Debate on Artificial Intelligence at InforSalud 2022, the Spanish National Congress on Information Technology in Healthcare

Artificial Intelligence was one of the main topics under discussion at the 25th National Congress on Information Technology in Healthcare (InforSalud 2022), organised by the Spanish Society of Information Technology in Healthcare between March 22 and 24 in Madrid, Spain. Under the banner of ‘Sharing data, information and knowledge in healthcare’, the event drew together healthcare experts and professionals involved in the digitalisation and innovation sectors from the different regional healthcare systems and research centres, as well as from the technology and healthcare industry.

Debate on the use of Artificial Intelligence in healthcare

The Director of the Department of Artificial Intelligence of the TIC Salut Social Foundation and Head of Medical Imaging and Pathology Systems of the Department of Health of the Generalitat de Catalunya, Carlos Gallego, took part in the symposium on the uses of this technology in the healthcare sector. In his talk, he highlighted the reasons why this technology is at present rarely used in clinical practice, despite the multitude of recent scientific publications and regular coverage in the media.

“There are three main factors that account for the difficulties in applying Artificial Intelligence to the healthcare system, for it is a very complex system and has many participants with many different responsibilities and functions; there is also a regulatory framework in place that involves following a certain process, just as a medical team would, and this makes it very difficult to implement changes along the whole line and, finally, we have yet to define a formal and scientific model that will enable us to develop applications”, explained Gallego.

During the debate, further hindrances for the implementation of Artificial Intelligence were also pointed out, such as the resistance to change among staff, the current limitations to the deployment of Big Data and the ‘simplistic’ use of this technology for carrying out specific classification tasks instead of regarding it as a complementary tool that can provide continuous support to professionals throughout the entire healthcare process. In this regard, Fernando Martín, a researcher at the Carlos III Institute in Madrid, warned that “when the issue of AI algorithms arises, we tend to emphasise its supposed superior efficiency over that of the medical specialist, but we should never compare an AI system with a real human being. What we should compare is the efficiency of a human using AI compared to someone who doesn’t have this added backup”.

Participants in the debate also stressed that Artificial Intelligence will be at the heart of the change and transformation of the healthcare system. Thus, the system will have to transition from studying the process to analysing the data, with a focus on safety and protection. Along these lines, the need to establish mechanisms for monitoring the ethical use of this technology and to cultivate the trust of medical staff regarding its effectiveness, reliability and ease of use was also highlighted. “We will have to establish ethical committees to ensure the proper use of algorithms, as is the case in other fields,” stated Maria Dolores Tomé, advisor to the Directorate General of Digital Health and Information Systems of the Ministry of Health, who also noted that the autonomous communities must “work together” and “be on good terms” in order to establish common ground and combine their efforts.

Federated learning environments

During the session on the interoperability of the National Health System, the head of System Planning of the Catalan Healthcare Service, Jordi Piera, remarked that “we need to establish federated learning environments through which we can use data from outside our own organisation and thus train artificial intelligence algorithms”. And for this to be successful, he explained, “the data must be standardised and we have to find a forum through which we can share the algorithms and use them collectively”.

Speaking in the same session, the Deputy Director General of the Digital Healthcare Services of the Ministry of Health, Maria Fe Lapeña, remarked that “the European funds we are receiving provide us with a unique opportunity to identify areas for improvement as well as enabling us to tackle bigger issues from the ground up: if something needs to be changed on an in-depth level, now is the time to identify it and to do so”.

Other talks on Artificial Intelligence at InforSalud 2022

The coordinator of the Department of Innovation and Digital Transformation of the Parc de Salut MAR, Jordi Martínez, explained how the application of a Natural Language Processing solution is helping to organise the information in text format that is stored in the medical records of patients at the Hospital del Mar. This project, according to Martínez, “is helping to instil greater awareness about the importance of data quality among our hospital staff”.

At a later point in this session, participants discussed the trend in the healthcare sector towards automated models of data intake and analysis. According to Julian Isla, from Microsoft, future technology will allow information to be automatically extracted from conversations between patients and healthcare professionals, and encoded in a standard language, with AI tools being used automatically to support healthcare professionals in their decision-making. This methodology is known as infused-AI and it can already be found in some email applications, which suggest words and partial or complete sentences when a user composes a message.

Participants in the congress also discussed some of the latest state-of–the-art technological architectures and approaches to effective data management, such as Data Lakehouse and Data Mesh. Speakers explained that one of the limitations in the deployment of Big Data in the healthcare environment is, as described above, the lack of quality of the data in patient records. According to Cristina Garcia, from CGM Clinical, “this problem can be solved through training aimed at educating users on the importance of data quality”.

Finally, a number of successful projects relating to the application of Artificial Intelligence in healthcare were presented. Most of these involved the use of Deep Learning techniques as a means of automatically detecting pathologies in radiological images and the use of image anonymisation techniques. It should also be noted that a considerable number of solutions that respond to specific needs such as, human resources management tools, support tools for motor rehabilitation, tools for recording and monitoring traumas, and tools for managing omics data, among others, are now moving towards the exploitation and application of Artificial Intelligence with the data they have generated. This phenomenon lends further support to the current trend in the healthcare system towards the deployment of artificial intelligence for improving healthcare.

For more information on other topics covered at the congress, you can consult the TIC Salut Social Foundation website.