Weng Guizhi, Hong Yanming, Luo Nan, Mukuria Clara, Jiang Jie, Yang Zhihao, Li Sha
College of Pharmacy, Jinan University, 601 West Huangpu avenue, 510632, Guangzhou, People's Republic of China.
Saw Swee Hock School of Public Health, National University of Singapore, Singapore, Singapore.
Eur J Health Econ. 2023 Mar;24(2):197-207. doi: 10.1007/s10198-022-01465-7. Epub 2022 May 10.
EQ-5D-3L has been used in the National Health Services Survey of China since 2008 to monitor population health. The five-level version of EQ-5D was developed, but there lacks evidence to support the use of five-level version of EQ-5D in China. This study was conducted to compare the measurement properties of both the EQ-5D-3L and EQ-5D-5L in quantifying health-related quality of life (HRQoL) burden for 4 different health conditions in China.
Participants from China were recruited to complete the 3L and 5L questionnaire via Internet. Quota was set to recruit five groups of individuals, with one group of individuals without any health condition and one group of generalized anxiety disorder (GAD), HIV/AIDS, chronic Hepatitis B (CHB), or depression, respectively. The 3L and 5L were compared in terms of response distributions, percentages of reporting 'no problems', index value distributions, known-group validity and their relative efficiency.
In total, 500 individuals completed the online survey, including 140 healthy individuals, 122 individuals with hepatitis B, 107 with depression, 90 individuals with GAD and 101 with HIV/AIDS. 5L also had smoother and less clustered index value distributions. Healthy group showed different response distributions to the four condition groups. The percentage of reporting 'no problems' decreased significantly in the 5L in all domains (P < 0.01), especially in the pain/discomfort dimension (relative difference: 43.10%). Relative efficiency suggested that 5L had a higher absolute discriminatory power than the 3L version between healthy participant and the other 4 condition groups, especially for the HIV/AIDS group when the 3L results was not significant.
The 5L version may be preferable to the 3L, as it demonstrated superior performance with respect to higher sensitivity to mild health problems, better relative efficiency and responses and index value distributions.
自2008年起，EQ-5D-3L已被用于中国国家卫生服务调查以监测人群健康状况。EQ-5D的五水平版本已开发出来，但在中国缺乏支持使用该五水平版本的证据。本研究旨在比较EQ-5D-3L和EQ-5D-5L在中国4种不同健康状况下量化健康相关生活质量（HRQoL）负担的测量属性。
招募来自中国的参与者通过互联网完成3L和5L问卷。设定配额招募五组个体，一组为无任何健康问题的个体，另外四组分别为广泛性焦虑障碍（GAD）、艾滋病毒/艾滋病、慢性乙型肝炎（CHB）或抑郁症患者。从应答分布、报告“无问题”的百分比、指数值分布、已知组效度及其相对效率等方面对3L和5L进行比较。
共有500名个体完成了在线调查，其中包括140名健康个体、122名乙型肝炎患者、107名抑郁症患者、90名广泛性焦虑障碍患者和101名艾滋病毒/艾滋病患者。5L的指数值分布也更平滑且聚类较少。健康组与四个疾病组的应答分布不同。在所有领域中，5L报告“无问题”的百分比均显著下降（P<0.01），尤其是在疼痛/不适维度（相对差异：43.10%）。相对效率表明，在健康参与者与其他4个疾病组之间，5L比3L版本具有更高的绝对辨别力，尤其是当3L结果不显著时，对于艾滋病毒/艾滋病组而言。
5L版本可能优于3L版本，因为它在对轻度健康问题的更高敏感性、更好的相对效率以及应答和指数值分布方面表现更优。
