eCOVID: Enhanced Monitoring and Guidance for Ambulatory Patients with COVID-19 using Wearables and Machine Learning
The objective of this project is to enable effective remote monitoring, digital triage and personalized guidance for ambulatory patients with COVID-19 using wearables and machine learning. The project has two components:
- Provide immediate relief and improve efficiency of overburdened healthcare systems by augmenting and eventually replacing the current practice of COVID-19 ambulatory care which relies on manual status monitoring of ambulatory patients through phone calls, and leads to premature or even late hospital visits by patients, both draining healthcare worker and PPE resources, and possibly resulting in non-optimal care. Our objective is to augment and replace the current system with an automated continuous remote monitoring and digital questionnaire-based patient data collection and reporting system, which allows COVID-19 health care team to perform a much better informed and orderly digital triage, providing data-driven personalized care suggestions through our app, phone calls and/or video visits, and ensuring timely hospitalization if needed while reducing unnecessary hospital visits. The above will lead to significant improvement in ambulatory COVID-19 patient care while at the same time enabling significant savings in human, PPE and equipment resources.
- While the remote monitoring based digital triage system will have timely clinical benefits for the ambulatory patients and healthcare systems, the data collected in the process can potentially reveal aggregate trends in those who recover or deteriorate further so as to inform how we think about COVID-19 ambulatory patient monitoring moving forward. We plan to develop machine learning (ML) and deep learning (DL) models to further transform multimodal raw data to actionable insights. Such insights can help better identify the symptoms, help determine which symptoms and vitals might be most important for a patient and predict the progression for each patient in a personalized way so as to enable personalized guidance, and help optimize preparedness and utilization of healthcare resources. For example, ML/DL models can provide insight into which specific variables are the best predictors of worsening/improving health in COVID patients and how these factors vary among patients. The above can lead to better diagnosis and treatment regimens as well as better forecasting and preparedness for COVID and related pandemics that we are unfortunately to experience in the future.
- e-COVID is deployed with our Healthcare Partners UC San Diego Health and Neighborhood Healthcare.
If you are a researcher, physician, healthcare provider, healthcare technology provider, or charitable foundation, and would like to participate in this project, please send email to firstname.lastname@example.org.