Objectives
Using Discrete Choice Experiments (DCEs), healthcare researchers can model population, patient, staff or provider preferences and choices. DCEs are rooted in Random Utility Theory (RUT), assuming that respondents are fully rational and process all provided information when choosing the alternative that maximizes their attribute-based utility. However, these behavioral premises do not always reflect observed behavior in choice experiments. The often complex, uncertain, and emotion-laden choice scenarios in health contexts can prompt heuristic-based simplifications, violating RUT assumptions. Although heuristics are recognized as behavioral violations relative to the normative RUT-based perspective, their impact is rarely modeled and limited guidance exists on how to do so.
Methods
Using three existing DCE studies, we demonstrate how researchers can identify and assess the impact of common RUT-violating heuristics in applied DCE studies. A structured interview guide and rating instrument were developed to identify heuristics a priori, based on clinician and researcher input. Latent-Class Models with restricted parameters were pre-specified to estimate heuristic impacts vis-à-vis preference heterogeneity.
Results
Our sensitivity analyses showed that up to 22% of the respondents likely applied particular heuristics. This impacted study results: For example, accounting for dominant decision-making (i.e., a class in which respondents ignored any attributes other than the risk of side effects), increased respondents’ average willingness-to-pay for antibiotics with a low contribution to antibiotic resistance by 15,63 Euros per treatment.
Conclusions
Ignoring heuristics that violate RUT assumptions biases preference estimates, marginal rates of substitution and uptake predictions in DCEs. We recommend assessing the likely presence and impact of heuristics in sensitivity analyses of future DCEs to ensure robustness and accuracy of results.
扫码关注我们
求助内容:
应助结果提醒方式:
