OBJECTIVES: To derive a US value set for the EQ-5D-5L from the preferences of the US adult general population based on the EuroQol Group’s international protocol.

METHODS: Quota-based sampling for age, gender, ethnicity, and race was employed to obtain a sample representative of the general US adult population. Trained interviewers guided respondents through composite time trade-off (cTTO) and discrete choice (DC) valuation tasks using the EuroQol Valuation Technology (EQ-VT) platform in face-to-face interviews. For the cTTO, 86 health states were divided into 10 blocks, and each respondent valued 9 health states plus the worst health state 55555. For the DC tasks, 196 pairs of health states were grouped into 28 blocks, and respondents were presented with 7 choice tasks. Responses were excluded if, despite interviewer efforts, there was failure to comprehend the task. Main effects models were examined using a Tobit model (assuming heteroscedasticity and censoring at -1) for cTTO data, conditional logistic regression for DCE data, and combined data in hybrid regression models.

RESULTS: Of 1135 respondents completing interviews, 1103 were included in the analysis. Sociodemographic characteristics of the included respondents were similar to the US general population. The main effects hybrid model provided utility scores ranging from -0.70 (health state 55555) to 1.00 (health state 11111). The utility score closest to full health was 0.947 (health state 11211). The largest utility decrement for a dimension was pain/discomfort and the smallest was usual activities.

CONCLUSIONS: These initial results for a US-based EQ-5D-5L value set can help inform economic evaluations and decision-making in US health systems as well as facilitate cross-country comparisons of health preferences based on an international protocol. Further research will examine alternative model specifications as well as the possibility of combining data collected face to face and using online panels.