Predictive Medicine is a Reality, Responsibility Remains with Doctors

You are mistaken if you think that only the doctor can see the condition of your heart, blood vessels, brain, kidneys, stomach, or any other organ… The device that displays the organ during the examination records various data that artificial intelligence (AI) simultaneously collects, analyzes, and suggests a diagnosis. Advances in computer science and the development of new technologies have made artificial intelligence indispensable in healthcare; AI algorithms and applications powered by support assist doctors in disease research and diagnostics. Although general standards for the use of artificial intelligence in medicine are still being defined, the potential of AI in healthcare is constantly growing, and artificial intelligence will soon become a key part of the digital health system that shapes and supports modern medicine. This will not only facilitate diagnosis and therapy selection in treating patients but will also compel doctors to engage in continuous improvement.

The speed of acquiring new knowledge

The speed of generating new knowledge in medicine is such that, according to IBM data, it doubled every 73 days less than five years ago. In 1950, medical knowledge doubled every 50 years, but at the beginning of the 21st century, thanks to new technology, there was a sudden leap. By 2010, the speed of generating new knowledge in medicine had decreased to just three and a half years, and by 2020 to an incredible 73 days.

This poses a significant challenge for healthcare professionals as it means they must constantly learn to keep up with the development of new technology and its applications in medicine to ensure effective patient care. In this context, the traditional model of memorization is no longer sufficient. Medical education must focus on fostering advanced skills, critical thinking, and the ability to engage in lifelong learning and application of information so that doctors can provide the best possible care to patients.

There is an increasing need for the use of technology, such as data analysis and decision support systems, to assist in the effective management and application of growing medical knowledge. However, using AI tools requires new knowledge, so artificial intelligence will not only change the way diagnostics are performed in healthcare but will also bring significant changes to the education of doctors.

The impact of AI

There are numerous ways in which artificial intelligence positively impacts medical practice, either by accelerating the pace of research or by helping doctors make better decisions based on big data analysis. For example, AI tools improve image analysis in radiology and pathology, leading to more accurate and faster diagnoses.

It is used to analyze X-rays and CT scans, magnetic resonance imaging, and other images of the body’s condition that a radiologist might miss, and whose results assist in making treatment decisions, medications, mental health, and other patient needs by providing them with quick access to information or relevant research. By processing large datasets, AI identifies patterns and biomarkers for diseases, thereby accelerating diagnosis and treatment planning.

Research has shown that artificial intelligence, powered by artificial neural networks, can be as effective as human radiologists in detecting signs of breast cancer and other diseases. In addition to helping identify early signs of disease, artificial intelligence also aids in managing the large number of medical images that doctors must track through a patient’s medical history, thus reducing the time required for diagnosis.

Moreover, during clinical trials, too much time is spent assigning medical codes and updating relevant patient datasets, and artificial intelligence can speed up this process by providing faster and smarter searches.

The example of two clients of IBM Watson Health (formerly part of IBM, now a company called Merative L.P.) shows that with artificial intelligence, they can reduce the time spent searching for medical codes by more than 70 percent.

Risk management

Additionally, machine learning models could be used to monitor patients’ vital signs in intensive care and alert doctors if certain risk factors increase. A medical device like a heart monitor tracks vital signs, and artificial intelligence collects data from it and looks for more complex conditions such as the onset of sepsis in the body. For instance, one IBM client developed a predictive artificial intelligence model for premature babies that is 75 percent accurate in detecting severe sepsis.

Among the advantages is that AI models can learn and provide patients with personalized recommendations in real-time, 24 hours a day. Instead of repeating information each time with a new person, the healthcare system could offer patients 24-hour access to a virtual AI assistant that would answer questions based on their medical history, preferences, and personal needs.

In addition to the advantages in analyzing and connecting a large number of medical data, artificial intelligence also enables rapid drug development, making research faster and cheaper. Artificial intelligence could help overcome many challenges of big data faced by life sciences. Integrating medical artificial intelligence into medical workflows can provide service providers with valuable context when making care decisions. A trained machine learning algorithm can help reduce research time by providing doctors with valuable evidence-based search results about treatments and procedures in real-time, at the moment of patient examination.

Furthermore, there are many potential ways in which artificial intelligence could reduce costs in healthcare. Some of the most promising opportunities include reducing medication errors, personalized virtual health assistance, preventing fraud, and supporting more efficient administrative and clinical processes.

Many patients need answers to questions outside of doctors’ working hours, and artificial intelligence can help provide all-day support through chatbots that can answer basic questions. However, artificial intelligence is not smart on its own; it was created by humans who ‘feed’ it information. While it is good for quickly gathering, screening data, and recognizing problems, in medicine, the doctor remains paramount, and artificial intelligence is merely an assistant that must ‘know’ how to read.

The future of health and telemedicine

Artificial intelligence also assists Croatian doctors, as discussed at the recently held conference ‘Next of Health & Telemedicine’. The theme of the mentioned conference held in Zagreb, organized by the Croatian Society for Telemedicine and the Croatian Society for Medical Robotics and Artificial Intelligence in Medicine, was how artificial intelligence (AI), robotics, and telemedicine affect treatment. Doctors spoke about the application of artificial intelligence in ophthalmology, cardiology, oncology, drug research, therapy… and the advantages and limitations of its use in medicine.

– When experts train artificial intelligence systems, they train them based on our foundational knowledge, on the algorithms we already have. There is an opportunity to expand so that artificial intelligence seeks data and recognizes the causes of diseases. Some rare diseases are not recognized during examinations, which is a great shame because they can be treated, and artificial intelligence can help with that. Additionally, the diversity of patients and the entire clinical context is something that often falls outside of these algorithms, and it seems to me that we must remain curious and talk to patients and obtain other data that could influence their diagnosis. In the end, the question remains of who is responsible. We are responsible for making decisions; the decision lies with the doctor. I do not think that artificial intelligence will replace us, but the problem is how to resolve some discrepancies with artificial intelligence because we cannot get an answer on how it arrived at a certain conclusion, said Dr. Vlatka Rešković Lukšić, a cardiologist at the Heart and Blood Vessel Clinic of KBC Zagreb.

Open possibilities

An important topic at the conference was the future of education under the influence of AI, discussed by Mislav Balković, rector of Algebra Bernays University and president of the Croatian Employers’ Association, and Prof. Dr. Sc. Stjepan Orešković from the Medical Faculty of the University of Zagreb, a member of the European Academy of Sciences and Arts and the majority owner of Bosqar Invest.

– The development of neuroscience on one side and big data on the other opens numerous possibilities. Although humans have many advantages over technology, one of the advantages of technology is rapid connectivity and updating, said Balković, explaining that education in the future will need to focus on non-routine jobs and very complex elements and tasks.

Orešković spoke about the curriculum 2040 – preparing doctors for intelligent healthcare.

– The traditional curriculum is based on memorization, not on systematic thinking and interpretation. This needs to change by learning how to solve the most difficult problems at any moment. If we train people in future medical education to think about problem-solving every day and train them for logical behavioral-cognitive capacities that enable that solution, then we are educating them for the future regardless of how many mechanical facts may change in the meantime or how much ChatGPT accumulates past knowledge. The curriculum 2040 must create not only competent clinicians but also reflective leaders who must be ethicists, scientists, interpreters, and communicators, Orešković explained, emphasizing that the capacity of artificial intelligence in healthcare is very large, but for the system to benefit from it, it needs experts who will continuously acquire and refine knowledge.