Development of a detection algorithm for statin-induced myopathy using electronic medical records

Area of Science:

• Pharmacovigilance
• Pharmacoepidemiology
• Drug Safety

Background:

• Electronic medical records (EMRs) offer potential for pharmacovigilance (PV) but require validated methods for adverse drug event (ADE) detection.
• Statin-induced myopathy (SIM) is a significant ADE, yet reliable pharmacoepidemiological detection methods are lacking.

Purpose of the Study:

• To develop and validate a highly selective algorithm for detecting SIM using EMRs.
• To assess the algorithm's predictive accuracy for SIM.

Main Methods:

• A three-step algorithm was developed using EMR data (prescriptions, labs, diagnoses) from 5109 statin users.
• Steps included event detection (creatine kinase increase, statin discontinuation), exclusion criteria, and creatine kinase (CK) time-course refinement.
• Positive predictive value (PPV) was calculated, comparing the algorithm to a diagnostic code-based approach.

Main Results:

• The algorithm identified 5 patients with suspected SIM, yielding a 0.1% frequency.
• All identified cases were confirmed as likely SIM by expert review, resulting in a 100% PPV.
• The proposed algorithm demonstrated significantly higher PPV (100%) compared to the diagnostic code approach (33.3%).

Conclusions:

• A novel EMR-based algorithm demonstrates high predictability for detecting statin-induced myopathy.
• Combining exclusion criteria, medical practice data, and CK levels enhances SIM prediction accuracy.
• Further validation in larger studies is recommended to confirm the algorithm's utility.