Association of Region and Hospital and Patient Characteristics With Use of High-Intensity Statins After Myocardial Infarction Among Medicare Beneficiaries

Question  What drives variation in high-intensity statin use after acute myocardial infarction among older adults?

Findings  In this cohort study of 139 643 Medicare fee-for-service beneficiaries hospitalized for myocardial infarction, postdischarge high-intensity statin use increased from 23.4% in 2011 to 55.6% in 2015. In multivariable-adjusted models, geographic region was more strongly associated with high-intensity statin use after myocardial infarction than hospital or patient characteristics.

Meaning  These findings suggest that large geographic treatment disparities in high-intensity statin use after myocardial infarction are poorly understood and require further research and intervention.

Importance  High-intensity statin use after myocardial infarction (MI) varies by patient characteristics, but little is known about differences in use by hospital or region.

Objective  To explore the relative strength of associations of region and hospital and patient characteristics with high-intensity statin use after MI.

Design, Setting, and Participants  This retrospective cohort analysis used Medicare administrative claims and enrollment data to evaluate fee-for-service Medicare beneficiaries 66 years or older who were hospitalized for MI from January 1, 2011, through June 30, 2015, with a statin prescription claim within 30 days of discharge. Data were analyzed from January 4, 2017, through May 12, 2019.

Exposures  Beneficiary characteristics were abstracted from Medicare data. Hospital characteristics were obtained from the 2014 American Hospital Association Survey and Hospital Compare quality metrics. Nine regions were defined according to the US Census.

Main Outcomes and Measures  Intensity of the first statin claim after discharge characterized as high (atorvastatin calcium, 40-80 mg, or rosuvastatin calcium, 20-40 mg/d) vs low to moderate (all other statin types and doses). Trends in high-intensity statins were examined from 2011 through 2015. Associations of region and beneficiary and hospital characteristics with high-intensity statin use from January 1, 2014, to June 15, 2015, were examined using Poisson distribution mixed models.

Results  Among the 139 643 fee-for-service beneficiaries included (69 968 men [50.1%] and 69 675 women [49.9%]; mean [SD] age, 76.7 [7.5] years), high-intensity statin use overall increased from 23.4% in 2011 to 55.6% in 2015, but treatment gaps persisted across regions. In models considering region and beneficiary and hospital characteristics, region was the strongest correlate of high-intensity statin use, with 66% higher use in New England than in the West South Central region (risk ratio [RR], 1.66; 95% CI, 1.47-1.87). Hospital size of at least 500 beds (RR, 1.15; 95% CI, 1.07-1.23), medical school affiliation (RR, 1.11; 95% CI, 1.05-1.17), male sex (RR, 1.10; 95% CI, 1.07-1.13), and patient receipt of a stent (RR, 1.35; 95% CI, 1.31-1.39) were associated with greater high-intensity statin use. For-profit hospital ownership, patient age older than 75 years, prior coronary disease, and other comorbidities were associated with lower use.

Conclusions and Relevance  This study’s findings suggest that geographic region is the strongest correlate of high-intensity statin use after MI, leading to large treatment disparities.

High-intensity statin therapy is a class IA indication after myocardial infarction (MI) for individuals 75 years or younger in the absence of safety and tolerability concerns; moderate-intensity statin therapy is recommended for most individuals older than 75 years.^1,2 Use of high-intensity statins after MI compared with use of lower-intensity statins decreases rates of non-fatal MI, deaths due to coronary heart disease, and ischemic stroke in randomized clinical trials.³ The proportion of US adults taking high-intensity statins after MI has increased in recent years, but this therapy remains underused.⁴

For many years, regional variation in care processes and health outcomes across the United States have been documented for cardiovascular diseases and many other health conditions.⁵ Between-facility variation in statin use has been found within hospital registries such as the Get with the Guidelines Registry,⁶ among Veterans Administration hospitals,⁷ and in registries of clinical outpatient practices.^8,9 Also, variation in high-intensity statin use associated with patient characteristics^10,11 and regional variation in pharmacy performance related to statin use have been reported.¹² The goal of the present study was to describe variation in high-intensity statin use by geographic region and explore the relative contributions of geographic region and hospital and patient characteristics in a contemporary cohort of patients discharged after an MI.

We conducted a retrospective cohort study of US adults with fee-for-service Medicare coverage who were hospitalized for MI. Hospitalizations for MI were identified as inpatient claims with an International Classification of Diseases, Ninth Revision, primary discharge diagnosis of 410.xx, except 410.x2.¹³ Medicare data do not capture information on medications when the beneficiary pays out of pocket. Therefore, in the primary analysis, we required beneficiaries to have a claim for a statin within 30 days of hospital discharge. For analysis of trends in high-intensity statin use over time, we included Medicare beneficiaries who were hospitalized for MI from January 1, 2011, through June 30, 2015 (eFigure 1 in the Supplement). For analysis of characteristics associated with high-intensity statin use, we included Medicare beneficiaries who were hospitalized for MI from January 1, 2014, through June 30, 2015 (eFigure 2 in the Supplement). We further required that beneficiaries (1) were 66 years or older on the day of discharge; (2) had a hospital length of stay of at least 1 night and no more than 30 days; (3) survived for at least 30 days after discharge; (4) had full fee-for-service Medicare coverage during the hospitalization, the year before the hospitalization, and the 30 days after hospitalization; (5) were not in a skilled nursing facility or hospice care during the 30 days after hospitalization; (6) did not have end-stage heart failure or end-stage renal disease; (7) had a claim for a statin within 30 days of discharge; (8) were treated for their MI in hospitals with available data from the 2014 American Hospital Association Survey¹⁴ and the 2014 Hospital Compare metrics¹⁵; and (9) received care at hospitals that had at least 10 beneficiaries 75 years or younger and at least 10 beneficiaries older than 75 years. In sensitivity analyses, we included beneficiaries with and without statin claims within 30 days after discharge. Further explanation of the inclusion and exclusion criteria is provided in the eMethods in the Supplement. After all inclusion and exclusion criteria were applied, 139 643 beneficiaries treated in 1437 hospitals were included in the analysis of trends. For analysis of characteristics associated with high-intensity statin use, 42 962 beneficiaries treated in 833 hospitals were included, representing 23.7% of potentially eligible beneficiaries with hospitalizations for MI (range, 20.8% in the Pacific region to 27.6% in the East South Central region) (eTable 1 in the Supplement).

The institutional review board of the University of Alabama at Birmingham and the Centers for Medicare & Medicaid Services Privacy Board approved this research. A waiver of informed consent was granted for the use of deidentified data. Race/ethnicity is reported as recorded in Medicare enrollment data and is included in these analyses owing to its importance in treatment disparities. Data required to replicate this research are available from the Centers for Medicare & Medicaid Services. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.

Using Medicare provider numbers, we linked the beneficiaries’ claims data to the 2014 American Hospital Association Survey and the 2014 Hospital Compare quality metrics. From these data sources, we selected characteristics that may be related to posthospitalization use of high-intensity statins. Information on hospital ownership type (non–federal government, not-for-profit, or for profit), hospital size (<100, 100-199, 200-299, 300-399, 400-499, or ≥500 beds), metropolitan area (yes or no), and medical school affiliation (yes or no) were obtained from the American Hospital Association Survey. From the Hospital Compare metrics, we obtained information on post-MI 30-day readmission rates, post-MI 30-day mortality rates, overall hospital rating, patient satisfaction, and 30-day risk-adjusted overall posthospitalization mortality comparison, all expressed as above or equal to the median vs below the median. In addition, we identified the US Census region for each hospital (East South Central, Mountain, West South Central, Middle Atlantic, South Atlantic, West North Central, East North Central, Pacific, or New England) (eFigure 3 in the Supplement).

From the Medicare enrollment files, we identified each beneficiary’s age, sex, race/ethnicity, and low-income status, defined as dual enrollment in Medicare and Medicaid or eligibility for a Part D subsidy. From the claims in the year before the hospitalization, we identified prior coronary heart disease, diabetes, stroke, heart failure, dementia, chronic kidney disease, and cancer. We additionally identified claims for skilled nursing facility stays and hospitalizations in the year before the hospitalization for the MI. We used the MI hospitalization claims to determine whether a beneficiary had a new diagnosis of heart failure and whether the beneficiary had a coronary stent placed during the hospitalization.

Beneficiaries were required to have a prescription claim for a statin within 30 days of discharge after their MI hospitalization in the primary analysis. Statin intensity was assessed for the first statin claim after discharge. In accordance with the American Heart Association/American College of Cardiology guidelines for the treatment of cholesterol levels, we considered doses of 40 to 80 mg of atorvastatin calcium and 20 to 40 mg of rosuvastatin calcium to be high-intensity statins.¹ All other statins and doses were considered of low to moderate intensity.

Data were analyzed from January 4, 2017, through May 12, 2019. We calculated the frequency of hospital and individual characteristics among beneficiaries whose first statin claim after discharge for an MI was high intensity vs low or moderate intensity. We then calculated the percentage of beneficiaries whose first statin claim after discharge was for a high-intensity statin over time. To test whether a statistically significant variation occurred across regions from 2011 to 2015, we used Poisson distribution mixed models with hospital-specific random intercepts and region-specific fixed effects. To evaluate whether the variation across regions differed over time, we added a region by calendar year interaction term. Among beneficiaries hospitalized for MI from 2014 to 2015, we used Poisson distribution mixed models with hospital-specific random intercepts to calculate adjusted risk ratios (RRs) for high-intensity statin use. Initial models included the region and hospital characteristics and the individual characteristics separately. In subsequent models, all these characteristics were included in a single model.

Because the models included variables with different scales and variables with multiple categories, the magnitude of the RRs cannot be directly compared. We therefore calculated semipartial R² values, a function of the χ² statistics, for each of the independent variables included in the fully adjusted models.¹⁶ The semipartial R² values measure the relative contribution of individual independent variables to the model, adjusted for the contribution of the other independent variables in the model, with higher values indicating greater contribution. Calculation of the semipartial R² values allows direct comparison of the relative contribution of individual variables to the overall R² value, which measures the multivariate association between the full set of independent variables and the outcome. Multivariate association ranges from 0 to 1, indicating a worthless and a perfect model, respectively. For categorical variables (eg, region), we summed the semipartial R² statistics for each level to estimate the overall semipartial R² value.

Because the recommendation for use of high-intensity statins after MI is stronger for individuals 75 years or younger than for individuals older than 75 years,^1,2 we stratified the analyses by these age groups. In previous investigations, receipt of high-intensity statin therapy before a hospitalization for MI was strongly associated with high-intensity statin therapy after the index hospitalization.⁴ As sensitivity analyses, we repeated the analyses described above for patients who did not take high-intensity statins before their index MI. In addition, we repeated the analyses without requiring a statin claim within 30 days of discharge. All analyses were conducted using SAS, version 9.3 (SAS Institute, Inc). Two-sided P < .05 was considered statistically significant.

Among the 139 643 fee-for-service beneficiaries (69 968 men [50.1%] and 69 675 women [49.9%]; mean [SD] age, 76.7 [7.5] years), the percentage of Medicare beneficiaries with a high-intensity statin claim after a hospitalization for MI increased from 23.4% in 2011 to 55.6% in 2015. The percentage with high-intensity statin claims increased across all strata defined by region, hospital characteristics, and patient characteristics (eTables 2 and 3 in the Supplement). In 2011, the use of high-intensity statins varied from 17.0% in the West North Central region to 37.3% in New England (P < .001 for variation across regions) (Figure 1). The West North Central region had the greatest increase in high-intensity statin use (from 17.0% in 2011 to 58.4% in 2015), and the West South Central region had the smallest increase (from 19.6% in 2011 to 44.3% in 2015) (P < .001 for interaction between region and time). In 2015, the West South Central region had the lowest use of high-intensity statins (44.3%) whereas New England had the highest (74.6%) (P < .001 for variation across regions).

Hospital characteristics by high-intensity statin claims after MI among Medicare beneficiaries (2014-2015) are shown in Table 1, and multivariable-adjusted modeling is shown in eTable 4 in the Supplement. Beneficiaries residing in New England were most likely to be treated with high-intensity statins after MI (1861 [73.5%]), whereas those in the West South Central region were least likely to receive high-intensity statins (1703 [40.6%]). Non–federal government ownership (2136 [54.1%]), larger hospital size (9498 [58.6%] for hospitals with at least 500 beds), and medical school affiliation (15 533 [56.4%]) were associated with a greater likelihood of claims for high-intensity statin therapy. High-intensity statin claims did not differ by metropolitan area or Hospital Compare metrics.

Beneficiary characteristics by high-intensity statin claims after MI among Medicare beneficiaries hospitalized for MI in 2014 to 2015 are shown in Table 2, with multivariable modeling presented in eTable 5 in the Supplement. Individuals 75 years or younger (12 481 of 21 575 [57.8%]), men (12 656 of 22 619 [56.0%]), and those who received a stent during their index hospitalization (14 929 of 24 968 [59.8%]) were more likely to have a high-intensity statin claim. In contrast, individuals with comorbidities including prior coronary heart disease (7198 of 15 099 [47.7%]), prior heart failure (2962 of 6758 [43.8%]), and history of dementia (830 of 2065 [40.2%]) and those with a hospitalization in the year before the index MI (5226 of 11 198 [46.7%]) were less likely to have a high-intensity statin claim than their counterparts without these characteristics.

Geographic region (eg, New England vs West South Central RR, 1.66 [95% CI, 1.47-1.87]; P < .001 across all categories), hospital ownership (eg, for-profit vs not-for profit RR, 0.86 [95% CI, 0.79-0.92]; P < .001 across all categories), hospital size (eg, <100 beds vs 200-299 beds RR, 0.95 [95% CI, 0.79-1.13]; P < .001 across all categories), and hospital medical school affiliation (RR, 1.11; 95% CI, 1.05-1.17) as well as younger patient age (RR, 0.87; 95% CI, 0.84-0.89), male sex (RR, 1.10; 95% CI, 1.07-1.13), receipt of a stent during the index hospitalization (RR, 1.35; 95% CI, 1.31-1.39), and lack of prior coronary heart disease (RR, 0.93; 95% CI, 0.90-0.96) and heart failure (RR, 0.95; 95% CI, 0.91-1.00) were associated with a high-intensity statin claim (Table 3). None of the Hospital Compare metrics were associated with a high-intensity statin claim. The multivariate association (ie, overall R²) between the full set of independent variables and high-intensity statin claim was 0.3146. Geographic region was the strongest correlate of high-intensity statin claim with a semipartial R² value of 0.2207, followed by hospital ownership (semipartial R², 0.0286), stenting during the index hospitalization (semipartial R², 0.0229), medical school affiliation (semipartial R², 0.0182), hospital size (semipartial R², 0.0070), and patient age (semipartial R², 0.0061) and sex (semipartial R², 0.0026) (Figure 2). Characteristics of the study population by region are shown in eTables 6 and 7 in the Supplement.

Models for beneficiaries aged 66 to 75 years are shown in eTable 8 in the Supplement; for those older than 75 years, in eTable 9 in the Supplement. In both age groups, geographic region was most strongly associated with high-intensity statin use, and hospital ownership, hospital size, medical school affiliation, male sex, and receipt of a stent during the index hospitalization were associated with greater use of high-intensity statins.

Analyses restricted to beneficiaries who had no high-intensity statin claim before their index hospitalization are presented in eTable 10 in the Supplement. Consistent with the main analysis, geographic region and hospital characteristics were more strongly associated with a high-intensity statin claim than were beneficiary characteristics. Similarly, in an analysis that pooled beneficiaries with or without a statin claim in the 30 days after hospital discharge (eTables 11-13 in the Supplement), geographic region was the most strongly associated with high-intensity statin claims.

High-intensity statin claims after MI varied widely by region, with greatest use in New England. In adjusted models, high-intensity statin claims also varied by hospital characteristics and beneficiary characteristics, but geographic region remained most strongly associated with a high-intensity statin claim after hospitalization for an MI among Medicare beneficiaries. Although high-intensity statin claims have increased in all regions since 2011, the increase has varied by region, and substantial disparities persist.

Heterogeneity in statin use by region has been reported by others in a variety of care settings. Kumar and colleagues⁹ published data from the outpatient Reduction of Atherothrombosis for Continued Health (REACH) Registry in 2009 demonstrating that statin use was highest in the Northeast and lowest in the Southern region of the United States among patients undergoing primary and secondary prevention therapy. Data from the Get with the Guidelines Initiative 2002 to 2010,⁶ an in-hospital quality improvement initiative for MI care with voluntary participation by selected hospitals, similarly showed the highest use of statins in the US Northeast and the lowest rate in the South. This report was limited to Hispanic patients but is of interest because few differences in clinical patient characteristics between regions were found, demonstrating that the regional differences were not primarily driven by characteristics of the patient population served.⁶ Lower rates of statin use in the southern United States were also present in the Get with the Guidelines Stroke initiative¹⁷ and in 2014 data from the EQuIPP database that measures performance among pharmacies serving Medicare beneficiaries through the Part D provisions.¹² None of these studies differentiated use by intensity of statin treatment. Consistent with the pattern of regional variation in claims for high-intensity statins seen in the present investigation, we previously reported that, among Medicare beneficiaries, discontinuation of statin therapy after MI was lowest in New England.¹⁸ The present study extends prior findings by showing that regional differences in high-intensity statin use after MI were present in 2011 before the publication of the 2013 American College of Cardiology/American Heart Association cholesterol guidelines¹ and have persisted through 2015 despite substantial increases in use of high-intensity statins in all regions since 2011.

Hospital characteristics shown to correlate with use of high-intensity statins in the present analysis are generally similar to those that have been reported to correlate with any vs no statin use by others.^19-21 Consistent with the present analysis, statin use was higher in teaching hospitals compared with nonteaching hospitals based on data collected in 1999 to 2000 as part of the Global Registry of Acute Coronary Events.¹⁹ The increase in the use of statin therapy after MI from 1992 to 2005 was also greater at teaching than nonteaching hospitals in Canada.²¹ Larger hospital size, which correlates with greater volume of patients with MI, has been associated with better adherence to MI performance measures and better outcomes.²⁰

Use of high-intensity statins after MI varied with patient characteristics. After controlling for region and hospital characteristics, receipt of a stent was the patient characteristic most strongly associated with the use of high-intensity statins. As in prior studies, men were more likely than women to receive high-intensity statins after MI.^22,23 In contrast, characteristics known to identify patients as being at very high risk of future events (prior coronary disease, diabetes, and chronic kidney disease)^24,25 were not associated with a high-intensity statin claim or were associated with a slightly lower likelihood of a high-intensity statin claim. Despite emphasis on matching intensity of therapy with patient risk in current guidelines,^1,2 data from the present study suggest that the treatment paradox of directing more intensive risk reduction therapies toward lower-risk patients persists in contemporary patient populations.^26,27

We believe that our present study has important strengths. Use of claims data from Medicare with its large sample size allows for stable estimates of high-intensity statin claims and provides a high degree of generalizability to older adults in the United States. By combining information from Medicare with data from the American Hospital Association Survey and Hospital Compare, we were able to simultaneously assess regional and facility variations and beneficiary characteristics. The present study also has limitations. Having a prescription claim for a high-intensity statin does not necessarily equate to taking the high-intensity statin, although self-reported use and pill-bottle review correlate well with pharmacy claims.^28,29 The present study lacks information about the health care professionals involved in inpatient and subsequent outpatient care and does not provide information about intended therapy (ie, prescriptions written). We relied on claims data and could not determine whether low- to moderate-intensity statin therapy was appropriate for individual patients based on lipid profiles and propensity for or prior experience of adverse effects of statins. However, such clinical characteristics are unlikely to vary sufficiently across the country to explain the large differences in high-intensity statin use by region. The present study is limited to adults 66 years or older with Medicare fee-for-service health insurance. Whether the results can be generalized to younger adults and those with commercial health insurance is unknown to date.

Among Medicare beneficiaries, geographic region, rather than patient and hospital characteristics, was the most closely associated with high-intensity statin use after MI, leading to large treatment disparities. Reasons for these persistent regional disparities are poorly understood and require further research and intervention.

Accepted for Publication: May 31, 2019.

Corresponding Author: Vera Bittner, MD, MSPH, Division of Cardiovascular Disease, University of Alabama at Birmingham, 701 19th St S, Lyons-Harrison Research Building, Room 310, Birmingham, AL 35294 (vbittner@uab.edu).

Published Online: July 24, 2019. doi:10.1001/jamacardio.2019.2481

Author Contributions: Ms Dai and Dr Levitan had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Bittner, Colantonio, Rosenson, Muntner, Monda, Levitan.

Acquisition, analysis, or interpretation of data: Bittner, Colantonio, Dai, Woodward, Mefford, Rosenson, Muntner, Monda, Jaeger, Levitan.

Drafting of the manuscript: Bittner.

Critical revision of the manuscript for important intellectual content: Bittner, Colantonio, Dai, Woodward, Mefford, Rosenson, Muntner, Monda, Jaeger, Levitan.

Statistical analysis: Dai, Woodward, Jaeger, Levitan.

Obtained funding: Muntner, Monda, Levitan.

Administrative, technical, or material support: Colantonio, Rosenson, Muntner, Levitan.

Supervision: Rosenson, Muntner.

Conflict of Interest Disclosures: Dr Bittner reported serving on the Executive Steering Committee of the ODYSSEY OUTCOMES trial (Sanofi) with personal compensation, serving as National Coordinator for STRENGTH (AstraZeneca), DalGene (DalCor Pharmaceutical), and CLEAR (Esperion Therapeutics, Inc), and serving as a local site investigator for ARTEMIS (AstraZeneca) and COMPASS (Bayer Healthcare), all contracted through the University of Alabama at Birmingham; and consulting for Sanofi. Dr Woodward reported consulting to Amgen, Inc, and Kyowa Hakko Kirin Co Ltd, with compensation. Dr Rosenson reported receiving research funding through the Icahn School of Medicine from Akcea Therapeutics, Inc, Amgen, Inc (with consulting fees paid to Icahn School of Medicine at Mount Sinai), The Medicines Company, and Regeneron Pharmaceuticals, Inc; serving as a research consultant for Amgen, Inc; receiving honoraria from Akcea Therapeutics, Inc, Amgen, Inc, and Kowa Pharmaceuticals America, Inc; receiving royalties from UpToDate, Inc; and owning stock holdings in MediMergent, LLC. Dr Muntner reported receiving research funding from Amgen, Inc, through the University of Alabama at Birmingham. Dr Monda reported being an employee and a stockholder of Amgen, Inc. Dr Kilgore reported receiving research funding from Amgen, Inc, through the University of Alabama at Birmingham. Dr Jaeger reported receiving research support through grant 15SFRN2390002 from the American Heart Association and grant R01 HL117323 from the National Heart, Lung, and Blood Institute. Dr Levitan reported receiving research funding from Amgen, Inc, through the University of Alabama at Birmingham; serving on Amgen, Inc, advisory boards; and consulting for a research project funded by Novartis International AG with compensation. No other disclosures were reported.

Funding/Support: This study was supported by a research grant from Amgen, Inc.

Role of the Funder/Sponsor: The sponsor provided input on the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript. The sponsor had no role in the decision to submit the manuscript for publication.

Additional Information: Dr Kilgore died October 17, 2018.