Bioinformatics and Biostatistics lab, Seoul National University

COVID-19 Nomogram

Predicting maximum clinical severity score using initial clinical information

Purpose of this nomogram is to visualize maximum clinical severity score(max CSS) prediction based on initial & epidemiological data from COVID-19 patients. COVID-19 nomogram visualizes effects of individual risk factors on the predicted probability. We are developing many models from different data sources and are planning to update and maintain them in this website.
* Mobile version is not yet supported. Using PC is highly recommended.

KDCA Public Model estimates probabilities of max CSS using three different binary classification models. It can be utilized, if a patient has initial laboratory results. This model was developed using KDCA public data.

KDCA Model1 estimates probabilities that the max CSS is severe or critical. Since there are variables that require medical diagnosis, it can be applied to hospitalized patients.

KDCA Model2 estimates probabilities that the max CSS is severe or critical. It differs from KDCA Model1, in which, Non-medical person can simply use questionnaire, to obtain the variables used in the model. The advantage is that it can be used immediately for initial confirmed patients at public health centers.

KDCA Model3 estimates probabilities that the max CSS is critical. This model is useful for detecting patients who need urgent care and immediate hospitalization.

Maximum clinical severity score is defined as the maximum value of the clinical severity score during hospitalization. The definition of clinical severity score can be found in the adjacent table (on the right).

Using eight stages of CSS, for analytical and practical purposes, we re-grouped them into four classes labeled mild, moderate, severe, and critical. All the prediction models are built on this redefined CSS of four classes.