STTR Phase I: User-Friendly Spirometer and Mobile App for Self-Management and Home Monitoring of Asthma Patients

这项小企业技术转让（STTR）第一阶段项目的更广泛影响/商业潜力将缩小医院和家庭之间在适当哮喘护理方面的差距。对肺部进行适当评估的最大障碍是昂贵的设备，价格从1000美元到3万美元不等，而且必须由熟练的实验室技术人员监督，以确保正确使用。这些医院肺功能测试可以在咳嗽或喘息等明显症状出现前几天早期发现肺功能下降，但目前还没有可靠的家庭评估选择。拟议中的技术将机器学习整合到一个易于使用的移动应用程序中，该应用程序可以复制训练有素的实验室技术人员的指导。结合可负担得起的消费级肺功能测量设备，该技术可以在医院外进行适当的肺部评估，在办公室就诊之间进行定期跟踪，并提供医生指导的建议，以减少不必要和昂贵的急诊和住院治疗。该项目仅在美国就为1000万哮喘儿童提供了预防性护理解决方案。这个解决方案需要一个引人入胜的移动应用程序游戏，由一个测量肺活量的手持设备控制，这些设备一起为父母提供一个行动计划，以防止他们的孩子？在哮喘症状出现之前。将与儿科肺科医生协商制定指南，以获得与在医院呼吸治疗师指导下进行的测试一样可靠的结果。一个大型的专家标记肺测量数据库将用于训练和测试神经网络模型，以检测和破译特征模式和相关性。如果收集到不精确的信息，应用程序的机器学习算法将确定故障原因，并为下一次尝试提出纠正措施。这确保只有正确收集的肺部测量数据才能触发行动建议。在此提案期结束时，申请人将在移动应用程序上拥有机器学习算法，该算法与训练有素的实验室技术人员亲自指导的效果相匹配，并评估拟议技术的可行性，以供第二阶段考虑。

The broader impact/commercial potential of this Small Business Technology Transfer (STTR) Phase I project will close the gap between the hospital and home for proper asthma care. The greatest hurdles for proper lung assessment are costly devices ranging from $1000 to $30,000, and mandatory oversight from a skilled lab technician to ensure proper use. These hospital lung function tests allow for early detection of declining lung performance up to several days in advance of visible symptoms like coughing or wheezing, but no option for reliable assessment in the home currently exists. The proposed technology incorporates machine learning into an easy to use mobile-app that replicates a trained lab technician's coaching. In combination with an affordable consumer device that measures lung function, the technology makes proper lung assessment accessible outside the hospital, for regular tracking between office visits supported with physician-guided suggestions for reducing unnecessary and costly emergency visits and hospitalizations. The proposed project offers a preventative care solution for 10 million children with asthma in the US alone. This solution entails an engaging mobile-app game controlled by a handheld device that measures lung capacity, which together provide parents with an action plan to prevent their child?s asthma symptoms before they occur. Guidelines will be developed in consultation with pediatric pulmonologists to gain results as reliable as tests conducted under the guidance of a respiratory therapist in the hospital. A large database of expert labeled lung measurements will be used to train and test the neural network model to detect and decipher feature patterns and correlations. If imprecise information is collected, the app's machine-learned algorithm will determine the cause of failure and suggest a corrective action for the next attempt. This ensures that only properly collected lung measurements trigger recommendations for action. By the end of this proposal period, the applicant will have a machine learning algorithm on a mobile-app that matches the efficacy of in-person coaching by a trained lab technician, as well as evaluate the feasibility of the proposed technology for Phase II considerations.