Methods

We conducted a retrospective cohort study of Medicare beneficiaries experiencing a CHD-related hospitalization between January 1, 2007 and December 31, 2009 using a national 5% random sample from the Centers for Medicare and Medicaid Services Chronic Condition Data Warehouse. Data were obtained from Medicare enrollment, inpatient, outpatient, skilled nursing facility, noninstitutional provider, home health, and prescription drug files. These data comprise Medicare parts A (inpatient), B (outpatient), and D (prescription drug coverage) insurance coverage. The institutional review board at the University of Alabama at Birmingham approved the study.

Our analyses included Medicare beneficiaries who experienced an acute myocardial infarction (AMI), coronary artery bypass graft or PCI in 2007, 2008, or 2009. AMI was defined by International Classification of Diseases, Ninth Edition, Clinical Modification ( ICD-9-CM ) code 410 .xx (except 410.x2 which indicates a subsequent episode of care) in any discharge diagnosis position in an inpatient file record. CABG was identified using current procedure terminology codes 33510 to 33536 or ICD-9-CM procedure codes 36.10-36, and PCI was identified using current procedure terminology codes 92980 to 92996 or ICD-9-CM procedure codes 00.66, 36.01-36.09. To be eligible for this analysis, beneficiaries had to meet the following criteria: (1) be ≥66 years on the day of hospital admission for their CHD event, (2) have a hospitalization duration <30 days, (3) have continuous “full coverage” for Medicare during the 365-day period before admission for their CHD hospitalization (i.e., the look-back period), (4) survive 365 days after hospital discharge with full Medicare coverage, (5) have no CHD diagnoses or coronary revascularization procedures during the look-back period, and (6) live in the United States and have valid birth/death dates. Full coverage was defined as enrollment in traditional Medicare fee-for-service (parts A and B) and part D Medicare coverage and not being enrolled in a Medicare Advantage plan. Of the Medicare beneficiaries meeting these criteria (n = 13,567), we excluded 3,424 who were not taking antihypertensive medication before their CHD event and 3,857 who did not fill a statin within 90 days of hospital discharge. Finally, to focus on patients initiating statins, we excluded 3,591 beneficiaries who had filled statins during the look-back period. After these exclusions, we included 2,695 Medicare beneficiaries in the current analyses.

We used the pharmacy-based proportion of days covered (PDC) to quantify antihypertensive medication adherence during the look-back period. PDC is the measure of adherence endorsed by the Pharmacy Quality Alliance. The 5 most commonly used classes of antihypertensive medications in US adults (angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, β blockers, calcium channel blockers, and diuretics) were identified using Medicare part D claims. For each antihypertensive class, the PDC was defined as the number of days between the first and last fill for which the patient had medication available to take divided by the number of days during this period. For primary analyses, we defined adherence based on the average PDC across antihypertensive classes. In secondary analyses, we used the antihypertensive medication class with the lowest PDC. Adherence was categorized as low (PDC <50%), medium (PDC 50% to <80%), and high (PDC ≥80%).

The 2 outcomes were statin discontinuation and low statin adherence in the 365 days after hospital discharge. Statin prescription fills were identified using Medicare part D claims and included simvastatin, atorvastatin, pravastatin, rosuvastatin, lovastatin, and fluvastatin. Statin discontinuation was defined as having no statin fills after the initial fill during the 365 days after hospital discharge. Statin adherence was calculated using the interval-based PDC. The interval-based PDC is useful when defining adherence to a single drug class. The denominator for the PDC was the number of days between a beneficiary’s first statin fill and 365 days after discharge. The numerator was the total number of days during this period with a statin available; if a prescription was filled earlier than the last day supplied by the previous fill, the overlapping supply was added to the end of the last prescription fill. Beneficiaries were considered to have statins available regardless of brand, dose, or switching types of statins. Low statin adherence was defined as a PDC <80%.

Demographic covariates included calendar year of the CHD event, age, gender, and self-reported race-ethnicity. Receipt of a low-income subsidy for Medicare prescription drug premiums was used as a marker of low socioeconomic status. Diabetes, renal disease, depression, and the Charlson index during the look-back period were defined using previously published algorithms. Total number of medications filled during the look-back period was captured using Medicare part D claims. Covariates assessed during the 90 days after hospital discharge included first type of statin filled, use of non-statin lipid-lowering medications (niacin, fibrates, bile acid sequestrants, or ezetimibe), being rehospitalized, having an outpatient cardiologist visit, and being admitted to a skilled nursing facility.

Characteristics of the study sample and the percentage of beneficiaries who discontinued statins were calculated by the level of antihypertensive medication adherence. The risk ratio for statin discontinuation was calculated for patients with low and medium versus high adherence to antihypertensive medications before hospitalization. In beneficiaries who did not discontinue their statin, we calculated the distribution of PDC for statins in the 365 days after hospital discharge. Also, we calculated the percentage of patients with, and risk ratio for, low statin adherence associated with antihypertensive medication adherence. Risk ratios were calculated using generalized linear models with a Poisson distribution, a log link, and a robust variance estimator. For both statin discontinuation and low statin adherence, risk ratios were calculated after initial adjustment for age, race-ethnicity, sex, and calendar year of the CHD event. A second model also adjusted for receipt of a low-income subsidy, diabetes, renal disease, depression, Charlson index, and number of medications taken during the look-back period. The third model also adjusted for the covariates from the 90 days after hospital discharge (as described earlier). Analyses were repeated using the lowest PDC for antihypertensive medication adherence before hospital admission. All analyses were conducted using SAS 9.2 (SAS Institute, Research Triangle Park, NC).

Results

Before their CHD hospitalization, Medicare beneficiaries included in the current analyses were taking an average of 1.9 classes of antihypertensive medications, 41%, 36%, and 23% were taking 1, 2, and ≥3 classes, respectively. Averaging PDC across classes, 63%, 29%, and 8% of beneficiaries had high, medium, and low adherence to antihypertensive medication, respectively. Whites and patients with a Charlson index of 0 had higher antihypertensive medication adherence before hospitalization ( Table 1 ). Receipt of a low-income subsidy was associated with lower antihypertensive medication adherence. Those with higher antihypertensive medication adherence before hospitalization were more likely to fill non-statin lipid-lowering drugs after discharge, whereas those with lower antihypertensive medication adherence were more likely to be rehospitalized within 90 days of discharge and to fill atorvastatin after discharge.

Overall, 6.9% of beneficiaries (n = 187) discontinued statins after a single fill. Beneficiaries with worse antihypertensive medication adherence based on average PDC before their CHD hospitalization were more likely to discontinue statins ( Figure 1 ). After multivariable adjustment, lower adherence with antihypertensive medication before hospitalization remained associated with an increased risk for statin discontinuation ( Table 2 ). In secondary analyses defining antihypertensive medication adherence as the lowest PDC across classes, lower antihypertensive medication adherence was associated with an increased risk for discontinuation of statins ( Supplementary Table 1 ).

Percentage of Medicare beneficiaries discontinuing statins in the year after initiation by level of adherence to antihypertensive medication before a CHD-related hospitalization.

Table 2

Risk ratios for statin discontinuation associated with average proportion of days covered for antihypertensive medication prior to a coronary heart disease event

Model	Average Proportion of Days Covered for Antihypertensive Medication
	≥80% (n = 1,990)	50% to <80% (n = 560)	<50% (n = 145)
1	1 (ref)	1.40 (1.00, 1.97)	2.41 (1.51, 3.85)
2	1 (ref)	1.37 (0.97, 1.92)	2.39 (1.50, 3.83)
3	1 (ref)	1.38 (0.98, 1.95)	2.41 (1.51, 3.87)

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Outcomes of a Pharmacoinvasive Strategy for Successful Versus Failed Fibrinolysis and Primary Percutaneous Intervention in Acute Myocardial Infarction (from the Strategic Reperfusion Early After Myocardial Infarction [STREAM] Study)
2. Meta-Analysis of the Effect of Renal Denervation on Blood Pressure and Pulse Pressure in Patients With Resistant Systemic Hypertension
3. Long-Term Survival for Patients With Acute Decompensated Heart Failure According to Ejection Fraction Findings
4. Usefulness of N-Terminal Pro-Brain Natriuretic Peptide Levels to Predict Success of Weaning from Intra-Aortic Balloon Pumping

Stay updated, free articles. Join our Telegram channel

Join