OBJECTIVE: to estimate the costs and effects of pamidronate vs placebo in prevention of bone complications in metastastic breast cancer in Poland. METHODS: The model for the Polish health-care context was developed, based on the use of clinical data from literature and local data on health-care resource utilisation and unit cost. Only direct medical costs were analysed. The perspective of health-care payers and time horizon of 18 months were taken. In the cost-effectiveness analysis the primary clinical outcomes for measuring success in the analysis were the avoided skeletal related events (SRE), which were defined as pathological fractures, irradiation and episodes of hypercalcaemia. In the cost-utility analysis QALY values measured in the analysis were patients preferences for alternative health outcomes, reported in the randomised trial. The cost-effectiveness treshold, calculated on the basis of cost of 1-year haemodialysis treatment (62 400 PLN, 1 USD=4,5 PLN), was suggested to asses if pamidronate therapy should be adopt. The one-way sensitivity analysis and treshold analysis were performed. RESULTS: The costs of avoiding of SRE for pamidronate vs placebo was 3420 PLN (taking into consideration lowest costs of treatment of pathological fractures) or, taking into consideration highest costs of treatment of pathological fractures, pamidronate treatment was dominant program (129 SRE avoided and 42 546 PLN/100patients/year saved). The cost/QALY gained was 45 021 PLN or pamidronate was the dominant program (9,8 QALY gained and 42 546 PLN/100 patients/year saved). The results were most sensitive to the costs of treatment of pathological fractures and number of SRE avoided. CONCLUSION: Comparing this outcomes with suggested threshold for cost-effectiveness it was concluded, that prophylactic pamidronate therapy for advanced breast cancer offers substantial benefit at reasonable cost to the Polish health care system. By assumption, that all pathological fractures would be treated with most costly methods – pamidronate might be dominant program.