32 Chapter 2 We planned subgroup analyses for age (children/adults), and type of allergen, since we expected that test characteristics and treatment effectiveness would differ in these groups. Sensitivity analysis was planned by excluding studies with a serious risk of bias. We planned to model the different elements to patient-important outcomes if the certainty of evidence in each element was at least moderate. If the evidence was less certain, we assumed modelling was not applicable. Certainty of evidence For each evidence element, we prepared GRADE evidence profiles (Appendix 3) using GRADEpro [19]. An overall rating of confidence in estimates of effect is relevant in CPG development. It is based on the critical outcome providing the lowest confidence [20]. The overall certainty of the evidence of the test-treatment strategy was defined as the weakest link in the chain of evidence [7]. Identification of challenges and proposal of solutions We made field notes for each methodological step lacking direct guidance on how to continue. All reported challenges were discussed between the authors, leading them to propose solutions that included a rationale for each practical/methodological choice. Results Consequences of test results The consequences of test results are presented in table 3. Literature search, selection and data synthesis The yield of the literature search and selection is presented per element of the testtreatment strategy in figure 1 (a, b). For the sub-question about diagnostic accuracy, three studies were included [21-23]. The search for test burden yielded no results. We included one systematic review about avoidance measures [24], two about antihistamines [25, 26], and three about corticosteroids [27-29]. We included seven cohort studies about natural course [30-36], and seven about the link between test result and management [37-43]. Study characteristics, including critical appraisal, are summarised in Appendix 2. GRADE evidence profiles with detailed judgements about the certainty of the evidence in the different comparisons are listed in Appendix 3.
RkJQdWJsaXNoZXIy MTk4NDMw