Identification of Heart Failure Events in Medicare Claims: The Atherosclerosis Risk in Communities (ARIC) Study

Area of Science:

• Cardiology
• Health Services Research
• Medical Informatics

Background:

• Medicare diagnostic codes are frequently used to identify heart failure (HF) hospitalizations.
• Accuracy of these codes for acute decompensated HF (ADHF) and chronic stable HF (CSHF) requires evaluation.

Purpose of the Study:

• To assess the accuracy of Medicare diagnostic codes in identifying ADHF and CSHF hospitalizations.
• To compare Medicare claims data with adjudicated medical records for HF identification.

Main Methods:

• Utilized linked Medicare Provider Analysis and Review (MedPAR) files and Atherosclerosis Risk in Communities (ARIC) study medical records (2005-2009).
• Used ARIC's adjudicated classification as the criterion standard for ADHF and CSHF.
• Matched 8,239 hospitalizations based on participant ID and discharge date, achieving a 68.5% match rate for fee-for-service (FFS) criteria.

Main Results:

• Excellent agreement (κ > 0.9) was found for HF codes in any position within matched records.
• Agreement was attenuated for primary diagnostic codes (κ < 0.8).
• Sensitivity of Medicare codes for identifying ADHF and CSHF was low, particularly for primary codes.

Conclusions:

• Matching hospitalizations between Medicare claims and abstracted medical records is incomplete.
• While Medicare HF codes show good agreement within matched records, their low sensitivity suggests potential underestimation of HF cases.
• Medicare data might overestimate the incidence of hospitalized ADHF or CSHF.