3103 — Frequency and Predictors of Appropriate Response to Elevated LDL Cholesterol Levels in Veterans with Cardiovascular Disease
Virani SS (Michael E. DeBakey Veterans Affairs Medical Center), Woodard LD
(Michael E. DeBakey Veterans Affairs Medical Center), Landrum CR
(Michael E. DeBakey Veterans Affairs Medical Center), Wang D
(Michael E. DeBakey Veterans Affairs Medical Center), Ballantyne CM
(Baylor College of Medicine), Petersen LA
(Michael E. DeBakey Veterans Affairs Medical Center)
Although the current cholesterol performance measures define good quality as low density lipoprotein cholesterol (LDL-C) levels < 100mg/dl in cardiovascular disease (CVD) patients, they provide performance measurement at one time point and do not inform whether an appropriate action was taken to manage elevated LDL-C levels. Our objective was to assess frequency and predictors of this appropriate response (AR).
We used administrative data to assess 22,357 CVD patients receiving care in one Veterans Affairs network for the first quarter of 2010. We determined the proportion of patients at LDL-C goal < 100 mg/dl, and the proportion of patients with uncontrolled LDL-C levels ( > 100 mg/dl) who had an AR. AR was defined as the initiation or dosage increase of a lipid lowering medication (LLM), addition of a new LLM, receipt of maximum dosage or > 1 LLM, or LDL-C reading < 100 mg/dl at 45 days follow-up. Logistic regression analyses were performed to evaluate facility, provider and patient characteristics associated with AR.
LDL-C levels were at goal in 69% of patients. Among the 31% not controlled, 47% had an AR. Controlling for clustering between facilities and patient’s illness severity, history of diabetes (OR 1.33, 95% CI 1.19-1.48), hypertension (OR 1.61, 95% CI 1.40-1.85), and a higher number of patients with hyperlipidemia in a provider’s panel (OR 1.02, 95% CI 1.01-1.03) were associated with AR. Patient’s age > 75 years (OR 0.35, 95% CI 0.30-0.40) and larger provider panel size (OR 0.98, 95% CI 0.96-0.99) were negatively associated with AR. Of note, facility type (teaching vs. non-teaching, p = 0.93), provider type (physician vs. non-physician, p = 0.25), specialist vs. non-specialist primary care provider (p = 0.40), and patient’s race (p = 0.41) were not predictors of AR.
Among Veterans with CVD and LDL-C above guideline recommended levels, less than half receive AR. Diabetic and hypertensive CVD patients are more likely to receive AR, whereas older Veterans with CVD receive AR less often likely reflecting providers’ belief of lack of efficacy from treatment intensification in older patients.
Our findings are important for quality improvement and policy making initiatives since they provide more actionable information compared with isolated LDL-C assessment as a quality indicator.