PP4 - Consumer Preferences for Precision Medicine Technologies: Evidence from A Discrete Choice Experiment

Abstract

Challenges to precision medicine implementation include a lack of evidence on preference-based utility in context to clinical validity and uncertain health outcomes. We address these challenges by eliciting patients’ preferences for the types of health and non-health outcomes arising from genomic testing. We used a sequential mixed methods approach. Qualitative research (n=20 patients) identified the following attributes: genetic test type (disease risk, treatment prediction), probability of having a biomarker, expert agreement on changing care based on genomics, quality of life gains, life expectancy gains with statistical uncertainty, and cost. Discrete-choice methodology enumerated preferences with a view to predict uptake and estimate willingness-to-pay (WTP). The population was a probability-based sample of the U.S. public. Error component mixed logit modeling enabled analysis of realistic substitution patterns. We validate our results by predicting the uptake of a 21-gene breast cancer recurrence risk test and comparing that with real-world estimates. 1124 representative members of the public who had recently seen a doctor completed the survey. Respondent’s had disutility if there was low medical expert agreement, if there was no or little change in quality or quantity-of-life, and as out-of-pocket costs increased. Respondents distinguished between high and low degrees of statistical uncertainty. Average WTP to decrease uncertainty was between $147 and $223, with higher WTP associated with resolving uncertainty for life-year gains of 3 months and lower WTP for life-year gains of 4 years. The predicted uptake of the recurrence risk test was 21% in 2004. As clinical utility was demonstrated, predicted uptake was 45%. The correlation between our predictions and published uptake was 0.90. Preferences for genomic testing are significantly influenced by medical expert agreement and the degree of scientific uncertainty surrounding outcomes. These findings can be used to project a trajectory of demand as well as inform translational research.