SBIR Phase I: Voice-based telehealth interface for symptom monitoring and screening for chronic and acute respiratory diseases, including COVID-19

这项小型企业创新研究（SBIR）第一阶段项目的更广泛影响/商业潜力是一种基于智能手机的新型方法，用于慢性和急性呼吸道疾病（包括COVID-19）的症状监测和筛查。目前评估呼吸系统疾病的方法不容易获得，或者不能扩展到筛查大规模人群。所提出的技术将使得能够检测和监测任何能够访问互联网连接的麦克风的人的呼吸系统疾病（例如，智能手机），使用语音作为指标。 该技术将在几分钟内进行简单的测试，并在几秒钟内提供结果，而无需专门的用户培训。预期成果是一个广泛的、实时的筛查、监测和加重警告系统，可以远程分析慢性和急性呼吸系统疾病患者的语音信号，包括COVID-19。这个小型企业创新研究（SBIR）第一阶段项目旨在开发基于语音的分类器，诊断COVID-19并监测疾病的严重程度。检测呼吸和语音中指示肺功能和呼吸系统疾病的声音生物标志物的现有算法将被扩展，以纳入COVID-19特征。将收集接受COVID-19阳性检测的患者的录音，以提取微签名并开发算法，以自动识别并将模式映射到临床发现和报告的症状。研究目标包括：（1）建立区分有症状和无症状患者的二元分类器;（2）多类分类器，（在未来预测）当提供一系列语音样本时患者症状的严重性的变化（3）为医生开发仪表板，其提供最新的报告以及横截面和纵向分析的可视化（4）提供肺功能指标和分类器的API，可集成到第三方IT基础设施中。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project is a novel smartphone-based method for symptom monitoring and screening for chronic and acute respiratory diseases, including COVID-19. Current methods of evaluating respiratory diseases are not easily accessible or do not scale to screen large populations. The proposed technology will enable detection and monitoring of respiratory diseases to anyone with access to an internet-connected microphone (e.g., smartphone), using voice as an indicator. The technology will administer simple tests in minutes and deliver results in seconds, without requiring specialized user training. The anticipated outcome is a widespread, real-time screening, monitoring and exacerbation warning system that remotely analyzes voice signals for patients with chronic and acute respiratory diseases, including COVID-19.This Small Business Innovation Research (SBIR) Phase I project seeks to develop voice-based classifiers that diagnose COVID-19 and monitor the severity of the disease. Existing algorithms that detect vocal biomarkers in breath and speech indicative of lung function and respiratory disease will be extended to incorporate COVID-19 signatures. Audio recordings from patients receiving a positive COVID-19 test will be collected to extract micro -signatures and develop algorithms to automatically recognize and map patterns to clinical findings and reported symptoms. The research objectives include developing: (1) A binary classifier that differentiates symptomatic and asymptomatic patients; (2) A multi-class classifier that correlates (in future predicts) changes in the severity of a patient’s symptoms when provided a series of voice samples (3) Developing a dashboard for physicians that provides up to date reports and visualizations of the cross sectional and longitudinal analytics (4) An API giving lung function metrics and classifiers available for integration into 3rd party IT infrastructure.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.