| Popis: |
Background Supporting evidence for medical test recommendations in clinical practice guidelines (CPGs) should not only include diagnostic accuracy, but also the downstream consequences of the test result on patient relevant outcomes. The aim of this study is to assess the methodology of developing of evidence-based CPGs about medical tests, and to explore determinants of best CPGs. Methods We performed a systematic document analysis and quality assessment of publicly accessible CPGs about three common diagnostic topics: C-reactive protein (CRP), colonoscopy and fractional exhaled nitric oxide (FeNO). Evaluation of the complete test-treatment pathway (diagnostic accuracy as well as downstream consequences on patient relevant outcomes (i.e. burden of the test, natural course, treatment effectiveness, and link between test result and administration of treatment)) was considered best practice for developing medical test recommendations. Results We retrieved 15 recommendations in 15 CPGs. The methodological quality of the CPGs varied from poor to excellent. Seven CPGs reported the use of the GRADE approach. In three CPGs, however, we could not find elements of that approach in the guideline documents. Ten recommendations considered diagnostic accuracy. Four of these were supported with a systematic review of the literature and rating of the certainty in the evidence. None of the CPGs evaluated all steps of the test-treatment pathway. Burden of the test was considered in three CPGs, but not with a systematic review of the evidence. Natural course was considered in two CPGs, and also not with a systematic review of the evidence. In three recommendations, treatment effectiveness was considered, which was supported with a systematic review and rating of the certainty in the evidence in one CPG. The link between test result and administration of treatment was not considered in any CPG. Conclusions CPGs with recommendations about medical tests do not seem to consider evidence about test consequences on patient relevant outcomes. This might be explained by reporting issues and challenging methodology. Future research is needed to investigate how to facilitate guideline developers in explicitly and reliably considering all steps of a test-treatment pathway when developing medical test recommendations. |
