What Matters Most to Patients with NSCLC When Choosing Anticancer Treatments in China? A Best–Worst Scaling Study

Jingyi Qiao,1,2 Shimeng Liu,1,2 Shiyi Bao,1– 3 Ying Tao,1,2 Jiaxin Zhao,1,2 Yingyao Chen1,2 1School of Public Health, Fudan University, Shanghai, People’s Republic of China; 2National Health Commission Key Laboratory of Health Technology Assessment (Fudan University), Shanghai, People’s Republic of China; 3Department of Teaching Management, Nanjing Medical University, Nanjing, Jiangsu, People’s Republic of ChinaCorrespondence: Shimeng Liu, School of Public Health, Fudan University, 130 Dongan Road, Xuhui, Shanghai, 200032, People’s Republic of China, Tel +8613046006196, Email [email protected] Yingyao Chen, School of Public Health, Fudan University, 130 Dongan Road, Xuhui, Shanghai, 200032, People’s Republic of China, Tel +8613564508981, Email [email protected]: This study aimed to evaluate treatment preferences for anticancer drugs among patients with non–small cell lung cancer (NSCLC), in order to provide evidence not only for clinical prescribing decisions but also for broader applications such as health insurance reimbursement and policy-making.Methods: A Best–Worst Scaling object‐case (BWS‐1) questionnaire was administered to NSCLC patients. Thirteen choice sets, each containing four attributes, were generated using Balanced Incomplete Block Design. Respondents were asked to identify the most and least important attribute within each set. The attributes include overall survival (OS), disease control rate, progression-free survival (PFS), dyspnea, pain, objective response, hemoptysis, fever, nausea and vomiting, monthly out-of-pocket expenditure, cough, mode of administration, fatigue. Relative attribute importance and preference heterogeneity were estimated using counting analysis and a conditional logit model (CLM).Results: A total of 102 NSCLC patients were enrolled. High concordance was observed between counting analysis and conditional logit results. In the CLM, OS (β=3.537, P< 0.01), disease control rate (β=2.025, P< 0.01), and PFS (β=1.574, P< 0.01) showed the strongest positive preferences. Fatigue (β=– 1.158, P< 0.01) and mode of administration (β=– 0.600, P< 0.01) were associated with lower relative importance. Cough showed a negative coefficient but was not statistically significant (β=– 0.185, P=0.11).Conclusion: This study is the first to explore the NSCLC patient medication preference using a BWS-1. The findings suggest OS, disease control rate, and PFS are prioritized by patients when choosing treatment regimens.Keywords: patient preferences, best–worst scaling, non–small cell lung cancer, drug therapy