Point-of-Care Differentiation of Kawasaki Disease from Other Febrile Illnesses

A doctor explains electrocardiogram procedure to a young girl

No specific diagnostic test currently exists for the diagnosis of Kawasaki disease (KD).

The Issue:

This study aimed to see if statistical learning on clinical and laboratory test patterns would lead to an algorithm for KD diagnosis.


Key Findings

  • The number of correctly diagnosed KD patients and FCs was 67 percent of all FCs and 81 percent of all KD patients in the training models.

  • The number of correctly diagnosed KD patients and FCs was 63 percent of all FCS and 74 percent of all KD patients in testing models.


This study resulted in a KD scoring metric and associated data system available online. Additionally, smartphone applications were developed to provide clinicians with point-of-care access for real-time diagnosis.

The researchers acknowledge that clinical experience and judgment are important aspects of diagnosis, yet this KD scoring metric is a good example of an inexpensive, easily accessible tool to improve point-of-care differentiation.

About the Study:

Demographic, clinical, and laboratory data collected from 276 patients with KD and 243 febrile controls (FCs), who shared some features of KD, were used to train multivariate models. The models were tested by 136 patients with KD and 121 FCs. All subjects were seen at Rady Children’s Hospital San Diego. Using density plots, KD scores were separated into three groups: low KD score group allowing accurate FC diagnosis, intermediate KD score group with inadequate KD/FC differentiation, and high KD score group allowing accurate KD diagnosis.