Wiebe S, Eliasziw M, Matijevic S
Department of Clinical Neurological Sciences, University of Western Ontario, Ontario, Canada.
Epilepsia. 2001 Jan;42(1):113-8. doi: 10.1046/j.1528-1157.2001.081425.x.
The study goal was to assess the magnitude of change in generic and epilepsy-specific health-related quality-of-life (HRQOL) instruments needed to exclude chance or error at various levels of certainty in patients with medically refractory epilepsy.
Forty patients with temporal lobe epilepsy and clearly defined criteria of clinical stability received HRQOL measurements twice, 3 months apart, using the Quality of Life in Epilepsy Inventory-89 and -31 (QOLIE-89 and QOLIE-31), Liverpool Impact of Epilepsy, adverse drug events, seizure severity scales, and the Generic Health Utilities Index (HUI-III). Standard error of measurement and test-retest reliability were obtained for all scales and for QOLIE-89 subscales. Using the Reliable Change Index described by Jacobson and Truax, we assessed the magnitude of change required by HRQOL instruments to be 90 and 95% certain that real change has occurred, as opposed to change due to chance or measurement error.
Clinical features, point estimates and distribution of HRQOL measures, and test-retest reliability (all > 0.70) were similar to those previously reported. Score changes of +/-13 points in QOLIE-89, +/-15 in QOLIE-31, +/-6.3 in Liverpool seizure severity-ictal, +/-11 in Liverpool adverse drug events, +/-0.25 in HUI-III, and +/-9.5 in impact of epilepsy exclude chance or measurement error with 90% certainty. These correspond, respectively, to 13, 15, 17, 18, 25, and 32% of the potential range of change of each instrument.
Threshold values for real change varied considerably among HRQOL tools but were relatively small for QOLIE-89, QOLIE-31, Liverpool Seizure Severity, and adverse drug events. In some instruments, even relatively large changes cannot rule out chance or measurement error. The relation between the Reliable Change Index and other measures of change and its distinction from measures of minimum clinically important change are discussed.
本研究的目标是评估在药物难治性癫痫患者中，为排除不同确定水平下的偶然因素或误差，通用及癫痫特异性健康相关生活质量（HRQOL）工具所需的变化幅度。
40例符合颞叶癫痫且临床稳定性明确标准的患者，使用癫痫生活质量量表89项版和31项版（QOLIE - 89和QOLIE - 31）、利物浦癫痫影响量表、药物不良事件、癫痫发作严重程度量表以及通用健康效用指数（HUI - III），在间隔3个月的时间里接受了两次HRQOL测量。获得了所有量表以及QOLIE - 89各子量表的测量标准误差和重测信度。使用雅各布森和特鲁克斯描述的可靠变化指数，我们评估了HRQOL工具为确定90%和95%确实发生了真实变化（而非因偶然因素或测量误差导致的变化）所需的变化幅度。
临床特征、HRQOL测量的点估计值和分布以及重测信度（均>0.70）与先前报道的相似。QOLIE - 89得分变化±13分、QOLIE - 31得分变化±15分、利物浦癫痫发作严重程度（发作期）得分变化±6.3分、利物浦药物不良事件得分变化±11分、HUI - III得分变化±0.25分以及癫痫影响得分变化±9.5分可排除90%的偶然因素或测量误差。这些变化分别相当于各工具潜在变化范围的13%、15%、17%、18%、25%和32%。
HRQOL工具中真实变化的阈值差异很大，但对于QOLIE - 89、QOLIE - 31、利物浦癫痫发作严重程度量表和药物不良事件量表而言相对较小。在某些工具中，即使相对较大的变化也不能排除偶然因素或测量误差。讨论了可靠变化指数与其他变化测量方法之间的关系及其与最小临床重要变化测量方法的区别。
