Clinical practice guidelines & test consequences 113 3 Discussion Our document analysis on a sample of 15 CPGs about CRP, colonoscopy and FeNO diagnostic tests revealed that none of these CPGs reported evidence on all components of the test-treatment pathway. Consideration of any test consequences on patient-relevant outcomes was described in only six CPGs (three CPGs considered burden of the test, two considered natural course of the disease of interest, and four considered treatment effectiveness). Systematic review of the literature, including a judgement of the certainty in the supporting evidence was only reported for four recommendations and covered diagnostic accuracy in all four cases and treatment effectiveness in one case. The importance of systematically evaluating test consequences for the purpose of developing CPGs has been recognised [5, 12, 13]. For instance, one could imagine that a certain diagnostic test might have limited value when it has no treatment consequences (e.g. no treatment available). Or, when comparing two tests with the same diagnostic accuracy to ascertain the same disease, one could recognize that differences in test burden may play an important role. This study suggests that implementation of the systematic evaluation of the value of a test is lagging behind. This also applies to CPGs that claim to use the GRADE approach. There seems to be a gap between following a methodologically robust approach and developing CPGs in practice. Two issues may explain that gap. First, guideline developers may have considered the downstream consequences of a diagnostic test but did not explicitly report these. It may not be strictly necessary to systematically evaluate all evidence components. However, we still recommend transparent documentation of choices made in the guideline development process. A guideline user should be able to read which elements of a test-treatment pathway were considered and how, and which were not considered and why. Second, performing systematic literature reviews of the complete test-treatment pathway – including assessment of the certainty in the evidence of test accuracy and downstream consequences – is complex and time-consuming. The use of the GRADE approach for the evaluation of diagnostic tests and test strategies is considered challenging [10, 11]. Strategies to facilitate the use of this approach, such as training of CPG panel members, may improve the application. Unfortunately, we could not determine factors that contribute to successful use of the GRADE approach, because we could not identify a ‘best practice’.
RkJQdWJsaXNoZXIy MTk4NDMw