Materials and Methods
Study design and population – A workshop (“AIT prescription workshop”, AIT-WS) has been organized with 10GP + 11AS at “Ospedale S. Pertini” (Rome, Italy) and with 18GP + 7AS at “Ospedale S. Maria degli Angeli” (Pordenone, Italy). The participants were recruited among those physicians collaborating with each center on a regular base concerning patients suffering from allergic diseases. Each workshop consisted of the following three phases: a) educational training; b) decision taking on clinical cases; and c) feedback survey.
Educational training – During the first part of the AIT-WS, the target, nature and methodology of the diagnostic tools (i.e. questionnaires, SPT, CRD, eDiary) were presented in comprehensive lectures. In detail, three lectures explained the general concepts, specific methodologies and clinical interpretation of the diagnostic tools. A fourth lecture was focused on the procedures for the following workshop module involving clinical cases.
Clinical cases – Among the @IT.2020 pilot study population (n=200), twenty clinical index cases (10 cases for each center) were selected in order to reproduce the local epidemiological scenario6 and to provide the widest spectrum of allergen(s) among the patients affected by moderate-to-severe SAR[Table E1] . The @IT.2020 pilot study population (n=200) has been described in detail elsewhere23. Briefly, 101 children (“Ospedale Sandro Pertini”, Rome), and 99 adults (“Ospedale S.Maria degli Angeli”, Pordenone) underwent a complete diagnostic allergy work-up, including a detailed assessment of the retrospective clinical history, skin prick testing (SPT), blood drawing for IgE determination against allergenic extracts and molecules, a prospective collection of clinical data via mobile phone application and an allergen-specific nasal provocation test (NAPT) with the extract of the AIT-candidate pollen(s) for a subgroup. Among the selected clinical cases, NAPT data were available for 18 of the 20 patients (90%). During the second part of AIT-WS, doctors were asked to express their therapeutic decision concerning their respective 10 clinical index cases.
Therapeutic Decision Taking - Each participating doctor filled a questionnaire reporting his/her own hypothetical AIT decisions on the base of the primary data progressively added. The first decision had to be taken based on the retrospective clinical history and SPT results. An independent second decision was then asked involving the clinical history, SPT and CRD results. Finally, the results of each patient´s symptom monitoring via eDiary during the past pollen season plus pollen counts were shown and doctors decided again on the hypothetical prescription of AIT for every patient.
Feedback survey– Finally, the doctors filled a questionnaire not only on the impact of each given diagnostic tool and its perceived benefits, but also on their role in the doctor´s hypothetical AIT prescription process. In addition, participants were asked to express their satisfaction level on the entire AIT-WS (tutorial, clinical cases and feedback survey) in terms of content and general organization[Figure E1] .
Algorithm for a potential clinical decision support system (CDSS) for pollen allergy– The @IT.2020-DSS tools are based on clinical data progressively considered in three steps: 1) clinical history and SPT (and/or serum sIgE) to allergen extracts; 2) IgE assays to pollen-derived molecular components (component resolved diagnostics, CRD); 3) electronic clinical diary (eDiary)24. In the first step, a list of potentially relevant allergens are selected considering the period of allergic symptoms reported by the patient (clinical history, seasonality of AR symptoms) and the SPT reactions (“traditional” diagnosis). In the second step, the list of allergens previously selected is restricted to those confirmed by IgE sensitization to their respective major allergenic proteins (Cup a 1 for cypress, Phl p 1 and/or Phl p 5 for grass, Bet v 1 for birch, Ole e 1 for Olive, Amb a 1 for ragweed, Art v 1 for mugwort, and Alt a 1 for Alternaria). Finally, the list of allergens considered after the second step (anamnesis+SPT+CRD) is further restricted to those whose pollination period, identified by the local aerobiologist, corresponded to moderate-to-severe and/or persistent allergic rhinoconjunctivitis symptoms prospectively registered by the patient during the same period. The three steps of the @IT2020-DSS algorithm can therefore be represented by a “pyramid” scheme for each allergen [Figure 1a ] generating a precision “target” when combined [Figure 1b ]. However, the algorithm does not exclude any obtained result based on the described exclusion scheme. In the rare case of positive test results occurring at an advanced stage (e.g. positive IgE to a major molecule of an allergenic source which had been previously excluded on the base of a negative SPT result), the respective allergen is being considered potentially relevant for the next step of the algorithm.
Statistics- Data were summarized as numbers (n) and frequencies (%) if they were categorical and as mean and standard deviation (SD) if quantitative. Percentages of correct hypothetical AIT prescription at each step and for each medical category were computed, taking as comparison reference, for each examined case, the most frequent AIT hypothetical prescription of allergen immunotherapy among allergy specialists at the final stage of CDSS (gold standard). Chi squared test, when conditions were respected or Fisher exact test was used to evaluate the association of categorical data between AS and GP groups. McNemar’s test was used to compare difference of frequency within each group. A p-value < .05 was considered statistically significant. Statistical analyses were performed with R Core Team (2014), version 3.2.3.