九色视频

Jean Nikiema, MD, PhD

Assistant Professor | School of Public Health
Department of Health Management, Evaluation and Policy Universit茅 de Montr茅al

The Seminars in Epidemiology organized by the Department of Epidemiology, Biostatistics and Occupational Health at the 九色视频 School of Population and Global Health is a self-approved Group Learning Activity (Section 1) as defined by the maintenance of certification program of the Royal College of Physicians and Surgeons of Canada. Physicians requiring accreditation, please complete the Evaluation Form and send to admincoord.eboh [at] mcgill.ca

WHEN: Monday, January 26, 2026, from 3:30-4:30 p.m.
WHERE: Hybrid | Onsite at 2001 九色视频 College, Rm 1140 |
NOTE: Jean Nikiema will present in-person

Abstract

Real-world data (RWD), especially information extracted from electronic health records, holds enormous potential for clinical decision support, cost analysis, and public health planning. Yet this potential is frequently limited by fragmented data pipelines, inconsistent clinical semantics, incomplete provenance, and uncertain fitness-for-use. The CIRCULATE consortium, paired with the Provem platform, were created to address these challenges through an integrated ecosystem that turns raw RWD into decision-ready assets.

This talk will describe how Provem and the CIRCULATE consortium operationalizes a pragmatic 鈥淎I for Health鈥� mindset: starting with data provenance and quality signals, enabling interoperability through ontology-based harmonization, and supporting clinical trajectory reconstruction (hospital stays, transitions, and longitudinal care pathways). We will present how pathway discovery (clustering, labeling, and phenotyping of stays/trajectories) is combined with knowledge-based standardization, and how a human-in-the-loop workflow supports validation, traceability, and clinical relevance of AI algorithms. Finally, we will highlight how the ecosystem creates a continuous transparency loop so that improvements to data governance and data production directly translate into more reliable analytics and AI capabilities.

Learning Objectives

At the end of this talk, attendees will be able to:

• Present the limitations of RWD for AI: data quality, bias, and recognize what 鈥渞ealistic expectations鈥� look like in clinical settings;
• Present the evaluation of a responsible AI system that identifies and convert patterns in EHR data into clinically useful information for decision support and cost analysis;
• Understand the operationalized responsible, and scalable analytics by combining data-driven pathway discovery with knowledge-based standardization while keeping human-in-the-loop for algorithm validation.

Speaker Bio

Jean-No毛l Nikiema is an Assistant Professor at the Universit茅 de Montr茅al School of Public Health (ESPUM), a regular researcher at the Centre de recherche en sant茅 publique (UdeM鈥揅IUSSS du Centre-Sud-de-l鈥櫭巐e-de-Montr茅al), a researcher with OBVIA (Sustainable Health Axis), and Co-Director of LabTNS, focused on digital transformation in health, and co responsible of the infrastructure axe at Qu茅bec Digital Health Network. His research centers on the real-world conditions required for successful health innovation, data quality, interoperability, governance, organizational impact, and uptake in practice settings, to support learning health systems aligned with operational constraints. His training integrates clinical, public health, and informatics perspectives: MD (Universit茅 Nazi Boni), MSc in Public Health-Medical Informatics (Universit茅 de Bordeaux), PhD in Public Health-Informatics and Health (Universit茅 de Bordeaux), followed by postdoctoral training at CHUM. Prior to joining ESPUM, he was research visiting scholar at the U.S. National Library of Medicine.