DSEN Abstract
CNODES Common Data Model (CDM) Pilot Project

A study conducted by the Canadian Network for Observational Drug Effect Studies (CNODES)

Summary

The Sentinel CDM was adapted for use in Canada at all six CNODES sites and successfully demonstrated with three pilot projects.
CDM queries have reduced the timelines by an estimated 60%.
Further work to be done to improve timelines, expand the data sources and analytical tools.

Key messages

The Sentinel CDM was successfully transformed and piloted-tested in six CNODES provincial sites.

Project Lead & Team

For more information, please contact info@cnodes.ca.

What is the issue?

CNODES has addressed many important questions regarding the use and safety of prescription drugs in Canada but these were necessarily time consuming.
While such studies are essential to evaluate the real-world impact of medications, decision-makers also often have less complex, pressing questions about how drugs are used in practice. Specifically, about use in combination with other drugs, their indications and the co-morbidities of patients who use them, and the extent of variation in these characteristics and practices across Canada.
Such descriptive studies could be done more efficiently by leveraging the tools and experience of the FDA Sentinel Initiative.

What was the aim of the study?

The purpose of this project was for CNODES to create and pilot-test a Canadian version of the Sentinel Common Data Model (SCDM) at multiple CNODES sites (Alberta, British Columbia, Manitoba, Nova Scotia, Ontario, and Sasktachewan).

How was the study conducted?

The CNODES CDM team established a multi-stakeholder advisory committee and an internal working group. Following this, they developed the Canadian common data model tables at multiple CNODES sites.
The CDM tables and Sentinel analytical packages were tested through multiple demonstration projects.
Outcome measures of feasibility and timeliness included successful adaptation of the Sentinel CDM; successful construction and quality assurance of the core CDM tables at each site; successful completion and pooling of the demonstration study results; and the timelines associated with study approval and conduct.

What did the study find?

We successfully adapted the Sentinel CDM for use in Canada and transformed the source data files at all six CNODES sites included in the pilot.
We developed the policies and procedures needed to successfully complete and pool the results of all three pilot studies.
The average time to study approval was 22 days overall (range 5 to 50 days).
Overall study timeline of approximately 80 days.
We estimated that our efforts have reduced the timeline for CDM-suitable queries by at least 60% to a maximum of 140 days, shorter for studies involving fewer sites.