Furber Gareth, Segal Leonie, Leach Matthew, Cocks Jane
Health Economics and Social Policy Group, University of South Australia, Playford Building, City East Campus, North Terrace, Adelaide, SA, 5000, Australia,
Qual Life Res. 2014 Mar;23(2):403-11. doi: 10.1007/s11136-013-0494-6. Epub 2013 Aug 13.
Quality of life mapping methods such as "Transfer to Utility" can be used to translate scores on disease-specific measures to utility values, when traditional utility measurement methods (e.g. standard gamble, time trade-off, preference-based multi-attribute instruments) have not been used. The aim of this study was to generate preliminary ordinary least squares (OLS) regression-based algorithms to transform scores from the Strengths and Difficulties Questionnaires (SDQ), a widely used measure of mental health in children and adolescents, to utility values obtained using the preference-based Child Health Utility (CHU9D) instrument.
Two hundred caregivers of children receiving community mental health services completed the SDQ and CHU9D during a telephone interview. Two OLS regressions were run with the CHU9D utility value as the dependent variable and SDQ subscales as predictors. Resulting algorithms were validated by comparing predicted and observed group mean utility values in randomly selected subsamples.
Preliminary validation was obtained for two algorithms, utilising five and three subscales of the SDQ, respectively. Root mean square error values (.124) for both models suggested poor fit at an individual level, but both algorithms performed well in predicting mean group observed utility values.
This research generated algorithms for translating SDQ scores to utility values and providing researchers with an additional tool for conducting health economic evaluations with child and adolescent mental health data.
当未使用传统效用测量方法（如标准博弈法、时间权衡法、基于偏好的多属性工具）时，诸如“转换为效用”之类的生活质量映射方法可用于将特定疾病测量指标的得分转换为效用值。本研究的目的是生成初步的基于普通最小二乘法（OLS）回归的算法，以将优势与困难问卷（SDQ，一种广泛用于儿童和青少年心理健康测量的工具）的得分转换为使用基于偏好的儿童健康效用（CHU9D）工具获得的效用值。
200名接受社区心理健康服务的儿童的照料者在电话访谈中完成了SDQ和CHU9D。以CHU9D效用值为因变量、SDQ分量表为预测变量进行了两次OLS回归。通过比较随机选择的子样本中预测的和观察到的组平均效用值来验证所得算法。
分别利用SDQ的五个和三个分量表对两种算法进行了初步验证。两个模型的均方根误差值（.124）表明在个体水平上拟合度较差，但两种算法在预测组平均观察效用值方面表现良好。
本研究生成了将SDQ得分转换为效用值的算法，为研究人员提供了一种利用儿童和青少年心理健康数据进行健康经济评估的额外工具。
