Unidad de Análisis y Generación de Evidencias en Salud Pública, Instituto Nacional de la Salud, Lima, Peru; Division of Cancer Epidemiology, McGill University, Montreal, Canada.
Institute for Clinical Effectiveness and Health Policy (IECS-CONICET), Buenos Aires, Argentina.
Value Health Reg Issues. 2021 Dec;26:56-65. doi: 10.1016/j.vhri.2021.02.001. Epub 2021 May 21.
Social health preference sets are necessary for conducting health economic evaluations. Values from other countries are often used when local sets are not available, which may alter the results. We aimed to evaluate the degree of variability of currently available country-specific value sets of the EuroQol EQ-5D instrument from South American countries (SAC).
We selected EQ-5D value sets from SAC and 2 reference countries. We obtained crosswalk value sets for the countries that use the EQ-5D-3L instrument. We compared the value sets with the Kruskal-Wallis test and then carried out pairwise comparisons with the Sign test. We also assessed correlations among the countries' value sets using the Spearman test. We calculated the absolute difference across countries for each health state, considering a difference of greater than 0.05 relevant.
The range of value sets varied greatly. The Peruvian value set had the widest range (1 to -1.076) and the lowest values (median: 0.055; interquartile range: -0.171 to 0.275). The Ecuadorian set had the highest values (median: 0.587; interquartile range: 0.443-0.704). The Peruvian value set also had the greatest proportion of health states (43.6%) with a negative value, and the Uruguayan set had the smallest proportion (0.9%). Differences among countries were significant in all cases, with the greatest difference between Ecuador and Peru (median difference: 0.495; 95% confidence interval: 0.515-0.528).
Social health preference sets varied greatly among SAC. Using non-local values could distort resource allocation decisions; hence, we recommend that countries obtain and use local value sets.
社会健康偏好集对于进行健康经济评估是必要的。当无法获得本地数据集时，通常会使用来自其他国家的数据，这可能会改变结果。我们旨在评估来自南美国家（SAC）的欧洲五维健康量表（EQ-5D）仪器的现有特定国家价值数据集的可变性程度。
我们选择了来自 SAC 和 2 个参考国家的 EQ-5D 值集。我们为使用 EQ-5D-3L 仪器的国家获取了交叉值集。我们使用 Kruskal-Wallis 检验比较了这些数据集，然后使用 Sign 检验进行了两两比较。我们还使用 Spearman 检验评估了这些国家的价值数据集之间的相关性。我们考虑了各国之间每个健康状态的绝对差异，认为差异大于 0.05 是相关的。
价值数据集的范围差异很大。秘鲁的价值集范围最广（1 到-1.076），数值最低（中位数：0.055；四分位距：-0.171 至 0.275）。厄瓜多尔的数据集具有最高的数值（中位数：0.587；四分位距：0.443-0.704）。秘鲁的价值集也有最大比例（43.6%）的健康状态具有负值，而乌拉圭的价值集比例最小（0.9%）。在所有情况下，各国之间的差异均具有统计学意义，厄瓜多尔和秘鲁之间的差异最大（中位数差异：0.495；95%置信区间：0.515-0.528）。
SAC 之间的社会健康偏好集差异很大。使用非本地值可能会扭曲资源分配决策；因此，我们建议各国获取和使用本地价值集。
