Role of artificial-intelligence powered digital-health in reducing strain on health-care-system: First of its kind experience in COVID-19 pandemic

ABSTRACT

Background: The second wave of COVID19-Pandemic was associated with massive case surge causing overwhelming of health-care-system. Early monitoring and identification of at-risk patients through digital mode, can reduce mortality and judicious use of hospitalresources. Objectives: To study the efficacy of CareShare®- application in predicting hospital-admission for homeisolated- mild-COVID19-patients (HIMCP). Methods: CareShare® an android-application, designed in collaboration, to follow-up HIMCP. A baseline-profile is entered on first contact and includes demography, comorbidities, and current symptoms. Application generated a baseline-symptom-score for follow-up. Subjects were automatically reminded to refill their symptom-severities on daily basis. Change in symptoms-score led to one of the three-flags Red (critical), Orange (review) and Green (Safe). Results: Over 2 months, 550 patients were screened and 485 were enrolled on the application-platform with 2511 entries. Most-common reason for missing out was lack of an android-phone. The mean age was 35.1±12 years with 40.8% being females and 82.1% from urban background. Average number of entries/ patients were 5.18±4.9 (1 to 22). 1072 daily entries were given orange-flag, 298 were reviewed on call and remaining were reclassified as falsewarnings. Out-of-298, 16 were called in hospital and 6 were admitted. 690 entries were given a red-flag, leading to urgent-review, and calling of 108. 86 cases were called for in-person review, leading to admission of 27 patients. Out of 33, 27 required ICU admission. There was no mortality in the study population. Conclusion: CareShare® is a reliable method of symptommonitoring of HIMCP taking pressure off overwhelmed health care system.