OBJECTIVE: To identify prescription patterns and costs of benign prostatic hyperplasia (BPH) therapy in real-life setting. METHODS: Records from the French primary care database Thales were screened to identify patients who initiated a therapy with either ?-blockers (A) [alfuzosin, tamsulosin, terazosin] or finasteride (F), or a lipido-sterolic extract of Serenoa repens (Px), between June 1996 and May 1997 and who were followed up for at least 12 months. Patients with a diagnosis of prostatic cancer were excluded. Endpoints were: one-year BPH related costs, concomitant medication costs and total health care costs according to the French National Health System. In order to adjust for differences of socio-economic and clinical characteristics between treatment groups, endpoints were analyzed using the propensity score method. RESULTS: 601 patients were identified in group A, 264 in group F and 315 in group Px. The median age was 67 years. Groups were unbalanced as to age (P<0.01), hypertension (P<0.05) and urological troubles (P<0.05): A were prescribed to younger patients with less hypertension, while F was prescribed to patients with less urological disturbances. Unadjusted annual total health care costs per patient were $ 1130 for group A, $ 1360 for group F and $ 1120 for group Px. Difference between A and Px remained after adjustment ($ 36), due to a lower price of Px and to reduced concomitant prescription costs. Difference between F and Px also remained after adjustment ($ 242), due to the previous reasons but also to a shorter treatment duration. CONCLUSION: in real-life clinical practice, annual total health care costs were lower when initiating a treatment for BPH with Serenoa repens than with finasteride and ?-blockers. “USE OF OUTCOMES RESEARCH (ECONOMIC, CLINICAL, HUMANISTIC) IN HEALTHCARE DECISION-MAKING”