OBJECTIVES: There is no EQ-5D-5L value set for any Central and Eastern European country. Our objective was to derive a tariff for the EQ-5D-5L in Poland.

METHODS: Quota sampling was used to achieve a representative sample of the Polish population with regard to age, sex, education, geographical region, and the size of locality. The study design followed the EQ-VT protocol (v2.0). Fifteen trained professional interviewers performed computer-assisted face-to-face interviews between June and October 2016. Each respondent valued ten health states using composite time trade-off (cTTO) and completed seven discrete choice experiment (DCE) tasks. Quality control was performed according to EQ-VT rules. All the flagged interviews were removed. The hybrid model was estimated in Bayesian setting, with non-informative priors, using JAGS/R environment.

RESULTS: Data from 1252 respondents (52.5% females, age 18 - 91 years) were available. In the estimation, we used: 11,480 TTO valuations and 8,764 DCE pairs. In TTO, in 10.7% experiments, the time was not traded, and eight respondents did not trade for any state. The average utility of 55555 state in TTO amounted to -0.433. In the final model, the estimated decrease of utility for level five amounts to: 0.341 (Mobility), 0.312 (Self-Care), 0.237 (Usual Activities), 0.592 (Pain/Discomfort), and 0.249 (Anxiety/Depression). In the final value set we got u(22222) = 0.889, u(33333) = 0.832, u(44444) = 0.375, u(55555) = -0.731 (compare to u(33333) = -0.523 in the Polish EQ-5D-3L tariff).

CONCLUSIONS: New value set is available and should be used in health technology assessment in Poland. It may be considered as a second best choice in other CEE countries, lacking their own directly measured or cross-walk value sets. The new tariff does not introduce revolutionary changes as compared to the previous EQ-5D-3L version but offers greater sensitivity to subtle health state changes thanks to five levels.