Verella J Tipan, Patek Stephen D
University of Virginia, Department of Systems and Information Engineering, Charlottesville, Virginia, USA.
J Diabetes Sci Technol. 2009 Mar 1;3(2):353-62. doi: 10.1177/193229680900300217.
Health care is a major component of the U.S. economy, and tremendous research and development efforts are directed toward new technologies in this arena. Unfortunately few tools exist for predicting outcomes associated with new medical products, including whether new technologies will find widespread use within the target population. Questions of technology adoption are rife within the diabetes technology community, and we particularly consider the long-term prognosis for continuous glucose monitoring (CGM) technology. We present an approach to the design and analysis of an agent model that describes the process of CGM adoption among patients with type 1 diabetes mellitus (T1DM), their physicians, and related stakeholders. We particularly focus on patient-physician interactions, with patients discovering CGM technology through word-of-mouth communication and through advertising, applying pressure to their physicians in the context of CGM device adoption, and physicians, concerned about liability, looking to peers for a general level of acceptance of the technology before recommending CGM to their patients. Repeated simulation trials of the agent-based model show that the adoption process reflects the heterogeneity of the adopting community. We also find that the effect of the interaction between patients and physicians is agents. Each physician, say colored by the nature of the environment as defined by the model parameters. We find that, by being able to represent the diverse perspectives of different types of stakeholders, agent-based models can offer useful insights into the adoption process. Models of this sort may eventually prove to be useful in helping physicians, other health care providers, patient advocacy groups, third party payers, and device manufacturers understand the impact of their decisions about new technologies.
医疗保健是美国经济的一个主要组成部分，并且大量的研发工作都指向了这个领域的新技术。不幸的是，用于预测与新医疗产品相关结果的工具很少，包括新技术是否会在目标人群中广泛使用。糖尿病技术领域充斥着技术采用的问题，我们特别关注连续血糖监测（CGM）技术的长期预后。我们提出了一种设计和分析代理模型的方法，该模型描述了1型糖尿病（T1DM）患者、他们的医生以及相关利益相关者采用CGM的过程。我们特别关注患者与医生的互动，患者通过口碑传播和广告发现CGM技术，在采用CGM设备的背景下向他们的医生施加压力，而医生则担心责任问题，在向患者推荐CGM之前会向同行了解该技术的总体接受程度。基于代理模型的重复模拟试验表明，采用过程反映了采用群体的异质性。我们还发现患者与医生之间互动的影响是代理性的。每个医生，比如说由模型参数定义的环境性质所影响。我们发现，通过能够代表不同类型利益相关者的不同观点，基于代理的模型可以为采用过程提供有用的见解。这种类型的模型最终可能被证明有助于医生、其他医疗保健提供者、患者倡导团体、第三方支付者和设备制造商理解他们对新技术的决策影响。
