Low levels of statin adherence have been documented in patients with coronary artery disease (CAD), but whether coronary revascularization is associated with improved adherence rates has yet to be evaluated. We identified all Medicare beneficiaries enrolled in 2 statewide pharmacy assistance programs who were ≥65 years old, who had been hospitalized for CAD from 1995 through 2004, and who had been prescribed statin therapy within 90 days of discharge (n = 13,130). Statin adherence was measured based on the proportion of days covered with statin therapy after hospital discharge, and full adherence was defined as proportion of days covered ≥80%. Statin adherence was compared in patients with CAD treated with medical therapy (n = 3,714), percutaneous coronary intervention (n = 6,309), or coronary artery bypass graft surgery (n = 3,107). Statin adherence significantly increased over the period of the study from 70.5% to 75.4% (p <0.0001). After hospitalization for CAD, patients treated with percutaneous coronary intervention and coronary artery bypass graft surgery had full adherence rates of 70.6% and 70.2%, respectively. Full adherence rates were significantly lower for patients treated with coronary revascularization compared to patients treated with medical therapy (79.4%, p <0.0001). Independent predictors of higher statin adherence included treatment with medical therapy, later year of hospital admission, white race, previous statin use, and use of other cardiac medications after CAD hospitalization (p <0.01 for all comparisons). In conclusion, in patients receiving invasive coronary treatment, statin adherence remains suboptimal, despite strong evidence supporting their use. Given the health and economic consequences of nonadherence, these findings highlight the need for developing cost-effective strategies to improve medication adherence after coronary revascularization.
Substantial attention has been focused on increasing statin prescription rates in hospital and at time of discharge for patients with coronary artery disease (CAD). Conversely, less interest has been paid to understanding and improving postdischarge adherence to these medications. Long-term studies of patients with CAD treated primarily with medical therapy have noted disappointingly low rates of adherence, ranging from 50% to 85% at 1 year to 45% to 50% at 5 years after statin initiation. However, no studies have assessed adherence to statin therapy in patients with high-risk CAD treated with percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG). Mechanically revascularized patients may be more adherent to preventive therapies given patient perceptions of the severity of their illness. Alternatively, after undergoing PCI or CABG, patients may have fewer ischemic symptoms and believe that subsequent preventive therapies are unnecessary. Therefore, the goal of this study was to compare adherence to statin therapy in patients hospitalized with CAD treated with coronary revascularization or medical therapy.
Methods
We assembled a cohort of Medicare beneficiaries with CAD by linking Medicare files describing all clinical encounters to complete medication-use data from the Pennsylvania Pharmaceutical Assistance Contract for the Elderly (PACE) and the New Jersey Pharmaceutical Assistance to the Aged and Disabled (PAAD) programs. During the period studied, PACE and PAAD provided prescription drug benefits to lower middle-income Medicare beneficiaries ≥65 years of age whose yearly earnings were above the threshold to qualify them for Medicaid. Participants paid copayments of $5 through $10 per prescription without any deductibles. The programs covered all medications that required a prescription and did not restrict which medications could be prescribed (i.e., the programs did not use formularies, preferred drug lists, or previous authorization programs). Data from PACE, PAAD, and Medicare was assembled into a relational database consisting of claims for all filled prescriptions and hospital and clinical characteristics for the patients in our cohort. These data sources have been used extensively to study population-based medication use. All traceable patient-specific identifying factors were transformed into anonymous coded study numbers to protect subjects’ privacy. This study was approved by the institutional review board of the Brigham and Women’s Hospital, Boston, Massachusetts.
We included all patients who were discharged alive from the hospital after admission for active CAD from January 1, 1995 through December 31, 2004. This included patients who were treated with medical therapy after myocardial infarction or unstable angina ( International Classification of Disease, Ninth Revision , codes 410.01 to 410.91 or 411), patients who underwent PCI (codes 36.01 to 36.09), and patients who underwent CABG (codes 36.1x or 36.2x). Validation studies have demonstrated that these codes have a specificity of ≥96% for the diagnosis of CAD. We used a hierarchical approach to define the 3 subgroups in this study. Patients admitted with active CAD who underwent CABG during the index hospitalization were assigned to the CABG subgroup, regardless of whether they received PCI. Patients with CAD who underwent PCI (but not CABG) during the index admission were allocated to the PCI subgroup. All other patients with CAD were classified as those treated with medical therapy. Because our focus was adherence to statin therapy during follow-up after CAD hospitalization, only patients who filled a statin prescription within 90 days of hospital discharge were included. The date of the first statin prescription fill after hospital discharge was defined as the index date for each patient (i.e., start of follow-up). Follow-up terminated on December 31, 2005. We excluded patients who died during the index hospitalization and patients who were not active users of either drug benefit program.
We determined patient co-morbidities by searching physician service claims and hospitalization records for relevant diagnostic codes in the 1-year period before the index date. In this manner, the following characteristics were identified: age at index date, year of hospitalization, gender, race, length of hospital stay, previous myocardial infarction or acute coronary syndrome, hypertension, diabetes mellitus, congestive heart failure, stroke, peripheral vascular disease, previous CABG, previous PCI, and chronic kidney disease. We assessed whether each patient used any statin in the 1-year period before the index CAD hospitalization. We also determined the total number of generic medications prescribed and use of the following specific medications in the 1-year period before and 30 days after CAD hospitalization: angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers, β blockers, calcium channel blockers, clopidogrel, fibrates, diuretics, nitrates, digoxin, and warfarin. Hospitals accredited by the Association of American Medical Colleges were classified as teaching hospitals. All other hospitals were classified as nonteaching hospitals.
We measured statin adherence using the proportion of days that patients had medication available to them after hospital discharge. This widely used measurement, proportion of days covered, was calculated by dividing the number of days of medication supplied between the first and last statin prescriptions during the study period (numerator) by the number of days between these 2 prescriptions plus the accumulated days supplied from the last prescription (denominator). We defined patients as being “fully adherent” if their proportion of days covered was ≥80%. We then compared rates of full statin adherence between patients admitted to hospital for CAD who were treated with medical therapy, PCI, or CABG using Fisher’s exact test. Baseline characteristics of patients treated with medical therapy, PCI, or CABG were compared using Student’s t tests, Fisher’s exact tests, or chi-square tests, as appropriate. Statistical significance was defined as a p value <0.05. Trends in average yearly full adherence were compared among the 3 subgroups using analysis of variance.
Multivariable logistic regression analysis was used to identify predictors of statin full adherence for the entire cohort. Indicator terms for the 3 subgroups were included in the models. To adjust for important clinical covariates, the following factors were also incorporated in the models: age, gender, race, year of index hospitalization, treatment in a teaching hospital, history of peripheral vascular disease, congestive heart failure, previous stroke, previous myocardial infarction, diabetes mellitus, statin use in the 1-year period before hospital admission, number of generic medications used in the 1-year period before hospital admission, and cardiac medication use (β blocker, angiotensin-converting enzyme inhibitor, or angiotensin II receptor blocker) within 30 days after hospital discharge. We then repeated our analysis to test whether statin adherence differed if we included only new statin users (i.e., those patients who did not fill a statin prescription in the 1-year period before their index hospitalization, n = 5,120). Odds ratios are reported with 95% confidence intervals. All analyses were performed using SAS 8.2 (SAS Institute, Cary, North Carolina).
Results
Our cohort consisted of 13,130 patients admitted to hospital with CAD who were treated with medical therapy (n = 3,714), PCI (n = 6,309), or CABG (n = 3,107). Mean age of the cohort was 76.7 ± 6.1 years, and 73.3% of patients were women ( Table 1 ). Patients treated with CABG were significantly younger, had a higher incidence of diabetes mellitus, and were more likely to have previously received statin therapy (p <0.05 for all comparisons). Patients treated with medical therapy were more likely to have a history of stroke (p <0.05). Patients treated with PCI had a high rate of clopidogrel use before hospitalization (i.e., clopidogrel prescription before elective PCI).
| Characteristics | Medical Therapy (n = 3,714) | PCI (n = 6,309) | CABG (n = 3,107) |
|---|---|---|---|
| Patient characteristics | |||
| Age (years), mean ± SD | 79.5 ± 6.9 | 75.9 ± 6.0 | 75.1 ± 5.3 ⁎ |
| Women | 77.4% | 73.3% | 68.4% ⁎ |
| White | 91.9% | 90.0% | 91.6% ⁎ |
| Previous myocardial infarction | 14.9% | 21.3% | 38.0% ⁎ |
| Previous percutaneous coronary intervention | 4.4% | 2.1% | 6.8% ⁎ |
| Previous coronary artery bypass graft surgery | 0.8% | 1.0% | 0% ⁎ |
| Heart failure | 64.1% | 33.4% | 45.2% ⁎ |
| Stroke | 7.8% | 4.7% | 6.0% ⁎ |
| Peripheral vascular disease | 5.5% | 5.2% | 5.2% |
| Hypertension | 84.7% | 89.0% | 91.1% ⁎ |
| Diabetes mellitus | 48.3% | 42.2% | 49.8% ⁎ |
| Chronic kidney disease | 34.1% | 22.6% | 26.3% ⁎ |
| Prehospital medication use (1 year) | |||
| Total number of medications, mean ± SD | 11.2 ± 7.6 | 11.0 ± 7.1 | 11.5 ± 8.1 ⁎ |
| Previous statin | 58.6% | 59.5% | 66.8% ⁎ |
| Angiotensin-converting enzyme inhibitor or angiotensin II receptor blocker | 61.4% | 50.8% | 51.8% ⁎ |
| Clopidogrel | 23.3% | 54.7% | 19.5% ⁎ |
| β Blocker | 68.1% | 66.2% | 72.6% ⁎ |
| Calcium channel blocker | 51.6% | 50.7% | 51.7% |
| Digoxin | 13.7% | 8.2% | 13.4% ⁎ |
| Diuretics | 13.5% | 11.8% | 10.8% ⁎ |
| Fibrate | 3.5% | 3.5% | 4.5% ⁎ |
| Nitrates | 67.2% | 63.1% | 60.4% ⁎ |
| Warfarin | 9.4% | 6.9% | 7.4% ⁎ |
| Posthospital medication use (30 days) | |||
| Total number of medications, mean ± SD | 6.0 ± 4.2 | 5.6 ± 3.8 | 5.6 ± 4.3 ⁎ |
| Angiotensin-converting enzyme inhibitor or angiotensin II receptor blocker | 44.8% | 38.8% | 31.4% ⁎ |
| Clopidogrel | 22.1% | 53.0% | 9.3% ⁎ |
| β Blocker | 56.0% | 56.0% | 57.4% |
| Calcium channel blocker | 22.9% | 27.6% | 16.2% ⁎ |
| Digoxin | 9.7% | 5.4% | 15.5% ⁎ |
| Diuretics | 5.1% | 4.8% | 3.5% ⁎ |
| Fibrate | 0.8% | 1.3% | 1.1% |
| Nitrates | 57.1% | 45.8% | 11.7% ⁎ |
| Warfarin | 8.0% | 5.5% | 8.0% ⁎ |
| Posthospital statin type | |||
| Atorvastatin | 30.5% | 36.7% | 33.9% ⁎ |
| Fluvastatin | 2.5% | 2.9% | 2.6% |
| Lovastatin | 8.1% | 6.4% | 6.8% |
| Pravastatin | 13.4% | 13.3% | 13.8% |
| Rosuvastatin | 0.3% | 1.0% | 0.4% |
| Simvastatin | 45.2% | 39.7% | 42.5% |
| Hospital characteristics | |||
| Teaching hospital | 48.8% | 70.1% | 76.5% ⁎ |
| Length of stay (days), mean ± SD | 6.8 ± 3.8 | 4.4 ± 4.3 | 10.0 ± 5.1 ⁎ |
During the period of the study, rates of full adherence significantly increased ( Figure 1 ) from 70.5% to 75.4% for the entire cohort (p <0.0001) and within the CABG and medical therapy subgroups individually (p <0.01 for the 2 comparisons). For the PCI subgroup, there was a nonsignificant trend of increased adherence over time (p = 0.15). After hospitalization for CAD, patients treated with PCI and CABG had full adherence rates of 70.6% and 70.2%, respectively ( Figure 2 ). Full adherence rates were significantly lower for patients treated with coronary revascularization compared to patients treated with medical therapy (79.4%, p <0.0001). Similar results were obtained when the cohort was limited to new statin users only, with full adherence rates significantly lower in patients treated with PCI (70.4%) and CABG (67.2%) compared to patients treated with medical therapy (78.7%, p <0.0001).
Multivariable logistic analysis was performed to identify independent predictors of statin full adherence for the entire cohort ( Table 2 ). Factors independently associated with higher statin adherence included later year of hospital admission, older patient age, white race, history of congestive heart failure, previous statin use, and cardiac medication use (β blocker, angiotensin-converting enzyme inhibitor, or angiotensin II receptor blocker) within 30 days after hospital discharge (p <0.01 for all comparisons). Compared to medical therapy alone, treatment with CABG or PCI independently predicted lower statin adherence after hospitalization for CAD, even after controlling for other factors influencing adherence (p <0.0001 for the 2 comparisons). Similar results were obtained when analysis was restricted to new statin users only ( Table 2 ).
