Study Shows AI Can Help Prevent Sepsis Deaths by 20%
A new study from researchers at Johns Hopkins University shows that an AI algorithm was able to detect early illness and prevent sepsis deaths by 20%.
To conduct the research, provider interactions with a sepsis early detection tool were developed at five hospitals over a two-year period.
Among 9,805 retrospectively identified sepsis cases, the early detection tool achieved high sensitivity with 82% of cases identified.
Sepsis is the body’s extreme response to an infection that triggers a chain reaction throughout the body.
It impacts vulnerable populations, such as newborns, pregnant women and people living in low-resource settings — with 85% of sepsis cases and sepsis-related deaths occurring in these settings.
The findings, published in Nature Medicine and Nature Digital Medicine on July 21, suggest that a machine learning program was able to warn thousands of health care providers about patients at high risk of sepsis, allowing them to begin treatment sooner.
According to data from the World Health Organization, sepsis kills 11 million people globally each year. Nearly 270,000 Americans die each year as a result of sepsis.
The WHO indicates there is an urgent need for better data as most published studies on sepsis have been conducted in hospitals and intensive care units in high-income countries.
To prevent sepsis, improved sanitation, water quality and availability and infection prevention measures must be available, which the WHO finds could prevent as many as 84% of newborn deaths due to sepsis.
Alexa can be reached at [email protected]
