The Government of Catalonia took its first steps today towards bringing artificial intelligence (AI) into the Catalan healthcare system. Catalan Minister for Health, Manel Balcells i Díaz, chaired the first meeting of the Health/AI Programme Steering Committee, made up of senior representatives from the Catalan health sector, professional clinicians and AI experts. Balcells said that "the Health/AI Programme is a strategic bid to ensure the robustness, ethical nature, equity and efficiency of AI-based health solutions in Catalonia". Solutions which, as he pointed out, "need to be vetted beforehand with the data at our disposal and comply with the provisions of the AI law being drafted by the European Union".

Balcells also stressed that "this innovative programme, which is up and running as of today, is unprecedented in Spain and gives us a structured way to implement artificial intelligence in the healthcare system", while adding that "properly applied AI can help us to improve patient diagnosis, efficiency and the sustainability of the system". Finally, he underlined that "this programme is one of the projects intended to transform the system", complemented by training for professionals.

The programme aims to create a fertile environment for innovation and research in the healthcare field through the development and implementation of AI solutions. As explained by the programme director, Dr José Ibeas, "the point of AI is to improve people's health and the sustainability of the healthcare system, while respecting personal rights and ensuring that the solutions we implement are vetted, safe and reliable".

Action lines under the Health/AI Programme

The programme will pursue three distinct lines of action. First, with the support of the Recovery, Transformation and Resilience Plan (PRTR) financed by the European Union with NextGenerationEU funds, it will coordinate the launch of public procurement challenges for AI tools that address unmet needs in the system. These tools will have to meet certain standards, set by the programme, to ensure the equity, scalability and quality of the proposed solutions.

Second, via the AI in Health Observatory, launched at the end of 2022, the programme aims to understand and disseminate comprehensive information on how AI is generated in the healthcare field, from conceptualisation to implementation, with an emphasis on the risks and opportunities that emerge as a result. Finally, it will offer support to the initiatives identified in the Observatory by developing reference guides for the implementation of AI with full guarantees for people and healthcare professionals, and providing a mentoring service.

AI experiences in Catalan healthcare

The AI in Health Observatory has identified over 100 AI solutions that are being developed or used in the Catalan healthcare system and research centres in Catalonia, and it continues to actively collect information. Of the initiatives identified, 47% are being carried out in hospital settings, 24% in primary healthcare centres and 8% in mental health facilities.

In 61% of cases, these tools are used in medical and surgical procedures, and in 16% they are used for management. By medical specialisation, the most reported AI solutions are used in medical oncology, totalling 20% of the cases, followed by general medicine (14%), radiology (12%), and family and community medicine (12%).

At the meeting to present the newly created Health/AI Programme Steering Committee, time was taken to speak about one of the initiatives that was identified in the Observatory and will receive support from the programme: the Hospital Clínic skin cancer diagnosis support platform. For more than five years, the dermatology team has been working with AI tools to support professionals in the diagnosis of melanomas and democratise this technology for the prevention and early screening of the disease among the general population. More specifically, they have developed algorithms to classify skin tumours using high-precision digital photography in the laboratory, and are working on predictive AI prognostic models for cancer patients.