Presentation (CTI)

OBJECTIVES: Advances in precision oncology, including biomarker testing, targeted therapies, immunotherapies and other novel therapies, have increased the complexity of decision-making. Decisions often occur with sequential pathways, where initial choices influence future treatment opportunities. Existing preference studies provide limited insight into how individuals value uncertainty in future treatment pathways. As behavioural economics and psychology suggest individuals are often risk averse, better methods are needed to capture how uncertainty influences preferences in complex oncology decisions. This research aims to examine how uncertainty is represented in oncology preference studies and develop a framework that captures preferences for pathway-related uncertainty.
METHODS: A targeted review of 164 published oncology preference studies assessed how uncertainty and treatment sequencing have been represented. Based on the findings, a pathway-based choice framework and integrated utility model were developed to capture sequential treatment decisions in which downstream treatment eligibility is uncertain at initiation. The model separates expected outcomes from pathway-related uncertainty and applies a probability weighting function to reflect subjective perceptions of uncertainty. Using preference data, this framework can estimate utility weights for decision-support tools comparing oncology treatment pathways.
RESULTS: Uncertainty was predominantly represented through probabilistic attributes attached to individual treatment profiles, with limited consideration of sequencing or pathway structure. The framework enables estimation of preferences for future treatment opportunities, pathway-dependent outcomes, and uncertainty beyond expected clinical outcomes. The illustrative application showed that pathway-related uncertainty can influence choices, with individuals discounting benefits delivered through uncertain downstream pathways. Varying risk aversion demonstrated that accepting uncertainty may require compensation through greater expected benefits or lower risks.
CONCLUSIONS: Current preference methods provide limited insight into how individuals value uncertainty. This framework offers a practical approach for studying complex oncology decisions and demonstrates that not accounting for risk preferences may distort conclusions about benefit-risk acceptability and predicted treatment preferences.