Despite these attempts, the drug errors are still occurring with disturbing regularity.
“I have read some studies that say mainly 90 % of anesthesiologists Dr. Kelly Michelsen, a Uddran colleague at UW Medicine and an anesthetic professor at Washington University, said that there was an error of a drug at some point in their career.
As a medical professional and a trainer, he was surprised to discover a mistake about making a mistake, and alerting anesthesiologists in the actual time, should be among the abilities of artificial intelligence. She said, “I was like, this looks like something that should not be very difficult for artificial intelligence to do.” “Ninety-nine percent of the medications we use are these 10-20 medications, and therefore my idea was that we could train artificial intelligence to identify and act as a second group of eyes.”
the study
Michaelsen focused on the vials ’exchange errors, which represent About 20 % Of all the errors of the drugs.
All medications that were injected into flasks carry signs, which are then transferred, come to a syringe that carries signs on a drug cart in the operating room. But in some cases, someone chooses the wrong vial, or the injection is classified incorrectly, and the patient is injected with the wrong drug.
In one of the notorious vials’ exchange error in particular, a 75 -year -old woman was injected into the Medical Center at Vanderbelt University in Tennessee with a deadly dose of paralyzed drugs instead of restored disorder, which led to her death and a subsequent criminal trial.
Mikesen believed that these tragedies can be prevented through “smart glasses”-adding a cosmetic that can be worked on behalf of the protective glasses that all employees wear during operations. By working with her colleagues in the Department of Computer Science at the University of Washington, she designed a system that can erase the immediate environment for injection and decision -making designations, read it and reveal whether they are identical.
She said: “He is enlarged on the poster and discovers, for example, propofols inside the syringe, but Oondsetron is inside the bottle, and thus produces a warning.” “Or that the two silences are identical, so everything is good, keep your day.”
The construction of the Michaelsen device and its team took more than three years, and half of it spent approval to use the pre -recorded video flows for anesthesia specialists in preparing medicines properly inside the operating room. Once the green light was given, I managed to train artificial intelligence on this data, along with additional shots – this time in preparing a laboratory – of the mistakes made.
“There are many problems related to exhaustion in the operating room, so we had to make sure that it works very well, and it can do an ideal work in discovering errors, and so on [if used for real] She said: “It will not give false warnings. For clear ethical reasons, we were unable to make mistakes on the purpose of the patients concerned, so we did so in the simulation operating room.”
in study It was published late last year, Mikesen stated that the device discovered the bottom of the bottle of the vials with a resolution of 99.6 %. All that remains is to determine the best way to transmit warning messages and can be ready for use in the real world, waiting for food and medicine removal. The study has not been funded by Amnesty International Technology.
