OBJECTIVES: To evaluate cost-utility of docetaxel (75 mg/m2 every 3 weeks) compared with best supportive care (BSC) or pemetrexed (500 mg/m2 every 3 weeks) as second-line therapy in advanced non-small cell lung cancer in the Polish setting. METHODS: A cost-utility approach was adopted, evaluating total direct costs incurred by the Polish National Health Fund (NHF), life years gained (LYG) and quality-adjusted life years (QALY). A micro-simulation Markov model was used to estimate utilities and costs of treatment. The first course of chemotherapy was assumed as the starting-time point of the model. Simulation was terminated at the time of the patient’s death or after two years following start of the treatment. It was assumed that patients would undergo 4 cycles of chemotherapy or less if progression occurred during treatment. In case of progression of disease chemotherapy would be terminated and patients would receive palliative care until death. Transition probabilities between health states were calculated based on a systematic review of RCTs. Health states utilities were taken from published literature. Costs were taken from NHF catalogue. Probabilistic sensitivity analysis was performed in order to estimate the probability that docetaxel was cost effective in Polish settings where threshold is about 91,000 polish zloty (PLN). RESULTS: Incremental costs for docetaxel compared with BSC was PLN57,498 per LYG and PLN105,956 per QALY. In 2-years time horizon docetaxel was PLN13,582 less costly than pemetrexed (CI95%: 8,224; 17,704). The probability of docetaxel cost effectiveness over BSC was 99.71% for LYG and 10.11% for QALY. The probability of docetaxel cost effectiveness over pemetrexed was 99.78% for LYG and 100% for QALY. CONCLUSIONS: Docetaxel seems to be cost effective in comparison with BSC and pemetrexed in the Polish setting.