OBJECTIVES: Health-related quality of life assessments are increasingly used in clinical decision-making. A phase III trial of pemetrexed versus docetaxel in second-line therapy for advanced non-small cell lung cancer (NSCLC) showed similar efficacy; however, key safety benefits were reported with pemetrexed. Following a pilot study in the United Kingdom (UK), an expanded study in the UK and France was conducted to determine patient value associated with individual toxicity profiles. METHODS: A discrete choice conjoint analysis was used to quantify patient preference and willingness-to-pay for chemotherapy. A review of trial data, along with expert opinion, identified clinically meaningful toxicities that were statistically significantly different. Levels of risk of: febrile neutropenia (requiring hospitalisation) and nausea, neuropathy, arthralgia/myalgia, alopecia (all grades), were evaluated in the pilot. Following the pilot, arthralgia/myalgia was removed and a sample size of 70 pre-treated NSCLC patients per country was calculated. Patients considered unique, randomly generated sets of 10 pair-wise choice scenarios representing levels for toxicity attributes plus cost, designed to elicit trade-offs. Logistic regression analysis was applied to the stated scenario preferences against the individual attribute levels. RESULTS: In the expanded study, patients (N=140) were predominantly male, mean age 61 years, and 60% Stage III, which is comparable to the pilot. Pemetrexed would be accepted in preference to docetaxel at zero cost, with a probability of 0.81 in the UK and 0.90 in France. The probability of choosing pemetrexed over docetaxel decreases with increasing cost; however, patient preference remains strong at 0.70 in the UK and 0.85 in France with a cost per cycle of £400 and 2500€, respectively. CONCLUSIONS: NSCLC patients showed clear preference for the enhanced toxicity profile with pemetrexed, which translates to valuable quality of life gains in the second-line setting. These data provide sensitive strength of preference measures. Additional country studies are planned.