SHF: Small: Collaborative Research: Designing a Patient-Oriented Prescription Language: An Executable Medical Algorithm for Gestational Diabetes Mellitus

医疗保健中可预防的错误是导致患者受伤和死亡的主要原因。尽管付出了相当大的努力和数十亿美元的支出，但电脑化尚未提高医疗保健的有效性或安全性。对于这个问题，PI带来了一个新的见解：处方就是程序。也就是说，他们假设处方编写是软件工程的一个失传已久的兄弟，使用了许多计算机编程构造，但没有计算机科学提供的广泛概念框架的好处。这个项目将通过利用软件工程和编程语言来验证这一假设，这些语言知道如何创建一种高度直观的、特定于领域的语言来构建和执行处方。然后，PIs将建立并验证一份以患者为中心的处方，以管理妊娠期糖尿病(GDM)患者的护理，GDM是一种每年影响60万女性的疾病。目前GDM管理的局限性包括患者依从性低、干预效果不佳、临床医生的时间负担和成本。使用来自真实患者的丰富数据的广泛档案，包括饮食、活动、血糖读数和胰岛素使用，PI将执行纯基于计算机的模拟，这将帮助他们确定他们的POP-GDM程序将如何响应真实世界的数据。他们还将在我们最先进的医疗模拟设施中使用真实的临床医生、演员患者以及来自智能手机、血糖仪和加速计的模拟输入进行模拟。PI预计这项工作将产生关于建立和维护多阶段程序、开发新类型的程序分析以及发现新的通用方法来开发促进机器和人类之间协作的高可靠性软件的多个新见解。

Preventable errors in healthcare are a leading cause of patient injury and death. Despite considerable effort and the expenditure of billions of dollars, computerization has yet to improve the efficacy or safety of healthcare. To this problem, the PIs bring a novel insight: a prescription is a program. That is, they hypothesize that prescription writing is a long lost sibling of software engineering, using many computer programming constructs but without the benefit of the extensive conceptual framework provided by Computer Science. This project will test this hypothesis by exploiting software engineering and programming languages know?how to create a highly?intuitive, domain specific language for building and executing prescriptions. The PIs will then build and validate a "patient oriented prescription" to manage the care of patients with Gestational Diabetes Mellitus (GDM), a disease affecting 600,000 women annually.Limitations of current GDM management include low patient adherence, modest efficacy of interventions, time?burden for clinicians, and cost. Using extensive archives of rich data from real patients, including diet, activity, glucose readings, and insulin use, the PIs will perform pure computer-based simulations that will help them determine how their POP-GDM program will respond to real-world data. They will also perform simulations with real clinicians, actor patients, and simulated inputs from smartphones, glucometers, and accelerometers in our state of the art medical simulation facility. The PIs expect this work to generate multiple new insights regarding the building and maintenance of multi-stage programs, the development of new kinds of analyses of programs, and the discovery of new general methods for developing high-reliability software that facilitates collaboration between machines and humans.