Using stated preference choice modelling to determine treatment preferences : investigating preferences for depression treatment

• Main Author: Vosper, Jane
• Published: University of Bristol 2010
• Subjects: 616.852706
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.559707

Background & Aim The PhD aimed to quantify preferences for depression treatments using Discrete Choice Experiments (DCEs). A secondary aim was to investigate the relationship between demographic variables, psychological variables and preferences. Methods & Analysis Two DCEs were designed and administered, investigating preferences for treatment of depression by: (1) a drug and (2) a physical activity intervention. The DCE designs were informed by focus groups and qualitative interviews. A best-worst scaling DCE was used for both studies. The physical activity intervention DCE was included in the baseline questionnaire of an RCT at the University of Bristol. Demographic information and psychological measures (including BDI scores) were analysed alongside the DCE. The drug treatment DCE was distributed as a postal questionnaire to a general population sample of 5000. Psychological measures of illness and medicine beliefs as well as the BDI were included in the questionnaire. Demographic data were also collected. Quantitative data were analysed primarily using conditional logistic regression. Results Results from the Physical activity DCE (for 152 patients) indicated that on average, patients particularly valued key aspects of the intervention, such as small goals, fitting activity into daily routine and having support over the intervention itself (being one giving choice of activity rather than exercise on prescription). Results from Drug study DCE (for 425 respondents) revealed a particular desire for no side-effects: large improvements in likely effectiveness are required to compensate respondents for non-zero risks of these. Heterogeneity analysis for both studies revealed effects of a number of demographic and psychological variables on preferences for attributes of depression treatment. Discussion Results of both studies are discussed in terms of their policy relevance and also from a methodological angle. The implications the results have on the use of DCEs in health care is considered.