Figure 5).

Algorithm for assigning evidence grades to diagnostic recommendations; (a) The gold standard. This can be either another test that is currently accepted as the gold standard or analysis of a representative cohort of patients who underwent the test of interest and are followed for a sufficient length of time that occurrence of the target outcome is likely if the diagnosis is present (with adjustment for covariates associated with prognosis); (b) Note that if follow-up of a cohort is not sufficiently long or complete enough to rule out diagnostic errors, or if data are not adjusted for covariates, this category would apply