Theoretical models can help to overcome experimental limitations to better our understanding of lung physiology and disease. While such efforts often begin in broad terms by determining the effect of a disease process on a relevant biological output, more narrowly defined simulations may inform clinical practice. Two such examples are phenotype-specific and patient-specific models, the former being specific to a group of patients with common characteristics, and the latter to an individual patient, in view of likely differences (heterogeneity) between patients. However, in order for such models to be useful, they must be sufficiently accurate, given the available data about the specific characteristics of the patient. We show that, for asthma in particular, this approach is promising: phenotype-specific targeting may be an effective way of selecting patients for treatment based on their airway remodelling phenotype, and patient-specific targeting may be viable with the use of a clinically-plausible dataset. Specifically we consider asthma and its treatment by bronchial thermoplasty, in which the airway smooth muscle layer is directly targeted by thermal energy. Patient-specific and phenotype-specific models in this context are considered using a combination of biobank data from ex vivo tissue samples, CT imaging, and optical coherence tomography which allows more detailed resolution of the airway wall structures.