2.1 Development of triggers for Anaphylaxis
ADR reports in our hospital from January 2014 to December 2021 were exported from the National Adverse Drug Reaction Monitoring System and retrospectively reviewed by a pharmacist and an allergist. Referring to the diagnostic criteria for anaphylaxis [1](Table 1) and the Technical Specifications and Evaluation Criteria for Common Serious Adverse Drug Reactions issued by the National Center for ADR, China, reviewers identified anaphylaxis cases and filled in the electronic case report forms, including diagnoses, department, characteristic medical orders, and progress notes,etal. Relevant data were analyzed and initial entries of triggers were formulated. An expert consultation questionnaire was designed according to the initial entries. The questionnaire comprises two parts, the first part was the basic information of experts, including major, education background, working years, professional title, etc. The second part was the evaluation of experts on the importance, familiarity, and judgment basis of trigger entries. We convened an 8-member multidisciplinary panel of experts involving one each of allergists, dermatologists, emergency specialists, cardiologists, intensivists, respiratory physicians, neurologists, and pharmacists, all members had rich experience in the diagnosis and treatment of anaphylaxis. All experts were asked to rate the importance and familiarity of each item on a 5-point Likert scale (with 1 meaning strongly disagree, 2 meaning agree, 3 meaning neutral, 4 meaning agree, and 5 indicating strongly agree). The judgment basis has 4 levels: theoretical analysis, practice, informed by domestic and foreign peers, and intuition. A score of 0.1-0.5 points was given according to the degree of influence on expert judgment. The highest score of 0.5 was obtained when the practice had a significant impact on expert judgment. Additionally, panel members were encouraged to submit free-text comments to clarify their responses to every question, suggest additional questions, or recommend modifications to exist questions. The indicators of the Delphi method include the positive coefficient of experts, the degree of expert authority, the degree of expert opinion concentration, and the degree of expert opinion coordination[14]. The expert positivity coefficient was expressed as the questionnaire recovery rate. The authority coefficient of experts (Cr) was determined by the judgment basis of the entries (Ca) and familiarity with the consultation content (Cs), where Cr = (Ca + Cs)/2, and Cr ≥ 0.7 was generally considered to be highly credible. The degree of expert opinion concentration was expressed by the mean value of importance score (Mj) and full score frequency (Kj) of trigger entries. The cut-off value of Mj and Kj = mean - standard deviation, and those with scores higher than the cut-off value were included. The degree of expert opinion coordination was expressed as the coefficient of variation (Vj). The cut-off value of Vj = mean + standard deviation and those with scores lower than the cut-off value were included. Entries that did not meet any of the three criteria were eliminated [15].