Atopic dermatitis is a common disease that severely impairs patients’ quality of life. In recent years, novel targeted therapies have emerged as an alternative to conventional treatments for the most serious cases of the disease. Dupilumab is one such medication that offers many patients a chance to improve their condition. Despite its high efficacy rate, not all patients improve after therapy. In addition, the drug is expensive and can lead to significant side effects. This paper defines a methodology for patient profiling based on the expected response to dupilumab treatment. Based on various scenarios, we built decision trees that are able to distinguish patients who respond to dupilumab treatment from those who do not, achieving acceptable values for specificity, sensitivity and accuracy. Once properly validated, the rules associated with these decision trees can be used as a tool to support medical decision-making and contribute to an initial screening of patients even before treatment.