COVID-19 Outcome Prediction and Monitoring Solution for Military Hospitals in South Korea: Development and Evaluation of an Application (Preprint)

Abstract

<sec> <title>BACKGROUND</title> <p>COVID-19 has officially been declared a pandemic. The spread of the virus is placing extraordinary and sustained demands on public health systems. There is speculation that the differences in mortality rates between regions are a result of an abundance and availability of medical resources. The selection of patients for diagnosis and treatment is essential in a situation where medical resources are scarce. Military personnel are especially at risk of such an infectious disease, and patient selection with an evidence-based prognostic model is critical.</p> </sec> <sec> <title>OBJECTIVE</title> <p>This study presents and assesses the usability of a novel platform to gather data and deploy a patient selection solution for COVID-19, as used in Korean military hospitals.</p> </sec> <sec> <title>METHODS</title> <p>The platform structure was developed to provide users the prediction results, while simultaneously using the received data to, in turn, enhance the prediction models. Two applications were developed: a patient’s application and a physician’s application. The primary outcome for the models was the need for oxygen supplements by patients with COVID-19. The outcome prediction model was developed using data from patients from four centers. A Cox proportional hazards model was developed. The outcome of the model for the patient’s application was the length of time from the date at hospitalization to the date at first usage of oxygen supplements. The patient’s demographic characteristics, past medical history, symptoms, social history, and body temperature were considered as risk factors. A usability study was conducted on the physician’s application, with the Post-Study System Usability Questionnaire answered by 50 physicians.</p> </sec> <sec> <title>RESULTS</title> <p>The patient’s application and physician’s application are deployed online for wider availability. A total of 246 patients from four centers were used to develop the outcome prediction model. A small percentage (7.32%) of the patients required professional care. The variables included in developing the prediction model were age, body temperature, pre-disease physical status, history of cardiovascular disease, hypertension, visits to a region with a known outbreak of COVID-19, and symptoms of chills, feverishness, dyspnea, and lethargy. The overall C-statistic was 0.963 (SE, 0.014), and the time-dependent area under the ROC curve ranged from 0.976 at 3 days to 0.979 at 9 days. The usability of the physician’s application is good, with the overall average response to the PSSUQ being 2.2 (SD 1.1).</p> </sec> <sec> <title>CONCLUSIONS</title> <p>The platform introduced in this study enables evidence-based patient selection in an effortless and timely manner, which is critical in the military. With well-designed user experience and an accurate prediction model, this platform may help save lives and contain the spread of the novel virus, COVID-19.</p> </sec>