Clinical Decision Support Tool and Rapid Point-of-Care Platform for Determining Disease Severity in Patients with COVID-19

Michael P McRae; Glennon W Simmons; Nicolaos J Christodoulides; Zhibing Lu; Stella K Kang; David Fenyo; Timothy Alcorn; Isaac P Dapkins; Iman Sharif; Deniz Vurmaz; Sayli S Modak; Kritika Srinivasan; Shruti Warhadpande; Ravi Shrivastav; John T McDevitt

The p-BNC, a point-of-care biosensor system with the capacity to learn, is adapted here for the application of COVID-19 disease severity. Figure 1 highlights the key diagnostic infrastructure required to complete the integrated biomarker assays as needed to establish the COVID-19 Severity Score. From a small amount of patient sample (~100 μL serum), the cartridge and instrument perform a multistep assay sequence to 'digitize biology' by converting fluorescence immunoassay signal into biomarker concentrations. Statistical learning algorithms trained on data of biomarker studies predict a spectrum of disease. The result is a single value score which can be displayed to patients and providers in a mobile health app or directly on the instrumentation completing the test. Previously, we published a general framework for implementing a point-of-care based clinical decision support system. 17, 22 Here, we have adapted these methods to the task of predicting mortality in patients with COVID-19. It should be emphasized that while the integrated testing and COVID-19 Severity Score reporting here articulated represent what is arguably the most efficient delivery mode, the scaling and regulatory approval for this test ecosystem will take several months to complete. With the imminent arrival of the peak of the COVID-19 pandemic, it is important to emphasize that the COVID-19 Severity Score can be generated immediately using biomarker measurements collected from multiple separate test platforms. We anticipate this contribution could have an immediate impact on COVID-19 patient management, and we plan to promptly distribute the COVID-19 Severity Score capabilities following additional model refinement and validation.