OBJECTIVES: To evaluate the use and costs of all health care provider services among 1000 patients with chronic asthma during a five year period. METHODS: Patients who had visited the Pulmonary Clinics of Helsinki University Hospital during the yrs 2000-2004 were invited to the study. The participants' consent was obtained and a comprehensive medical history collected from all health care providers who had treated the participant during the last 5 years. Data included all physician recorded symptoms, signs, adverse drug effects, complications, diagnostic test results, given treatments, and procedures as unstructured text. Using advanced linguistic analysis methods, we identified for each event both cost related attributes (location, personnel, type of contact, urgency) and clinical attributes (age, gender, BMI, smoking, lung functions, co-morbidities, and prescribed medication) potentially explaining the differences across their clinical outcomes. We developed a medication independent score for the stability of asthma based on physician recorded observations. Membership [0.1] of an asthma related contact was defined as a probability-like function P. RESULTS: Our regression model is based on 65,698 health care contacts explaining a significant proportion (43%) of all direct non-pharmacological costs of the patients. Fifty-six percent of the contacts were asthma related. An asthma related event occurred on average 6 times per year with an average cost of €256 (based on average costs in Finland). Unexpectedly the use of prescription antihistamines resulted less frequent patient interaction with the health care system and potentially in better asthma control. Smoking, adverse drug events, co-morbidities, existing allergy symptoms increased asthma related expenditures significantly. Use of antihistamines, normal FVC in spirometry, and male gender decreased the costs. CONCLUSIONS: This empirical model of health care utilization can be used to explain differences in health care utilization among patients with different co-morbidities and different treatment protocols.