OBJECTIVES: To understand how the communication of risk in a discrete choice experiment (DCE) affects respondents’ choices.
METHODS: An online pilot DCE was designed to understand the preferences of female members of the public (recruited via an internet panel provider) for a breast screening programme described by three attributes (probability of detecting a cancer, risk of unnecessary treatment, and outof-pocket cost) each with four levels. Two versions were used that presented the risk attributes (probability of detecting a cancer and risk of unnecessary treatment per 100 women screened) as: (A) a percentage or (B) a percentage and risk image (icon array). The DCE was blocked into four surveys, each containing 10 choice sets. The design, generated using Ngene, included an internal validity test through the inclusion of a dominant choice set. The DCE data were analysed using conditional logit models.
RESULTS: 62 women completed the DCE (31 for each version A and B); all were currently eligible for screening under the current NHS programme. All coefficients, but no interactions, were significant and had the expected signs. Of the respondents who received the percentages only version, almost 20% failed the validity test (compared to only 3% of those who received the risk image). Probability of detecting a cancer was the most important attribute. Willingness-to-pay (WTP) for an additional cancer detected was £175 for respondents presented with the risk image, compared to only £152 in the percentages only group. Similarly, WTP to avoid an unnecessary treatment was £59 for respondents presented with the risk image compared to only £19 in the percentages only group.
CONCLUSIONS: This pilot study highlights the impact attribute framing can have on respondents’ choices in a DCE. The use of risk images also resulted in fewer “irrational” responses implying respondents had a greater understanding of the task.
