Trends in Use and Expenditures for Brand-name Statins After Introduction of Generic Statins in the US, 2002-2018

Key Points Question How did expenditures for statins change after market exclusivity ended and generic statins became available? Findings Using 17 years of the Medical Expenditure Panel Study data, this survey study of generic competition among statins found that the end of market exclusivity was associated with $925.60 of annual savings per individual and $11.9 billion in savings for the US. Meaning Full generic competition of statins was associated with significant cost savings across all major payers within the US health care system.


Introduction
The high and increasing health care expenditure in the US is a national problem. In 2019 alone, national health care spending increased 4.6%. 1 Retail prescription drug spending reached $369.7 billion (10% of total health care spending), 1 resulting in immense public outcry. 2 Improving the affordability of prescription drugs has become a focus of both the scientific and policy communities. 3,4 Generic competition is one important policy tool to stem rising prescription drug expenditures. 3 In the US, the entry of generic drugs into the market typically starts after the market exclusivity period of a brand-name drug ends. 4 A generic drug is a pharmaceutically equivalent alternative to a brand-name drug that is made by another manufacturer and is sold at a fraction of the cost. 7,8 According to the Center of Drug Evaluation and Research estimates, the average generic drug price is 49% of the corresponding brand-name drug. 9 When 6 or more generic competitors exist, the reduction is up to 95%. 5 Statins are a class of lipid-lowering medication used to prevent and treat atherosclerotic cardiovascular disease (ASCVD). 6 In the US, heart disease is the leading cause of death for men and women across most racial and ethnic groups. 7 From 2014 to 2015, heart disease resulted in $351.3 billion in annual direct and indirect costs. 8 Per the American Heart Association guidelines, statin therapy has become the pharmaceutical cornerstone of ASCVD prevention and treatment, with an emphasis on high-intensity statin use for patients with clinical ASCVD or high ASCVD risk. 6 The 7 available statins are classified into 3 levels of intensity (low, moderate, and high) based on pharmacologic potency and dosage. 9 Before the end of their market exclusivity, Zocor (simvastatin) and Lipitor (atorvastatin) were 2 of the top-selling medicines in the world. 10 Furthermore, Crestor (rosuvastatin) was 1 of the 3 most costly drugs for the Medicare Part D program in 2015. 11 Between 2002 and 2013, the overall rate statin use among adults older than 40 years increased from 17.9% to 27.8%, with a total cost amounting to $16.9 billion in 2012-2013. 12  Prior studies 21,35 have shown that the introduction of generic competition was associated with (1) the reduction of Lipitor prescriptions in a large private insurer and (2) a large increase in the number of prescriptions filled for generic simvastatin in statewide Medicaid programs. Nevertheless, to our knowledge, the economic impact of generic competition for all available statins has not been comprehensively examined. In this survey study, we examine use and expenditure trends for all available statins, using data obtained on private health insurance, public health insurance (ie, Medicaid and Medicare), and out-of-pocket (OOP) spending. In addition, we examine the association between generic statin competition and relevant use and cost savings.

Methods Data
A 17-year panel survey study was performed using Medical Expenditure Panel Survey data. The Medical Expenditure Panel Survey is an annual nationwide representative survey of the noninstitutionalized US population. In this study, we included data from January 1, 2002, to December 31, 2018. Full-year consolidated data and prescribed medicine files were merged to construct the analytical data sets. 13 The Medical Expenditure Panel Survey is a publicly available and deidentified data file; per the US Department of Health and Human Services guidelines, this study was exempted from institutional review broad approval. Written consent was obtained from survey participants who were contacted for interviews and to contact their particular clinicians and pharmacies. This analysis and report followed the American Association for Public Opinion Research (AAPOR) 12-item reporting guideline (eTable 4 in the Supplement).

Study Design
A panel difference-in-differences design was used to evaluate the outcome of end of market exclusivity on the annual number of brand-name generic statin purchases and expenditures. The first difference refers to the difference in the number of purchases and expenditures of each brandname statin before and after the end of market exclusivity. The second difference refers to the difference between a given brand-name statin that ends market exclusivity in a given year and all other brand-name statins with market exclusivity yet to end at that time. This same design was used to analyze the generic statins (eAppendix 1 in the Supplement). Seven brand-name statins and the 5 generic statins available as of 2018 were included in the study. The preperiod trends were tested and determined to meet the parallel trends assumptions (eFigures 5-8 in the Supplement).

Data Processes
Both national-and individual-level analyses were performed. We merged prescription drug event files with full-year consolidation files to obtain the number of statin purchases and corresponding expenditures at the individual level (n = 58 354) (eTable 3 in the Supplement). We then converted individual-level data to national-level data by aggregating the annual number of purchases and expenditures by each statin class. A final analytical panel-data structure with nationally representative estimates was then created based on statin class (n = 204). More detailed information about the data processes and sample selection is presented in eAppendix 1 and eFigure 1 in the Supplement.

Annual Number of Statin Purchases and Expenditures
The main outcome variables were annual number of statin purchases and expenditures.
Expenditures referred to what was paid for statin purchases by payers and individuals combined. We calculated total annual expenditure as well as mean payments by private insurance, Medicaid, Medicare, and OOP spending. The expenditure estimates were presented in 2018 US dollars (USD) based on Consumer Price Index prescribed medicines (ie, the price index for prescription drugs) (eAppendix 1 in the Supplement). 14

Key Independent Variable
The Multum Lexicon provided by the Agency for Healthcare Research and Quality was used to identify the class of statin. 15 The level-1 drug therapeutic category of 358 identified metabolic agents, the level-2 code of 19 specified antihyperlipidemic agents, and the level-3 code of 173 classified 3-hydroxy-3-methylglutaryl coenzymes (HMG-CoA reductase inhibitors). The recorded drug name was used to specify whether a statin was branded or generic.

Covariates
Age, sex (male and female), race and ethnicity (self-reported as Asian, Black or African American, and White), health insurance coverage (whether a person had private or public health insurance, or did not have any kind of coverage in the past year), usual source of care (defined as whether participants had a particular medical professional, doctor's office, clinic, or other place where they would usually go if sick or in need of advice about health), marital status (currently married or otherwise), geographic region (Northeast, Midwest, South, and West), and cardiovascular disease (ie, angina, congenital heart disease, stroke, and myocardial infarction) were included in the regression analysis. 6 For the analysis at the national level, all covariates were presented as a percentage, other than age, which was presented as a mean.
generalized linear model was developed to estimate national-level annual expenditures. To estimate models of individual-level expenditures, which have a large number of zeros, we used a 2-part model-first part logit and second part generalized linear model. [16][17][18] The modified Park test was performed to confirm the choice of γ family as appropriate (eTable 1 in the Supplement). Sampling weights and strata were implemented to generate nationally representative estimates (eAppendix 1 in the Supplement). A 2-sided P value <.05 was considered statistically significant. All statistical analyses were conducted from November 1, 2020, to March 30, 2021, using SAS, version 9.4 (SAS Institute Inc) and STATA, version 16.1 (StataCorp LLC).

Trends of Statin Use and Expenditures
After the end of market exclusivity, the number of brand-name statin purchases decreased across branded statins (Figure 1

Association Between End of Market Exclusivity and Statin Purchases and Expenditure
The end of market exclusivity was associated with substantial cost savings for individuals and the US as a nation. Nationally, the number of brand-name statin purchases decreased by 90.9% (95% CI, 56%-98%) following the end of market exclusivity ( Table 2). Individually, a 27.4% (95% CI, 13%-40%) decrease in the number of brand-name statin purchases was associated with the end of market exclusivity (

The Benefits of Generic Competition
Generic competition is considered the most effective tool for reducing prescription prices in the US. 19 In 2017, although accounting for just 10% of total prescriptions, brand-name drugs represented 77%   In line with the findings of existing studies on high blood pressure medications such as generic substitution of angiotensin 1 and on the substitution of generic oral contraceptives, 2 the current study found that generic entry was associated with a great reduction of brand-name statin purchases. This reduction will likely continue to grow, especially for high-intensity statins. Following the release of the 2018 American College of Cardiology/American Heart Association Guideline on the Management of Blood Cholesterol recommendations, 6,9 high-intensity statins (ie, atorvastatin and rosuvastatin) became the most dominant statin class in the US, accounting for 62.83% of the market share in 2018 (Table 1). As shown in prior studies, 21,22 full generic competition (defined as having 4 or more generic manufactures for a certain statin) reduced brand-name atorvastatin prescriptions significantly within 1 large private health insurance company and saved the US health care system approximately $1.9

JAMA Network Open | Public Health
billion. Our results indicate that full generic competition of all statins occurred within 1 to 2 years and resulted in an 88.1% reduction in total annual brand-name statin expenditures.
Comparable reductions were also observed for the low-to moderate-intensity statins. Sales of Zocor declined dramatically after the end of market exclusivity. 23   The cost savings for private insurance was greater than other payers (Medicaid, Medicare, and out of pocket, F score = 11.79; P < .01).

Barriers and Incentives to Generic Entry
Brand-name manufactures have an incentive to engage in strategies to delay the introduction of generic drugs and reduce competition. 26 Policies to encourage generic entry, such as the 1984 Hatch-Waxman Act, can promote lower prices. 21,22,27,28 For instance, the total annual expenditure for atorvastatin rose in 2016 ( Figure 2) because of an increase in the manufacturer's price; in 2017, however, the annual expenditure decreased after a drop in the manufacturer's price because of the competition from additional generics.

Generic Competition Among Different Major Payers
Generic competition has had different impacts across payers, some of which are able to negotiate prices. 29 The current study found that generic entry was associated with a 98.6% reduction in Medicare spending on statins, which is higher than that estimated in prior studies. 30 (Figure 3). 33 The observed increase in Medicare spending on brand-name and generic statins is reflective of this shift.
Medicaid programs experienced modest cost savings after the introduction of generic competition.
Possible reasons for the relatively smaller reduction include the price spike observed in 20% of generic drugs covered under Medicaid between 2014 and 2017 34 as well as other structural barriers. 35 Our cost savings estimates for Medicaid are likely conservative because the rebates received by payers (ranging from 45% to 76% between 2015 and 2018) are unaccounted for. 36,37 The complicated rebate system embedded in pharmacy benefit managers could also have resulted in conservative national estimates of the private health insurance cost savings. 38 Private insurers incurred the highest savings, a result possibly attributable to the generic competition in commercial health plans and high OOP spending. 39 Our findings on reductions in OOP expenditures are comparable to other estimates. These relatively large savings likely partially reflect a relatively steep brand-name statin list price previously faced by cash-paying patients who held little bargaining power compared with large private and public insurers, as well as the trend toward high deductibles among the privately insured. In light of the evidence that people with lower incomes purchase generic medications more often, 40 generic competition reduces disparities and increases access to affordable medication. However, the mean duration of market exclusivity of statins is 13.1 years, a time period that greatly exceeds other market exclusivity federal statutory minimums, such as the 5-to 7-year minimum for small molecules, and perpetuates disparities in access. 4 Figure 1 and Figure 2). Nevertheless, we decided to not modify the original data owing to lack of sufficient information. Third, the cost-saving estimates for Medicaid are likely conservative because the analysis did not account for rebates. Fourth, the influence of other newly available treatment options was not evaluated. For instance, the introduction of ezetimibe and proprotein convertase subtilisin/kexin type 9 (PSCK9) inhibitors as adjuvant therapy has