Statin Cost-Effectiveness Comparisons Using Real-World Effectiveness Data: Formulary Implications

Similar documents
Effectiveness of statins in Medicare-eligible patients and patients < 65 years using clinical practice data*

Cost-Effectiveness of Rosuvastatin Compared with Other Statins from a Managed Care Perspective

Reducing low-density lipoprotein cholesterol treating to target and meeting new European goals

Andrew Cohen, MD and Neil S. Skolnik, MD INTRODUCTION

How would you manage Ms. Gold

Comments from AstraZeneca UK Ltd

Dyslipidemia in the light of Current Guidelines - Do we change our Practice?

Cost-effectiveness of evolocumab (Repatha ) for hypercholesterolemia

Setting The setting was primary care. The economic study was carried out in the UK.

Extensive evidence exists that aggressive. Differences Between Clinical Trial Efficacy and Real-world Effectiveness REPORTS

The cost-effectiveness of a new statin (rosuvastatin) in the UK NHS Palmer S J, Brady A J, Ratcliffe A E

Prescription Switching and Reduced LDL-C Goal Attainment

Supplement Table 2. Categorization of Statin Intensity Based on Potential Low-Density Lipoprotein Cholesterol Reduction

Managed Care Trends in Statin Usage GARY R. BAZALO, MS, MBA

Spending in the United States for prescription drugs was $216.7

Hae Sun Suh, B.Pharm., Ph.D. Jason N. Doctor, Ph.D.

NATIONAL INSTITUTE FOR HEALTH AND CLINICAL EXCELLENCE. Overview

rosuvastatin, 5mg, 10mg, 20mg, film-coated tablets (Crestor ) SMC No. (725/11) AstraZeneca UK Ltd.

4/7/ The stats on heart disease. + Deaths & Age-Adjusted Death Rates for

Comparison of Original and Generic Atorvastatin for the Treatment of Moderate Dyslipidemic Patients

Setting The study setting was secondary care. The economic analysis was conducted in the UK.

Changing lipid-lowering guidelines: whom to treat and how low to go

Cholesterol Management Roy Gandolfi, MD

Cost-effectiveness of evolocumab (Repatha ) for primary hypercholesterolemia and mixed dyslipidemia.

Cost effectiveness of statin therapy for the primary prevention of coronary heart disease in Ireland Nash A, Barry M, Walshe V

CLINICAL. Determinants of Compliance With Statin Therapy and Low-Density Lipoprotein Cholesterol Goal Attainment in a Managed Care Population

STATIN UTILIZATION MANAGEMENT CRITERIA

LOW-DENSITY LIPOPROTEIN CHOLESTEROL GOAL ATTAINMENT AMONG MALAYSIAN DYSLIPIDEMIC PATIENTS

ATP IV: Predicting Guideline Updates

Disclosures No relationships (not even to an employer) No off-label uses. Cholesterol Lowering Guidelines: What now?

SOLVAY GROUP London Morning Meeting. June 26, 2009

Lipid Panel Management Refresher Course for the Family Physician

Costing report: Lipid modification Implementing the NICE guideline on lipid modification (CG181)

An update on lipidology and cardiovascular risk management. Lipids, Metabolism & Vascular Risk Section - Royal Society of Medicine

NATIONAL INSTITUTE FOR HEALTH AND CLINICAL EXCELLENCE

Is Lower Better for LDL or is there a Sweet Spot

Drug Class Monograph

The new guidelines issued in PRESENTATIONS... Future Outlook: Changing Perspectives on Best Practice

CVD risk assessment using risk scores in primary and secondary prevention

Common Repatha Documentation Requirements for Patients With Primary Hyperlipidemia and Established CVD 1,2

See Important Reminder at the end of this policy for important regulatory and legal information.

Learning Objectives. Patient Case

Issue date: January Review date: November Statins for the prevention of cardiovascular events. Technology Appraisal 94

Medical evidence suggests that

Pharmaceutical Help to Control Cholesterol

Coronary heart disease (CHD) has. Clearfield The National Cholesterol Education Program Adult Treatment Panel III guidelines

JAMA. 2011;305(24): Nora A. Kalagi, MSc

UnitedHealthcare Pharmacy Clinical Pharmacy Programs

Dyslipedemia New Guidelines

Update on Dyslipidemia and Recent Data on Treating the Statin Intolerant Patient

A Model for Comparing Unnecessary Costs Associated with Various Prescription Fill-Quantity Policies: Illustration Using VA Data

Statins in the Treatment of Type 2 Diabetes Mellitus: A Systematic Review.

PCSK9 Agents Drug Class Prior Authorization Protocol

4 th and Goal To Go How Low Should We Go? :

What do the guidelines say about combination therapy?

The updated guidelines from the National

Supplementary Online Content

ESC Geoffrey Rose Lecture on Population Sciences Cholesterol and risk: past, present and future

HDL-C. J Jpn Coll Angiol, 2008, 48: NIPPON DATA80, MEGA study, JELIS, dyslipidemia, risk assessment chart

UnitedHealthcare Pharmacy Clinical Pharmacy Programs

Data Alert. Vascular Biology Working Group. Blunting the atherosclerotic process in patients with coronary artery disease.

This is a lipid lowering drug strategy which should only be used within an overall lifestyle and clinical management strategy.

Cost-effectiveness of pravastatin for primary prevention of coronary artery disease in Japan Nagata-Kobayashi S, Shimbo T, Matsui K, Fukui T

Novel PCSK9 Outcomes. in Perspective: Lessons from FOURIER & ODYSSEY LDL-C. ASCVD Risk. Suboptimal Statin Therapy

03/30/2016 DISCLOSURES TO OPERATE OR NOT THAT IS THE QUESTION CAROTID INTERVENTION IS INDICATED FOR ASYMPTOMATIC CAROTID OCCLUSIVE DISEASE

2013 ACC AHA LIPID GUIDELINE JAY S. FONTE, MD

Review of guidelines for management of dyslipidemia in diabetic patients

Introduction. Objective. Critical Questions Addressed

PCSK9 Inhibitors and Modulators

Marshall Tulloch-Reid, MD, MPhil, DSc, FACE Epidemiology Research Unit Tropical Medicine Research Institute The University of the West Indies, Mona,

HYPERLIPIDEMIA IN THE OLDER POPULATION NICOLE SLATER, PHARMD, BCACP AUBURN UNIVERSITY, HARRISON SCHOOL OF PHARMACY JULY 16, 2016

CVD Risk Assessment. Lipid Management in Women: Lessons Learned. Conflict of Interest Disclosure

CLINICAL OUTCOME Vs SURROGATE MARKER

Statins ARE Enough For The Prevention of CVD! Professor Kausik Ray Imperial College London, UK

... CPE/CNE QUIZ... CPE/CNE QUESTIONS

Medical Policy An independent licensee of the Blue Cross Blue Shield Association

ARBITER 6-HALTS HDL And LDL Treatment Strategies

ISCHEMIC VASCULAR DISEASE (IVD) MEASURES GROUP OVERVIEW

Supplementary Online Content

Integrating Effectiveness and Safety Outcomes in the Assessment of Treatments

Nearly 62 million people in the. ... REPORTS... New Therapeutic Options in the National Cholesterol Education Program Adult Treatment Panel III

The JUPITER trial: What does it tell us? Alice Y.Y. Cheng, MD, FRCPC January 24, 2009

Improving Medication Adherence through Collaboration between Colleges of Pharmacy and Community Pharmacies

Effective Treatment Options With Add-on or Combination Therapy. Christie Ballantyne (USA)

Saving Lives, Saving Families: The health, social and economic advantages of detecting and treating familial hypercholesterolaemia (FH)

Abstract. Background. Methods

Making War on Cholesterol with New Weapons: How Low Can We/Should We Go? Shaun Goodman

JUPITER NEJM Poll. Panel Discussion: Literature that Should Have an Impact on our Practice: The JUPITER Study

Ezetimibe and SimvastatiN in Hypercholesterolemia EnhANces AtherosClerosis REgression (ENHANCE)

( Diabetes mellitus, DM ) ( Hyperlipidemia ) ( Cardiovascular disease, CVD )

What have We Learned in Dyslipidemia Management Since the Publication of the 2013 ACC/AHA Guideline?

In May 2001, the National Cholesterol. Effective Management of Patients With Dyslipidemia REPORT. Robert J. Lipsy, PharmD

Setting The setting was secondary care. The economic study was carried out in the USA.

Approach to Dyslipidemia among diabetic patients

In 2001, the National Cholesterol Education Program

Early Clinical Development #1 REGN727: anti-pcsk9

New ACC/AHA Guidelines on Lipids: Are PCSK9 Inhibitors Poised for a Breakthrough?

STATINS FOR PAD Long - term prognosis

Nicole Ciffone, MS, ANP-C, AACC Clinical Lipid Specialist

Source of effectiveness data The effectiveness evidence came from a review of published studies and the authors' assumptions.

Transcription:

Volume 11 Number 7 2008 VALUE IN HEALTH Statin Cost-Effectiveness Comparisons Using Real-World Effectiveness Data: Formulary Implications Robert L. Ohsfeldt, PhD, 1 Sanjay K. Gandhi, PhD, 2 Kathleen M. Fox, PhD, 3 James M. McKenney, PharmD 4 1 Texas A&M Health Science Center, College Station,TX, USA; 2 AstraZeneca Pharmaceuticals LP, Wilmington, DE, USA; 3 University of Maryland School of Medicine, Baltimore, MD, USA; 4 National Clinical Research, Inc., Richmond,VA, USA ABSTRACT Objective: To compare the effectiveness and costeffectiveness among generic and branded statins in routine clinical practice. Methods: Retrospective database study of patients, 18+, who were newly prescribed statin therapy. Statin effectiveness and cost-effectiveness in reducing low-density lipoprotein cholesterol (LDL-C) and attaining LDL-C goals were evaluated. Results: Of 10,421 eligible patients, % LDL-C reduction was significantly greater (P < 0.001) with rosuvastatin (-31.6%) than other statins (-13.9 to -21.9%). Percentage of patients at moderate/high risk attaining LDL-C goal was higher (P < 0.001) for rosuvastatin (76.1%) versus other statins (57.6 72.6%). Rosuvastatin was more effective and less costly than atorvastatin. Among generic statins, simvastatin required >61% discount to branded price to achieve similar cost-effectiveness as generic lovastatin. Conclusions: In clinical practice, rosuvastatin is more effective and less costly in lowering LDL-C and LDL-C goal attainment compared with atorvastatin. Simvastatin was more cost-effective compared with lovastatin if >61% discount to branded price was achieved. Keywords: ATP III goal attainment, cost-effectiveness, LDL reduction, real-world effectiveness, rosuvastatin. Introduction Coronary heart disease (CHD) continues to be the leading cause of mortality and morbidity in the United States [1] and is one of the top five most costly health conditions to US employers, with total annual costs of $130 billion [2,3]. Evidence-based guidelines issued by the National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATP III) underline the importance of hyperlipidemia treatment with an aggressive LDL-C goal of <100 mg/dl for high risk patients [4]. In the changing statin marketplace and in multitier formulary systems, it is important for managed care plans to identify and utilize the most cost-effective options among generic and branded statins. Lovastatin (LOV) became available as a generic statin in 2002. Pravastatin (PRV) and simvastatin (SMV) followed with generic introduction in April and June 2006. Because there are more choices for generic statins, health-care administrators are considering various statin formulary changes to optimize pharmacy budget spend. An optimal statin formulary would provide adequate clinical flexibility to maximize clinical benefit Address correspondence to: Robert L. Ohsfeldt, Texas A&M Health Science Center, Department of Health Policy and Management, College Station, TX 77843-1266, USA. E-mail: rohsfeldt@srph.tamhsc.edu 10.1111/j.1524-4733.2008.00354.x to patients with efficient use of health-care resources. So it is important for health plan administrators to identify the most cost-effective statin that may be a generic statin for tier one and a more effective branded statin that may be placed in tier two. A formulary structure such as this can meet the objective of appropriate clinical choices, as well as cost-effective use of statins. On the other hand, efficient health-care delivery systems may consider using generic agents for patients at low risk and a branded statin for patients at higher risk of CHD. Nevertheless, there is limited real-world evidence available comparing the effectiveness and cost-effectiveness of the specific generic and branded statins to guide the formulary decision. Clinical trials have clearly demonstrated that statin therapy lowers LDL-C by 45% to 63% with rosuvastatin (RSV) and 25% to 60% with other statins [5 7]. Moreover, randomized trials have shown that approximately 86% to 94% of patients attain NCEP ATP III LDL-C goal with rosuvastatin therapy; fewer attain goal with the other statins [6,7]. Nevertheless, observational studies from chart review of primary care physician practice have revealed lower rates of statin effectiveness and the number of patients achieving treatment goals. For example, only 23% to 48% of high-risk patients were reported to have attained their LDL-C goal in these studies [8 10]. This paradox between clinical trial-reported statin efficacy and actual clinical practice statin effectiveness highlights 2008, International Society for Pharmacoeconomics and Outcomes Research (ISPOR) 1098-3015/08/1061 1061 1069 1061

1062 Ohsfeldt et al. the need to better understand the comparative effectiveness of statins in a real-world setting. The objectives of the present study was to estimate the effectiveness of rosuvastatin to other statins, cost-effectiveness of RSV compared with atorvastatin (ATV) (branded statins), and among SMV, PRV, and LOV (generic statins) in patients treated in routine clinical practice. Cost-effectiveness comparisons for branded and generic statins were separately conducted to aid in the formulary decision-making within a typical US health plan or pharmacy benefit management organization where formulary structure is comprised of generic agent(s) in tier one and branded agent(s) in tier two. Methods A retrospective study was conducted utilizing the General Electric Medical System (GEMS) electronic medical records database of patients who were treated in outpatient physician practices in the United States. The objectives of the study were to examine the effectiveness of statin monotherapy in the clinical practice setting by comparing RSV with ATV, SMV, PRV. and LOV for reducing LDL-C and LDL-C goal attainment. The analysis was conducted to determine whether the most efficacious (LDL-C lowering and NCEP ATP III goal attainment) statin (RSV) as observed in clinical trials is the most effective statin in the real-world setting as well. Similarly, the analysis determined if the most efficacious (LDL-C lowering and NCEP ATP III goal attainment) generic statin (SMV), as observed in clinical trials, is also more effective than other generic statins (PRV, LOV) in the real-world clinical practice. Cost-effectiveness was assessed among branded statins and generic statins separately. LOV was selected as the reference generic statin because it had a more stable acquisition cost because it has been on the market longest. Patients receiving fluvastatin, the least efficacious statin, were too few for meaningful analysis (<4% of the statin users), and were excluded from the analysis. Recently, a fixed dose combination of SMV and ezetimibe became available; however, it was not included in this study because the number of eligible patients receiving this fixed dose combination was relatively small for reliable effectiveness estimates. Patients who were newly prescribed statin therapy during August 2003 to May 2005 and had no prior prescription for any dyslipidemic medication, including bile acid sequestrants, fibrin, niacin, ezetimibe, or a statin, in the preceding 12 months were included in the study. Titration of statin therapy was allowed, but patients switching to other statins during the study period were excluded. Patients had to be continuously enrolled (i.e., active in physician s practice) for a minimum of 15 months; 12 months prior to and 3 months postinitiation of statin monotherapy. Additionally, patients were required to have a minimum of 90-day supply of statin therapy (either a 90-day prescription or three 30-day prescriptions) and lipid results within 90 days prior to and greater than 30 days after initiating statin therapy. The lipid values closest to the date of statin therapy initiation was defined as the baseline lipid measure. The follow-up lipid value was defined as the average of all lipid measures during the follow-up period, from 30 days after initiation of statin therapy to the date of the last statin prescription at time of discontinuation or end of study (i.e., August 2005). Therapy discontinuation was defined as the lack of a prescription or refill order within 1.5 times the prescription days supply. Thus, if a 30-day statin supply was ordered, then the prescription must be refilled or a new order written within 45 days of the initial prescription to consider the patient persistent on statin therapy. Two effectiveness outcomes were assessed: 1) percent reduction in LDL-C; and 2) percentage of patients attaining NCEP ATP III LDL-C goal. The outcome measures were computed for each individual statin. For LDL-C goal attainment assessment, patients were stratified based upon NCEP CHD risk groups [4]. CHD and CHD risk equivalent was defined as myocardial infarction, ischemic heart disease, acute coronary syndrome, cerebral vascular accident, transient ischemic attack, peripheral vascular disease, abdominal aortic aneurysm, angina pectoris, atherosclerosis, and diabetes mellitus. The GEMS database did not contain information on inpatient procedures. A count of risk factors was done to assign patients not assigned to the CHD or CHD risk equivalent category to moderate or low CHD risk. Moderate CHD risk patients were defined by the presence of two or more CHD risk factors, including current cigarette smoking, hypertension diagnosis or blood pressure 140/90 mmhg, low high-density lipoprotein cholesterol (HDL- C) < 40 mg/dl, and age 45 for men and 55 for women. Low-risk patients were defined as those with one or no CHD risk factors. Because inpatient procedure data and data from nonprimary care settings were not available in the database, there was potential that some high-risk patients were misclassified into the moderate risk category due to missing information on inpatient procedures. The fact that physicians started these patients with LDL-C < 130 mg/dl on a statin treatment was considered a strong indicator of their underlying high risk status or a more aggressive LDL-C target goal (<100 mg/dl). The LDL-C goal was defined as <100 mg/dl for high-risk patients, as well as moderate-risk patients who had statin therapy started, and baseline, untreated LDL-C levels <130 mg/dl. Moderate risk patients with LDL-C 130 mg/dl at baseline were assigned a goal of

Cost-Effective Generic and Branded Statins 1063 <130 mg/dl. Low-risk patients were assigned a target goal of <160 mg/dl. Effectiveness Analysis Multivariate linear and logistic regression analyzes were conducted to adjust the effectiveness estimates for baseline characteristics of age, sex, smoking, hypertension, CHD, systolic blood pressure, baseline LDL-C, and therapy duration. The LDL-C goal attainment analyzes were stratified by CHD risk level, high plus moderate risk, and low risk. Logistic regression models were used to estimate the probability of attaining goal with different statins, and the sample mean of predicted probabilities was used as an estimate of the expected rate of goal attainment adjusted for patient characteristics. All statistical analyzes were conducted using SAS Version 9.1 (SAS Institute, Inc., Cary, NC). Cost-Effectiveness Analysis A decision analysis, model-based cost-effectiveness analysis was employed to estimate the incremental cost relative to incremental effectiveness for the three group comparisons. First, cost-effectiveness was computed for the branded statins, RSV and ATV. Second, costeffectiveness was estimated comparing LOV to SMV and PRV. LOV was used as the reference generic statin because it has been a generic product the longest, with a stable market price. The third comparison estimated cost-effectiveness between the most cost-effective branded statin and the most effective generic statin while accounting for rebates on branded statins and different patient copayments between branded and generic statins. The analysis employed a payer perspective with direct medical costs only and time horizon of 1 year. Resource-use parameters were based upon actual utilization observed in the study cohort and included annualized statin costs and the cost of titration (including additional physician visits and drug wastage). The costs of lipid panel and safety monitoring tests were not included because their frequencies were nearly identical across the statin user groups. Routine physician visits were also not included because there was no anticipated difference in utilization among the statins [11]. Unit costs were based on Medicare reimbursement rates for physician visits associated with titration of the statins (measured as the average of the 2006 Medicare Physician Fee Schedule for 99212 and 99213, or $60) [12]. Unit drug costs were based on the wholesale acquisition cost (WAC) of statins as of May 2006 [13]. The base case drug price was $2.63 for RSV (all doses), $2.31, and $3.30 for ATV (10 and 20 mg and higher doses, respectively), $1.87, $2.51, and $4.37 for SMV (5, 10, and 20 mg or higher doses, respectively), $2.86, $2.90, and $4.26 for PRV (10, 20 and 40 mg or higher doses, respectively), and $0.62, $1.10, and $1.97 for LOV (10, 20, and 40 mg, respectively). Patient-level effectiveness parameters were taken from the multivariate regression models, and included percentage reduction in LDL-C and achievement of ATP III LDL-C goal. All cost-effectiveness models were estimated using TreeAge Pro Healthcare (TreeAge Software, Inc., Williamstown, MA). Key parameters were subjected to extensive sensitivity analyzes [14,15], with particular attention to the drug cost inputs, given uncertainty about future generic statin prices as competition among generic manufacturers intensifies. A threshold analysis [16] was performed for willingness-to-pay equal to zero to determine the price point at which SMV or PRV have the same cost per unit of effectiveness as LOV utilizing generic LOV pricing. Because SMV and PRV are new generic agents, competitive pricing has not occurred yet. Nevertheless, this landscape will change as more manufacturers produce these generic products, driving the price lower among all generic statins. Thus, the threshold analysis provided the WAC discount price needed for SMV and PRV to be as cost-effective as LOV. Also, different levels of patient copayment ($5 generic/$15 brand, $10 generic/$25 brand, and $10 generic/$30 brand) were applied in the sensitivity analysis to determine the actual threshold price for generic SMV from a payer perspective after accounting for patient costs. Results The eligible study population consisted of 10,421 patients: 541 on RSV, 6383 ATV, 2232 SMV, 705 PRV, and 560 LOV. Several characteristics differed among the patients on statins. RSV patients were younger (P < 0.002) and had a higher baseline LDL-C (P < 0.001) than other statin patients (Table 1). Because the baseline LDL-C level was higher for RSV patients, these patients were further from attaining ATP III goals and required a greater LDL-C reduction than other statin patients. RSV patients also had a significantly shorter length of statin therapy (183 days) than other patients (239 259 days), P < 0.001. The average daily dose of statin was significantly lower (P < 0.05) for RSV (11.6 mg) versus other statins (17.2 34.8 mg) (Table 1). Change in Lipid Parameters The unadjusted and adjusted percent change in LDL-C for each statin is reported in Table 2. RSV patients had the greatest observed decrease in LDL-C, with an average -32.5% reduction compared with -14.3% to -20.4% for other statin patients. Multivariate analyzes adjusting for baseline differences in age, sex, smoking, hypertension, CHD, systolic blood pressure, baseline LDL-C, and treatment duration also demonstrated that RSV patients had a significantly larger decrease (-31.6%) in LDL-C than patients using

1064 Ohsfeldt et al. Table 1 Baseline characteristics of patients treated in clinical practice by statin therapy type Characteristics Rosuvastatin Atorvastatin Simvastatin Pravastatin Lovastatin Number of subjects 541 6383 2232 705 560 Age, years, mean SD 62 12 63 12* 67 11* 67 11* 68 11* Male, % 45% 48% 51% 47% 41% Coronary heart disease, % 11% 11% 11% 10% 10% Diabetes, % 6% 5% 4% 3% 5% Hypertension, % 45% 40% 41% 44% 45% Initial dose, mg, mean (SD) 11.6 (6.2) 17.2 (12.8)* 23.9 (11.4)* 34.8 (14.7)* 24.1 (11.2)* Therapy duration, mean (SD) 182.6 (116.7) 258.5 (159.1)* 253.9 (157.9)* 253.0 (156.7)* 238.6 (155.4)* Baseline lipid parameters LDL-C, mean SD 155.5 48.8 131.0 45.7* 126.8 43.8* 133.3 39.8* 145.1 40.1* TC, mean SD 229.7 47.1 209.2 46.9* 203.1 43.6* 209.8 39.2* 223.0 41.8* HDL-C, mean SD 49.7 13.6 51.3 13.5* 52.2 14.1* 52.2 13.8* 54.1 14.7* TG, mean SD 170.5 92.8 151.3 82.6* 147.0 80.3* 145.1 74.4* 147.2 77.9* *P-value < 0.05 for comparison to rosuvastatin. LDL-C, low-density lipoprotein cholesterol;tc, total cholesterol; HDL-C, high-density lipoprotein cholesterol;tg, triglycerides. other statins (-13.9% to -21.9%). Total cholesterol change among the statins was significantly greater for RSV users (observed mean change: -21.8% 19.7, adjusted mean change: -18.9% 10.5) compared with other statins (Table 2). Among generic statins, the observed 18.0% reduction in LDL-C for SMV was significantly greater than PRV (-14.3%, P < 0.001), but was not different from LOV (-20.0%) (Table 2). The observed LDL-C reduction for PRV was significantly less than LOV (P < 0.001). After adjustment for baseline characteristics, average percent LDL-C reduction remained significantly greater for SMV compared with PRV (P < 0.001). The adjusted percent LDL-C change among SMV patients was not different from LOV patients. PRV patients had significantly less LDL-C reduction after adjustment than LOV patients. NCEP Goal Attainment After adjustment in baseline characteristics, the odds of attaining NCEP ATP III LDL-C goal among the moderate/high-risk group were greatest with RSV users (Table 3). The estimated percent of patients attaining ATP III LDL-C goal was significantly higher (P-value 0.004) among RSV patients (76.1%) than among ATV (72.6%), SMV (65.0%), PRV (57.6%) and LOV (61.7%) patients (Table 3). There was no significant difference in goal attainment between rosuvastatin and other statins in the low-risk group, except for PRV (90.5% of RSV attained goal vs. 79.5% of PRV, P < 0.05). Focusing on generic statins, LDL-C goal attainment was significantly greater among SMV patients than PRV or LOV patients (P < 0.001, Table 3). For SMV, 65.0% of patients attained LDL-C goal compared to 57.6% of pravastatin and 61.7% of lovastatin patients. There were significantly more LOV patients attaining LDL-C goal than PRV (P < 0.001). Cost-effectiveness Branded statins. In the base-case analysis of drug costs (WAC pricing) and titration costs (physician visits), RSV had the lower annualized cost ($1008) compared with ATV ($1179) after adjusting for differences in patient characteristics (Table 4). Utilizing the adjusted LDL-C reduction (adjusted for baseline characteristics), RSV had lower adjusted annualized cost per unit of LDL-C reduction ($35). For incremental costeffectiveness, RSV dominated ATV for LDL-C reduction (Table 4). For LDL-C goal attainment, focusing on patients in the moderate/high-risk group, the annualized cost per additional moderate/high-risk patient attaining goal was lower for RSV ($1401) (Table 5). Table 2 Mean change in lipid parameters by statin type Lipid parameter Rosuvastatin N = 541 Atorvastatin N = 6383 Simvastatin N = 2232 Pravastatin N = 705 Lovastatin N = 560 % LDL-C change -32.5 (27.7) -20.4* (28.1) -18.0* (27.3) -14.3* (23.2) -20.0* (21.0) % LDL-C change adjusted -31.6 (16.8) -21.9* (16.8) -19.1* (16.8) -13.9* (16.8) -18.0* (16.8) % Total Cholesterol change -21.8 (19.7) -14.6* (18.8) -11.1* (17.5) -10.5* (14.5) -13.1* (14.4) % Total cholesterol change adjusted -18.9 (10.5) -14.5* (10.5) -11.9* (10.5) -10.3* (10.5) -11.2* (10.5) *P-value 0.001 as compared to rosuvastatin. P < 0.001 for lovastatin compared to pravastatin. P < 0.001 for simvastatin compared to pravastatin or lovastatin. Mean change adjusted for age, sex, smoking, hypertension, CHD, systolic blood pressure, baseline LDL-C, and treatment duration. LDL-C, low-density lipoprotein cholesterol; CHD, coronary heart disease.

Cost-Effective Generic and Branded Statins 1065 Table 3 Odds ratios and 95% confidence intervals for attaining ATP III LDL-C goal among patients with high/moderate CHD risk ATP III LDL-C goal attainment* Rosuvastatin N = 309 Atorvastatin N = 2698 Simvastatin N = 929 Pravastatin N = 354 Lovastatin N = 303 Odds ratio (95%CI) 0.72 (0.53 0.96) 0.49 (0.38 0.64) 0.26 (0.20 0.36) 0.31 (0.23 0.44) Adjusted percentage attain LDL-C goal 76.1 (8.2) 72.6 (8.3) 65.0 (9.3) 57.6 (10.1) 61.7 (9.5) *All comparisons are with rosuvastatin and adjusted for baseline differences in age, sex, smoking, hypertension, CHD, systolic blood pressure, percent LDL-C reduction needed to reach goal, and treatment duration. P < 0.001. P < 0.001 for simvastatin compared with pravastatin or lovastatin. P < 0.001 for lovastatin compared with pravastatin. ATP III, Adult Treatment Panel III; LDL-C, low-density lipoprotein cholesterol; CHD, coronary heart disease. RSV also dominated (less costly and more effective) ATV for LDL-C goal attainment in high and moderaterisk patients. Generic statins. For generic statins, LOV, as the least expensive generic statin, was utilized as the reference and the threshold discount needed to achieve equivalent cost-effectiveness across the generic statins was estimated. PRV must be discounted by 71% (1- threshold discount) of current WAC price to achieve equal cost-effectiveness as LOV, and SMV must be discounted by 61% to achieve equivalent costeffectiveness as LOV for lowering LDL-C (Table 4). Similarly, for moderate/high-risk patients, PRV must be discounted by 65.6% and SMV by 62% of current WAC price to achieve equivalent cost-effectiveness as LOV for ATP III LDL-C goal attainment (Table 5). Most cost-effective branded statin versus most effective generic statin. It is assumed that as more manufacturers of generic SMV emerge, its price may decline steeply, making it more economically attractive to health plans. As a result, a comparison of the most cost-effective branded statin, RSV, to the most effective (and soon to be much cheaper) generic statin, SMV, may be of significant interest to decision-makers. As noted above, given that drug acquisition costs are a substantial component of overall drug treatment costs, cost-effectiveness assessments are highly dependent on acquisition cost assumptions. If the generic price discount from the current WAC for SMV is allowed to vary, a threshold discount can be identified, at which RSV is preferred to generic SMV by a payer willing to pay nothing for incremental effectiveness (willingness-to-pay = 0). As shown in Table 6, Table 4 Base case cost per patient per LDL-C reduction for branded and generic statins for all patients Statin therapy Effectiveness % LDL-C change* Annualized cost ratio Mean cost-effectiveness Incremental cost-effectiveness ratio Branded statins Rosuvastatin -31.6% $1008 $35 Atorvastatin -21.9% $1179 $54 Dominated Generic statins Threshold discount Lovastatin -18.0% $632 $35 1.00 Pravastatin -13.9% $480 $35 0.286 Simvastatin -19.1% $661 $35 0.389 *Mean change adjusted for age, sex, smoking, hypertension, CHD, systolic blood pressure, baseline LDL-C, and treatment duration. Nondrug costs were on average $7 $25/year. Threshold discount was the price at which generic simvastatin or generic pravastatin have the same cost per unit of effectiveness as lovastatin for willingness-to-pay = 0. LDL-C, low-density lipoprotein cholesterol; CHD, coronary heart disease. Table 5 Base case cost per patient attaining ATP III LDL-C goal for patients of moderate/high CHD risk Statin therapy Effectiveness % attain Mean cost-effectiveness LDL-C goal* Annualized cost ratio Incremental cost-effectiveness ratio Branded statins Rosuvastatin 76.1% $1066 $1401 Atorvastatin 72.6% $1195 $1646 Dominated Generic statins Threshold discount Lovastatin 61.7% $632 $1025 1.00 Pravastatin 57.6% $591 $1026 0.344 Simvastatin 65.0% $667 $1027 0.379 *Adjusted for baseline differences in age, sex, smoking, hypertension, CHD, systolic blood pressure, percent LDL-C reduction needed to reach goal, and treatment duration. Nondrug costs were on average $7 $25/year. Threshold discount was the price at which generic simvastatin or generic pravastatin have the same cost per unit of effectiveness as lovastatin for willingness-to-pay = 0. ATP III, Adult Treatment Panel III; LDL-C, low-density lipoprotein cholesterol; CHD, coronary heart disease.

1066 Ohsfeldt et al. Table 6 Threshold generic simvastatin price for different patient copayment levels Patient Copayment focusing on LDL-C reduction, generic SMV would have to be discounted by at least 58.3% [that is, (1-0.417) 100] to be preferred when generic/ branded copays are $10/$25. The threshold discount decreases as the generic/branded copay differential narrows and increases if the differential expands the threshold discount is 54.7% given $5/$15 copay tiers, and is 61.8% given $10/$30 copay tiers. Shifting the focus to LDL-C goal attainment among moderate and high-risk patients, the threshold discount is 59.2% given generic/branded copays of $10/$25, compared with 58.1% for $5/$15 copays and 61.6% for $10/$30 copays. In the low-risk group, considering no difference in LDL-C goal attainment between RSV and SMV, SMV would be cost-effective if the average dose weighted cost is below that of RSV. Discussion Threshold generic simvastatin price as compared to current WAC* LDL-C reduction ATP III LDL-C goal $5 generic/$15 brand 0.453 0.419 $10 generic/$25 brand 0.417 0.408 $10 generic/$30 brand 0.382 0.384 *Generic simvastatin price threshold above which rosuvastatin is preferred at willingness-to-pay = 0. Moderate/high-risk group. WAP, wholesale acquisition cost;atp III,Adult Treatment Panel III; LDL-C, low-density lipoprotein cholesterol; CHD, coronary heart disease. RSV was more effective than other statin monotherapies in lowering LDL-C (-31.6% change) and total cholesterol (-18.9% change) in patients treated in a real-world clinical setting. Importantly, RSV was also more effective in getting moderate/high-risk patients to attain NCEP ATP III LDL-C goal as compared with other statins at an average daily dose lower than other statins. A significantly greater proportion of moderate and high-risk patients using RSV (76%) were estimated to achieve LDL-C goal than those using other statins (57.6 72.6%). RSV provided greater treatment effectiveness at lower treatment costs than ATV for patients in clinical practice in both LDL-C reduction and NCEP goal attainment. Thus, in base case models, RSV dominated ATV, because ATV was less effective and more costly than RSV. The implications of these findings are important for managed care decision-makers. Payers may achieve substantial cost savings and greater effectiveness by using RSV rather than ATV, depending upon the actual acquisition prices. Thus, our findings provide evidence for formulary decisions for the most effective and cost-effective branded statin. For generic statins, ATV and LOV had similar LDL-C reduction, but SMV had significantly more patients attaining LDL-C goal than LOV or PRV. Thus, SMV was the more effective generic agent. Regarding the more cost-effective generic, SMV generic price would need to be discounted by about 61% to 62% of current WAC price for it to be costeffective for payers to utilize as the most effective and least costly generic statin. Furthermore, our findings indicate that comparing the most cost-effective branded (RSV) and generic (SMV) statins, RSV was more cost-effective than generic SMV until SMV WAC price was discounted by at least 55% to 62% for cost/ldl-c reduction, and by at least 58% to 62% for cost/atp III goal attainment among high/moderate-risk patients. The price of generic SMV has fallen as more generic manufacturers enter the market. Assuming that approximately 10 generic manufacturers of SMV are anticipated by the end of 2007, based on an FDA analysis of the association between relative price and number of generic manufacturers [17], the price of generic SMV may fall to 26% of its branded price (or a 74% discount). Anecdotal information indicates that some payers may even be purchasing generic SMV for $0.06 to $0.10 per tablet [18]. At this discount, SMV is less costly but less effective than RSV. This has important implications for physicians and decision-makers, because for patients who need a less effective agent, a generic agent like SMV may be optimal, but patients who need a more effective agent to reduce LDL-C and attain goal, then RSV may provide an attractive option from both an effectiveness and cost-effectiveness viewpoint. ATV will go off patent in several years, and this may result in the availability of ATV at a reduced cost. In that future scenario, using the same methodology as employed for generic SMV, RSV would be more costeffective than off-patent ATV until ATV WAC price was discounted by at least 42% to 56% for cost per LDL-C reduction, and by at least 39% to 49% for cost per LDL-C goal attainment, depending upon copayment level. The cost-effectiveness estimates (branded and generic) were based on aggregate statin doses and aggregated prices across doses. Because some statins have higher prices for higher doses, this was accounted for in the analysis. Thus, health plans interested in drawing comparisons to these estimates may need to examine costs across doses by weighing the cost by dose. Dose-to-dose comparisons were not feasible due to small number of patients with different doses. Several quality of care initiatives utilize the NCEP LDL-C goals, thus it is important that MCOs focus on lipid management to assist patients in attaining their LDL-C goal. The National Committee for Quality Assurance Health Plan Employer Data and Information Set benchmarks the quality of care provided by managed care organizations for patients after an acute cardiovascular event based upon the proportion of

Cost-Effective Generic and Branded Statins 1067 patients achieving NCEP goals [19]. Additionally, several pay-for-performance programs are currently rewarding physicians who treat their patients intensively and get them to NCEP target goals (http:// www.iha.org/p4poview.html, last accessed on January 30, 2006) [20]. This study is unique because it is based on actual clinical practice. The LDL-C reduction and NCEP goal attainment were estimates among patients from different physician offices on their usual care regimen rather than patients selected and randomized in a clinical trial. Our effectiveness estimates represent the actual findings from patients treated in the community; largely by primary care physicians (85% of practices were primary care). Results from our study are similar to investigations that quantified the efficacy [21 23] and cost-effectiveness [11,24] of statin therapy from clinical trial data. RSV has been shown to be the most efficacious monotherapy statin in the clinical trial setting [25], and we found RSV to be the most effective of all monotherapy statins in clinical practice. Similar to our findings, Benner [11] found RSV to be the most cost-effective, and dominated ATV, PRV, and SMV under the base case and several alternative scenarios using data derived from clinical trial data. Bullano showed that a greater percent of patients reached NCEP ATP III LDL-C goal with RSV compared with all other statins for high and moderate-risk patients [26]. Our study confirmed this greater effectiveness of RSV for LDL-C reduction and goal attainment as compared with other statins. The study findings should be interpreted in light of several limitations. The study population is comprised of patients treated by physicians who utilize an electronic medical record system (GEMS) in their routine clinical practice. Although the population includes over 3000 physicians across the United States, there may be differences in clinical practice patterns between physicians who utilize GEMS and physicians who are not electronically equipped. Moreover, the study pharmacy data is the physician prescription order and not pharmacy dispensing data indicating that the prescription was filled. Hence, the assumption was made that all statin prescription orders were filled by the patient. To eliminate bias from this assumption, all patients had to have at least 90 days of statin therapy to be included in the study (i.e., at least three statin prescriptions or one prescription with three refills). Nevertheless, the rate of therapy compliance is unknown and may increase the variance in the lipid values. Also, combination statin therapy was not included in this study because of the study time period (before combination therapy was widely available) and small number of patients prescribed combination therapy. An additional limitation is the lack of inpatient procedure data in the GEMS database. Patients with angioplasty, coronary artery bypass graft, and other revascularization procedures are classified as high-risk patients according to NCEP ATP III guidelines [4]. These patients could not be distinguished in our dataset, thereby reducing the size and composition of our high-risk group. To compensate for the lack of procedure data, the high and moderate-risk groups were combined. Costs were estimated using published WAC for drugs and Medicare reimbursement rates for physician visits and lab tests. These costs may not reflect the actual costs negotiated by managed care organizations, but wide variation in these parameters were tested. Costs associated with adverse drug events, and routine physician visits were excluded, but these were not expected to differ across treatment groups and therefore would not have affected the incremental cost-effectiveness analysis. The analysis had a 1-year time horizon and thus examined short-term endpoints (LDL-C reduction and ATP III goal attainment) instead of long-term outcomes such as cost per life-year gained or cost per avoided CVD event. A longer follow-up period or analytic model would be required to estimate the long-term outcomes. A meta-analysis of 14 randomized trials of statin therapy indicated that statins safely reduced the 5-year incidence of major coronary events, coronary revascularization, and stroke by about one-fifth per mmol/l (39 mg/dl) reduction in LDL-C [27]. This analysis found an approximately linear relationship between the absolute reductions in LDL-C achieved and the proportional reductions in the incidence of coronary and other major vascular events, and those larger LDL-C reductions produced larger reductions in vascular disease risk. These results cannot be directly applied to the current study findings because it is not a randomized trial and there are obvious differences in inclusion criteria and follow-up period between the trials included in the meta-analysis as compared with the current observational study. Nevertheless, one can use the results of the meta-analysis to arrive at a general guidance of the expected benefits in terms of avoided cardiovascular events. Extrapolating to the study s 50.5 mg/dl average absolute reduction in LDL-C for patients with RSV, it could be estimated that continued RSV therapy would reduce major cardiovascular events by one-fourth. Similar estimates can be arrived at for other statins in this study or for differences in LDL-C reduction observed between different statins. In conclusion, the findings indicated that RSV is more effective and less costly than ATV, providing evidence for the branded statin of choice for formulary inclusion in second tier. As generic prices for SMV and PRV stabilize, it is likely that generic SMV will be the most cost-effective generic statin based on the better effectiveness and low cost of generic SMV [18]. To facilitate effective and efficient management of patients with dyslipidemia, a tiered formulary could include generic SMV or PRV in first tier (depending upon level

1068 Ohsfeldt et al. of discount to WAC at multisource generic pricing) and RSV in second tier. Compared with ATV, RSV provides additional clinical benefits (better percentage LDL-C reduction and goal attainment) and also results in cost savings. Compared with generic SMV, RSV offers substantial additional LDL-C reduction and NCEP goal achievement, but at an additional cost. Further understanding of the value that clinicians and payers place on the additional clinical benefits (getting additional percentage LDL-C reduction and additional patients to goal) would determine the economic attractiveness of RSV as compared with generic SMV for management of dyslipidemia in this at-risk patient population. Source of financial support: This research was sponsored by AstraZeneca LP. References 1 American Heart Association. Heart Disease and Stroke Statistics: 2005 Update. Dallas, TX: American Heart Association, 2004. 2 Goetzel RZ, Hawkins K, Ozminkowski RJ, Wang S. The health and productivity cost burden of the top 10 physical and mental health conditions affecting six large US employers in 1999. J Occup Environ Med 2003;45:5 14. 3 American Heart Association. Heart Disease and Stroke Statistics: 2003 Update. Dallas, TX: American Heart Association, 2003. 4 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. 5 Cannon CP, Braunwald E, McCabe CH, et al. Intensive versus moderate lipid lowering with statin after acute coronary syndromes. N Engl J Med 2004; 350:1495 504. 6 Berne C, Siewert-Delle A, URANUS study investigators. Comparison of rosuvastatin and atorvastatin for lipid lowering in patients with type 2 diabetes mellitus: results from the URANUS study. Cardiovasc Diabetol 2005;4:7. 7 Schuster H, Barter PJ, Stender S, et al. Effects of switching statins on achievement of lipid goals: Measuring effective reductions in cholesterol using rosuvastatin therapy (MERCURY I) study. Am Heart J 2004;147:705 13. 8 Quilliam BJ, Perez HE, Andros V, Jones P. Quantifying the effect of applying the NCEP ATP III criteria in a managed care population treated with statin therapy. J Manag Care Pharm 2004;10:244 50. 9 Harley CR, Williams SA, McDonough KL, et al. Cholesterol management in a population of managed care enrollees. J Clin Outcomes Manage 2003;10: 147 54. 10 Straka RJ, Taheri R, Cooper SL, et al. Assessment of hypercholesterolemia control in a managed care organization. Pharmacotherapy 2001;21:818 27. 11 Benner JS, Smith TW, Klingman D, et al. Costeffectiveness of rosuvastatin compared with other statins from a managed care perspective. Value Health 2005;8:618 28. 12 2006 National Physician Fee Schedule Relative Value File. Available from: http://www.cms.hhs.gov/ providers/pufdownload/rvudown.asp [Accessed May 6, 2006]. 13 First DataBank National Drug Data File. Available from http://www.firstdatabank.com/ [Accessed May 6, 2006]. 14 Doubilet P, Begg CB, Weinstein MC, et al. Probabilistic sensitivity analysis using Monte Carlo simulation: a practical approach. Med Decis Making 1985;5:157 77. 15 Briggs AH, O Brien BJ, Blackhouse G. Thinking outside the box: recent advances in the analysis and presentation of uncertainty in cost-effectiveness studies. Ann Rev Public Health 2002;23:377 401. 16 Miller W, Robinson LA, Lawrence RS, eds. Valuing Health for Regulatory Cost-Effectiveness Analysis. Washington, DC: National Academies Press, 2006. 17 IMS Health, IMS National Sales Perspective (TM), 1999 2004, extracted February 2005. 18 Rubenstein S. Why generic doesn t always mean cheap. Wall St J March 13, 2007, p. 1. 19 National Committee for Quality Assurance: HEDIS 2004. Health Plan Employer Data and Information Set. Technical Specifications. Washington, DC: National Committee for Quality Assurance, 2003:374. 20 Gosfield AG. Pay-for-performance (P4P): transitional at best. Manag Care 2005;14:64 9. 21 The Long-Term Intervention with Pravastatin in Ischaemic Disease (LIPID) Study Group. Prevention of cardiovascular events and death with pravastatin in patients with coronary heart disease and a broad range of initial cholesterol levels. N Engl J Med 1998;339:1349 57. 22 Heart Protection Study Collaborative Group. MRC/ BHF Heart Protection Study of cholesterol lowering with simvastatin in 20,536 high-risk individuals: a randomized placebo-controlled trial. Lancet 2002; 360:7 22. 23 McKenney JM, Jones PH, Adamczyk MA, et al. Comparison of the efficacy of rosuvastatin versus atorvastatin, simvastatin, and pravastatin in achieving lipid goals: results from the STELLAR trial. Curr Med Res Opin 2003;19:689 98. 24 Morrison A, Glassberg H. Determinants of the costeffectiveness of statins. J Manag Care Pharm 2003; 9:544 51. 25 Jones PH, Davidson MH, Stein EA, et al. Comparison of the efficacy and safety of rosuvastatin versus atorvastatin, simvastatin, and pravastatin across doses (STELLAR trial). Am J Cardiol 2003;93:152 60. 26 Bullano MF, Wertz DA, Yang GW, et al. Effect of rosuvastatin compared with other statins on lipid

Cost-Effective Generic and Branded Statins 1069 levels and National Cholesterol Education Program goal attainment for low-density lipoprotein cholesterol in a usual care setting. Pharmacotherapy 2006;26:469 78. 27 Baigent C, Keech A, Kearney PM, Cholesterol Treatment Trialists (CTT) collaborators. Efficacy and safety of cholesterol-lowering treatment. prospective meta-analysis of data from 90,056 participants in 14 randomised trials of statins. Lancet 2005;366:1267 78.