Medical AI Databases and Resources - OMARI

Female blonde pharmacist helps and assists her young aprrentice trainee during studies of medical books — Faculty of Medicine

Research and Innovation - RO

Medical AI Databases and Resources

The Ottawa Medical AI Research Institute (OMARI) supports medical AI innovation by sharing curated datasets and practical tools. This section includes data sources, platforms, compute services and other resources to help researchers build, test and apply AI in health.

Databases and Resources

The Ottawa Medical AI Research Institute (OMARI) has compiled a list of platforms, tools and compute resources that support medical AI projects and help researchers manage, store and process data. We have included information about the resources, how to access them and any prerequisites.

ARCHIMEDES Platform – University of Ottawa Heart Institute

ARCHIMEDES (Advanced Research Collaboration for Health Integration, Medical Exploration, and Data Synthesis) is a health data platform that provides centralized access to curated, multi-modal brain-heart datasets, advanced analytics tools and high-performance computing. The platform supports secure, collaborative health research, and can help research groups manage their own datasets through a data governance platform that supports compliance with Tri-Council data sharing requirements.

To learn more or request access, contact [email protected] or visit the Archimedes platform website.

International Data Access Tools Repository – ADR UK

The International Data Access Tools Repository is a publicly accessible repository of tools and resources that support the discovery and access of administrative and population-level data across multiple countries. Hosted by Administrative Data Research UK (ADR UK), it brings together data catalogues, metadata tools, algorithm inventories and guidance on good practices for accessing and using administrative data, including health-related data. The repository does not host datasets itself. Instead, it helps researchers identify relevant data sources and understand access pathways, governance requirements and conditions for use, which vary by dataset and data holder. Access is open. Requirements for data access depend on the linked dataset and may include registration, applications or data-sharing agreements with the original data custodian.

To learn more, visit the International Data Access Tools Repository website.

The Ottawa Medical AI Research Institute (OMARI) provides access to databases and datasets that support the development, evaluation, and application of artificial intelligence in health. This section brings together data sources identified by experts and contributors in the field, along with information on dataset type, access conditions, prerequisites, and links to documentation or contacts.

Description	No. sources	Longitudinal	Accessibility
Statistics Canada Research Data Centre (RDC) Microdata Catalogue: The RDC program provides secure access to Statistics Canada’s individual-level microdata for health and social research. The catalogue can be filtered by keywords like “health.” Data type: Survey, administrative, population health, social determinants Funding source: Government of Canada Data source: Statistics Canada	Multiple	Yes	Apply for access
Google Research: Google shares datasets that may be useful to researchers across various fields, including health. Use the keyword “health” to help narrow your search. Data type: EHR, genomic, medical imaging, insurance, community, survey Funding source: Google Data source: Various	Multiple	No	Open
UK Biobank: UK Biobank is a biomedical database with genetic and health data from 500,000 UK participants, available to approved researchers studying major diseases. Data type: EHR, genomic, medical imaging, sensors, labs Funding source: National Health Service (NHS) Data source: Collaborative (data linkage and harmonization)	Multiple	Yes	Apply for access
Project Data Sphere: Open access to patient-level oncology datasets from completed clinical trials, enabling researchers to explore and analyze cancer data to advance scientific discovery. Data type: Clinical trials, oncology, patient-level Funding source: CEO Roundtable on Cancer Consortium Data source: Project Data Sphere, LLC	Multiple	No	Open after registration/agreement
Surveillance, Epidemiology, and End Results (SEER) Program: SEER provides information on cancer statistics in an effort to reduce the cancer burden among the U.S. population. Data type: Registry, oncology, patient-level Funding source: unknown Data source: Surveillance Research Program (SRP) in NCI's Division of Cancer Control and Population Sciences (DCCPS)	Multiple	No	Open after registration with a valid institutional email, completed application form, data user agreement and acknowledgement of data limitations
OpenML datasets: Datasets provide training data for machine learning models Data type: Biomedicine and imaging, text and NLP, sensor and signal, network traffic and security, finance and economics, social sciences and surveys Funding source: open source / community Data source: OpenML.org	Multiple	No	Open
GEMINI Database (Canada): GEMINI is a hospital-based database with administrative and clinical data from more than 2 million admissions across over 30 Canadian hospitals, updated regularly. Data type: Administrative and clinical EHR data, labs, imaging, pharmacy, vitals Funding source: Partner hospitals; Canadian Institutes of Health Research (CIHR) Data source: Contributed by participating hospitals; standardized and harmonized with CIHI data (Discharge Abstract Database [DAD] and National Ambulatory Care Reporting System [NACRS])	Multiple	Yes	Apply for access (project proposal and approval required)
MIMIC-III Clinical Database: MIMIC-III is a freely available database of de-identified health data from over 50,000 ICU patients admitted to Beth Israel Deaconess Medical Center between 2001 to 2012. Data type: EHR (vitals, labs, medications, diagnoses, imaging reports, clinical notes) Funding source: National Institutes of Health (NIH), MIT Laboratory for Computational Physiology, Beth Israel Deaconess Medical Center Data source: Beth Israel Deaconess Medical Center	Single	Yes	Apply for access
MIMIC-IV: MIMIC-IV is the updated release of the MIMIC database, providing de-identified EHR data from Beth Israel Deaconess Medical Center, organized into modules (hospital-wide, ICU, and clinical notes). Data type: EHR (admissions, demographics, diagnoses, procedures, lab results, medications, radiology, free-text notes) Funding source: National Institutes of Health (NIH), MIT Laboratory for Computational Physiology, Beth Israel Deaconess Medical Center Data source: Beth Israel Deaconess Medical Center	Single	Yes	Apply for access via PhysioNet (training and data use agreement required)