OBJECTIVES: The study aimed to quantify treatment preferences of women with advanced or metastatic breast cancer in Germany.

METHODS: Patients completed a DCE including 5 attributes: ‘Time to Progression(TTP)/Overall Survival(OS)’, ‘Response to Treatment’, ‘Level of Fatigue’ and ‘Risk of Infection’. Treatment choices were analysed using a latent class analysis to estimate preference heterogeneity.

RESULTS: N=233 patients advanced or metastatic breast cancer completed the survey. Four subpopulations were found. For the attributes ‘Response’, ‘Infection’, and ‘Fatigue’ better levels were preferred over worse levels across all classes. Regarding ‘infection’, only in class 2 the middle level ‘moderate’ yielded a lower value (coef.: -0.29) than the worst level ‘high (coef.: -0.06). The linear preference order is found for the levels of the attribute ‘TTP/OS’ in classes 1 and 4, e.g. ‘5 months without progression at 50 months OS’ is less preferred than ‘25 months without progression at 30 months OS’. So 25 months without progression is rated higher than 50 months OS. In contrast, in classes 2 and 3 the level ‘5 months without progression at 50 months OS’ is more preferred than ‘25 months without progression at 30 months OS’. When looking at the two dimensions ‘5 months’ and ‘25 months’ without progression separately, there is a clear ascending order of preference.

CONCLUSIONS: No significant differences in preferences could be determined based upon age, even when an accumulation of one age group was observed in one class. Within these classes, different weightings for the treatment characteristics can be seen: in one class, as assumed, the outcome is most important. For another group, ‘Fatigue’ and TTP/OS are nearly on the same level of importance. The third group bases its assessment mainly on ‘Response’ and ‘Fatigue’. All other treatment characteristics are practically unimportant. Surprisingly, one class of respondents seems to have based their choices solely on fatigue.