AI-guided clinical case simulator becomes the winning project of the first edition of GenAI Health Hack

More than 100 participants in 13 teams took part in the first edition of the GenAI Health Hack, which was held from 14 to 16 November at the Sea Towers space of NTT Data, organised by Hospital Clínic Barcelona, on Health/AI Programme of the Department of Health of the Government of Catalonia and the Chair of Precision Oncology of the UB, in collaboration with MIT Critical Data.

For three days, professionals from various disciplines, such as health professionals, data specialists, engineers, as well as lawyers and experts in healthcare law and management, among others, have worked on the development of solutions based on generative AI to deal with challenges such as patient education, structuring clinical reports, oncology findings and health data analysis.

On the first day of the hackathon, talks and round tables were held with experts, open to all health professionals interested in the applications of generative AI in this field. After three intense days, on the afternoon of 16 November, the participants presented their solutions to a jury that evaluated the projects according to criteria such as difficulty, prevention of bias and regulatory compliance.

“Generative artificial intelligence is revolutionising the health sector, thanks to its extraordinary ability to understand and process natural language. With the organisation of this pioneering event we seek to drive innovation and collaboration between experts in AI and health. Our goal is to make it a key annual event where the development of innovative and sustainable solutions to improve medical care is encouraged.” explained the organisers of the first edition of the GenAI Health Hack.

The clinical case simulator that revolutionises the learning of health professionals

The Pizza Team was awarded first prize for developing a dynamic AI-guided clinical case simulator. Specifically, it is an advanced chatbot which allows doctors, nurses, residents and other healthcare professionals to practice solving medical cases using a conversational system that simulates real patients, in real time, with realistic interactions and multiple paths to a diagnosis.

The panel of judges, made up of experts from the Hospital Clínic and the TIC Salut Social Foundation, and members of MIT Critical Data, as well as doctors and researchers from the scientific field of other prestigious international universities, highlighted this solution for its ability to encourage learning from error, offering open and variable routes in the resolution of clinical cases, and to allow professionals to learn in a safe and controlled environment. With multiple ways to solve clinical cases (based on cases from the Hospital Clinic, published clinical cases and various medical books) this software stimulates critical thinking and autonomous decision-making, preparing professionals to deal with real-world situations. The panel of judges assessed the originality and difficulty of the solutions, the metrics of the model, the study of biases, as well as the policies to mitigate risks. The team was comprised Rosalía Cayuela, resident anaesthesiologist at the Hospital Clínic; Nuria González, Joana Ros, Laia Colomé and Marc Biosca, biomedical engineers from the University of Barcelona. Among the promoters were Gerard Espinosa, Director of Teaching at the Hospital Clínic and internal medicine doctor, and Ramses Marrero, anaesthesiologist at the Hospital. Eusun Lee, Data Scientist at Seoul National University Hospital, acted as mentor.

BotiClin OncoGenIA and NutrIA were the two finalist projects.

The hackathon also recognised two other finalist projects for their innovative proposals. Among them, BotiClin, a virtual assistant for the Hospital Clínic portal, which aims to facilitate access to truthful information for different users. The system integrates multiple data sources in several languages and allows different forms of information input (text, voice and calls). Their ethical approach ensures that responses are secure, always providing a special warning for emergencies and sets up a suicide prevention system. In addition, BotiClin avoids offering diagnoses or treatments, ensuring that the information provided is for informational purposes only, and issues warnings against self-medication when the use of medications is mentioned.

OncoGenIA is another outstanding solution that addresses the optimisation of oncology patient data management, collecting key information prior to the first visit and following European oncology treatment guidelines to recommend appropriate treatments. In addition, it offers support on side effects and specific patient needs. In addition, OncoGenIA quickly identifies candidates for experimental clinical trials, optimising the selection process and contributing to the development of new therapies.

Finally, the last GenAI Health Hack finalist project was NutrIA, an application designed to improve the quality of nutritional care for patients in prehabilitation programmes. This tool uses artificial intelligence to capture and analyse conversations between a nutritionist and patient, allowing the patient’s specific nutritional needs to be classified automatically. Thanks to this technology, NutrIA is able to adapt dietary recommendations taking into account not only food preferences, but also the patient’s allergies and comorbidities, all with minimal interaction on the part of the nutritionist. This solution facilitates more personalised and efficient care, saving time for healthcare professionals and improving decision-making in patient health management.

Reflection and debate on AI in the health sector

The Hackathon started with a full morning of talks and round tables with experts in generative artificial intelligence applied to health.

The day began with a welcome on behalf of the institution by Raquel Sánchez-Valle, Medical Director of Hospital Clínic Barcelona; Aleix Prat, head of the Institute of Cancer and Blood Diseases at Hospital Clínic Barcelona, and Joan Guanyabens, director of the TIC Salut Social Foundation.

During the sessions, prominent specialists such as Leo Anthony Celi, Senior Research Scientist at MIT Critical Data; Juan Castro, GenAI team leader at NTT DATA; Giovanni Angeloti, MIT Critical Data expert; Xavier Borrat, Head of the Clinical Informatics Service at Hospital Clínic Barcelona; Paula Petrone, ISGlobal Data Scientist; Montserrat Esquerda, PhD in Medicine at the Autonomous University of Barcelona (UAB), specialist in Paediatrics and director of the Borja Institute of Bioethics-URL; and Manuel Ramos, specialist in Autoimmune Diseases at the Clinic and director of the Postgraduate programme on the Application of Generative AI in the Methodology in Research and Teaching in Health Sciences at the University of Barcelona (UB), among others, addressed issues such as bias in AI, the transformation of biomedical processes and the impact of generative AI on medical care. The talks offered a space for reflection and debate on the challenges and opportunities of AI in the healthcare field.

It should be noted that organisations such as Amazon Web Services, IDIBAPS and the Barcelona Supercomputing Centre (BSC) were partners of the event, which was also sponsored by Logicalis, Avvale, NTT Data Spain, Siemens, My Cloud Door, Olympus, Vodafone and Roche.

• Nota de premsa GenAI Health Hack

  • 328.7 Kb

    
  •