Evaluating Patient Preferences for Clinical Trial Endpoints in Early-Stage Cancer: A Discrete Choice Experiment in Canada

To quantify preferences and trade-offs for non-overall survival (OS) endpoints among Canadian adults treated for early-stage cancers, participants completed ten choice sets in a discrete choice experiment. Standard-of-care (SoC) was compared to new treatment profiles defined by five attributes: five-year OS, two-year disease advancement, pathological complete response (pCR), and short- and long-term side effects. SoC attribute levels remained constant across respondents, except OS, which varied. For the new treatment, all attributes except OS (fixed as unknown) varied. Data were analyzed using logistic regression and presented as odds ratios (ORs) with 95% confidence intervals (CIs). Among 103 adults treated for early-stage gastrointestinal (n = 40), breast (n = 32), or lung (n = 31) cancer, median age was 59 (Q1, Q3: 43, 75) years; 46.6% were female. Each 25% decrease in SoC OS was associated with higher odds of choosing the new treatment (OR 3.49; 95% CI 2.82–4.31; p < 0.01). Reductions in disease advancement (OR 1.55; 1.26–1.91; p < 0.01) and mild or no short-term side effects versus severe (OR 6.67; 4.35–10.00) significantly increased selection odds; long-term side effects and pCR showed modest, non-significant influence. When OS data were available for SoC, OS strongly influenced decisions; without OS data for new treatments, participants prioritized disease control and short-term tolerability.