OBJECTIVES: As research aims to provide cures for cancer, this study was performed to determine the characteristics driving the value that the general population would attach to curative treatments for different cancer types.

METHODS: Following a literature search and in-depth interviews with members of the general public, five attributes important to cancer treatment were identified, with 4 levels each: incidence of the cancer, patient age at diagnosis, morbidity associated with disease and treatment, patient comorbidities and cancer mortality with treatment. The discrete choice experiment used an efficient sequential design of 4 blocks with 6 choice sets each. Two additional choice sets involving dominant alternatives were presented to identify participants who misunderstood the exercise. Respondents with and without experience of cancer were recruited through a panel of the French general population. A mixed multinomial logit model was used to determine relative preferences of each attribute.

RESULTS: There were 1041 respondents with an average age of 51.8, ranging from 18-84 years of age. 536 (51.5%) of respondents were male. 249 respondents had cancer experience. The model was applied to responses from 870 (83.6%) of respondents who selected both dominant alternatives. The most important attributes were age, with marginal utilities of: -3.8, -1.9 and -0.3 for >65, 45-64, and 20-44 versus <20 years of age, respectively, mortality (3.7, 2.2, and 2.2 for 9/10, 8/10 and 5/10 versus 2/10 mortality) and incidence (2.6, 1.7, and 1.1 for 10/100, 5/100, and 1/100 versus 1/1000). Comorbidities had a smaller impact (1.8, 0.7, and 0.9 for 50%, 30%, 10% versus no serious comorbidities), whereas the preferences for treatment morbidity was inconsistent (1.0, -0.1 and -1.0 for light, severe and moderate versus no treatment morbidity).

CONCLUSIONS: The stated preferences demonstrate the age of diagnosis, incidence and mortality of cancer have the greatest impact on preference for making curative cancer treatments available.