Quality data in health: an essential requirement for Artificial Intelligence

The cycle of sessions of the XI R&D&I ICT Social Health Conference 2021, provided numerous examples of projects in the field of health and artificial intelligence (AI). The third session, held on 27 October, began with a lecture by Xavier Ferràs, professor of Operations, Innovation and Data Science Management at ESADE. In Ferràs’s opinion, Artificial Intelligence (AI) is changing the way we see the world. His assessment was global, providing the current geopolitical perspective: the two superpowers of China and the United States.

(Photo: from left to right, Karma Peiró (moderator), Xavier Ferràs, Anna Escalé, Carles Ramió, Carolina Garcia and Raimon Dalmau)

The next person to speak was Anna Escalé, primary care doctor at EAP Navàs-Balsareny and SAP Bages-Berguedà-Solsonès. She presented the pilot study of the use of AI as a support and diagnostic tool for skin lesions in primary care. “7.6% of the population presents with skin lesions annually. One more reasons to use an AI tool. The initial hypothesis is that machine learning will allow us to diagnose faster and reduce unnecessary referrals to specialists,” Escalé explained.

Carles Ramió, Professor of Political Science at Pompeu Fabra University, focused his presentation on the aging population. “This will result in extra economic costs equivalent to 12% of Spanish GDP in social and health care alone. This is unaffordable,” he explained. In his expert opinion, AI can help balance out this situation. He also stressed that it is just an instrument and can only be put to the uses that “the public administrations are capable of taking on.”

Carolina Garcia-Vidal, senior specialist in Infectious Diseases at the Hospital Clínic de Barcelona – described the experience of applying AI to improve decision-making in real-time. Personalised medicine and data quality are going to be key in the coming years. The success story was based on a classification of data from patients with Covid-19 from Hospital Clínic de Barcelona to help doctors during an emergency. The lessons learned have been applied in a Belgian and a Dutch hospital.

Finally, Raimon Dalmau, Head of the Information Systems, ICT and Data Department (CIO) at the Medical Emergencies System, gave his presentation. He spoke about the advanced data processing by the 061 Salut Respon and Transport Sanitari Urgent services for the development of predictive models that allow for the optimisation of resources in the health and logistics field.

Strategic organisation

During the debate, demands from public administrations to increase data governance was brought up. Other issues that arose included the need to train professionals in technical knowledge, budget investment for AI research, time and perspective, and building the trust of citizens.

“We lack a framework of political, legal and economic support for Catalan healthcare,” stressed the Clínic researcher. “We need to improve the quality of medical data.” In the ethical sense, Raimon Dalmau highlighted the issue of personal data privacy because “there are private company tools, and the question is whether they should be used in the public health system”.

It was concluded that we are in a historical moment, in which it is difficult to determine the border between technology and individual rights. All the speakers agreed that we are in a revolution more than just a transformation, one which is expected to solve major current structural problems.

(Follow this link to see session 3 in full)

Strategies for adopting AI

The fourth and last session of the Conference took place on 28 October. Josuè Sallent, director of TIC Salud Social, presented the TIC Salud Social foundation’s Health/AI programme, and strategies for the adoption of AI in the sector. “We don’t want black boxes; we want tools capable of understanding how an AI has made a decision. We must be absolutely transparent and give full explanations. We need to integrate algorithms into the system. And train the health sector in the technology. In addition to communication, explaining very well to the public what we are doing”. According to Sallent, the Health/AI Programme is getting started. And he called for it to be critically evaluated in order to improve it.

Next, the various speakers presented their experiences and projects. Salvador Pedraza, regional clinical director of radiology and nuclear medicine at los Hospitales de Girona; Juliana Ribera, COO of Amalfi Analytics; Pau Rodriguez, CEO of Methinks AI; Pere Romero, head of the ophthalmology service at the Hospital Universitari Sant Joan de Reus; and David Reifs, technical director at the Seidor innovation department and CEO of Skilled Skin (Clinicgram).

(Photo: from left to right, Josué Sallent, Salvador Pedraza, Juliana Ribera, Karma Peiró, Pau Rodríguez, Pedro Romero, David Reifs)

During the round table, the main challenges were brought up and the data and regulatory problems, such as issues of anonymity and transparency, were highlighted. Public confidence in the technology companies was mentioned, but so too was the fear that people have of seeing personal data being used in a healthcare setting.

It was stressed that AI must support health professionals, and a machine can never make an automated diagnosis. “The main risk I see is like solving a Rubik’s cube. You have one side that is regulation, the other quality data, the other investment, etc. And if you solve it, for it to work, it falls within the legal framework, if you respect ethical principles, manage to explain it to patients, and convince investors, then you have solved the cube. It is not easy”, commented Pau Rodríguez.

It was concluded by committing to making Catalonia a leader in the technological field, calling on administrations to streamline bureaucratic processes with AI and to establish a government body in charge of digitisation and another specialised body to resolve regulatory issues with greater procedural traceability, and finally calling for greater cooperation between experts in the field.

With this session, the 11th R+D+i Tic Salut Social 2021 Conference ended, held online on 20, 21, 27 and 28 October, and organised by the TIC Social Health Chair of Vic University – Universitat Central de Catalunya (UVic-UCC), the TIC Salut Social Foundation (Ministry of Health of the Generalitat de Catalunya), the Consorci Hospitalari de Vic, Fundació Hospital de la Santa Creu de Vic, the EAP/ABS Centelles, the EAP Vic, the Institut Català de la Salut-Gerència Catalunya Central, the Althaia Foundation and FORES. It was also supported by the Seidor company.

 (Follow this link to see session 4 in full)