OBJECTIVES: Sensitization to pollen is common in the Mediterranean Area, and patients are frequently treated with allergen-specific immunotherapy (SIT); in SIT the precise identification of the disease-eliciting allergen is a requisite for long-lasting therapeutic effect. SIT is prescribed following the EAACI Guidelines using skin prick test (SPT) (Alvarez-Cuesta, 2006). Nowadays, molecular allergology (MA) gives more accurate information about the true patient sensitization, and the usefulness of adding MA to SPT for SIT-indication is demonstrated in (Sastre, 2012).   Non-adherence to SIT is a major hurdle, contributing to poor clinical outcomes: 39% of pediatric patients receive SIT for less than 6 months and only 16% for >3 years (Hankin, 2008; similar data for adults in Senna, 2010). In this study, we analyze this multi-dimensional phenomenon merging all the measures available in the literature in one HE model: 1) data on high inaccuracy of guideline-based diagnostics (Sastre); with 2) the hypothesis of improvements by MA; with 3) QoL measures from (Ciprandi, 2010); with 4) SIT-adherence (Hankin and Senna); with 5) SIT monitoring by MA. METHODS: A total of 141 patients with allergic rhino-conjunctivitis and/or asthma sensitized to pollen from a complex pollen area (Spain) were SPT tested, and subsequently MA was performed using a microarray-based panel of 96 allergens (ImmunoCap ISAC®). The model covers a 3-year timeframe; the outcomes include the number of SITs correctly prescribed, and cost-effectiveness (CE). Uncertainty was addressed with sensitivity analysis. RESULTS: Results show that adding MA to SPT reduces SIT prescriptions by at least 20% (substantial costs saving), and allows for a more targeted (i.e. more successful with a higher QoL) SIT. CONCLUSIONS: Our results based on available data from multiple studies show that MA usage for SIT prescription is CE, and can increase SIT-adherence. We recommend a clinical trial capturing all the multi-dimensional aspects of this phenomenon should be conducted for model validation purposes.