Discrete Choice Experiments in Public Health

Abstract

One approach to improve public health is to implement preventive programs that have been proven effective and cost-effective. For any preventive program to be successful, it is of paramount importance that a large majority of the target population participates. Unfortunately, it is not self-evident that high participation rates will be attained once preventive programs become available. Identification of characteristics that influence individuals’ participation decisions enable understanding why people decide (not) to participate in effective preventive programs. Evidently, some characteristics, such as high cost, may hamper participation, while others, such as high effectiveness, may have positive influence on participation rates. Discrete Choice Experiments (DCEs) are relatively new in the field of public health but have acquired an established position for studying decision-making behavior in marketing and transportation. The DCE methodology is a stated preference method that requires respondents to choose between different sets of program characteristics. This method enables researchers to determine preferred program characteristics, to quantify the importance of those program characteristics relative to each other and to determine potential uptake rates. In this dissertation, we explored if DCEs can be used to study individuals’ preferences concerning characteristics of specific preventive programs. In the first part of this thesis, we used DCEs to determine individuals’ preferences for three specific preventive interventions in three different public health domains, i.e. healthy lifestyle, vaccination and screening. We concluded that, all associated measures such as the relative importance of attributes, trade-offs between attributes, willingness-to-pay and potential participation rates could be estimated. Those results provide specific recommendations for program developers and policy makers to facilitate the development, implementation and marketing of effective preventive interventions. In part two of this thesis, we studied three aspects regarding the validity of the preferences that were measured by a public health related DCE. We concluded that the vast majority of the population is very well capable of completing public health related DCEs and the results could be used to predict engaging behavior. However, DCEs in their current form might not measure preferences validly among all members of the target population. Therefore, the extent to which alternative questioning methods, such as interviews or online interfaces, may improve the ability of older respondents and respondents with a lower educational level and/or health literacy to complete a DCE accurately should be verified. In the third part of this thesis, we explored if and concluded that presenting choice tasks in different ways influences respondents’ decision-making behavior and DCE study outcomes. Based on additional research the DCE method should be adapted and existing guidelines for best practice DCEs should be complemented. In conclusion, DCEs can be used to measure individuals’ preferences for public health interventions. There are several options to improve this method specifically for the public health setting. By combining additional research, knowledge from other scientific disciplines, well-designed experiments and advanced analytical models, our capacity to measure individuals’ preferences and to understand individuals’ decisions will increase further. Therefore, DCEs are a promising method to contribute to the improvement of the overall public health.