Models must be able to calculate the risk of a coronapatient be in intensive care or in the respirator.

Artificial intelligence should try to evaluate more accurately which patients with the coronavirus, which ends up in the intensive care unit or in the worst case in the respirator.

computer Scientists at the University of Copenhagen is in the process of building these models, where all data about the patient into a supercomputer.

One of them, who is behind, is professor and head of department at the Department of computer Science Mads Nielsen.

He explains that the whole sygdomshistorikken for the individual patient will be taken.

– Now we try to take hold of their entire medical history to see if we can see a pattern in the patients. So we can see which diseases have the greatest risk to end up in intensive care, says Mads Nielsen.

the Project is a collaboration between research groups from copenhagen university hospital and Bispebjerg Hospital, and it is still in the very early stage.

It is the goal that the first coarse model is ready in a few weeks. The model should for example be able to calculate how high of a risk each patient has of having in the respirator, and when it will happen, ” explains the professor.

– It will never be completely finished, because every time the more knowledge, the better.

– We hope that we the next few weeks can bring knowledge to the table, which means that we can help Denmark to understand where we are in the coronakrisen, says Mads Nielsen.

I currently use the Danish Health and medicines authority and Statens Serum Institut (SSI) a number of overall static models to find out how many patients who may have needs for intensive treatment.

According to Mads Nielsen is the goal in the first place, that the model must be more accurate than the overall statistical model, which is now.

It is the somewhat maligned Sundhedsplatform, which has made it possible to make computermodellerne. It is due to the patients ‘ data is digital, which makes it easier to put them into the machine, but it takes time.

– There are thousands of variables on each patient, as the machine must understand in the right way. It takes time for us, and, therefore, is not just ready in the morning, says Mads Nielsen.

It is also not only the current coronakrise, where the artificial intelligence can be helpful.

It is possible to retrain the machine to other types of diseases. The professor also hopes that the machine can help the countries in the world, there are shorter along in their coronaforløb.

In the first place will computermodellerne could be used in the Capital Region, but the goal is that they should be implemented in the whole country.

/ritzau/