A systematic review of methods for determining cross-sectional active medications using pharmacy databases

Area of Science:

• Pharmacoepidemiology
• Health Informatics

Background:

• Pharmacy dispensing databases are crucial for assessing medication use and clinical outcomes.
• Accurate identification of active medications at a specific time point is essential for research and quality measurement.

Purpose of the Study:

• To systematically review methods for assessing medications in use at a specific time point using pharmacy dispensing databases.
• To compare the performance of different approaches for identifying active medications.

Main Methods:

• A systematic literature search identified studies comparing pharmacy database methods with non-automated data sources.
• Two investigators independently reviewed study abstracts and full-text articles.

Main Results:

• Twenty-nine studies used fixed look-back periods (84-730 days), and 14 used medication-on-hand approaches.
• Sensitivity varied widely: 48%-93% for fixed look-back and 35%-97% for medication-on-hand.
• Head-to-head comparisons showed medication-on-hand sensitivity was 7% lower than fixed look-back.

Conclusions:

• The accuracy of methods for identifying active medications from pharmacy databases varies significantly.
• Further direct comparisons are needed to determine the accuracy of these methods across different populations, medication classes, and databases.