Innsbruck Institute of Patient-centered Outcome Research (IIPCOR), Innsbruck, Austria.
Italian Group for Adult Hematologic Diseases (GIMEMA), Data Center & Health Outcomes Research Unit, Rome, Italy.
J Clin Epidemiol. 2021 Sep;137:31-44. doi: 10.1016/j.jclinepi.2021.03.015. Epub 2021 Mar 20.
The aim was to investigate the relative validity of the preference-based measure EORTC QLU-C10D in comparison with the EQ-5D-3L in myelodysplastic syndromes (MDS) patients.
We used data from an international multicentre, observational cohort study of MDS patients. Baseline EORTC QLU-C10D and EQ-5D-3L scores were used and index scores calculated for Italy, Australia, and the UK. Criterion validity was established by Spearman and intraclass correlations (ICC) and Bland-Altman plots. Construct validity was established by the instruments' ability to discriminate known groups, i.e. groups whose health status is expected to differ.
We analyzed data from 619 MDS patients (61.1% male; median age 73.8 years). Correlations between theoretically corresponding domains were largely higher than between unrelated domains. ICCs and Bland-Altman plots indicated moderate to good criterion validity. Ceiling effects were lower for the QLU-C10D (4.7%) than for the EQ-5D-3L (22.6%). The EQ-5D-3L failed to discriminate known-groups in two and the QLU-C10D in one of the comparisons; the QLU-C10D's efficiency in doing so was higher in clinical known-groups. Results were comparable between the countries.
The QLU-C10D may be suitable to generate health utilities for economic research in MDS. Responsiveness and minimal important differences need yet to be established.
本研究旨在比较 EORTC QLQ-C10D 与 EQ-5D-3L 在骨髓增生异常综合征（MDS）患者中的相对有效性。
我们使用了一项 MDS 患者国际多中心观察性队列研究的数据。使用了基线 EORTC QLQ-C10D 和 EQ-5D-3L 评分，并计算了意大利、澳大利亚和英国的指数评分。采用 Spearman 相关系数和组内相关系数（ICC）以及 Bland-Altman 图来确定效标效度。通过这些工具区分已知组别的能力（即预期健康状况不同的组别）来确定结构效度。
我们分析了 619 例 MDS 患者的数据（61.1%为男性；中位年龄为 73.8 岁）。理论上相应的领域之间的相关性大多高于不相关的领域。ICC 和 Bland-Altman 图表明具有中度至良好的效标效度。QLU-C10D 的天花板效应（4.7%）低于 EQ-5D-3L（22.6%）。在两项比较中，EQ-5D-3L 无法区分已知组别，而 QLU-C10D 则在一项比较中无法区分；在临床已知组别中，QLU-C10D 这样做的效率更高。结果在各国之间具有可比性。
QLU-C10D 可用于生成 MDS 经济研究中的健康效用值。尚需确定其反应度和最小临床重要差异。
