Preventing Blood Poisoning With AI

Blood poisoning is the third most common cause of death in Germany – particularly affecting intensive care units in hospitals. Around 35 percent of patients affected by sepsis die. Especially for people with a weakened immune system, sepsis can be fatal if it is not treated in time. Medical professionals often have a very narrow window of opportunity to respond to sepsis.

Reducing the risk of sepsis

Now Telekom and the Hamburg-based start-up Telehealth Competence Center Analytics (TCC Analytics) want to reduce the risk of sepsis with an AI solution. The solution relieves the burden on medical professionals by alerting them to the threat of sepsis in a patient. After all, the earlier sepsis is treated, the greater the chances of survival. “The algorithm is constantly learning. It helps doctors assess whether patients are at risk,” says TCC Analytics CEO Professor Christian Storm.

Vital signs data of intensive care patients are collected in the hospital via standard interfaces and analyzed by an AI algorithm in the Open Telekom Cloud. This self-learning algorithm has been developed and trained by TCC Analytics. It can predict the individual risk around ten hours before the onset of sepsis and opens up the possibility of preventing sepsis. Clinics do not need to collect any parameters beyond the usual vital signs. A dashboard shows physicians the sepsis risk at a glance so they can act quickly.

The solution is currently being piloted at two hospitals. TCC Analytics and Telekom are in talks with other hospitals.