Comparison of different valuation methods for population health status measured by the EQ-5D in three European...

The purpose of this study was to analyze and compare different valuation methods for population health status measured by the EuroQol-5D (EQ-5D) in three European countries.

A representative survey of the noninstitutionalized population aged 18 and above was conducted in three European countries (Germany, The Netherlands, and Spain). A total of 11,932 respondents were interviewed using the EQ-5D self-classifier. Health state values based on community preferences (EQ-5D index) were calculated for each country using four different value sets: national value sets based on the time trade-off (TTO) and the visual analogue scale (VAS), the UK TTO-based value set and the European VAS-based value set. Linear regression analysis was conducted to evaluate the factors associated with different EQ-5D index scores depending on the value set used. Loss of quality-adjusted life-years (QALYs) was calculated for each country using the four value sets by multiplying the age and gender-specific values with the respective population size.

In all countries, means of all EQ-5D index scores were higher for men than women, and decreased with age. Index scores calculated using the national value set based on TTO were higher than those calculated using the UK TTO-based value set and, also, slightly higher than those calculated using the European VAS-based value set or the national value set based on the VAS. The mean loss of QALYs estimated for Germany per inhabitant varied between 0.062 (national value set based on TTO) and 0.094 (European VAS-based value set). In The Netherlands, the mean loss of QALYs per inhabitant ranged from 0.090 (national value set based on TTO) to 0.125 (national value set based on VAS). In Spain, the mean loss of QALYs per inhabitant ranged between 0.072 (national value set based on TTO) and 0.085 (European VAS-based value set).

In general, the differences among countries and valuations were rather small; nevertheless, some important variations should be taken into account while applying different valuation methods to the EQ-5D descriptive system. The associations between sociodemographic variables and health state scores remained the same across countries regardless of which value sets were used. Using different valuation methods lead to different QALY losses. To overcome this problem in international surveys aimed to compare health state scores or QALYs, it is advisable to use a single valuation method, making these scores comparable.