Guyatt Gordon H, Osoba David, Wu Albert W, Wyrwich Kathleen W, Norman Geoffrey R
Department of Clinical Epidemiology and Biostatistics, McMaster University and Health Sciences Center, Hamilton, Ontario.
Mayo Clin Proc. 2002 Apr;77(4):371-83. doi: 10.4065/77.4.371.
One can classify ways to establish the interpretability of quality-of-life measures as anchor based or distribution based. Anchor-based measures require an independent standard or anchor that is itself interpretable and at least moderately correlated with the instrument being explored. One can further classify anchor-based approaches into population-focused and individual-focused measures. Population-focused approaches are analogous to construct validation and rely on multiple anchors that frame an individual's response in terms of the entire population (eg, a group of patients with a score of 40 has a mortality of 20%). Anchors for population-based approaches include status on a single item, diagnosis, symptoms, disease severity, and response to treatment. Individual-focused approaches are analogous to criterion validation. These methods, which rely on a single anchor and establish a minimum important difference in change in score, require 2 steps. The first step establishes the smallest change in score that patients consider, on average, to be important (the minimum important difference). The second step estimates the proportion of patients who have achieved that minimum important difference. Anchors for the individual-focused approach include global ratings of change within patients and global ratings of differences between patients. Distribution-based methods rely on expressing an effect in terms of the underlying distribution of results. Investigators may express effects in terms of between-person standard deviation units, within-person standard deviation units, and the standard error of measurement. No single approach to interpretability is perfect. Use of multiple strategies is likely to enhance the interpretability of any particular instrument.
可以将建立生活质量测量可解释性的方法分为基于锚定的方法和基于分布的方法。基于锚定的测量需要一个独立的标准或锚定，其本身是可解释的，并且与所探索的工具至少有中度相关性。可以进一步将基于锚定的方法分为以人群为重点的方法和以个体为重点的方法。以人群为重点的方法类似于结构效度验证，依赖于多个锚定，根据整个人群来界定个体的反应（例如，一组得分为40的患者死亡率为20%）。基于人群方法的锚定包括单个项目的状态、诊断、症状、疾病严重程度和对治疗的反应。以个体为重点的方法类似于效标效度验证。这些方法依赖于单个锚定并确定分数变化中的最小重要差异，需要两个步骤。第一步确定患者平均认为重要（最小重要差异）的分数最小变化。第二步估计达到该最小重要差异的患者比例。以个体为重点方法的锚定包括患者内部变化的总体评分和患者之间差异的总体评分。基于分布的方法依赖于根据结果的潜在分布来表达效应。研究人员可以用人与人之间的标准差单位、个体内部的标准差单位和测量标准误来表达效应。没有一种单一的可解释性方法是完美的。使用多种策略可能会增强任何特定工具的可解释性。
