Use of Medicare Claims Data for the Identification of Myocardial Infarction: The Reasons for Geographic And Racial...

Area of Science:

• Cardiovascular research
• Health services research
• Epidemiology

Background:

• Accurate identification of myocardial infarction (MI) is crucial for patient care and research.
• Medicare claims are a widely used data source for epidemiological studies.
• The validity of Medicare claims for identifying MI requires careful assessment.

Purpose of the Study:

• To assess the accuracy of Medicare claims in identifying myocardial infarction (MI).
• To determine the positive predictive value (PPV) and sensitivity of Medicare claims for MI detection.
• To evaluate the utility of Medicare claims for investigating MI-associated characteristics.

Main Methods:

• Utilized data from 9951 Medicare beneficiaries aged 65+ from the Reasons for Geographic And Racial Differences in Stroke study (2003-2012).
• Compared MI cases identified through study adjudication (gold standard) with those identified via Medicare inpatient claims using ICD-9 codes 410.x0 or 410.x1.
• Calculated PPV and sensitivity for claims with MI codes in any and primary discharge positions.

Main Results:

• Medicare claims with an MI code in any position had a PPV of 84.3% and a sensitivity of 49.0%.
• Claims with an MI code in the primary position showed a higher PPV (89.7%) but lower sensitivity (40.1%).
• Sensitivity varied significantly based on MI type, sex, and in-hospital vs. out-of-hospital occurrence, with lower rates for type 2/4-5 MIs and in-hospital events.

Conclusions:

• Medicare claims demonstrate high positive predictive value but limited sensitivity for identifying myocardial infarction.
• Despite limitations in sensitivity, Medicare claims are valuable for research into individual-level factors associated with MI.
• Findings highlight the need for careful interpretation of MI data derived from administrative claims.