The Health AI Observatory identifies nearly a hundred Artificial Intelligence algorithms in SISCAT (Catalonia's comprehensive health system for public use) and research centres in Catalonia

21 DECEMBER 2022

The TIC Salut Social Foundation has launched the Observatory of Artificial Intelligence in Health (Health AI Observatory), which has identified nearly 100 Artificial Intelligence algorithms in the Catalan Health System (SISCAT) and research centres in Catalonia.

The Observatory is part of the Artificial Intelligence in Health Programme (Health/AI Programme) at the Department of Health. Its aim is to find out about and spread the application of Artificial Intelligence in this field in a cross-cutting way, from conceptualisation to implementation, having an impact on the risks and the opportunities that arise.

Núria Abdón, the head of the Health AI Observatory at the TIC Salut Social Foundation, explains “The Observatory was born out of the desire to set a benchmark for knowledge transfer regarding innovations in the field of Artificial Intelligence. It also promotes the adoption of this technology and acts as a liaison between the health system and the rest of the parties involved”.

To set up the Health AI Observatory, work sessions were held with different areas of the Department of Health, the Catalan Health Service, the Catalan Institute of Health, the Health Quality and Evaluation Agency of Catalonia (AQuAS), the Observatory of Ethics in Artificial Intelligence of Catalonia (OEIAC), the Catalan Hospital Union, the Health and Social Consortium of Catalonia, the Catalan Association of Health Entities (ACES), the Official Association of Doctors, the Official Association of Nurses of Barcelona, the Catalan Society of Family and Community Medicine (CAMFIC), and the Family and Community Nursing Association of Catalonia (AIFICC).


Artificial Intelligence algorithms in health identified

One of the first studies that the Observatory performed identifies Artificial Intelligence algorithms in health used by SISCAT and research centres in Catalonia (figure 1). Analysis of nearly 100 initial algorithms that SISCAT and health research centres have reported shows that they are developing or using an average of 2.3 Artificial Intelligence algorithms.

By degree of maturity, more than 40% of the algorithms are in the most advanced stages and are applied in a real-world environment. This is the case, for example, for the DigiPATICS project of the Catalan Institute of Health. This provides diagnostic support tools to professionals based on digitised samples of pathological anatomy. Various Artificial Intelligence algorithms are included. These algorithms allow biomarkers such as Her2, Ki67, oestrogen receptors and progesterone receptors in breast cancer to be quantified, among other use cases.

At the other end of the algorithm development cycle, 35% of those identified are still in the initial laboratory stages, prior to proof of concept, and the remaining nearly 25% are in intermediate validation stages (figures 2 and 3).

In the care sector, hospitals are leading the application of Artificial Intelligence algorithms. These account for 50% of all those identified, followed by Primary Care with 20%, Mental Health with 8%, and Social Health Care with 6% of the identified algorithms, among others (figure 4).

The medical specialty that stands out the most in the use of Artificial Intelligence algorithms is oncology. In this area there are solutions such as Athena Care, in which the Clinical Campus and its researchers participate. This is at an intermediate level of maturity. It is an Artificial Intelligence cognitive tool that enables health professionals to carry out diagnosis and management focused on people with skin cancer.

Apart from oncology, the Health AI Observatory has identified other areas with greater use of algorithms in health, such as cardiology, family and community medicine, endocrinology and nutrition, and pathological anatomy (figure 5). By pathology, just as for specialties, Artificial Intelligence algorithms are most used for neoplasms, followed by endocrine diseases, diseases of the circulatory system, congenital anomalies, mental disorders, infectious diseases, and diseases of the digestive system and the nervous system (figure 7).


Benefits of using Artificial Intelligence in Health

Identifying the algorithms, which is still in the data gathering stage, will make it possible to know the degree of development and implementation of Artificial Intelligence in the field of health, to share and inform about good practices, and to promote the adoption of this technology in Catalonia, among others. The study will also be extended to other institutions and organisations.

Susanna Aussó, the head of the Artificial Intelligence Area of the TIC Salut Social Foundation, points out the benefits of implementing tools based on Artificial Intelligence in the health system: “Artificial Intelligence solutions will improve care for citizens, as they provide tools to support diagnosis, prognosis and follow-up of diseases, as well as the management and administration of health resources and public health.” However, Aussó stresses that “to make this possible, a systemic approach is essential to ensure fairness throughout the territory, transparency, security and ethical values. With this aim in mind, the Observatory and the Health/AI Programme will be key players in facing these challenges”.

Carlos Gallego, the Director of the Health/AI Programme, highlights that “with the launch of the Health AI Observatory promoted by the TIC Salut Social Foundation, and the first challenge concerning Artificial Intelligence-based solutions to support the process of diabetic retinopathy in primary care, coordinated by the Health Quality and Evaluation Agency of Catalonia (AQuAS), we have ended the year by deploying two key strategic lines to advance the promotion and development of people-focused AI in the field of health.”


Next steps of the Health AI Observatory

The Health AI Observatory will regularly update its web page with indicators concerning the algorithms used in the field of health and will publish reports and reference guides for the sector. The next to be published will be a report on state-of-the-art virtual assistants, a maturity study on Artificial Intelligence in Health in research centres in Catalonia, a guide to explainability in Artificial Intelligence, and another on good practices for code development in Artificial Intelligence solutions in health.

Link to the Observatory video presentation