His rapid growth artificial intelligence opened up a discussion about professions that are “at risk” of being replaced by systems AI (artificial intelligence). “Doctors and nurses are not included in them,” experts emphasize.

One of the key applications of artificial intelligence systems concerns clinical decision support. With the ability to quickly analyze large amounts of data, AI algorithms can uncover patterns in various health conditions and help healthcare professionals make definitive diagnosis and treatment decisions. “Now the biggest application of AI systems in healthcare is image analysis. For example. with CT, mammography, etc. it can detect malignancies at stages that the human eye cannot distinguish. But in no case can this replace a doctor. notes on “K” Professor of Biology/Genetics – Nanomedicine, Faculty of Medicine EKPA, Maria Gazulis. Professor of Health Policy, Dean of the Faculty of Social and Political Sciences, University of Peloponnese, Kyriakos Souliotis, underlines in “K” how “it must be recognized that with the complexity caused by the combination of scientific research, rapid technological development, the availability of large amounts of data, and the expansion of digital functions in many areas of health systems, artificial intelligence becomes by definition necessary. This condition, however, does not mean that it can evolve to replace the human factor, i.e. with automated decision making based on diagnostic and therapeutic algorithms. They are useful provided that they are evaluated by physicians who will make decisions in collaboration with patients. In short, it’s a system that thinks very fast but doesn’t make decisions.”

“Diagnostic and therapeutic algorithms are useful provided they are evaluated by clinicians.”

Last week World Health Organization warned against the use of artificial intelligence, stressing, among other things, that the data it uses could be biased or misused. Ms. Gazulis points out that “as developers create AI systems, there are several risks and issues: risk to patients due to system errors, to privacy, and more.”

Professors emphasize the need to form an institutional framework as soon as possible that will regulate all critical points of the application of artificial intelligence in health. “The last thing we want is to lose the opportunity to take advantage of artificial intelligence, because we do not have a ready-made “environment” for its application,” emphasizes Mr. Souliotis. According to him, our country must move quickly in the field of digitization of files in order to keep up in this area. While we are talking about artificial intelligence and feeding data into systems, in Greece the doctor still requests information about previous treatments and test results from the patient himself.

FOR

Early prediction, faster development of new drugs

A tool with enormous potential in terms of improving the ability to diagnose diseases, supporting the clinical decision for the most effective treatment, as well as easier patient access to the healthcare system, characterizes healthcare professionals, artificial application systems.

Such as notes in “K” Mr. Kyriakos Souliotis, “The potential provided by artificial intelligence has many advantages, such as processing large databases in a short period of time and, based on this, supporting clinical decision making. Making treatment decisions based on real data on the outcomes of a large number of patients guarantees an increased likelihood of achieving the eternal goal of “the right treatment for the right patient”, while the benefits also extend to the economic functioning of health systems.

In addition, the digitization of all medical information of citizens and the ability to search and immediately find available structures and medical services will help improve the processes and conditions for citizens to access the healthcare system, such as, for example. in the intensive care unit, but also in the standardization of the use of medical services, which leads to an optimal allocation of resources.”

Ms. Maria Gazulis emphasizes that “although this field is quite new, artificial intelligence can play an important role in the healthcare system.”

The digitization of medical information and the ability to search for available healthcare structures will improve access to the healthcare system.

According to the professor, the most impressive application of medical AI is that “human” medical workers cannot do it yet. It can be used for image analysis such as x-ray and MRIto help physicians identify diseases and plan treatment. “For example,” he notes, “artificial intelligence algorithms can detect signs of cancer on mammograms with a high degree of accuracy, which can help doctors diagnose and plan treatment faster.”

Artificial intelligence systems can give a timely forecast. According to Ms Gazulis, electronic health records and other patient data can be analyzed using artificial intelligence to predict which patients are at risk of developing certain diseases. This can help doctors intervene early, before the condition becomes more serious, and can also help healthcare organizations allocate resources more efficiently.

Its scope is also the development of new drugs. Artificial intelligence can be used to study data on drug interactions and side effects, and to predict compounds that will be more effective in treating certain conditions. This could speed up the process of drug discovery and development, which could eventually lead to new treatments for patients.

Remote Diagnostics

AI can share the experience and work of experts to complement health care providers such as a general practitioner in a remote location. Ms. Gazulis explains that “for example, many artificial intelligence programs use images of the human eye to make diagnoses that would otherwise require the intervention of an ophthalmologist.”

AGAINST

Risk of bias and information leakage

Ensuring the confidentiality and availability of reliable and representative data is a key challenge, as well as a prerequisite for the use of artificial intelligence systems in healthcare.

Mr. Kyriakos Souliotis states that “Undoubtedly, this is something relatively new in healthcare, and there are reasonable concerns about the risks, for example, from submitting incorrect data to AI systems or leaking confidential and confidential information about the health of citizens. In addition, concern was expressed about the risk of gradually replacing the human factor with health services, with multiple implications for employment and the quality of health care.

As Ms. Maria Gazulis notes, “When it comes to the application of artificial intelligence in healthcare, privacy is a major concern. Patient data consists of highly sensitive personal information (PII), such as medical history, identification and billing information, which is protected by regulatory guidelines under the GDPR and HIPAA.

“Lack of transparency can make it difficult for doctors and other healthcare professionals to trust the results of an artificial intelligence system.”

But when the data is uploaded to the cloud for analysis by AI systems, there are open security concerns.” According to the professor, the risk also concerns the availability of data. “Training AI systems requires large amounts of data from sources such as electronic health records, pharmacy records, insurance company records, or consumer-generated information such as fitness trackers or purchase history. But medical data is often problematic and fragmented,” he notes, and continues: “Patients often turn to different health care providers and change insurance companies, which leads to the separation of data across multiple systems and multiple formats. This fragmentation increases the risk of error, reduces the completeness of datasets, and increases the cost of data collection, which also limits the types of entities that can develop effective healthcare AI.”

inequalities

In the same context, there is also the risk of AI bias and inequality in health. As Ms. Gazulis points out, AI systems learn from the data they train on and can account for bias based on that data. “For example, if the available data for AI is collected primarily at academic medical centers, the resulting AI system will know less about patients from populations that do not normally attend academic medical centers. Even if AI systems learn from accurate, representative data, problems can arise if this information reflects underlying biases and inequities in the healthcare system.”

Lack of transparency is also a problem. “Many artificial intelligence systems are considered “black boxes” because it is difficult to understand how they came to a particular solution. The lack of transparency can make it difficult for doctors and other healthcare professionals to trust the results of the AI ​​system,” emphasizes Ms Gazulis.