“Artificial intelligence can help the healthcare from an organizational and medical point of view, for example, by supporting patient pathway, from diagnosis and treatment to operations related to the financial settlement of health services. In addition, AI algorithms can optimize patients’ time and the money spent on healthcare,” says Kornowska.
Not only patients benefit, but also healthcare professionals. “AI reduces doctors’ workload by taking over repetitive and time-consuming tasks that could be digitalized. AI will impact the quality of patient care, for example, by detecting automatically abnormal test results more. It will increase the effectiveness of patient treatment and optimize doctors’ working time. And it’s already happening”, she says.
ICT&Health interviewed Ligia Kornowska, who is also the leader of the AI in Health Coalition, a nonprofit organization established to advocate for the safe and transparent implementation of artificial intelligence algorithms in health care.
AI attracts a lot of attention, but its use in medicine has faced many problems so far. So which are the most significant barriers?
Problems with AI implementation in medicine can be divided into three categories: knowledge-related, technological, and legal.
Some doctors and CEOs of medical facilities still do not trust solutions supported by AI algorithms – most probably due to insufficient knowledge about the available AI solutions and the lack of clear guidelines on how to check their effectiveness. However, some algorithms have successfully supported work in medical facilities and have been validated for some time. Even so, AI implemented in a hospital still comes as a surprise. It is an additional challenge whenever a new algorithm is integrated within the hospital.
Knowledge, technology and legislation – these are key obstacles for AI development.
Moreover, groundbreaking solutions often neglect the user experience – a doctor is only willing to use a given IT tool when it is intuitive and easy to use. It cannot be the case that a doctor spends more time using a given algorithm than assessing tests results on their own.
Apart from the know-how-related and technological challenges, there is also a legal barrier. Efforts are underway to regulate AI. I mean, for example, the EU AI Act. But many issues remain unresolved. Liability for the result, the patient’s rights confronted with AI, verification of the effectiveness – these are just a few examples.
Which machine learning and AI solutions fascinate you the most?
I’m fascinated by solutions that can improve the effectiveness of diagnosing and treating a patient. Algorithms that can increase cancer detection at an early stage by several or even over ten percent save people’s lives! In fact, they can save hundreds of thousands of patients!
I’m also fascinated by AI-driven solutions which citizens can use individually. In this way, care is shifted to our homes – individuals can therefore focus on prevention, while digital tools allow them to take more responsibility for their health.
So far, few hospitals use innovations such as clinical decision support systems or algorithms which help assess medical images, even though AI has already started to become more common in these areas. Do you think this situation can change, bearing in mind that health care is facing different challenges right now?
Definitely yes. Digital innovations can help solve some of the problems healthcare is facing. But, of course, the implementation of every breakthrough technology generates initial costs. These are financial costs and the costs of training staff, changing habits, and transforming the patient treatment model – and we must be aware of that.
However, when the initial investment is made, it starts to pay off. In what way? For example, by optimizing the workflow. This is extremely important mainly due to staff shortages – one of the biggest problems in healthcare. Technology also improves treatment effectiveness, which in turn leads to financial optimization. However, before we start to benefit from AI, we need to solve the problems mentioned before in order to enable widespread adoption of AI in medicine.
Technological progress is so fast that many doctors can’t keep up with it. It’s not surprising – work schedules are full, there are enormous staff shortages and the COVID-19 pandemic highlighted other problems that have existed for a long time. So why. Why is AI still so important?
Artificial intelligence can help doctors spend more time with their patients and treat them more effectively. In medicine, AI is compared to the discovery of penicillin.
Like telemedicine, which showed its value during the COVID-19 pandemic, AI can help us deal with “old and unsolved” problems. Nevertheless, I hope that the implementation of AI will be systematic and gradual so that we can avoid issues related to an immediate change in the model of service delivery. Everyone – doctors, patients and healthcare managers– needs to be adequately prepared for such transformation.
To sum up, how can we use new digital technologies, such as AI, to make health care innovative?
We need to coordinate legislative initiatives on AI and other technological innovations. We need to open access to data because the lack of data in healthcare is the biggest obstacle to the development of AI. We need to educate patients, medical staff, and healthcare managers and show the capabilities of AI. Artificial intelligence is not only advanced mathematical algorithms – it can help treat patients.
We also need to increase funding for the development of AI algorithms. When we compare the European Union to the United States or China, it becomes evident that the EU’s funds to develop AI are several times lower.
Nevertheless, I’m optimistic. Five years ago, we had 40 AI algorithms for medicine usage worldwide. In 2020 – over ten times more. It shows that AI is becoming fact, and it’s not a science-fiction any more.
