One should not just read what one believes: the nearly irresolvable issue of producing truly objective, evidence-based guidelines for the management of differentiated thyroid cancer

Abstract

summary of some of the cited publications introduces eachrecommendation; comprehensive tabular overviews ofsupporting data are not given. The reader therefore cannoteasily assess the thoroughness with which the literatureregarding controversial topics was considered. In particular,the ATA guidelines’ summaries often fail to mentionevidence against the recommendation.Further, whether greater weight is accorded to publishedmeta-analyses such as those by Sawka et al. [12, 13] than tosingle-centre observational studies is not specified in theguidelines’ methodology description. Meta-analyses ofcourse may show effects that individual studies are toosmall to detect; for example, the meta-analysis by Sawka etal. [13] regarding the effect of RRA on recurrence anddisease-specific mortality considered six studies addressingrecurrence at distant sites. None of the studies individuallyincluded enough distant recurrences for a 95% confidenceinterval (CI) for risk reduction that excluded the zero value.However, pooling the results from all six studies, the riskreduction was 2% and the 95% CI 4–1%. A significantlylower number of new metastases over the course of thedisease after