Background Chronic diseases are a significant health concern in Australia. Understanding community preferences for health screening services is vital for enhancing service delivery. We conducted a study to determine community preferences for health screening services for chronic diseases in Australia using a discrete choice experiment (DCE). This paper aims to present the development of the final DCE design using priors estimated from a survey. Methods A DCE was conducted in Australia. An online survey was administered to a general Australian population over 18. The final attribute list of five attributes with three levels each was designed. A D-efficient design with 30 pair-wise choice tasks was developed using a fractional factorial design. A pre-test was conducted to assess comprehension and understanding of the online DCE survey. The pilot survey aimed to compute priors (i.e. coefficients) associated with attributes. A multinomial logit model was used to analyse the pilot DCE data. Results The survey included 30 choice tasks in three blocks, with 119 participants responding. The best DCE design was selected based on D-error, with a lower D-error indicating the most efficient design. The pilot survey results indicated a strong preference for highly accurate screening tests, with coefficients for 85% and 95% accuracy being positive. Coefficients estimated from the pilot survey were used as priors to design the DCE choice tasks for the main survey. The final DCE design showed a notable improvement in the attribute level overlap compared to the design used for the pilot survey. Conclusions A rigorous approach was taken to develop a DCE survey that could effectively determine the preferences of the community for health screening services. The resulting DCE design consisted of 30 choice tasks presented in pairs and was deemed efficient enough to gather comprehensive information in the final survey.
Read full abstract