We discussed this and other topics with Norbert Schellong, the director of Havířov Hospital, a contributory organisation of the Moravian-Silesian Region. Since 2015, Norbert Schellong has been engaged in the development and research of distance medicine in the Czech Republic, and he was awarded the title of Manager of the Year 2020 in 2021.
As one of the foremost innovative hospitals in the Czech Republic, in which areas are you currently researching or implementing artificial intelligence?
To begin with, I would like to adjust this statement. In the Czech Republic, institutions like IKEM and certain university hospitals are actively involved in scientific research, including the advancement of artificial intelligence. Together with technical universities and the private sector, they are creating intriguing solutions that are sure to be highly beneficial in the future.
As a regional hospital that has also gained recognition as an official scientific research institution for medical research and development (R&D), our role is to create technological solutions – such as those involving artificial intelligence – in partnership with small and mediumsized startups.
Another area we are engaged in is the testing and piloting of projects and technical solutions, even if we are not directly involved in their development or research. We are gradually embedding an innovative corporate culture into our organisational DNA, which we take great pride in. With the help of our medical and non-medical teams, we have made significant progress and are achieving our strategic objective: demonstrating that a regional hospital can also play a significant role in healthcare innovation.
In addition to artificial intelligence, we are deeply engaged in telemedicine. We are among the leading hospitals in the Czech Republic in this field, having incorporated telemedicine into our regular operations for the past four years. In the realm of artificial intelligence, it is well-known that we were the first hospital in Europe to test and help improve the evaluation algorithms of the Carebot solution for analysing X-ray images in radiology. Before this initiative, we had implemented AI to detect nosocomial infections. More recently, AI has also been utilised in colonoscopy procedures and the diagnosis of heart diseases.
In the area of colonoscopy, imagine the cockpit of a fighter jet with squares on the windshield that guide the pilot in locating a target. Similarly, a gastroenterologist can utilise artificial intelligence to highlight polyps or other suspicious growths during a gastrointestinal examination, which could potentially be tumours and might otherwise be overlooked at a very early stage. This illustrates how innovation and artificial intelligence can assist in early cancer detection.
In this area, we are collaborating with the solution’s developer to secure appropriate research grants to aid in developing a similar solution for gastroscopy (the upper intestinal tract). Additionally, we have applied for funding together with Carebot for a research project focused on using artificial intelligence in thoracic diagnostics, utilising CT scans.
I would like to focus on its application in cardiology diagnostics. Could you elaborate a bit more on how AI assists in the field of heart disease?
In cardiology, we benefit from two innovative solutions. We strongly support the Czech startup KardiAi. Engineers, in collaboration with Professor Skála from the Olomouc University Hospital, have created AI tools capable of detecting atrial fibrillation and other cardiac issues through continuous ECG monitoring.
You might think this is not very cutting-edge since ECG recorders have been available for 30 years. However, the key difference with KardiAi is that it transforms a standard Polar sports chest strap, which is primarily designed for measuring heart rate, into a single-channel ECG using AIpowered software. This is where things get interesting. Current threechannel ECG recorders are still not comfortable enough for patients to wear almost continuously – 24 hours a day, several days a week, or even weeks each year. Typically, these require attaching electrodes and carrying a bulky recording unit, meaning continuous ECG is mainly used for patients with diagnosed heart conditions, such as during outpatient examinations or after heart failure. In these cases, monitoring can occur during normal life but usually only lasts a few days, rarely extending to two weeks.
In contrast, this single-channel ECG recorder can be worn nearly all the time, even while sleeping. It appears to be suitable not only for chronically ill cardiac patients who wish to monitor their health and share data with their doctors over the long term but also for those suspected of having heart disease. Havířov Hospital is the first in the Czech Republic to test this technology on obese patients with a BMI of 30 or higher, as well as those with other risk factors like a family history of heart disease, smoking, or an unhealthy lifestyle.
The second innovative solution in cardiology comes from the Slovak startup PMcardio, which is currently in a six-month testing phase. The results of this study are eagerly awaited, as they could lead to fundamental changes in heart disease detection and greatly enhance the organisation of care, particularly in diagnosing myocardial infarction.
At first glance, PMcardio appears to be a straightforward application. A doctor or nurse uses a mobile phone to capture an image of any ECG recording typically produced in outpatient clinics. Artificial intelligence then assesses the ECG within seconds, and initial tests indicate that it is up to 30% more accurate than the average doctor at diagnosing myocardial infarction.
In practice, we sometimes see patients whose infarction is not immediately apparent from the ECG. These patients arrive at the hospital experiencing shortness of breath or chest pressure, but their ECG appears “normal”. Even laboratory tests might not reveal any acute issues, so these patients are sent home after a brief observation period. In some instances, however, symptoms only manifest several hours later, resulting in delayed treatment.
Another issue is pre-hospital triage. The ambulance service might evaluate a patient’s risk using a standard ECG and decide to transport them to a regional hospital, such as ours. However, after additional examination, the patient may need to be transferred directly to a specialised facility, like the catheterisation lab at AGEL Hospital in Třinec or the University Hospital in Ostrava. With a more precise ECG diagnosis, the patient could be taken straight to a specialised hospital from the beginning. Any delay,
however, could have serious consequences.
This is where artificial intelligence can be crucial. With its capability to analyse large volumes of ECG recordings, it can identify even minor abnormalities that doctors might overlook. This could greatly enhance the quality and safety of care for patients with acute cardiac conditions.
Can artificial intelligence improve communication with patients and enhance the effectiveness of treatment?
Certainly, we will witness the emergence of significant AI tools in this area as well. Some are already being implemented, such as intelligent chatbots available 24/7 to engage with patients in place of human operators. These tools are not intended for communication in emergency medical situations but rather for offering general information. Educating and informing patients is a crucial responsibility of all regional hospitals.
As websites are enhanced and hospitals communicate through social media, the demand for continuous patient care in the digital environment is increasing.
This is where artificial intelligence plays a role, as it can swiftly integrate with an institution’s information network, like a healthcare provider’s website, and communicate with patients with high accuracy. These tools are extremely beneficial because they significantly assist patients in finding information, such as pre-operative preparation, risks associated with specific diseases, or childbirth preparation. We are currently testing AI in the area of childbirth preparation, and the results are looking very promising.
Currently, there are voice assistants, whose role is expected to grow, eventually replacing traditional switchboards. We are still exploring this domain. However, what fascinates me greatly and partially aligns with this area is AI-based technology designed to relieve doctors (and other professionals) of administrative tasks. Over the past two years, we have been diligently working to find an effective tool for the dictation of automatic transcription into clinical information systems, such as for outpatient records and inpatient ward rounds. There are even companies interested in developing such technology collaboratively with us. I consider this area to be one of the most crucial, perhaps even more so than the aforementioned diagnostic tools.
Thank you for the interview.
