LAUSR

PURPOSE This study aimed to evaluate the cost-effectiveness of osimertinib vs docetaxel and bevacizumab in third-line treatment in EGFR T790M resistance mutation advanced non-small cell lung cancer in China from the perspective of the health care system. METHODS To explore modeling uncertainty, 2 different model methods (a Markov model and a partitioned survival [PS] model) were developed to simulate costs and health outcomes during a lifetime. Both models consisted of 3 health states: progression-free survival, postprogression survival, and death. Efficacy and safety data of osimertinib vs docetaxel and bevacizumab in patients who had acquired EGFR T790M resistance mutation were derived from a key head-to-head clinical trial. Cost and utility values were derived from local charges, the literature, the China Drug Bidding Database, and patients' health care documents. Two scenario analyses and sensitivity analyses were performed to explore the robustness of the results. FINDINGS In the Markov model, compared with docetaxel and bevacizumab, osimertinib yielded 0.69 additional quality-adjusted life-years (QALYs) at an additional cost (in US dollars) of $17,311 for an incremental cost-utility ratio (ICUR) of $25,463 per QALY. In the PS model, osimertinib yielded an additional 0.69 QALYs with an incremental cost of $17,827 for an ICUR of $25,951 per QALY. From the Markov model, the ICUR was $29,416 per QALY in scenario 1 and $25,543 per QALY in scenario 2. From the PS model, the ICUR was $30,264 per QALY and $25,947 per QALY for scenarios 1 and 2, respectively. In the probabilistic sensitivity analysis, osimertinib treatment had a 21%-63% probability of being cost-effective at a willingness-to-pay threshold of $9777 to $29,330 per QALY (1-3 times the gross domestic product per capita). IMPLICATIONS The findings from the present analysis suggest that osimertinib could be cost-effective vs docetaxel and bevacizumab in third-line treatment in EGFR T790M resistance mutation advanced non-small cell lung cancer in China.