Study objective: To investigate determinants of health-care utilization in patients with obstructive sleep apnea syndrome (OSAS).
Design: Case-control prospective study with OSAS patients and a control group. We compared 218 patients with OSAS to those of age-, gender-, geographically-, and family physician-matched control subjects from the general population, matched 1:1 (chi2 = 0.999).
Participants: All participants were members of Clalit Health Care Services, a health maintenance organization in the southern region of Israel. All OSAS patients underwent nocturnal polysomnography studies. Indexes of health-care utilization 2 years prior to the polysomnography were analyzed.
Measurements and results: Health-care utilization was 1.7-fold higher (p < 0.001) in the OSAS patients due to more hospitalization days (p < 0.001), consultations (p < 0.001), and cost for drugs (p < 0.05), particularly those for the cardiovascular system. In comparison to men, women consumed significantly more health-care resources (p < 0.001). OSAS patients < or = 65 of age years consumed 2.2-fold more health-care resources than control subjects (p < 0.001). Polysomnography findings and OSAS severity and body mass index (BMI) did not predict health-care utilization, using multivariate logistic regression analysis. Age > 65 (odds ratio [OR], 2.2; p < 0.04) and female gender (OR, 2.0; p < 0.05) were the leading elements predicting the most costly OSAS patients. Arbitrarily dividing the OSAS group by cost of health-care utilization, the upper 25% (n = 55) of patients who were the "most costly" consumed sevenfold more health-care resources than the lower 75% of the patients. This was due to higher comorbidity, ie, 10 to 30% more hypertension, ischemic heart disease, diabetes mellitus, and pulmonary disease.
Conclusions: OSAS patients are heavy users of health-care resources. Age > 65 years and female gender were the leading elements predicting the most costly OSAS patients, and not necessarily patients with a high BMI and classic OSAS severity indexes.