Accounting for Preference Heterogeneity in Discrete Choice Experiments

A new publication co-authored by OHE’s David Mott provides an overview of the state of practice for accounting for preference heterogeneity in discrete choice experiments. The publication, which was a collaboration between members of the ISPOR Health Preference Research Special Interest Group, reports the results of a literature review and an expert survey.

What are discrete choice experiments (DCEs)?

DCEs are a stated preference method that can be used to elicit people’s preferences for a wide range of different topics by examining the trade-offs that they are willing to make. In health, DCEs have been used to explore a wide range of different questions (Soekhai et al., 2019). For example, OHE researchers have recently used the method to explore how people with experience of infertility value different aspects of Assistive Reproductive Therapy (Skedgel et al., 2021). Another OHE-led study compared the preferences of adults and adolescents for different EQ-5D-Y health states (Mott et al., 2021). OHE researchers also have an ongoing study with the Acute Leukemia Advocates Network to explore the preferences of people with acute leukaemia for different treatment characteristics.

What is preference heterogeneity?

Put simply, preference heterogeneity means that peoples’ preferences differ. Baseline models used to analyse DCE data actually assume the opposite – that preferences do not differ between respondents – which is unrealistic. It is therefore important for researchers to capture preference heterogeneity in their analyses to avoid biased results. An important distinction here is the difference between explained and unexplained heterogeneity. Explained heterogeneity is where preferences differ by some observable characteristics, such as the age or gender of respondents. An examination of explained heterogeneity can be an important aspect of a preference study – as noted in the recently published recommendations of the IMI PREFER project. In contrast, unexplained heterogeneity is not explained by observable characteristics. The methods for accounting for each type of heterogeneity differ.

How has preference heterogeneity been accounted for in practice?

The literature review found that many health preference studies have attempted to account for explained or unexplained preference heterogeneity, and that attempts to account for it appear to have increased over time. The former has involved the inclusion of interaction terms in models, whereas the latter has involved the use of more complex modelling approaches, such as latent class and mixed logit models. However, the expert survey indicated that health preference researchers are concerned about the rising complexity of models, and that there is some confusion and disagreement around the best approaches to use.

What were the implications of the study?

Preference heterogeneity was widely deemed important to account for, and many researchers have done so in practice. However, it was also felt that more pragmatic guidance would be beneficial to assist health preference researchers in future.

The manuscript was published in Value in Health and can be found here.

If you’d like to find out more about DCEs and how they could be used to address your research questions, please get in touch with Chris Skedgel or David Mott.

Citation

Vass, C., Boeri, M., Karim, S., Marshall, D., Craig, B., Ho, K.-A., Mott, D. J., Ngorsuraches, S., Badawy, S. M., Mühlbacher, A., Gonzalez, J. M., & Heidenreich, S. (2022). Accounting for Preference Heterogeneity in Discrete-Choice Experiments: An ISPOR Special Interest Group Report. Value in Health, 25(5), 685–694.

Related research

Mott, D. J. (2018). Incorporating Quantitative Patient Preference Data into Healthcare Decision Making Processes: Is HTA Falling Behind? The Patient, 11(3), 249–252.

Mott, D. J., Chami, N., & Tervonen, T. (2020). Reporting Quality of Marginal Rates of Substitution in Discrete Choice Experiments That Elicit Patient Preferences. Value in Health, 23(8), 979–984.

Mott, D. J., Hampson, G., Llewelyn, M. J., Mestre-Ferrandiz, J., & Hopkins, M. M. (2020). A Multinational European Study of Patient Preferences for Novel Diagnostics to Manage Antimicrobial Resistance. Applied Health Economics and Health Policy, 18(1), 69–79.

Skedgel, C., Wailoo, A., & Akehurst, R. (2015). Societal preferences for distributive justice in the allocation of health care resources: a latent class discrete choice experiment. Medical Decision Making, 35(1), 94–105.