THL, National Institute for Health and Welfare, Helsinki, Finland.
Health Econ. 2010 Jun;19(6):670-82. doi: 10.1002/hec.1504.
This study concerns the choice of primary dental service provider by consumers. If the health service delivery system allows individuals to choose between public-care providers or if complementary private services are available, it is typically assumed that utilisation is a three-stage decision process. The patient first makes a decision to seek care, and then chooses the service provider. The final stage, involving decisions over the amount and form of treatment, is not considered here. The paper reports a discrete choice experiment (DCE) designed to evaluate attributes affecting individuals' choice of dental-care provider. The feasibility of the DCE approach in modelling consumers' choice in the context of non-acute need for dental care is assessed. The aim is to test whether a separate two-stage logit, a multinomial logit, or a nested logit best fits the choice process of consumers. A nested logit model of indirect utility functions is estimated and inclusive value (IV) constraints are tested for modelling implications. The results show that non-trading behaviour has an impact on the choice of appropriate modelling technique, but is to some extent dependent on the choice of scenarios offered. It is concluded that for traders multinomial logit is appropriate, whereas for non-traders and on average the nested logit is the method supported by the analyses. The consistent finding in all subgroup analyses is that the traditional two-stage decision process is found to be implausible in the context of consumer's choice of dental-care provider.
本研究关注消费者对初级牙科服务提供者的选择。如果卫生服务提供系统允许个人在公共医疗服务提供者之间进行选择，或者如果有补充的私人服务可用，那么通常假定利用是一个三阶段决策过程。患者首先决定寻求医疗服务，然后选择服务提供者。最后阶段，涉及治疗的数量和形式的决策，在这里不考虑。本文报告了一项离散选择实验（DCE），旨在评估影响个人选择牙科护理提供者的属性。评估 DCE 方法在模拟非急性牙科护理需求背景下消费者选择的可行性。目的是测试单独的两阶段逻辑模型、多项逻辑模型还是嵌套逻辑模型最适合消费者的选择过程。估计了间接效用函数的嵌套逻辑模型，并测试了包容性价值（IV）约束以进行建模影响。结果表明，非交易行为会影响适当建模技术的选择，但在某种程度上取决于所提供的情景选择。得出的结论是，对于交易者来说，多项逻辑模型是合适的，而对于非交易者和平均而言，嵌套逻辑模型是分析支持的方法。在所有子组分析中一致的发现是，在消费者选择牙科护理提供者的情况下，传统的两阶段决策过程被认为是不可信的。
