Artificial intelligence outperforms doctors in intensive care units
Artificial intelligence outperforms doctors in intensive care units
In the future,Artificial Intelligencewill have an important role in the medical sector, especially since we are now seeing some successful experiments with artificial intelligence in diagnosing diseases, and this is by analyzing images with great accuracy, and checking whether they show disease symptoms or not. The obstacle facing this technology now is the difficulty of training artificial intelligence to examine patients’ changing conditions over time. Therefore, proposing an appropriate treatment would be difficult, but this hurdle has been overcome by researchers at the Technological University of Vienna in cooperation with the Medical University of Vienna.
With the help of a huge amount of data from intensive care units in various hospitals, an artificial intelligence technology has been developed that can suggest treatments for patients who need intensive care due to a serious condition known as sepsis. Analyzes show that artificial intelligence has actually been able to outperform the quality of human decisions in recommending treatment for these conditions.
For those who do not know, sepsis is a serious condition that results from a group of reactions outside the patient’s body, as a result of picking up germs, and this prompts the immune system to release chemicals; In order to confront these germs, but if the immune system’s reaction is exaggerated, it can then cause infections in the bloodstream or the affected organ, which may ultimately lead to sepsis.
Making the best use of data
In the intensive care unit, a lot of different data is collected around the clock, as patients are constantly subject to medical monitoring. So the research team wanted to verify whether or not it is possible to use this data better than human capabilities.
Professor Clemens Hitzinger from the Institute for Scientific Computing and Analysis at the Technological University of Vienna, who is also co-director of the Center for Artificial Intelligence and Machine Learning (CAIML) at the University of Vienna, says: “Medical staff make their decisions on the basis of well-established medical rules, and most of the time these medical staff are able to be aware of the variables that they must take into account; In order to provide the best care. Despite this, a computer can take into account large amounts of information more easily and with greater ability than a human, which can – in some cases – lead to better decisions regarding proposed treatments.”
The computer as a means of planning
Hitzinger says: “In our project, we used a form of machine learning called reinforcement learning, which is a type of machine learning by interacting with the surrounding environment, which allows artificial intelligence to perform a set of complex operations. This is compared to the previous role of artificial intelligence in the medical field, which was limited to classifying a group of images to track the condition of a tumor, for example, but the technology we developed can predict the state that the patient can reach, and it is a technology that there was little research on in the field. Medical”.
By using this new technology, the computer can become an agent with decisions it makes on its own. To determine whether the patient is well or not. The secret behind this new research is to use a hypothetical computer reward and punishment system. If the patient improves, the computer will reward him, and if his condition deteriorates, he will be punished. Therefore, the computer working with this technology develops over time an autonomous task to maximize this virtual reward by taking actions that improve the patient’s condition. In this way, comprehensive medical data can be used to determine a strategy with a high probability of success.
Artificial intelligence is better than humans
Professor Oliver Kimberger from the Medical University of Vienna says: “Sepsis is one of the most common causes of death in intensive care medicine, and it poses a major challenge for doctors and hospitals, as early detection and appropriate treatment are essential for the patient to survive, but – unfortunately – there have been few medical breakthroughs in this field, which makes the search for new treatments and methods more urgent; This is why investigating the extent to which AI can help improve medical care is particularly interesting, and the use of machine learning models and other AI techniques is an opportunity to improve the diagnosis and treatment of sepsis, and this will ultimately increase the patient’s chances of survival.”
The analysis shows that the capabilities of artificial intelligence are actually superior to those of humans in this particular area, as recovery rates are now higher using the AI strategy, compared to purely human decisions; In one of the studies of this research team, the recovery rate in terms of mortality within 90 days (90-day mortality) was increased by about 3%.
The above – of course – does not mean that the doctor should leave all medical decisions in the intensive care unit to the computer; The expected role of artificial intelligence is to act as an additional device alongside the beds, where the medical staff can refer to it and compare their own assessment with the artificial intelligence’s suggestions.
Legal and ethical aspect
This new research raises many legal and ethical questions, the most prominent of which is: Who will bear responsibility for any mistakes made by artificial intelligence? This leads us to another, more complex question: What if the AI made a correct decision, and then the human chose another treatment option, and the patient was harmed – would the doctor then face the accusation that it would have been better to trust the AI? Because he decides based on huge amounts of data? Or is it a human right to ignore computer advice at all times?
The research project shows that artificial intelligence can indeed be used successfully in clinical medicine, using the technology we have today, but there is still an urgent need to discuss the social framework and legal rules. In order to regulate the use of artificial intelligence in the medical sector; Because this sector has a special nature; A simple decision could mean life, whether saved or lost.
