May 14, 2011Pharmacoepidemiology and Drug Safety
Validity of claims‐based definitions of left ventricular systolic dysfunction in Medicare patients
Why the study?
Are claims-based definitions valid for identifying left ventricular systolic dysfunction in Medicare patients with heart failure or myocardial infarction?
Population
1072 Medicare patients with ejection fraction recorded for a heart failure or myocardial infarction…
Comparison
Thirteen claims-based definitions for systolic… vs Recorded ejection fractions (gold standard)
Design
Cohort
Authors
Discussion
Loading...
Member takes
Implication
High-specificity digoxin-based claims definitions may aid restricted HF/MI cohorts; leaves open substantial misclassification bias in effectiveness studies.
Key Points
- The study evaluates the accuracy of claims data definitions for diagnosing left ventricular systolic dysfunction in heart failure and myocardial infarction patients.
- Assessed 1072 Medicare patients with recorded ejection fractions from 1999-2006.
- Developed 13 claims-based definitions using ICD-9 codes and medication use.
- Calculated sensitivity, specificity, and positive predictive values based on recorded ejection fractions as the gold standard.
- Definitions based on digoxin use had positive predictive values between 76% and 84% and specificity over 97%, but low sensitivity (6%-14%).
- With varying ejection fraction cutoffs, specificity dropped by 3% and positive predictive values decreased by 52%.
- Use of these definitions in hypothetical studies risks underestimating treatment effectiveness and introduces significant misclassification bias.
Structured PICO
Are claims-based definitions valid for identifying left ventricular systolic dysfunction in Medicare patients with heart failure or myocardial infarction?
P
Population
1072 Medicare patients with ejection fraction recorded for a heart failure or myocardial infarction hospitalization linked with pharmacy data and national registries in 1999-2006
I
Intervention
Thirteen claims-based definitions for systolic dysfunction using single or combination of ICD-9 diagnosis codes and cardiovascular medications
C
Comparator
Recorded ejection fractions (gold standard)
O
Outcome
Sensitivity, specificity, and positive predictive values (PPVs) of claims-based definitions
Claims-based definitions for systolic dysfunction have excellent specificity and good PPV but low sensitivity, making them useful for cohort identification by restriction but prone to misclassification bias.
Limitations
- Limited generalizability
Cite This Study
Li et al. (2011) studied this question.