We have heard about the problems of medical facilities with blood deficiency for a long time. This problem is noticeable not only in Poland. The ambitious goal: 100% availability and 0% disposal of blood in the world – and artificial intelligence is to help in its implementation – is set by AIDA Diagnostics. I talk to Mateusz Pawełczuk, CEO and co-founder of AIDA Diagnostics, about the history, current activities and plans for the future.
1. As AIDA Diagnostics you operate in many areas, so I would like to start broadly. Where did the idea to deal with AI solutions in medicine come from?
Mateusz Pawełczuk, CEO & Co-founder of AIDA Diagnostics: The combination of AI and medicine came naturally in my case. I wanted to be a programmer most of my life, but by the end of high school, I turned to medicine. During my studies, I earned some extra money at the Copernicus Science Center, a place with which I worked for a longer time, and instead of putting on an apron right after graduation, I had the opportunity to organize a festival on artificial intelligence. At that time, I was living with Michał Janiszewski, my partner, who shared his passion for technology, was a doctor and, like me, wanted to build something on a larger scale. We were seasoned in classic business by conducting original workshops in schools, but we really wanted to enter the world of AI. Later, we encountered challenges related to blood donation and haemotherapy, and the AI seemed to be able to address most of them.
2. I wonder if the subject of blood is hard to deal with in the realities of healthcare organizations in Poland? Where are the greatest difficulties?
MP: In Poland, we have started a real digitization of health care relatively recently. Healthcare professionals rarely believe technology can help them, but rather see it as a nuisance. When we collaborated with hospitals, we often ran into blood banks that operated on paper, and the lack of data or even a computer is holding back innovation. Unfortunately, the announcement of the introduction of a modern cloud-based tool supported by machine learning scares more than makes us happy. Even experts appreciate rigid guidelines more than mysterious AI systems. Hospitals often operate in the mode of continuous firefighting, which effectively delayed our implementation. Ultimately, in our system, it doesn’t matter for a hospital whether it rolls 1000 or 2000 units. In the end, everything will be refunded by the National Health Fund.
Currently, we only cooperate with the Regional Blood Donation and Blood Treatment Centers, here we also face technological challenges, but the need for the best possible blood management in these institutions is so strong that we manage to overcome all of them successively.
3. What does AIDA Diagnostics do? What is your goal and why is your activity innovative?
MP: AIDA Diagnostics supports the decisions of those responsible for blood. We mainly use machine learning to implement our mission, but we also use other technologies. Our ultimate ambitious goal is “100% availability and 0% disposal of blood worldwide”. We are the only one of two companies in the world that successfully applies AI in the service of blood, and the only one with solutions adapted to all stages of this process. The best proof of our innovation is the reactions of blood stations around the world when they see our system. From Sydney, Australia to Ontario Canada, all agree that they would like to stop producing blood components on a gut-based basis and start using the tools our company offers.
4. You also brought your product to Qatar. How much experience does operating in such a market bring? Can we learn anything about the organization of the health care system? Or maybe they should learn from us?
MP: Qatar is a very specific place, a very small country, but very rich, with great ambitions and a different history with its neighbors. What we could learn from them is how they store medical data. However, here we are talking about a few large hospitals run by one Hammad family, so in such conditions it is easier to integrate information.
When it comes to blood management, Poland is doing better and here the Catharics can learn a lot from us. Despite the fact that markets are different, they all strive to make the best possible use of donated blood in a voluntary gesture of honor. There are many international organizations at the European and global level that take care of the flow of knowledge between different countries.
5. If you could describe your product in a few sentences, but in such a way as to put the patient and his benefits first. How would you do that?
MP: Thanks to our product, the blood needed to save lives can be obtained from donors and processed before it is urgently needed. Patients usually do not receive whole blood in the hospital, but only certain blood components, for those with low hemoglobin levels, we give concentrated red blood cells, in case of some bleeding we will administer platelet concentrate, others will require plasma. By providing information on the anticipated needs for individual blood components, we help blood donation stations be better prepared for the future. Thanks to this, for example, there is no need to cancel operations that always have to be secured with blood. It sounds a bit like fortune-telling from a ball, but it turns out that many phenomena can be efficiently described by mathematical models.
6. Finally, I would like to ask about three areas in which – in your opinion – artificial intelligence in medicine will develop the most. What could you indicate?
MP: First of all, image analysis, i.e. all kinds of tools supporting radiologists or other specialists using virtually any method of imaging. There are algorithms that are great for ultrasound, others interpret computed tomography very well, and subsequent results of magnetic resonance imaging. In the future, they will be an indispensable aid in diagnosis, allowing human brains to focus on the rare cases in which AI cannot cope.
Another area where AI allows us to break new boundaries is epidemiology and infectious diseases. In one bag I throw here algorithms that predict mutations of viruses, as well as those that detect an anomaly in social media that heralds a pandemic. This will happen a bit under the pressure of the COVID-19 situation, and some algorithms are doing extremely well in this field.
As the third area, I would mention technologies of natural language understanding. By combining them with medical knowledge, we could reach the level of real health assistants with whom we will be able to enter into deeper discussions, such as a basic doctor / dietitian / psychologist in our smartphone.