Taylor William J
University of Otago Wellington, Wellington, New Zealand.
Curr Opin Rheumatol. 2016 Mar;28(2):117-21. doi: 10.1097/BOR.0000000000000259.
Conjoint analysis of choice or preference data has been used in marketing for over 40 years but has appeared in healthcare settings much more recently. It may be a useful technique for applications within the rheumatology field.
Conjoint analysis in rheumatology contexts has mainly used the approaches implemented in 1000Minds Ltd, Dunedin, New Zealand, Sawtooth Software, Orem UT, USA. Examples include classification criteria, composite response criteria, service prioritization tools and utilities assessment. Limitations imposed by very many attributes can be managed using new techniques. Conjoint analysis studies of classification and response criteria suggest that the assumption of equal weighting of attributes cannot be met, which challenges traditional approaches to composite criteria construction. Weights elicited through choice experiments with experts can derive more accurate classification criteria, than unweighted criteria. Studies that find significant variation in attribute weights for composite response criteria for gout make construction of such criteria problematic.
Better understanding of various multiattribute phenomena is likely to increase with increased use of conjoint analysis, especially when the attributes concern individual perceptions or opinions. In addition to classification criteria, some applications for conjoint analysis that are emerging in rheumatology include prioritization tools, remission criteria, and utilities for life areas.
选择或偏好数据的联合分析在市场营销领域已应用40多年，但在医疗环境中的应用则是近年来才出现。它可能是一种在风湿病领域应用中有用的技术。
风湿病领域的联合分析主要采用新西兰达尼丁1000Minds有限公司、美国犹他州奥勒姆的锯齿软件公司所实施的方法。实例包括分类标准、综合反应标准、服务优先级工具和效用评估。使用新技术可以处理众多属性带来的限制。分类和反应标准的联合分析研究表明，无法满足属性等权重的假设，这对构建综合标准的传统方法提出了挑战。通过与专家进行选择实验得出的权重，比未加权的标准能得出更准确的分类标准。关于痛风综合反应标准的属性权重存在显著差异的研究，使得此类标准的构建存在问题。
随着联合分析使用的增加，对各种多属性现象可能会有更好的理解，尤其是当属性涉及个人认知或观点时。除了分类标准外，风湿病领域正在出现的联合分析的一些应用包括优先级工具、缓解标准和生活领域的效用评估。
