PAHO to Host Webinar on AI Applications in ICD-11 Morbidity and Mortality Coding

The Pan American Health Organization (PAHO) has announced an upcoming international webinar scheduled for September 10, 2026, focused on the integration of artificial intelligence (AI) into the use of the International Classification of Diseases, 11th Revision (ICD-11). The session aims to demonstrate how AI tools can analyze clinical documentation, such as hospital discharge summaries, to assist health professionals in identifying relevant conditions and proposing cause-of-death sequences.

This initiative is part of a broader effort within the Community of Practice on Civil Registration and Vital Statistics (CRVS) in Latin America and the Caribbean. By leveraging AI to support the application of ICD-11, PAHO seeks to strengthen the quality and integration of health data, ultimately facilitating more informed decision-making across the region's health information systems.

The webinar will feature presentations from experts including Daniel Luna, Charly Otero, and Juan Carlos Díaz, who will showcase AI applications designed to assist in the complex process of cause-of-death certification. These tools are engineered to suggest ICD-10 or ICD-11 codes and facilitate the review of alternative terminology based on clinical summaries.

PAHO emphasizes that these AI applications are intended to function strictly as support tools rather than replacements for human expertise. The organization maintains that all proposals generated by AI must be reviewed, modified, and validated by a responsible health professional. This distinction is critical, as the AI-assisted process is separate from the formal statistical coding and selection of the Underlying Cause of Death, which must adhere to established World Health Organization (WHO) mortality coding rules.

The event is framed within the context of the digital transformation of CRVS and Information Systems for Health (IS4H). The overall objective of this series is to bolster technical and institutional capacities across the Americas, ensuring that ICD-11 serves as a central semantic standard. By improving the integration of morbidity and mortality data, PAHO aims to enable the interoperability required for modern digital health infrastructure.

Speakers for the session bring extensive experience in health informatics and data engineering. Daniel Luna, who heads the Department of Health Informatics at Hospital Italiano de Buenos Aires, will join Charly Otero, a specialist in digital health and interoperability, and Juan Carlos Díaz, a Data Engineering Advisor at PAHO. The session will be moderated by Myrna Martí, an IS4H Advisor at PAHO, and will include simultaneous interpretation in English, Spanish, Portuguese, and French.

While the PAHO webinar focuses on practical implementation, recent scientific research highlights the broader potential of AI in this field. A study published in Nature in September 2025 demonstrated the efficacy of generative transformer models, such as the 'Delphi-2M' model, in learning lifetime health trajectories. Trained on data from 0.4 million UK Biobank participants and validated on 1.9 million Danish individuals, this model can predict the rates of over 1,000 diseases simultaneously.

These transformer-based models, which function similarly to large language models by processing sequences of health events, offer the ability to sample synthetic future health trajectories. This capability provides estimates of potential disease burden for up to 20 years. Researchers noted that while these models provide significant insights into temporal dependencies between disease events, they also highlight biases learned from training data, reinforcing the necessity for human oversight in clinical applications.

A central theme of the upcoming PAHO session is the discussion of benefits and limitations regarding the use of AI in public health. The speakers are expected to address critical considerations such as information quality, semantic standards, ethics, data governance, and transparency. As AI tools become more prevalent in clinical settings, PAHO is prioritizing the establishment of frameworks that ensure human oversight and accountability.

These discussions align with the broader challenges identified in the field of health informatics, where the integration of diverse data sources—including self-reports, primary care records, and hospital admissions—can introduce biases. By promoting the active exchange of good practices, PAHO aims to navigate these complexities, ensuring that the adoption of AI in health systems remains aligned with ethical standards and the ultimate goal of improving patient outcomes and public health security.