Could a computer-assisted model help predict sepsis?

Sepsis is a leading cause of death in hospitals. Could the first symptoms be detected with the help of a computer-badisted model?

A computer-badisted model has been shown to be promising using the data collected regularly to identify early symptoms of sepsis.

The new study entitled National Computer Aided Early Warning Score to Predict the Risk of Sepsis After Emergency Medical Hospitalization: Model Development Study and External Validation, was published in CMAJ (Journal of the Canadian Medical Association).

Sepsis

Early detection is crucial to prevent deaths from sepsis. Every hour of delay is badociated with a seven percent reduction in survival, but these delays are common.

There are currently several scores for identifying patients with sepsis, including the National Early Warning Score (NEWS), used in National Health Service (NHS) hospitals in the United Kingdom.

CNews

The National Computer Assisted Early Warning (CNEWS) score has been developed to determine if it is possible to improve the accuracy of the prediction of sepsis.

The cNEWS score can trigger Screening for sepsis, usually within 30 minutes of admission, once the information collected systematically has been entered electronically into the medical record.

cNEWS could be carefully presented to hospitals with an appropriate and evaluated IT infrastructure.

Using the computer-badisted model to detect symptoms of sepsis

Professor Mohammed A. Mohammed, of the University of Bradford, UK, said: "The main benefit of these computer models is that they are designed to incorporate data present in the patient record. , can be easily automated and do not represent any additional burden for the hospital. staff to gather additional information. "

Professor Mohammed explained, "These risk scores should support, rather than replace, clinical judgment. We hope that they will raise awareness about sepsis by providing additional information on this serious disease. The main benefit of these computer models is that they are designed to incorporate data present in the patient's chart, can be easily automated and do not impose any additional burden on hospital staff for collection. additional information. ".