Applying GRADE for diagnosis 37 2 ‘fewer symptoms or remission’. Follow-up varied from 2 to 23 years. Complete remission varied from 12% to 72%, while ‘fewer symptoms or remission’ varied from 47% to 55%. The certainty of the evidence was very low because of (very) serious risk of bias, inconsistency (in the studies about remission), indirect evidence (since we assumed patients in the included studies might have used medication) and imprecise results. Link between test results and management We included seven studies reporting the link between test results and clinical management [37-43]. All studies reported about medication compliance, which we used as indirect measure for the link between a positive test result and effective management. Three studies included children [40, 41, 43], whilst the other studied adults with AR [37-39, 42]. In children, medication adherence varied from 12.5% to 70% [40, 41, 43]. In adults, self-reported good adherence varied from 28% to 87% [37-39, 42]. Weight of the medication consumed showed a lower compliance than self-reported adherence [38]. Frequent reasons for non-adherence were forgetfulness, fear of side effects, belief that medication was no longer needed and belief that the medication was ineffective [42]. The certainty of evidence was very low, mostly due to indirectness: we assumed adherence/compliance was an indirect measure of the link between test result and clinical management. Also, most studies suffered from a high risk of bias. Overall certainty of the evidence and overall result The overall certainty of the evidence was very low. This implies that there is a very uncertain evidence base for the value of sIgE blood testing as an add-on test to history taking compared to history taking alone in patients suspected of having AR in primary care. Challenging issues in applying GRADE and suggested solutions Challenging issues and our suggested solutions are tabulated in table 4. A further explanation per element of the test-treatment strategy is stated below. Lack of evidence We noticed a lack of high certainty of evidence in all elements of the test-treatment strategy. A solution might be to conduct more high-quality research.
RkJQdWJsaXNoZXIy MTk4NDMw