D1.216 - AIT prescription drivers: a factor analysis of data from the CHOICE survey

Current AIT guidelines and studies are based on safety, efficacy and effectiveness of available products. On the other hand, there are no definitive recommendations nor studies on AIT product selection. The main goal of the present study was to evaluate real-world prescription patterns and drivers, using a factor analysis approach, in a large global cohort.

The CHOICE study is a real-life prospective, multicenter, international, observational, cross-sectional web-based survey, performed in 20 countries, distributed in 8 geographical regions, between November 2019 and April 2024. Participating physicians completed standardized questionnaires for each AIT prescription. Based on previous CHOICE publications, thirteen clinical, patient-related, and product-related parameters were scored using Likert scales and then analyzed using exploratory and confirmatory factor analysis. Once latent factors were identified, they were further analyzed through multivariate mixed logistic regression models, using country and physician nested within country as random intercepts.

We collected data from 467 physicians, prescribing AIT for a total of 11,550 patients. (60.0% SCIT and 40.0% SLIT). Three latent factors explaining 52% of variance were identified: 1 – Scientific and Clinical Evidence-Based; 2 – Resource and Access Optimization; and 3 –Tailored for the Patient and their Lifestyle. When country was included as a random effect, all three factors were associated with SLIT prescription. After adjustment for both country and physician, only Factor 2 and 3 remained significant. Factor 2 (Resource and Access Optimization) was associated with reduced SLIT prescription (OR 0.42; 95% CI 0.31–0.59), while Factor 3 (Resource and Access Optimization) showed a strong association with SLIT prescription (OR 10.90; 95% CI 7.20–16.57). Model discrimination was excellent (AUC 0.913). Variability attributable to country (ICC 0.86) and physician (ICC 0 0.74) was high.

Our results show that patient-centered considerations markedly increase the likelihood of SLIT use and that main variability on AIT drivers of prescription and selection depend on countries and physicians experience, thus suggesting a need for harmonized, patient-based recommendations.