类风湿关节炎患者健康效用值测量研究——从HAQ-DI到EQ-5D-5L的映射

Health utility measurement for rheumatoid arthritis:mapping HAQ-DI score onto the EQ-5D-5L utility score

目的构建适合我国类风湿关节炎患者人群的健康评估问卷残疾指数得分(HAQ-DI)到欧洲五维五水平健康量表(EQ-5D-5L)效用值得分的映射模型,为开展类风湿关节炎临床治疗方案的卫生经济学评价,进一步优化我国卫生资源配置提供科学依据。方法选取2017年8月1日-2019年7月31日期间在中山大学孙逸仙纪念医院首次就诊的674例类风湿关节炎患者的横断面调查数据,收集受试者的一般人口学信息、疾病相关指标及HAQ-DI、EQ-5D-5L等量表得分。将样本随机划分至训练集(n=472)和验证集(n=202);训练集中采用最小二乘法回归(OLS)模型、截断回归(Tobit)模型构建HAQ-DI量表到EQ-5D-5L效用值得分的映射函数式,并验证模型的预测能力。结果纳入HAQ-DI得分、EQ-5D视觉模拟量表(EQ-VAS)得分以及医生总体评价(PrGA)的模型报告了最高的调整R;(0.745)和最低的赤池信息准则(AIC,-276.559)、均方根误差(RMSE,0.129)及平均绝对误差(MAE,0.092)。基于OLS方法构建的映射方程为:EQ-5D=0.73-0.208×HAQ-DI+0.003×EQ-VAS-0.02×PrGA(调整R;=0.745,RMSE=0.127,MAE=0.085)。结论纳入HAQ-DI得分、EQ-VAS得分及PrGA的模型拟合效果最好,通过HAQ-DI得分、EQ-VAS得分及PrGA构建的映射公式预测EQ-5D-5L效用值得分具有一定的可行性与实际意义。

Objective To construct the mapping model for the five level version of EuroQol five dimensions questionnaire(EQ-5D-5L)utility values using the health assessment questionnaire disability index(HAQ-DI)among rheumatoid arthritis patients in China,and provide scientific basis for carrying out health economics evaluation among rheumatoid arthritis,and optimizing health resource allocation in policy making.Methods A total of 674 patients with rheumatoid arthritis were selected for a questionnaire survey when they were firstly treated from August 1,2017 to July 31,2019 in Sun Yat-Sen Memorial Hospital,and their general demographic information,disease-related indicators and HAQ-DI,EQ-5D-5L scale scores were collected.Patients were randomly divided into training set(n=472)and validation set(n=202).The mapping function between HAQ-DI scores and EQ-5D-5L utility value were constructed using the ordinary least squares(OLS)model and the Tobit model in the training set,and the predictive accuracy of the model was evaluated in the validation set.Results The model with HAQ-DI score,visual analogue scale of EuroQol five dimensions questionnaire(EQ-VAS)score and provider global assessment of disease activity(PrGA)reported the highest adjusted R~2(0.745)and the lowest akaike information criterion(AIC,-276.559),root mean square error(RMSE,0.129)and mean absolute error(MAE,0.092).The mapping equation constructed based on the OLS method was EQ-5D=0.73-0.208×HAQ-DI+0.003×EQ-VAS-0.02×PrGA(adjusted R~2=0.745,RMSE=0.127,MAE=0.085).Conclusions The model fitting effect turned to be the best when HAQ-DI score,EQ-VAS score and PrGA were included.It was feasible and practical to predicte EQ-5D-5L utility value through the mapping formula constructed by HAQ-DI score,EQ-VAS score and PrGA.