Network Control of Diabetes: Aligning Artificial Pancreas Design with Physiology

描述(申请人提供)：该项目是来自弗吉尼亚大学(UVA)和斯坦福大学的医生、数学家和工程师的跨学科和跨机构的努力，致力于开发和测试一种新的网络控制人工胰腺(AP)概念：通过在模块化设计中将工程控制功能与生理和行为过程相结合来优化1型糖尿病患者的血糖。将这一概念落实到临床实践的载体将是糖尿病助手(DIAS)--第一个用于门诊闭环控制(CLC)试验的[全球]智能手机AP系统。DIAS结合了本地患者服务(例如低血糖预防、建议、CLC)和全球服务(例如实时远程监测、远程医疗)。该系统的开发将利用15年的广泛理论工作、《中图法》研究、行为干预和尖端工程设计。然而，我们的基本理念是直截了当的：为了取得成功，CLC需要包括一个强大的安全系统，以降低低血糖风险，并根据每个人的日常行为和基础/给药模式学习控制算法。(1)技术开发将包括：(I)结构性软件设计，基于我们的医疗Android原型操作系统(存放在FDA主文件2109中)，并添加一个系统管理程序，以实现关键的CLC功能与其他非关键过程(如正常的智能手机操作)的无缝协同执行；(Ii)通过纳入专门用于预防低血糖的模块，使咨询和CLC算法与关键生理过程保持一致；(Ii)通过整合专门用于预防低血糖的模块，优化基础/丸剂的行为，以反映每个患者的日常模式；以及实时微调胰岛素的给药，以安全地强化治疗。(2)临床试验将包括：(I)研究1展示了使用CLC后低血糖意识的逆转，这将是使用我们的低血糖预防统一安全系统(USS弗吉尼亚号)1个月的结果；(Ii)研究2部署了一个由本地和全球咨询模块组成的咨询系统，该系统将通过将每个人的日常基础/给药情况与个人行为模式相匹配来优化它们。(Iii)研究3将安全、咨询和CLC模块整合到一个全面的网络控制系统中。具体地说：USS将预防低血糖--最佳糖尿病控制的主要障碍--从而允许完全集成的CLC进行强化治疗，它将调整胰岛素的输送，以微调每个人的行为优化配置。DIAS技术将主要由UVA团队开发。所有的临床研究都将在 在弗吉尼亚大学(专注于成人患者)和斯坦福大学(专注于儿科人群)是平行的。最终，该项目将最终完成并广泛测试模块化便携式AP系统，使其成为 为临床部署做好准备。

DESCRIPTION (provided by applicant): This project is an interdisciplinary and inter-institutional effort of physicians, mathematicians, and engineers from the University of Virginia (UVA) and Stanford University, dedicated to development and testing of a new network-control artificial pancreas (AP) concept: optimization of glycemia in type 1 diabetes by alignment of engineering control functions with physiologic and behavioral processes in a modular design. The vehicle for implementation of this concept into the clinical practice will be the Diabetes Assistant (DiAs) - the first [worldwide] smart-phone AP system used in outpatient closed-loop control (CLC) trials. DiAs combines local patient services (e.g. hypoglycemia prevention, advice, CLC), and global services (e.g. real-time remote monitoring, telemedicine). The system development will utilize 15 years of extensive theoretical work, CLC studies, behavioral interventions, and cutting-edge engineering designs. Nevertheless, our basic philosophy is straightforward: in order to be successful, CLC needs to include a robust safety system reducing the risk for hypoglycemia and control algorithms learning from each individual's daily behavioral and basal/bolus patterns. (1) Technology development will include: (i) Structural Software design, based on our prototype Medical Android operating system (deposited in FDA Master File 2109), and adding a system hypervisor to enable seamless co-execution of critical CLC functions together with other non-critical processes (e.g. normal smart phone operation); (ii) Alignment of Advisory and CLC Algorithms with key physiological processes by incorporating modules specific to prevention of hypoglycemia, behavioral optimization of basal/bolus profiles to reflect each patient's daily patterns; and real-time fine-tuning of insulin delivery to safely intensify treatment. (2) Clinical Trials will include: (i) Study 1 demonstrating reversal of hypoglycemia unawareness with CLC, which will result from 1-month use of our hypoglycemia-preventing Unified Safety System (USS Virginia); (ii) Study 2 deploying an Advisory System comprised of local and global advisory modules, which will optimize each person's daily basal/bolus profiles by matching them to individual behavioral patterns. (iii) Study 3 bringing together Safety, Advisory, and CLC modules into a comprehensive network-control system. Specifically: the USS will safeguard against hypoglycemia - the primary barrier to optimal diabetes control - thereby allowing intensified treatment by Fully-Integrated CLC, which will adjust insulin delivery to fine-tune each person's behaviorally-optimized profile. The DiAs technology will be primarily developed by the UVA team. All clinical studies will be carried out in parallel at UVA (focusing on adult patients) and Stanford (focusing on pediatric population). Ultimately, this project will finalize and test extensively a modular portable AP system, making it ready for clinical deployment.