Bernardo Duque Neves , an internist at Hospital da Luz Lisboa, earned his Ph.D. in Biomedical Engineering from the Instituto Superior Técnico (IST) at the University of Lisbon . His dissertation, titled “Leveraging Data Science for Multimorbidity Measurement and Risk Prediction from Clinical Data,” was unanimously approved by the committee with honors. Supervised by Mário Gaspar da Silva (IST-UL), the thesis was evaluated by a panel that included Rema Padman (Heinz College, Carnegie Mellon University, USA), Luís Filipe Ribeiro de Azevedo (Faculty of Medicine, University of Porto), and João Miguel Sanches (Chair) and Rui Henriques, both from IST. The work analyses electronic clinical records from more than 900,000 patients treated at Hospital da Luz Lisboa between 2006 and 2022, and presents four main contributions : An algorithm for phenotyping 55 chronic diseases, validated using Large Language Models (LLMs); A comparison of multimorbidity measurement methods, showing a 30–50% improvement in the ability to predict hospital visits; The adaptation of foundational clinical record models (CLMBR-T) to the hospital context, improving the detection of rare events such as unplanned hospitalisations; A direct comparison between physicians, machine learning models and LLMs in the same prognostic task, highlighting paths for clinical–algorithmic complementarity. Bernardo Neves is an internal medicine physician at the Internal Medicine Department of Hospital da Luz Lisboa, a member of the Value-Based Healthcare Management Directorate at Luz Saúde, a lecturer at the Faculty of Medicine of the Portuguese Catholic University, and a member of the Portuguese Medical Association’s Artificial Intelligence Committee.
