OBJECTIVES: Neuroendocrine tumors (NET) are relatively rare, usually slow-growing malignant tumors. So far there are no data on the patient preferences regarding the therapy of NET. This empirical study aimed at the elicitation of patient preferences in the drug treatment of NET. METHODS: Based on qualitative patient interviews and an AHP, patient-relevant outcomes were analyzed and weighted using a discrete choice experiment (DCE). Six patient-relevant attributes were included. Patients were recruited with the help of a NET support group. For the DCE an experimental design (3*3 + 3*6 Design) was created using Ngene. The selected design consisted of 84 choices, which were divided into 7 blocks. Participants were randomly assigned to these blocks. The analysis of DCE included random parameter logit models, conditional logit models, and latent class models. RESULTS: N = 275 NET-patients (51.6% female, mean age 58.4 years) participated. Under the chosen decision model the preference analysis  within the random parameter logit model, taking into account the 95% confidence interval, showed a predominance for the attribute “overall survival" (coef: 1,568). The attributes "response to treatment" (coef: 0,617) and "stabilization of tumor growth" (coef: 0,547) followed. However, the side effects "nausea/vomiting" and "diarrhea" are considered of relatively equal importance (coef: 0,544 / 0,413). The analysis of possible subgroup differences using latent class analysis revealed three preference patterns. The adverse event "occurrence of abdominal pain" had a significant effect in just one class. CONCLUSIONS: The results thus provide evidence about how much influence a treatment capacity has on therapeutic decision.  The preference measurement showed that "overall survival" has the strongest influence on the therapeutic decision. The preference analysis also made it clear that the participants weight the outcome attributes higher than the side effects. Thus it becomes clear that a mono-criterial decision would not fully reflect the patient benefits.