Developing attributes for discrete choice experiments in health: a systematic literature review and case study of alcohol misuse interventions

Discrete choice experiments (DCEs) become increasingly popular to value outcomes for health economic studies and gradually gain acceptance as an input into policy decisions. Developing attributes is a key aspect for the design of DCEs, as their results may misguide decision-makers if they are based on an inappropriate set of attributes. However, the area lacks guidance, and current health-related DCE studies vary considerably in their methods of attribute development, with the consequent danger of providing an unreliable input for policy decisions. The aim of this article is to inform the progress toward a more systematic approach to attribute development for DCE studies in health. A systematic review of the published health-related DCE literature was conducted to lay the foundations for a generic framework which was tested in a case study of alcohol misuse interventions. Four stages of a general attribute development process emerged: (i) raw data collection; (ii) data reduction; (iii) removing inappropriate attributes; and (iv) wording. The case study compared and contrasted a qualitative and mixed-methods approach for the development of attributes for DCEs in the area of alcohol misuse interventions. This article provides a reference point for the design of future DCE experiments in health.