Physicians’ risk and time preferences and their willingness to detect and disclose genomic secondary findings: A multiple price list experiment

To quantify physicians’ risk and time preferences and explore the association between the preferences and their willingness to detect and disclose secondary findings (SFs) derived from genome-scale sequencing. We designed a web-based survey incorporating a multiple price list (MPL) as the instrument for risk and time preference measurement. The estimation was under the expected utility theory (EUT) and rank dependent utility (RDU) frameworks, respectively. The isoelastic and power function utility models were applied. We received responses from 87 physicians, among whom 46 completed the questionnaire (a completion rate of 52.9%). We observed positive risk-aversion coefficients under EUT (0.33, 95% CI 0.15–0.51) and RDU (0.51, 95% CI 0.32–0.71), suggesting that physicians were generally risk-averse. Respondents were likely to underestimate probabilities of low to moderate levels, and slightly overestimate high-level probabilities. Physicians supporting the detection and disclosure of SFs had a larger risk-aversion coefficient and a smaller discounting parameter than the non-supporters, suggesting that they were more risk-averse and discounted future utility less. Assuming heterogeneous risk perception, we found respondents underestimated low/moderate risk and slightly overestimated high risk. This study indicates that physicians who are risk-averse and discount future utility slightly are willing to detect and return SFs. The findings contribute to the debate surrounding SF disclosure and generate implications for shared decision-making in clinical genome-scale sequencing. • Physicians' risk and time preferences influenced their decisions on genomic SFs disclosure. • Physicians favouring SFs disclosure were more risk-averse than those opposing such detection and disclosure. • Physicians with lower time discounting rates showed greater willingness to return SFs to patients. • Physicians tended to underestimate low to moderate levels and slightly overestimate higher probabilities.