Institutional, provider, and patient correlates of low-density lipoprotein and non–high-density lipoprotein cholesterol goal attainment according to the Adult Treatment Panel III guidelines

Prevention and Rehabilitation

Institutional, provider, and patient correlates of low-density lipoprotein and non–high-density lipoprotein cholesterol goal attainment according to the Adult Treatment Panel III guidelines Salim S. Virani, MD, a,b LeChauncy D. Woodard, MD, MPH, a Cassie R. Landrum, MPH, a Kenneth Pietz, PhD, a Degang Wang, PhD, a Christie M. Ballantyne, MD, b and Laura A. Petersen, MD, MPH a Houston, TX

Background The aim of this analysis was to identify the proportion of coronary heart disease (CHD) patients achieving guideline-recommended low-density lipoprotein cholesterol (LDL-C) and non–high-density lipoprotein cholesterol (non–HDL-C) goals and to identify correlates of dual goal attainment. Methods

We analyzed patient, provider, and facility characteristics for 21,801 CHD patients in one Veterans Affairs Hospitals Network.

Results Low-density lipoprotein cholesterol goal attainment was 80%, but optional LDL-C goal attainment was 41%. Of patients with triglycerides ≥200 mg/dL, 51% attained both LDL-C and non–HDL-C goals. Correlates of higher dual goal attainment included older age (65-74 years: odds ratio [OR] 1.47, 95% CI 1.28-1.69), diabetes (OR 1.33, 95% CI 1.16-1.53), obesity (OR 1.25, 95% CI 1.04-1.50), a higher number of primary care visits (OR 1.04, 95% CI 1.04-1.05), and mild increase in illness severity of patients in provider's panel (OR 1.20, 95% CI 1.0008-1.46), whereas African American patients were less likely to achieve dual lipid goals (OR 0.63, 95% CI 0.48-0.82). Receipt of care from physician (vs nonphysician) or specialist (vs primary care) provider, number of patients in provider's panel, and percentage of patients in provider's panel with diagnosis of hyperlipidemia were not associated with dual goal attainment. Conclusions

A large proportion of CHD patients attained LDL-C goal, but optional LDL-C goal attainment was low. Patients with elevated triglycerides had poor attainment of dual LDL-C and non–HDL-C goals, suggesting a treatment gap. Factors associated with dual goal attainment may identify interventions needed to improve future guideline adherence. (Am Heart J 2011;161:1140-6.)

Low-density lipoprotein cholesterol (LDL-C) is the primary treatment target for cholesterol management in the National Cholesterol Education Program Adult Treatment Panel III (ATP III) guidelines. The 2001 ATP III guidelines1 recommend a goal of LDL-C b100 mg/dL in patients with coronary heart disease (CHD). In a 2004

update,2 an optional LDL-C goal of b70 mg/dL was recommended in patients with CHD and very high risk (multiple major risk factors, severe and poorly controlled risk factors, multiple risk factors of the metabolic syndrome, or acute coronary syndromes). The ATP III guidelines established non–high-density lipoprotein

From the aHealth Policy and Quality Program, Michael E. DeBakey VA Medical Center

120) at the request of Veterans Integrated Service Networks 1, 12, and 23. Dr Petersen

Health Services Research and Development Center of Excellence; and Section of Health Services Research, Baylor College of Medicine; Houston, TX, and bSection of Cardiovascular Research, Baylor College of Medicine; and Center for Cardiovascular Disease Prevention, Methodist DeBakey Heart and Vascular Center; Houston, TX. This work was supported by Investigator Initiated Research funding by Merck and Co, Inc, as well as the Houston VA Health Services Research & Development Center of Excellence (grant HFP90-020). Dr Virani is supported by a Department of Veterans Affairs Health

was a Robert Wood Johnson Foundation Generalist Physician Faculty Scholar (045444) and an American Heart Association Established Investigator Awardee (0540043N) at the time this work was conducted. Submitted November 23, 2010; accepted March 15, 2011. Reprint requests: Salim S. Virani, MD, Health Services Research and Development (152), Michael E. DeBakey Veterans Affairs Medical Center, 2002 Holcombe Blvd, Houston, TX 77030.

Services Research and Development Service (HSR&D) Career Development Award (CDA09-028). This work was also supported in part by VA HSR&D PPO 09-316 (PI LeChauncy D. Woodard, MD, MPH), VA HSR&D IIR 04-349 (PI Laura A. Petersen, MD, MPH), andNIH R01 HL079173-01 (PI Laura A. Petersen, MD, MPH) and a VA contract (Project XVA 33-

E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2011.03.023

American Heart Journal Volume 161, Number 6

Virani et al 1141

Table I. Baseline characteristics of patients with CHD Variables, n (%) unless otherwise indicated Age (y), mean ± SD Men Race White African American Unknown Hypertension Diabetes Body mass index ≥30 kg/m2 Current smoking Acute coronary syndrome in the last 24 m Statin use‡ Simvastatin Pravastatin Lovastatin Fluvastatin Simvastatin/ezetimibe Atorvastatin Rosuvastatin Fibrate use‡ Niacin use‡ Fish oil use‡ Ezetimibe use‡,§ Bile acid sequestrant use‡ Patients using any lipid-lowering medication Patients using lipid-lowering medications from N1 class

All CHD patients (n = 21801)

CHD patients who attained LDL-C goal (n = 17432)⁎

CHD patients who attained LDL-C and non–HDL-C goals (n = 1918)†

65.2 ± 7.23 21608 (99.1)

65.8 ± 7 17313 (99.3)

64.23 ± 6.98 1912 (99.7)

17851 2290 1660 18526 9446 10307 5086 1422 17529 13094 718 796 75 750 531 2223 1946 2813 1078 795 242 18420 5167

14523 (83.4) 1624 (9.3) 1285 (7.4) 14876 (85.3) 7931 (45.5) 8380 (48.1) 3821 (21.9) 1071 (6.1) 14540 (83.4) 11191 (64.2) 405 (2.3) 618 (3.5) 51 (0.3) 643 (3.7) 427 (2.4) 1635 (9.4) 1544 (8.9) 2307 (13.2) 838 (4.8) 546 (3.1) 126 (0.7) 15086 (86.5) 4206 (24.1)

1643 (85.7) 116 (6.0) 159 (8.3) 1721 (89.7) 1169 (60.9) 1190 (62.0) 480 (25.0) 152 (7.9) 1638 (85.4) 1274 (66.4) 37 (1.9) 67 (3.5) 0 66 (3.4) 40 (2.1) 189 (9.9) 346 (18.0) 327 (17.0) 190 (9.9) 53 (2.6) 5 (0.3) 1725 (89.9) 675 (35.2)

(81.9) (10.5) (7.6) (84.9) (43.3) (47.3) (23.3) (6.5) (80.4) (60.1) (3.3) (3.7) (0.3) (3.4) (2.4) (10.2) (8.9) (12.9) (4.9) (3.6) (1.1) (84.5) (23.7)

⁎ In study patients overall. † In study patients with elevated triglycerides (candidates for LDL-C and non–HDL-C lowering per the ATP III guidelines). ‡ Within 100 days of the most recent qualifying visit in the study interval. § Isolated ezetimibe use. CHD patients using simvastatin/ezetimibe combination are listed in the statin category.

cholesterol (non–HDL-C) as a secondary treatment target in patients with elevated triglycerides. However, studies have shown that non–HDL-C goal attainment is as important as LDL-C goal attainment to improve cardiovascular outcomes.3-5 Whereas factors associated with goal attainment for LDL-C have been reported in the literature,6,7 those associated with dual goal attainment (LDL-C and non– HDL-C) are not well described. Thus, the aim of our current analysis was to identify the proportion of CHD patients achieving guideline-recommended levels for both LDL-C and non–HDL-C and to identify the patient, provider, and facility characteristics associated with attainment of both LDL-C and non–HDL-C goals.

Methods Study population and definition of variables We examined data on veterans with a history of CHD in fiscal year (FY) 2008 (October 1, 2007, through September 30, 2008) who were treated in the Veterans Affairs (VA) Health Care system in the Midwest region, which includes 3 states. We identified veterans as having CHD using International Classification of

Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes for unstable angina (411.xx-414.xx with the exception of 413.xx, 36.01-36.09, 36.10-36.19, 36.2, 36.31, 36.32, 36.39, 36.99) or myocardial infarction (410.xx, 412), or Current Procedural Terminology codes for percutaneous coronary intervention or coronary artery bypass grafting (92980-92982, 92984-92998, 33510-33519, 33521-33523, 33533-33536, 33572, 33530). To be included in the CHD cohort, a patient needed to have 2 outpatient diagnosis codes or 1 inpatient diagnosis code for unstable angina, or 1 code for myocardial infarction, percutaneous coronary intervention, or coronary artery bypass grafting. Patients with a history of metastatic cancers and those receiving hospice care were excluded. We used lipid values from the most recent lipid panel within 12 months before or 14 days after the patient's most recent qualifying visit during the study interval. Low-density lipoprotein cholesterol levels were calculated except for cases with triglycerides N400 mg/dL, for which directly measured LDL-C levels were used if available. Non–high-density lipoprotein cholesterol levels were calculated as total cholesterol minus HDL-C levels. Patient characteristics extracted included patient's age, race, history of diabetes (ICD-9-CM codes 250.xx, 357.2, 366.41), hypertension (ICD-9-CM codes 401-405), and mean Diagnostic Cost Group Relative Risk Score (RRS), a ratio of predicted cost

American Heart Journal June 2011

1142 Virani et al

to the mean cost of a population of veterans that is used as a measure of a patient's illness burden.8,9 For example, a patient with RRS = 5 is expected to be 5 times as costly as an “average” patient, whereas a patient with RRS = 0.5 is expected to be half as costly. The RRS was calculated using Diagnostic Cost Group release 6.2 software. Lipid-lowering medication use was determined for up to 100 days before the qualifying visit, as this period would represent a reasonable amount of time to reflect the patient's most current lipidlowering medication regimen.

Outcome definition and analyses We first examined the percentage of all CHD patients who attained the LDL-C goal and those with elevated triglycerides (≥200 mg/dL) who attained both LDL-C and non– HDL-C goals per the ATP III treatment guidelines1 (b100 and b130 mg/dL, respectively). Subsequently, we used stepwise multivariate logistic regression modeling to ascertain which facility, provider, and patient characteristics were associated with dual LDL-C and non–HDL-C goal attainment. These variables were selected based on whether they could affect combined goal attainment for LDL-C and non–HDL-C. The facility and provider variables included the type of facility (teaching or nonteaching), type of provider (physician or nonphysician), specialist versus primary care provider, number of patients in provider's panel, proportion of patients in provider's panel with a diagnosis of hyperlipidemia (ICD-9 codes 272.0, 272.2, 272.4), proportion of patients in specified age categories in provider's panel, and mean RRS of patients in provider's panel (as an indicator of illness burden of patients in provider's panel). Patient characteristics included patient's race across different age categories, history of diabetes, history of hypertension, and mean RRS (as an indicator of patient's burden of illness severity). Separate multivariable regression analyses were performed for facility/provider characteristics and patient characteristics, as some of the variables in each model could be highly correlated (eg, proportion of patients aged 65-74 years in a provider's panel or mean RRS of patients in a provider's panel [provider-level characteristics] and proportion of patients aged 65-74 years or mean RRS of patients [patient-level characteristics]). Finally, clustering of patients at the facility level was added to the adjustment models. We also assessed attainment of the optional goals for very high risk patients defined in the 2004 ATP III update2 (LDL-C b70 mg/dL; in patients with elevated triglycerides, non–HDL-C b100 mg/dL). All analyses were conducted with SAS version 9.1.3 (SAS Institute Inc, Cary, NC) and STATA version 11 (StataCorp, College Station, TX). The protocol was approved by the Institutional Review Boards at Baylor College of Medicine and the Michael E. DeBakey VA Research and Development Committee. The study was supported through an investigator-initiated research grant funded by Merck & Co, Inc. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper, and its final contents.

Results Our initial cohort included 25,528 CHD patients. We excluded 185 patients because of the presence of chronic

Table II. Proportion of CHD patients attaining goals for LDL-C and, in patients with elevated triglycerides, combined goals for LDL-C and non–HDL-C Attainment of ATP III1 goals,⁎ n (%)

CHD patients (n = 21801)

Overall cohort with LDL-C at goal Subset of patients with triglycerides ≥200 mg/dL Both LDL-C and non–HDL-C at goal LDL-C at goal Non-HDL-C at goal

17432 3781 1918 2808 1918

(80%) (17.3%) (51%) (74%) (51%)

⁎ LDL-C goal b100 mg/dL; in patients with triglycerides ≥200 mg/dL, non–HDL-C goal b130 mg/dL; non–HDL-C = total cholesterol − HDL-C.

liver disease, as escalation of lipid-lowering medications, especially statins, could be contraindicated in these patients. Among the remaining 25,343 patients, 22,991 (90.7%) had lipid panels in FY 2008. We excluded 634 patients (2.5%) with incomplete lipid panel results (missing values for total cholesterol, HDL-C, or both, precluding non–HDL-C calculation) and 556 patients with lipid panels before a diagnosis of CHD was made (as their LDL-C and non–HDL-C goals may be different before the CHD diagnosis). After these exclusions, our final cohort included 21,801 CHD patients. Baseline characteristics of all CHD patients, those who attained LDL-C b100 mg/dL, and those who attained both LDL-C and non–HDL-C goals are shown in Table I. Most patients were male and white. Many also had hypertension, diabetes, or obesity. As expected, statin use was common (80.4%). Most patients were on generic simvastatin; the use of on-patent high-potency statins (atorvastatin, rosuvastatin) or the combination of statin/ezetimibe was low. As expected, the group that attained both LDL-C and non–HDL-C goals had more intensive treatment regimens; greater proportions of these patients used statins (85.4% vs 80.4% in the overall cohort), any lipid-lowering medication (89.9% vs 84.5% overall), or lipid-lowering medications from N1 class (35.2% vs 23.7% overall). Goal attainment for LDL-C and non–HDL-C is shown in Table II. The proportion of CHD patients attaining the LDL-C goal of b100 mg/dL was high (80%). In the subset of patients with elevated triglyceride levels of ≥200 mg/dL (n = 3,781; 17.3% of the cohort), LDL-C goal attainment was again high (74%); but combined goal attainment for LDL-C and non–HDL-C was only 51%. Table III describes provider and facility characteristics associated with goal attainment for both LDL-C (b100 mg/dL) and non–HDL-C (b130 mg/dL) after multivariable regression analyses both with and without clustering at the facility level. Overall, the variance secondary to clustering between facilities was very small (b0.0001 for the provider/facility-level model) and nonsignificant (P N .99), suggesting that the random effect of clustering was small when covariates were

American Heart Journal Volume 161, Number 6

Virani et al 1143

Table III. Provider and facility characteristics associated with attainment of both LDL-C and non–HDL-C goals (n = 3781) after multivariate regression analyses Odds ratio (95% CI), P Variable Care at a teaching facility (nonteaching facility as referent) Physician provider (nonphysician provider as referent) Specialist provider (nonspecialist provider as referent) No. of patients in provider's panel (per increase of 100 patients) Percentage of patients in provider's panel with diagnosis of hyperlipidemia (per 10% increase) Mean age of patients in provider's panel (b65 y as referent) 65-74 y ≥75 y Mean RRS of patients in provider's panel (RRS b1 as referent) 1-1.5 N1.5

Without clustering 1.21 (0.99-1.47), 0.87 (0.75-1.01), 0.91 (0.59-1.40), 1.01 (0.99-1.03), 1.004 (0.99-1.01),

.05 .07 .66 .33 .34

With clustering 1.22 0.87 0.90 1.01 1.004

(0.99-1.48), .05 (0.75-1.01), .07 (0.58-1.40), .66 (0.99-1.03), .33 (0.99-1.01), .34

1.10 (0.95-1.27), .21 1.01 (0.35-2.93), .98

1.10 (0.94-1.28), .20 1.02 (0.35-2.2.94), .98

1.21 (1.008-1.45), .04 1.04 (0.83-1.30), .73

1.20 (1.008-1.46), .04 1.04 (0.82-1.31), .73

Table IV. Patient characteristics associated with attainment of both LDL-C and non–HDL-C goals (n = 3781) after multivariate regression analyses Odds ratio (95% CI), P Variable Race (white as referent) African American Unknown Patient's age (b65 y as referent) 65-74 y 75-85 y History of diabetes History of hypertension Mean RRS of patients (b1 as referent) 1-2 N2 Body mass index (b25 kg/m2 as referent) ≥25-29.9 kg/m2 ≥30 kg/m2 No. of primary care visits during the study period

included in the regression models. Patients with mildly increased illness burden (mean RRS 1-1.5) were more likely to attain dual lipid goals (odds ratio [OR] 1.20, 95% CI 1.008-1.46). Care at a teaching facility (OR 1.22, 95% CI 0.99-1.48) was a borderline-significant correlate of dual goal attainment, whereas having a physician primary care provider (as compared with a nonphysician primary care provider, eg, nurse practitioner) was a borderline-significant characteristic associated with a lower attainment of dual cholesterol goals (OR 0.87, 95% CI 0.75-1.01). Of note, receipt of care by a specialist provider, provider's panel size, and a higher percentage of patients with a diagnosis of hyperlipidemia in a provider's panel were not associated with dual cholesterol treatment goal attainment.

Without clustering

With clustering

0.65 (0.50-0.83), .001 0.85 (0.69-1.06), .14

0.63 (0.48-0.82), .001 0.84 (0.67-1.04), .11

1.46 (1.28-1.70), b.001 1.67 (1.21-2.31), .002 1.33 (1.16-1.53),b.001 1.21(0.98-1.49), .08

1.47 1.66 1.33 1.21

1.15 (0.96-1.37), .14 0.90 (0.76-1.06), .22

1.14 (0.96-1.38), .14 0.90 (0.76-1.07), .22

1.20 (0.97-1.48), .09 1.24 (1.03-1.48), .02 1.04 (1.02-1.06), b.001

1.20 (0.97-1.49), .08 1.25 (1.04-1.50), .02 1.04 (1.01-1.05), b.001

(1.28-1.69), (1.20-2.31), (1.16-1.53), (0.98-1.50),

b.001 .002 b.001 .07

Table IV describes patient characteristics associated with dual goal attainment. Diabetic patients (OR 1.33, 95% CI 1.16-1.53), patients aged ≥65 years (OR 1.47, 95% CI 1.28-1.69), and obese CHD patients (body mass index ≥30 kg/m2: OR 1.25, 95% CI 1.04-1.50) were more likely to attain dual cholesterol targets, whereas African American patients (OR 0.63, 95% CI 0.48-0.82) were significantly less likely to attain dual cholesterol targets. A higher number of primary care visits during FY 2008 was a modest but significant correlate of dual goal attainment (OR 1.04, 95% CI 1.01-1.05). Similar to the provider/ facility-level regression model, the random variance secondary to the clustering of patients between facilities had a very small and nonsignificant effect on the overall model prediction (variance = 0.038, P = .13).

American Heart Journal June 2011

1144 Virani et al

Table V. Proportion of very high risk CHD patients attaining optional LDL-C goal and, in patients with elevated triglycerides, optional goals for both LDL-C and non–HDL-C Attainment of ATP III update2 optional goals,⁎ n (%)

Very high risk CHD patients† (n = 12952)

Overall very high risk cohort with LDL-C at goal Subset of very high risk patients with triglycerides ≥200 mg/dL Both LDL-C and non–HDL-C at optional goals LDL-C at optional goal Non–HDL-C at optional goal

5261 (41%) 2712 (21%) 354 (13%) 1091 (40%) 355 (13%)

⁎ Optional LDL-C goal b70 mg/dL; in patients with triglycerides ≥200 mg/dL, optional non–HDL-C goal b100 mg/dL; non–HDL-C = total cholesterol − HDL-C. † Very high risk CHD defined as one of the following: (a) CHD plus diabetes; (b) CHD plus continued smoking; (c) CHD plus multiple risk factors of the metabolic syndrome (especially high triglycerides ≥200 mg/dL plus non–HDL-C ≥130 mg/dL with HDL-C b40 mg/dL); (d) CHD with history of acute coronary syndrome.

We also calculated the proportion of very high risk patients who attained the optional LDL-C goal of b70 mg/ dL and, in the subset with triglycerides ≥200 mg/dL, the optional non–HDL-C goal of b100 mg/dL (Table V). A total of 12,952 patients met the ATP III definition of very high risk for CHD. Of these, 2,712 patients (21% of the very high risk cohort) with triglycerides ≥200 mg/dL were eligible for the combined optional targets for LDL-C and non–HDL-C. Optional LDL-C goal attainment in very high risk patients was low (41%); and of the subgroup with elevated triglycerides, an even smaller percentage (13%) attained the combined optional goals of LDL-C b70 mg/dL and non–HDL-C b100 mg/dL.

Discussion In this analysis of CHD patients, a large proportion achieved the LDL-C goal of b100 mg/dL per the ATP III guidelines; but attainment of the optional goal of LDL-C b70 mg/dL in very high risk CHD patients was low. Of the CHD patients with elevated triglycerides, only about half achieved both LDL-C and non–HDL-C goals; and far fewer achieved the combined optional goals. We showed that providers with patients with mild to moderate increase in illness burden were more likely to have patients who attained dual cholesterol targets. We also found that older patients and patients with diabetes were more likely to achieve dual treatment goals. Conversely, a treatment gap for cholesterol management remained in African American CHD patients. Our results showed that LDL-C goal attainment to b100 mg/dL in CHD patients has improved. In a 2003 survey,6 62% of CHD patients achieved LDL-C goal of b100 mg/dL, although this survey enrolled the top 26% statin prescribers in the United States for 2002 and was likely an overestimation of goal attainment for LDL-C in clinical practice. In the Lipid Treatment Assessment Project 2

survey, which enrolled patients between September 2006 and April 2007, nearly 70% of CHD patients in the United States were reported to have achieved LDL-C goal of b100 mg/dL; but only 30% of those with CHD and ≥2 risk factors achieved the optional goal of LDL-C b70 mg/dL.7 Similarly, at 12-month follow-up of acute coronary syndrome patients enrolled in the Medications Applied and Sustained Over Time registry from January 2006 through September 2007, 71% and 31% of patients attained the respective goals.10 Our results show a continued improvement for LDL-C goal attainment in CHD patients (80% for FY 2008), which may be explained by improved guideline dissemination to providers and greater availability of highpotency statins. Apart from these, goal attainment might be higher in the VA health care system because of better access to care, efficient use of electronic health records allowing better care coordination, use of clinical reminders, presence of decision support algorithms, and continuous emphasis on measuring lipids and attaining LDL-C levels of b100 mg/dL in CHD patients as a quality improvement mandate.11 The overall high compliance with measuring lipid panels in CHD patients (90%) in the current study is better than the recently reported results from the American College of Cardiology and National Cardiovascular Data Registry's Practice Innovation and Clinical Excellence program, in which 74% of patients with coronary artery disease had lipid panels assessed.12 It is important to note that, in the VA system, the use of highpotency on-patent statins (ie, atorvastatin, rosuvastatin) or the combination of simvastatin and ezetimibe requires clear documentation of a lack of efficacy or an intolerance to generic formulary-based statins. Despite this, goal attainment was higher than in previous studies, indicating that a high LDL-C goal attainment rate in CHD patients can be obtained with the use of generic formulary-based statins and reserving the use of high-potency on-patent statins or combinations with ezetimibe for CHD patients needing further LDL-C reduction on top of the maximum tolerated doses of formulary-based statins or patients who show an intolerance to formulary-based statins. Although improvements in goal attainment for LDL-C are encouraging, we showed that combined attainment of both LDL-C and non–HDL-C goals per the ATP III guidelines (51%) as well as the update (13%) remains poor. The reasons may include lack of institutional mandates to calculate and report non–HDL-C levels (a system-based issue), lack of provider familiarity with the guidelines, lack of awareness of the importance of non– HDL-C levels, inability to calculate non–HDL-C levels in a busy clinical practice, and the presence of clinical inertia, which in this case might include a failure to intensify statin therapy or add other classes of lipid-lowering medications when indicated. Providers may also feel that they have adequately managed their patients' cholesterol

American Heart Journal Volume 161, Number 6

by attaining LDL-C goals alone because LDL-C goal attainment is routinely considered and reported as a marker of quality of care, whereas non–HDL-C goal attainment is not. Our results showed that receipt of care by a physician provider or a specialty provider does not necessarily improve cholesterol goal attainment, even after adjusting for patient illness severity. Although residual unmeasured confounding is possible, these results suggest that a teambased approach for chronic disease management, with teams comprised of both physician and nonphysician providers, may be more efficient than the current system of health care delivery, which depends heavily on physician providers. In addition, we found that a mild increase in the illness burden of patients in a provider's panel was associated with higher attainment of guidelinerecommended cholesterol targets. This likely reflects an increased awareness of providers to implement evidencebased therapies in patients with increased but not overwhelmingly high illness burden. Our study supports previous findings that more complex patients are more likely to receive guideline-recommended care.9,13,14 Our analysis indicated that a higher number of primary care visits was an independent correlate of dual goal attainment, likely because more visits may afford more time for the provider to focus on preventive measures after other competing acute medical issues have been addressed. We also found that obese CHD patients were more likely to attain dual targets, which could be because providers likely recognize the importance of reaching dual targets in these high-risk patients. Conversely, a treatment gap for cholesterol management remained in African American CHD patients; this finding supports prior research that indicated African American patients were less likely to receive guideline-recommended care, suggesting another potential area of quality improvement.15 Certain limitations should be considered when interpreting our findings. Our results are derived from observational data rather than clinical trial data; therefore, bias and residual confounding are possible, although we tried to decrease the possibility of bias by establishing strict definitions for each variable included in our analyses. Furthermore, because this study reflects the experience of one region in the VA Health Care system, with a predominantly male population, our findings may not be generalizable; however, the VA facilities used for our analysis included both rural and urban facilities. In addition, we did not have accurate data on time from CHD diagnosis to the most recent lipid panel in most patients and therefore could not account for this variable. Given the administrative nature of the database, we were also unable to determine what proportion of CHD patients did not attain LDL-C and non–HDL-C goals because of adverse drug reactions (including myalgias from statin use), poor response to lipid-lowering medica-

Virani et al 1145

tions, or poor compliance. Similarly, we were unable to determine the impact of patients' physical activity or fitness level on goal attainment, as these variables are not well described in our administrative database. In conclusion, we showed that although attainment of the ATP III guidelines' goal for LDL-C in CHD patients has improved, combined goal attainment for LDL-C and non–HDL-C remains poor. Similarly, attainment of optional LDL-C goal per the ATP III update remains poor. We also identified several facility-, provider-, and patient-level characteristics associated with dual goal attainment, which have implications for future research as well as quality improvement initiatives. These include increased use of nonphysician primary care providers and more intense targeting of younger and African American CHD patients.

Disclosures Dr Virani has received honoraria from Abbott Laboratories and research grants from Merck and Co, Inc, and National Football League. Dr Ballantyne has received grant/research support from Abbott, AstraZeneca, GlaxoSmithKline, Merck, Sanofi-Synthelabo, Schering-Plough, and Takeda; is a consultant for Abbott, Amylin, BristolMyers Squibb, Kowa, Merck/Schering-Plough, Metabasis, NicOx, Novartis, Pfizer, Resverlogix, Roche, SanofiSynthelabo, Schering-Plough, and Takeda; is on the speakers' bureau for Merck/Schering-Plough, Pfizer, and Schering-Plough; and has received honoraria from Abbott, AstraZeneca, GlaxoSmithKline, Merck, Merck/ Schering-Plough, Kowa, Novartis, Pfizer, Sanofi-Synthelabo, Schering-Plough, and Takeda. All other authors declare no relationships with industry. The views expressed in this article are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs.

Acknowledgements The authors would like to thank Mark Kuebeler, MS, of the Michael E. DeBakey VA Medical Center Health Services Research and Development Center of Excellence for his programming effort on this manuscript and Ms Kerrie C. Jara for her editorial assistance.

References 1. Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive summary of the third report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA 2001;285:2486-97. 2. Grundy SM, Cleeman JI, Bairey Merz CN, et al. Implications of recent clinical trials for the National Cholesterol Education Program Adult Treatment Panel III guidelines. Circulation 2004;110:227-39.

1146 Virani et al

3. Blaha MJ, Blumenthal RS, Brinton EA, et al. The importance of non– HDL cholesterol reporting in lipid management. J Clin Lipidol 2008;2: 267-73. 4. Kastelein JJ, van der Steeg WA, Holme I, et al. Lipids, apolipoproteins, and their ratios in relation to cardiovascular events with statin treatment. Circulation 2008;117:3002-9. 5. Emerging Risk Factors Collaboration. Major lipids, apolipoproteins, and risk of vascular disease. JAMA 2009;302:1993-2000. 6. Davidson MH, Maki KC, Pearson TA, et al. Results of the National Cholesterol Education (NCEP) Program Evaluation ProjecT Utilizing Novel E-Technology (NEPTUNE) II survey and implications for treatment under the recent NCEP Writing Group recommendations. Am J Cardiol 2005;96:556-63. 7. Waters DD, Brotons C, Chiang CW, et al. Lipid Treatment Assessment Project 2: a multinational survey to evaluate the proportion of patients achieving low-density lipoprotein cholesterol goals. Circulation 2009;120:28-34. 8. Petersen LA, Pietz K, Woodard LD, et al. Comparison of the predictive validity of diagnosis-based risk adjusters for clinical outcomes. Med Care 2005;43:61-7. 9. Petersen LA, Woodard LD, Henderson LM, et al. Will hypertension performance measures used for pay-for-performance programs penalize those who care for medically complex patients? Circulation 2009;119:2978-85.

American Heart Journal June 2011

10. Melloni, C, Shah, BR, Ou, FS, et al. Lipid-lowering intensification and low-density lipoprotein cholesterol achievement from hospital admission to 1-year follow-up after an acute coronary syndrome event: results from the Medications ApplIed aNd SusTAINed Over Time (MAINTAIN) registry. Am Heart J 2010; 160:1121-9, 1129 e1. 11. Hynes DM, Perrin RA, Rappaport S, et al. Informatics resources to support health care quality improvement in the Veterans Health Administration. J Am Med Inform Assoc 2004;11:344-50. 12. Chan PS, Oetgen WJ, Buchanan D, et al. Cardiac performance measure compliance in outpatients: the American College of Cardiology and National Cardiovascular Data Registry's PINNACLE (Practice Innovation And Clinical Excellence) program. J Am Coll Cardiol 2010;56:8-14. 13. Harman JS, Edlund MJ, Fortney JC, et al. The influence of comorbid chronic medical conditions on the adequacy of depression care for older Americans. J Am Geriatr Soc 2005;53:2178-83. 14. Higashi T, Wenger NS, Adams JL, et al. Relationship between number of medical conditions and quality of care. N Engl J Med 2007;356: 2496-504. 15. Peterson ED, Shah BR, Parsons L, et al. Trends in quality of care for patients with acute myocardial infarction in the National Registry of Myocardial Infarction from 1990 to 2006. Am Heart J 2008;156: 1045-55.