PDF(Pubmed)
Abstract：
OBJECTIVE: Estimates of minimally important differences (MID) can assist interpretation of data collected using patient-reported outcomes (PRO), but variability exists in the emphasis placed on MIDs in health technology assessment (HTA) guidelines. This study aimed to identify to what extent information on the MID of a commonly used PRO, the EQ-5D, is required and utilised by selected HTA agencies.
METHODS: Technology appraisal (TA) documents from HTA agencies in England, France, Germany, and the US between 2019 and 2021 were reviewed to identify documents which discussed MID of EQ-5D data as a clinical outcome assessment (COA) endpoint.
RESULTS: Of 151 TAs utilising EQ-5D as a COA endpoint, 58 (38%) discussed MID of EQ-5D data. Discussion of MID was most frequent in Germany, in 75% (n = 12/16) of Gemeinsamer Bundesausschuss (G-BA) and 44% (n = 34/78) of Institut für Qualität und Wirtschaftlichkeit im Gesundheitswesen, (IQWiG) TAs. MID was predominantly applied to the EQ-VAS (n = 50), most frequently using a threshold of > 7 or > 10 points (n = 13). G-BA and IQWiG frequently criticised MID analyses, particularly the sources of MID thresholds for the EQ-VAS, as they were perceived as being unsuitable for assessing the validity of MID.
CONCLUSIONS: MID of the EQ-5D was not frequently discussed outside of Germany, and this did not appear to negatively impact decision-making of these HTA agencies. While MID thresholds were often applied to EQ-VAS data in German TAs, analyses were frequently rejected in benefit assessments due to concerns with their validity. Companies should pre-specify analyses of continuous data in statistical analysis plans to be considered for treatment benefit assessment in Germany.
METHODS: Technology appraisal (TA) documents from HTA agencies in England, France, Germany, and the US between 2019 and 2021 were reviewed to identify documents which discussed MID of EQ-5D data as a clinical outcome assessment (COA) endpoint.
RESULTS: Of 151 TAs utilising EQ-5D as a COA endpoint, 58 (38%) discussed MID of EQ-5D data. Discussion of MID was most frequent in Germany, in 75% (n = 12/16) of Gemeinsamer Bundesausschuss (G-BA) and 44% (n = 34/78) of Institut für Qualität und Wirtschaftlichkeit im Gesundheitswesen, (IQWiG) TAs. MID was predominantly applied to the EQ-VAS (n = 50), most frequently using a threshold of > 7 or > 10 points (n = 13). G-BA and IQWiG frequently criticised MID analyses, particularly the sources of MID thresholds for the EQ-VAS, as they were perceived as being unsuitable for assessing the validity of MID.
CONCLUSIONS: MID of the EQ-5D was not frequently discussed outside of Germany, and this did not appear to negatively impact decision-making of these HTA agencies. While MID thresholds were often applied to EQ-VAS data in German TAs, analyses were frequently rejected in benefit assessments due to concerns with their validity. Companies should pre-specify analyses of continuous data in statistical analysis plans to be considered for treatment benefit assessment in Germany.
摘要：
目的：对最小重要差异（MID）的估计可以帮助解释使用患者报告结果（PRO）收集的数据，但是在卫生技术评估（HTA）指南中对MID的强调存在差异。本研究旨在确定常用PRO的MID信息在多大程度上，EQ-5D,是选定的HTA机构所需要和使用的。
方法：来自英国HTA机构的技术评估（TA）文件，法国,德国,对2019年至2021年的美国进行了审查，以确定讨论EQ-5D数据MID作为临床结果评估（COA）终点的文件。
结果：在151个使用EQ-5D作为COA终点的TA中，58(38%)讨论了EQ-5D数据的MID。MID的讨论在德国最为频繁，在Gesundheitswesen的GemeinsamerBundesausschuss（G-BA）的75％（n=12/16）和质量研究所的44％（n=34/78），（IQWiG）TA。MID主要应用于EQ-VAS(n=50)，最常使用>7或>10点的阈值（n=13）。G-BA和IQWiG经常批评MID分析，特别是EQ-VAS的MID阈值的来源，因为他们被认为不适合评估MID的有效性。
结论：EQ-5D的MID在德国以外并不经常被讨论，这似乎并没有对这些HTA机构的决策产生负面影响。虽然MID阈值通常应用于德国TA的EQ-VAS数据，由于担心分析的有效性,在获益评估中经常被拒绝.公司应预先指定统计分析计划中的连续数据分析，以考虑在德国进行治疗效益评估。
方法：来自英国HTA机构的技术评估（TA）文件，法国,德国,对2019年至2021年的美国进行了审查，以确定讨论EQ-5D数据MID作为临床结果评估（COA）终点的文件。
结果：在151个使用EQ-5D作为COA终点的TA中，58(38%)讨论了EQ-5D数据的MID。MID的讨论在德国最为频繁，在Gesundheitswesen的GemeinsamerBundesausschuss（G-BA）的75％（n=12/16）和质量研究所的44％（n=34/78），（IQWiG）TA。MID主要应用于EQ-VAS(n=50)，最常使用>7或>10点的阈值（n=13）。G-BA和IQWiG经常批评MID分析，特别是EQ-VAS的MID阈值的来源，因为他们被认为不适合评估MID的有效性。
结论：EQ-5D的MID在德国以外并不经常被讨论，这似乎并没有对这些HTA机构的决策产生负面影响。虽然MID阈值通常应用于德国TA的EQ-VAS数据，由于担心分析的有效性,在获益评估中经常被拒绝.公司应预先指定统计分析计划中的连续数据分析，以考虑在德国进行治疗效益评估。
