OBJECTIVES: Study participants may inconsistently respond to discrete choice experiment (DCE) questions, especially when presented with many choice tasks. Numeracy skill (familiarity with numbers) is believed to exert an influence on decisions in many contexts, especially when the attributes of the alternatives are expressed numerically. The objective of this study was to explore the link between numeracy skills and choice consistency in a DCE study of kidney transplantation.

METHODS: A cross-sectional DCE was conducted to assess time and risk preferences for kidney transplantation in Italy. Three-item standardized numeracy questions were also included. Data were collected from 250 patients on the waiting list for a kidney transplant in Italy. A potential effect of numeracy skill on choice consistency was investigated using heteroscedastic multinomial and mixed logit models that allow error variance to depend on patients’ numeracy skill.

RESULTS: Patients showed a low level of numeracy: only 25% of respondents were able to answer all three questions correctly, while most patients (44%) answered 2/3 correctly. A positive and statistically significant effect of numeracy skill on choice consistency was found. Taking choice consistency into account resulted in different willingness to wait (WTW) estimates for changes in the attributes. For example, on average, patients were willing to wait 32 months longer for a standard infectious risk kidney compared to augmented risk, reduced to 26 months for those who answered the three numeracy questions correctly. These results suggested that pooling lower and higher numeric patients together might lead to overestimated WTW. Also, these relationships were statistically significant when the socio-demographic characteristics of these patients were controlled.

CONCLUSIONS: The study highlighted the importance of incorporating short numeracy questions in DCEs to determine choice consistency. Further research in larger samples, diverse populations, and disease contexts is needed to substantiate these results.