Unmet Need
Globally, there have been 73,275,943 confirmed cases of COVID-19, including 1,650,348 deaths reported to the World Health Organization (WHO) (WHO, 2020). The symptomatic infection ranges from mild to critical. Severe illness can occur in healthy individuals of any age, but it predominantly occurs in adults with advanced age or certain underlying medical comorbidities. Many prediction models have been proposed to identify patients with increased risk of infection, severe illness, and mortality from the disease. The models are based on epidemiologic, clinical, and laboratory factors such as comorbidities, sex, and lymphocyte count; however, most of the studies evaluating these models are limited by risk of bias, as well as unclear and varied prediction horizons (Wyants, et al., 2020). No model has been prospectively evaluated or validated for clinical management (UpToDate, 2020). Without validated models, healthcare systems struggle with triaging patients and appropriately allocating limited healthcare resources during outbreaks. Therefore, there is a strong need for effective prediction tools of COVID-19.
Technology Overview
Johns Hopkins researchers have created a series of prediction models for identifying severe disease progression and mortality in COVID-19 patients. The proposed prediction model outperforms other available models by using longitudinal data to refine the predication. With the proposed predication model, frontline healthcare workers can be more informed as they confront the ongoing challenges in decision making and logistical planning in caring for COVID-19 patients.
Stage of Development
Model development is complete and refinement is in progress.
