Lancsar Emily, Swait Joffre
Centre for Health Economics, Faculty of Business and Economics, Monash University, Melbourne, Australia,
Pharmacoeconomics. 2014 Oct;32(10):951-65. doi: 10.1007/s40273-014-0181-7.
External validity is a crucial but under-researched topic when considering using discrete choice experiment (DCE) results to inform decision making in clinical, commercial or policy contexts. We present the theory and tests traditionally used to explore external validity that focus on a comparison of final outcomes and review how this traditional definition has been empirically tested in health economics and other sectors (such as transport, environment and marketing) in which DCE methods are applied. While an important component, we argue that the investigation of external validity should be much broader than a comparison of final outcomes. In doing so, we introduce a new and more comprehensive conceptualisation of external validity, closely linked to process validity, that moves us from the simple characterisation of a model as being or not being externally valid on the basis of predictive performance, to the concept that external validity should be an objective pursued from the initial conceptualisation and design of any DCE. We discuss how such a broader definition of external validity can be fruitfully used and suggest innovative ways in which it can be explored in practice.
在考虑使用离散选择实验（DCE）结果为临床、商业或政策背景下的决策提供信息时，外部效度是一个关键但研究不足的主题。我们介绍了传统上用于探索外部效度的理论和测试，这些理论和测试侧重于最终结果的比较，并回顾了这一传统定义在健康经济学以及应用DCE方法的其他领域（如交通、环境和营销）中是如何进行实证检验的。虽然这是一个重要组成部分，但我们认为，对外部效度的调查应该比最终结果的比较广泛得多。在此过程中，我们引入了一种与过程效度密切相关的新的、更全面的外部效度概念，这使我们从基于预测性能简单地将一个模型描述为具有或不具有外部效度，转变为认为外部效度应该是从任何DCE的初始概念化和设计阶段就应追求的目标这一概念。我们讨论了如何有效地使用这种更广泛的外部效度定义，并提出了在实践中探索它的创新方法。
