OBJECTIVES: Periodontal diseases are treated with different approaches. The treatments available include simple oral hygiene practices, laser therapy, periodontal surgery and non-surgical practices. The aim of the treatments is reducing destruction of the periodontium, stopping bone loss and preventing tooth loss. However, it remained unclear which of the patient-relevant outcomes determine the benefit perceived by patients. The aim of this preference study was to quantify subgroup-specific preferences and characteristics in the sample.

METHODS: A DCE with six attributes were conducted as an online survey. Attributes represented morbidity, side effects and mode of administration of a periodontal disease treatment. Respondents were asked to choose between two treatment alternatives and the status quo alternative. Choosing the status quo meant opting out or deciding against one of the offered treatment alternatives and accepting possible consequences such as severe tooth loss of a non-treatment. A latent class model (LCM) was used to analyze the DCE data.

RESULTS: The final dataset contained N=610 respondents. A three-class model was used for the analysis. Respondents in class 1 were averse to tooth loosening and tooth loss. By far, the remaining attributes were of less important in this class. Class 2 respondents also preferred no tooth lost but were also averse to pain during the treatment and pain in the everyday life. For respondents in class 3 it was most important not to have pain during the treatment. All respondents consistently preferred no pain and mode of administration. Preference heterogeneity was especially observed in the magnitude of attribute importance.

CONCLUSIONS: Latent class models identify preference heterogeneity among respondents and therefore can help decision makers optimize health care delivery.