Association Discovery Approach in Healthcare Big Data to Identify Drug Safety and Drug Repurposing Signals

Area of Science:

• Pharmacovigilance
• Data Mining
• Health Informatics

Background:

• Big data analytics are crucial for evaluating drug safety and effectiveness.
• Association discovery is a data mining technique for identifying relationships in high-dimensional datasets.

Purpose of the Study:

• To implement association discovery in longitudinal healthcare data for drug safety and repurposing.
• To identify positive and inverse associations between drug use and clinical outcomes.

Main Methods:

• Utilized a 10% sample of Australian Pharmaceutical Benefits Scheme data (2014-2024).
• Employed the Magnum Opus association discovery tool to analyze prescription claims.
• Defined drug use via Anatomical Therapeutic Chemical (ATC) codes and identified conditions using indicator drugs.

Main Results:

• Identified associations between drugs and coronary artery disease, type 2 diabetes, and epilepsy.
• Reproduced known adverse drug events and protective effects.
• Found potential novel drug safety and repurposing signals, alongside confounding factors.

Conclusions:

• Association discovery offers a novel approach to identifying drug safety and repurposing signals from healthcare big data.
• Further validation is required for newly identified associations.
• Methodological comparisons with other data science approaches were discussed.