.Feasibility of implementationNo obstacles to feasibility identifiedNo obstacles to feasibility identifiedNo obstacles to feasibility identifiedExpected impact of implementationUnclearUnclearMay result in some practical changes in conclusions because uncertainty is more fully modeled compared with equal (“fixed”) effects modelsCongruence with context-specific requirementsHighHighHighOverall confidence that the recommendation is a mandatory