Tsiplova Kate, Pullenayegum Eleanor, Cooke Tim, Xie Feng
Dalla Lana School of Public Health, University of Toronto, Toronto, Canada.
The Hospital for Sick Children, Toronto, Canada.
Qual Life Res. 2016 Dec;25(12):3009-3016. doi: 10.1007/s11136-016-1336-0. Epub 2016 Jun 15.
The purpose of the study is to estimate the EQ-5D-derived health utilities associated with selected chronic conditions (hypertension, heart disease, arthritis, asthma or COPD, cancer, diabetes, chronic back pain, and anxiety or depression) and to estimate minimally important differences (MID) based on the Commonwealth Fund Survey of Sicker Adults in Canada.
We used a cross-sectional survey of 3765 sick adults in Canada conducted in 2011 by the Commonwealth Fund. Health utilities were calculated for the entire sample and for each of the eight chronic health conditions. An ordinary least squares regression was used to estimate the utility decrement associated with these conditions with and without adjustment for socio-demographic factors. The MIDs were estimated using the anchor- and distribution-based methods.
The adjusted utility decrement varied from 0.028 (95 % confidence interval (CI) -0.049, -0.008) for diabetes to 0.124 (95 % CI -0.142, -0.105) for anxiety or depression. The anchor-based MID for the entire group was 0.044 (95 % CI 0.025, 0.062), and the distribution-based MID for the entire group was 0.091. The condition-specific MIDs using the distribution-based method ranged from 0.089 for cancer to 0.108 for asthma or COPD.
The MID estimated by the distribution-based method was larger than the MID estimated by the anchor-based method, indicating that the choice of method matters. The impact of arthritis, anxiety or depression, and chronic back pain on health utility was substantial, all exceeding or approximating the MID estimated using either anchor- or distribution-based methods.
本研究旨在评估与特定慢性病（高血压、心脏病、关节炎、哮喘或慢性阻塞性肺疾病、癌症、糖尿病、慢性背痛以及焦虑或抑郁）相关的EQ-5D健康效用值，并根据英联邦基金会对加拿大患病成年人的调查来评估最小重要差异（MID）。
我们采用了英联邦基金会于2011年对加拿大3765名患病成年人进行的横断面调查。计算了整个样本以及八种慢性健康状况各自的健康效用值。使用普通最小二乘法回归来估计与这些状况相关的效用降低情况，同时对社会人口学因素进行了调整和未调整两种情况的分析。使用基于锚定和基于分布的方法来估计MID。
调整后的效用降低值范围从糖尿病的0.028（95%置信区间（CI）-0.049，-0.008）到焦虑或抑郁的0.124（95%CI -0.142，-0.105）。整个组基于锚定的MID为0.044（95%CI 0.025，0.062），基于分布的MID为0.091。使用基于分布方法的特定疾病MID范围从癌症的0.089到哮喘或慢性阻塞性肺疾病的0.108。
基于分布方法估计的MID大于基于锚定方法估计的MID，这表明方法的选择很重要。关节炎、焦虑或抑郁以及慢性背痛对健康效用的影响很大，均超过或接近使用基于锚定或基于分布方法估计的MID。
