Observational Studies of Drug Safety in Multi-Database Studies: Methodological Challenges and Opportunities

Area of Science:

• Pharmacoepidemiology
• Health Services Research
• Drug Safety

Background:

• The Canadian Network for Observational Drug Effect Studies (CNODES) is a collaborative network of researchers and databases.
• CNODES functions as a collaborating center within the Drug Safety and Effectiveness Network.
• Its primary role involves conducting observational drug safety studies based on stakeholder-submitted queries.

Purpose of the Study:

• To explore methodological opportunities and challenges in distributed pharmacoepidemiology networks.
• To examine the practical application of these networks through a case study.
• To assess the utility of administrative data for detecting drug safety signals.

Main Methods:

• Utilized a case study focusing on proton pump inhibitors and community-acquired pneumonia hospitalizations.
• Addressed challenges in study design, conduct, and data aggregation across individual sites.
• Leveraged the distributed nature of the network to explore opportunities for detecting heterogeneity and increasing sample sizes.

Main Results:

• Distributed networks like CNODES enable the detection and quantification of significant safety signals using administrative data.
• Significant methodological challenges exist in designing, conducting, and combining data from distributed studies.
• Opportunities include the ability to analyze large sample sizes and assess heterogeneity, which are often infeasible for single studies.

Conclusions:

• CNODES and similar networks are valuable for identifying drug safety concerns.
• Methodological advancements are crucial to address the complexities of distributed data in pharmacoepidemiology.
• The growing scale and research scope of these networks necessitate ongoing methodological development.