The collaboration between the Haukeland University Hospital and Holberg EEG has led to pioneering development of artificial intelligence (AI) in the healthcare system.
For the first time, artificial intelligence can be used for interpretation Electroencephalogram.
An electroencephalogram (EEG) is a test of brain activity, often used to screen for epilepsy. It is also important to follow up patients with epilepsy.
According to the World Health Organization, around 50 million people worldwide suffer from epilepsy.
Patients will benefit from this. We can screen more patients because it takes less time to interpret an EEG, says Torhild Næss Vedeler, director of the neurological clinic at Hawkland University Hospital.
With the help of the new artificial intelligence model “SCORE-AI”, it will be possible to examine and diagnose epilepsy faster and more accurately.
We will not get more health workers and we will not get more money. Therefore, we must work with innovation, and that means we use fewer resources where possible, such as using artificial intelligence to interpret EEGs, Fiedler says.
the new era
One of the most reputable medical journals in the world, JAMA Neurologyto the breakthrough as a “new era”.
– The fact that it was published there shows that it is appreciated and noticed, says Harald Orlin.
In addition to being a Senior Clinical Neurophysiologist at Hawkland University Hospital, he is medically responsible for Holberg EEG.
– This is the first time you see that you can interpret an EEG using artificial intelligence as well as interpreting the best experts. He continues that they think we are pioneers.
Aurlien has worked with EEG since the mid-1990s. Using artificial intelligence to read these surveys will provide more security.
– When I interpret an EEG as an expert, sometimes I’m not sure what it means. Then I discuss with colleagues. This AI will become like a support player constantly watching over my shoulder.
Many get the wrong diagnosis
The result will be faster decisions and more accurate answers. In addition, Orlin believes that the use of artificial intelligence will contribute to reducing the number of people who receive the wrong diagnosis.
Many people are diagnosed with epilepsy without actually having it.
One-third of patients admitted to highly specialized epilepsy hospitals do not have epilepsy. With this new method, we hope there will be fewer misdiagnoses, he says.
When people get a wrong diagnosis, a common reason is that the EEG is being interpreted incorrectly, Orlin explains.
– It is undesirable to get an incorrect diagnosis. This can have consequences, for example, that you lose your driver’s license, have to take medicines with side effects or lose the possibility to practice certain professions.
EEG is also an expertise that is not available in many parts of the world. Artificial intelligence will be able to fulfill this need.
– It should not replace anything
At the Haockland University Hospital in Bergen, about 4,000 patients come in every year and need an EEG.
Holberg EEG has entered into an agreement with the largest supplier of EEG in the world. The plan is for the model to be released to the entire world within the next year.
– We’re very proud that we’re sitting here today, and have developed a world-leading model, says Oyvind Gulbrandsen, general manager of Holberg EEG.
The AI model was developed on the basis of a huge database full of EEGs that were done between 2014 and 2020.
– What sets Holberg EEG apart is our database. No one has a similar database, either in terms of size or quality.
He stresses that artificial intelligence should not replace doctors.
– Decision must be supported. It should not replace anything, Gulbrandsen says, but should be a tool and help for professionals.
– Many people fear artificial intelligence, but you also have to use the new technology that is available, he adds.
Holberg EEG will be the first in the world to offer this kind of artificial intelligence.
– This is just the beginning, says Gulbrandsen.
“Web specialist. Lifelong zombie maven. Coffee ninja. Hipster-friendly analyst.”