Artificial Intelligence To Help Intensive Care Doctors
A team of systems engineers from the University of Sheffield is developing an intelligent computer system which imitates a doctor's brain to make treatment decisions for intensive care patients. The system will take some of the workload from emergency medical teams by monitoring patients' vital signs and then evaluating and administering the right amounts of different drugs needed: a job usually carried out by specialist medical doctors.
The team, led by Professor Mahdi Mahfouf in the University of Sheffield's Department of Automatic Control and Systems Engineering, is pioneering the intelligent decision-support system which, in effect, duplicates the decision making processes of specialist medical doctors in Intensive Care Units (ITU).
The system models all the possible interactions between different drugs and patients' bodies, and then makes intelligent decisions about the best way to treat patients during heart bypass operations, and post-operatively in the ITU. This unique system can decide on the types and quantities of drugs to give to patients in a matter of seconds. This will help doctors provide effective treatment for patients, whilst allowing them to concentrate on as many other important tasks as possible.
Professor Mahdi Mahfouf of the University of Sheffield explains that it is the system's ability to learn, adapt, and make informed decisions which is unique: "This new system not only monitors and treats critical patients, but it can also learn from the experiences of medical staff, who can override the machine at any time. If overridden, the system assimilates the doctor's input and uses the new information to make decisions about similar cases in the future.
"This system is not intended to replace the work that doctors do in intensive care units. However it will provide them with invaluable assistance by evaluating the complex interactions of different drugs which are needed to treat patients and protect them against the danger of septic shock."
The research is being funded via two grants from the EPSRC totalling