OBJECTIVES: Personalized diabetes management can be seen as a kind of "standard process" that provides both physicians and patients with a guideline. However, patients living with type II diabetes differ in their preferences and values. How can these differences be analyzed and described?

METHODS: A discrete choice experiment (DCE) was applied to identify preferences for a personalized diabetes management in patients with type II diabetes mellitus. Six attributes were included. The statistical data analysis used random effect logit models.

RESULTS: N=227 patients (66.1% male) were included. The latent class analysis shows three heterogeneous classes, each showing clearly different weights of the therapeutic properties. This results in a clear threefolding: for 1/3 of the respondents the change of the long-term blood sugar (HbA1c value) is the top objective. Another third is solely interested in the short-term effectiveness of the therapy in the sense of the occurrence of severe hypoglycemia per year. The last third of the interviewees finally focuses on the follow-up with regard to cardiovascular events. Overall, there were five structural and personality traits which have an influence on the probability of the class membership.

CONCLUSIONS: Overall, the effectiveness of a care program is the most important from the perspective of the patient and avoiding severe hypoglycemia has the greatest influence on the choice. However, the analysis of possible subgroup differences by means of latent class analysis, found that three preference patterns exist within the sample that vary significantly. This raises the question on how to best tailor personalized management approaches to the preferences of specific populations in order to improve the acceptability of and adherence to care programs.