1

Innovative Approaches to Coverage using CER

Methods and Applications for

Comparative Effectiveness Research

ITHS InstituteITHS Institute

Sept. 28, 2010

Lou Garrison, Ph.D., PORPP, University of Washington

Josh Carlson, Ph.D., PORPP, University of Washington

1

Acknowledgments

• Collaborators:  Sean Sullivan, Adrian Towse, David Veenstra, Peter Neumann, Rick Carlson

• Sponsors:

– Novartis

– Johnson & Johnson

– GSKGSK

– GE Healthcare

– Roche

– Pfizer

– Eli Lilly

– Sanofi‐Aventis

– Abbott Laboratories

– Amgen

2

Session Objectives

1. Discuss the relationship between CER and coverage decisions made by public and private payers

2. Present a taxonomy and examples for understanding how i CER i h d i ipayers are using CER in the coverage decision process, 

especially CMS in the US

3. Present two UK cases illustrating the use of CER in coverage and pricing decisions 

3

2

Introduction

4

What is CER?

According to Federal Coordinating Council (2009):

Comparative effectiveness research is the conduct and synthesis of research comparing the benefits and of different interventions and strategies to prevent, diagnose, treat and monitor health conditions in ‘real world’ settings. The purpose of this research is to improve health outcomes by developing and disseminating evidence‐based information to patients, clinicians, and other decision‐makers, responding to their expressed needs, about which interventions are most effective for which patients under specific circumstances.’’

Three key points:

• ‘Comparative’ means that a new intervention should be compared with the most common alternative, i.e. typically the existing ‘standard of care’. 

• ‘Effectiveness’ is to connote a focus on real‐world outcomes as opposed to the ‘efficacy’ outcomes demonstrated in experimental clinical settings. 

• The overall aim is to inform a wide range of decision makers.

5

What is coverage?

• Coverage—whether a particular medical service, product, procedure, etc. will be reimbursed or provided to the members of a health insurance plan

• Reimbursement—The extent to which cost of covered service is reimbursed.reimbursed.

Coverage can be on a continuum

6

3

Some Important Economic Terms and Conceptsfor Thinking About CER

• “Market failure”—when “free markets” do not provide an “optimal” allocation of resources—often when the conditions for a free market are not met.– For example, public goods, externalities, informational asymmetry, uncertainty, 

etc. —patent protection to incentivize investment and risk‐taking 

• “Public good”—a good for which one person’s usage or consumption does not keep other from using it, e.g., national defense.other from using it, e.g., national defense.– The free market can be expected to undersupply public goods.– We address this with interventions, such as “intellectual property” and public 

subsidies.– There can be a “free rider” problem.

• Information and scientific advances can be public goods—even “global public goods”– The whole world has a stake in innovative pharmaceutical R&D.– Economists agree that “differential pricing” would be an improvement:  the 

challenge is how to implement it.

7

8

Why the interest in performance‐based risk‐sharing? 

• Health care production is complex:Economists think of it as a “Health Production Function”:

– Output = f(Inputs)

– Health = H(hospital stays, doctors visits, drugs, OTHER)

Th h i i t th “ ti t’ t” i i i• The physician acts as the “patient’s agent” in organizing and advising on this process.

• Historically, either (1) all of these inputs were provided by a system of care or (2) each input is purchased on a fee‐for‐service basis—or some mix of these.

But the process is so complex that no one has offered guarantees of good outcomes. 

9

4

10

Greater interest in “pay for performance”

1. In the U.S, for health plan processes and for physician services.

2. Internationally, for branded drugs.

11

Why the new interest? What has changed?

Bioclinical Health Outcomes Framework:  Which Outcomes to Measure?

Improvements in Surrogate Co‐Morbidities:

•Glucose tolerance

Weight Loss

Long‐Term Improved Clinical Outcomes:

•Cardiovascular/

Better Health Outcomes:

•Length of Life

•Cholesterol

•Blood Pressure

•CerebrovascularEvents

•Quality of Life

Example: Obesity Disease-Treatment Model

12

5

Basics:  The Pervasiveness of Uncertainty

• Drugs are approved, launched, and reimbursed under conditions of uncertainty, affecting many key parameters:

– Efficacy (heterogeneity)

– Effectiveness in real world

– Risks (safety)

– Models, including links between surrogate markers and long‐term outcomes

– Cost‐effectiveness

– Budget impact.

1. VariabilityUncertainty (=Risk)

2. Gathering more evidence to reduce uncertainty is costly.

13

Definitions

• Performance-based agreement:– An agreement between a payer and a pharmaceutical, device, 

or diagnostic manufacturer where the coverage, price, and/or revenue received is related to the future performance of the product in either a research or real‐world environment.

• Risk-sharing agreement (de Pouvourville, 2006):g g ( )– “a contract between two parties who agree to engage in a

transaction in which there are uncertainties concerning its final value. Nevertheless, one party has sufficient confidence in its claims of either effectiveness or efficiency that it is ready to accept a reward or a penalty depending on the observed performance.” (emphasis added).

We use these terms interchangeably.14

The Historical Risk‐Sharing “Equilibrium”

• Risk to manufacturer: we operate with a blockbuster financing model for R&D.– Intellectual property—patent protection to incentivize investment and risk‐

taking – There is no ex ante clause to share innovation cost or to purchase drugs. 

• Risk to payer: The payer negotiates a price and/or use.  The payer bears the risks of making a bad buy (i e when incremental health– The payer bears the risks of making a bad buy (i.e., when incremental health benefits are not worth the additional cost). 

– The payer is free to collect post‐launch data.  Manufacturers  will only do this if it is in their competitive interests. 

• Individual countries strike different types of deals with manufacturers– Range of country environments:   negotiated prices < ‐‐ > free pricing– All of this provides an incentive for manufacturers to seek highest justifiable 

price at launch. Manufacturers would like to price for future (larger) indications.

15

6

Key Elements of Performance‐Based Risk‐Sharing Arrangements

1. There is an agreement about a program of data collection to reduce uncertainty about the expected cost‐effectiveness of the drug (or device or diagnostic). 

2. The coverage, price, and/or revenue is linked to the outcome of this program of data collection This may be prospective orthis program of data collection. This may be prospective or retrospective. 

3. It can be about health outcomes and cost‐effectiveness or about budgets. 

4. These arrangements provide a different distribution of risk as between the payer and the manufacturer than “conventional” arrangements*. 

* de Pouvourville EJHE, 2006  

16

Why the “sudden” interest?Fad or trend? 

• Understanding the cause should be helpful to predict long‐term adoption and impact.

• Two general explanations:

1) Innovation—it’s a new invention.1)  Innovation it s a new invention.

2)  Environment has changed.

• Some “trends” may be fads:

– Example:  disease management (carve‐outs)

• Did not fulfill original promise.  Why?

17

Cost Pressures

• Increasing drug expenditures attributed to:

– Use of high cost biopharmaceuticals for common, chronic conditions (RA, Asthma, Psoriasis, etc)

– Expensive, combination biopharmaceutical treatments in l d f doncology and infectious disease

– Prescribing beyond evidence and approved indications

– Other factors (aging population, fewer resources, etc)

Real cost of new branded drugs is rising—fewer approvals and higher R&D spending.

18

7

19

Payer Response to Increasing Cost Pressures

• Public and private sector payers are facing these challenges with various cost‐control instruments and management strategies:

– Increasing patient co‐payments

– Pre‐use authorization (targeting appropriate patients and appropriate use)

– Quantity and dose limitations

– Specialty pharmacy vendors

– Benefit restrictions (e.g. generic‐only benefits)

– Denial of coverage

Working hypothesis:  Performance‐based agreements are a market response to increasing cost pressures: manufacturers have incentive (esp. in US) to push prices to limit of willingness to pay, and payers are pushing back.

20

Other Related Developments and Discussions

• Value‐based pricing and reimbursement (UK Office of Fair Trading)

• Real‐world data (ISPOR Task Force, Netherlands)

• Drug safety and benefit‐risk assessment (FDA; EMEA)

• Conditional licensing (MHRA, Cooksey Review)

• Personalized medicine/tailored treatments

• Comparative effectiveness research 

• Financing R&D and fostering innovation

21

8

Two Analytic Approaches

• Empirical Practice (inductive)

– What examples have arisen in the real world?  What do their characteristics tell us about potential factors?

• Economic Theory (deductive)

– What does theory say about the historical risk‐sharing between manufacturers and payers?  Why might this have changed?

22

Coverage with Evidence Development in Landscape of Performance‐Based 

Arrangements:  Taxonomy and Review of Cases

23

Agenda

• Review of performance‐based arrangements

– Taxonomy of arrangement types

– Characterization of arrangements to date

– Observations and patterns

• Examples of Coverage with Evidence Development at CMS

24

9

Review of Performance‐Based Arrangements: Methods

• Sources: 

– PubMed, Scrip, Embase, and Google

– Experts and peers

• 20 year time frame

• Included: Health outcomes based agreements: price, level, or nature of reimbursement are tied to measures ultimately related to patient quality or quantity of life. 

• Excluded: non‐outcomes based models including price volume agreements, market share agreements, utilization caps 

• Develop a taxonomy of agreements:

– Inductive approach

– Refined using a modified Delphi approach with experts in the area.

25

6

8

10

12

of

Ag

ree

me

nts

Performance‐based schemes by year

> 50 schemes

0

2

4

Nu

m. o

Year

CED CTC PLR FU

26

CED: Coverage with evidence development; CTC: Conditional treatment continuation; PLR: Performance linked reimbursement; FU: Financial or utilization based agreements

Health outcomes‐based schemesNon‐outcomes based schemes 

Performance‐linked reimbursement (PLR)

Population level  Patient level  Conditional coverage

Performance‐based schemes between health care payers and manufacturers

Clinical Endpoint

[Ex: Bortezomib in UK]

Intermediate Endpoint

[Ex: Simvastatin in US]

Pattern or process of care

[Ex: OncotypeDx in US (United Healthcare)]

Only in research

[Ex: Cochlear implants 

in US (CMS)]

Only with research 

[Ex: Risperidone in France]

Market share

Manufacturer funded treatment 

initiation

Outcomes guarantee 

Price volume

Utilization caps

Coverage with evidence 

development (CED)

Conditional treatment continuation (CTC)

[Ex: Alzheimer’s drugs in Italy]

27

10

Taxonomy Definitions:  Conditional Treatment Continuation

• Conditional treatment continuation : continuation of coverage for individual patients is conditioned upon meeting short‐term treatment goals

– Example: Alzheimer’s drugs provided free by manufacturer in Italy.  If treatment goals are met after 3 months, treatment is continued with drug costs reimbursed by national health service. 

• Conditioning continued coverage on short‐term treatment goals helps ensure that only patients that benefit from treatment remain on treatment. 

• Key component: availability of a reliable measure of short‐term response, benefit, and/or continued benefit (e.g. tumor response) that is acceptable to both parties. 

28

Taxonomy Definitions:  Performance‐Linked Reimbursement

• Performance‐linked reimbursement: reimbursement level for covered products is tied, by formula, to the measure of clinical outcomes in the “real world”; 

– Outcomes guarantees: manufacturer provides rebates, refunds, or price adjustments if their product fails to meet the agreed upon outcome targets

• Example: J & J agreed to reimburse the NHS in either cash or product for patients who do not respond (Response measure: 50% decrease in serum p p ( pM protein) after 4 cycles of treatment with Velcade.  Responding patients receive additional 4 cycles.

• PLR arrangements are likely to arise when a manufacturer has sufficient confidence in their product and their claims about its attributes such that they are willing to accept a lower reimbursement level if it underperforms.  

– This may be especially relevant in areas with multiple competing products.  

29

Performance‐Linked Reimbursement

• Manufacturers develop value story: effectiveness, safety, QoL, adherence, etc.

– Payers may desire more evidence to support manufacturer’s claimsexpensive.

• Direct costs 

• Lost revenues due to delays in market access.  

Manufacturers may prefer to provide payers a guarantee for certain– Manufacturers may prefer to provide payers a guarantee for certain outcomes in place of additional product research.  

• Alternative mechanism to provide discounts without changing list prices.  

– For example, the list price for a 3.5 mg vial of Velcade is £760 in the U.K.

– With a rebate for patients that do not respond to treatment after four cycles (approximately 40% did not respond in the pivotal trial), the effective price paid during the first four cycles by the NHS is closer to £300 per vial—yet the list price remains the same

30

11

Taxonomy Definitions:  Conditional Coverage‐‐CED

• Conditional coverage: coverage is granted conditional on the initiation of a program of data collection

– Coverage with evidence development:  coverage is conditioned on collection of additional population level evidence, from pre‐specified study, to support continued, expanded, or withdrawal of coverage

• Only in research: coverage conditional on individual participation in y g p presearch (i.e. only patients participating in the scientific study are covered)

– Example: CMS may cover cochlear implantation for treatment of hearing loss when the provider is participating in, and patients are enrolled in, an approved clinical trial

• Only with research: coverage conditional on agreement to conduct a study that informs the use of the medical product in the payer patient population

– Example: France's ministry of health agreed to cover Risperidone if J&J performed studies to evaluate whether it helps patients stay on their medications.  31

Coverage with Evidence Development

• Creates a middle ground between coverage and no coverage for products that appear promising, but may not yet have the evidence base to support full coverage. 

• Five countries were found to have implemented a total of 34 CED schemes.  

– Sweden with 15 (44%), 

– U.K. with 10 (29%), 

– U.S. with 7 (21%)

– France with 1 (3%) 

– Australia with 1 (3%)  

• Only in research:

– US and UK (with the exception of the MS scheme in the U.K.) applied to devices and the off‐label use of pharmaceuticals.  

• Acceptable research participation: approved clinical trial or patient registry.  

• Only with research:

– Sweden, Australia, and France

• Data collection related to long‐term effectiveness, the products’ use and impact in actual clinical practice, and the products’ cost‐effectiveness.

32

CED at CMS

• On July 12, 2006 CMS released a guidance document titled National Coverage Determinations with Data Collection as a Condition of Coverage: Coverage with Evidence Development (CED). 

– Circumstances under which CMS would issue a national coverage determination (NCD) requiring, as a condition of coverage, collection of additional patient data to supplement standard claims data.additional patient data to supplement standard claims data.

• The purpose of CED is to generate data on the utilization and impact of the item or service evaluated in the NCD, so that Medicare can 

– Document the appropriateness of use of that item or service in Medicare beneficiaries under current coverage; 

– Consider future changes in coverage for the item or service; 

– Generate clinical information that will improve the evidence base on which providers base their recommendations to Medicare beneficiaries

33https://www.cms.gov/CoverageGenInfo/03_CED.asp

12

CED process

• CMS makes NCDs based on requests for national coverage from individuals or entities that identify an item or service as a potential benefit (or to prevent potential harm) to Medicare beneficiaries. 

– Medicare beneficiary, 

– manufacturer, 

– provider, 

– supplier, 

– medical professional association, 

– health plan, or 

• CMS may also initiate an NCD internally.

34https://www.cms.gov/mcd/ncpc_view_document.asp?id=8

CED process: What is covered?

• Coverage with Study Participation (CSP)

– Approved Registries

• CMS will pay for the clinical costs of patient care for Medicare beneficiaries for which data collection is required. 

• CMS will not provide financial support for registry development and maintenance.

– Approved Research studies

• CMS may pay for the investigational and or routine clinical costs of an item or service for a trial in an NCD requiring CED as described in the revision of the Clinical Trial Policy. 

• CMS will not fund the non‐clinical research costs.

35https://www.cms.gov/mcd/ncpc_view_document.asp?id=8

Active CMS CED decisionsNational Coverage Determination CED

1. Cochlear Implantation None scheduled at present

2. Chemotherapy for Colorectal Cancer A. National Cancer Institute: Clinical Trials Covered Under the Medicare Anti‐Cancer Drug National Coverage Decision

3. PET (FDG) for Brain, Cervical, Ovarian, Pancreatic, Small Cell Lung, and Testicular Cancers

B. National Oncologic PET Registry

4. Implantable Cardioverter Defibrillators C. American College of Cardiology‐National Cardiovascular Data Registry

5. PET (FDG) for Dementia and Neurodegenerative Diseases

D. Medicare‐Approved PET for Dementia Trial

36https://www.cms.gov/CoverageGenInfo/03_CED.asp

13

CMS: Decision Memorandum for Pharmacogenomic Testing to Predict Warfarin Responsiveness

• Available evidence does not demonstrate that pharmacogenomic testing of CYP2C9 or VKORC1 alleles to predict warfarin responsiveness improves health outcomes in Medicare beneficiaries. 

– Not reasonable and necessary under §1862(a)(1)(A) of SSA • Approved for CED: 

– Covered only when provided to Medicare beneficiaries who are enrolled in a y pprospective, randomized, controlled clinical study

• Clinical study must address specific research question: 

– What is the frequency and severity of the following outcomes, compared to subjects whose warfarin therapy management does not include pharmacogenomic testing?

• Major/minor hemorrhage

• Thromboembolisms

• Mortality

• What’s covered?

– Cost of test, routine, and non routine (protocol‐driven) medical care costs

37http://www.cms.gov/mcd/viewtrackingsheet.asp?from2=viewtrackingsheet.asp&id=224&

UK Case Examples

38

UK Patient Access Schemes (PPRS Section 6)

• Patient Access Schemes will facilitate earlier patient access for medicines that are not in the first instance found to be cost and clinically effective by NICE– This is a “fast track” route back to NICE– Schemes can be financially based or outcome based.

14

UK Patient Access Schemes

Expected value: rebate

Proven value: price increase

OutcomesRisk-share

Company seeks

agreement to a price subject to extra evidence

Financially Based

Arrangements

Company seeks agreement to a later increase in price subject to a re-review of

Outcomes are measured, and price adjusted

or cash transfers made

Company offersdiscount/rebatelinked to (eg):• patient no.s

Outcomes-Based Arrangements

as agreed with NICE and a

rebate/reduced price if

evidence is adverse

the drug using extra evidence as agreed by

NICE.

given outcomes

anticipated in the terms of the scheme

patient no.s• patient type• response• dose no.s

Process: Use in advance of or after NICE appraisal at launchLimit use after ACD to avoid gamingPost NICE resubmission is confidential until recommended by NICECost per QALY disclosed but not effective price

Example 1: Velcade for Multiple Myeloma in UK

• Efficacy/Effectiveness:

– Single phase III trial:  Strong efficacy (surrogate endpoints: TTP and response).  Trial stopped early.

– Limited data on overall survival: Secondary endpoint, trial stopped early.

• Budget impact/patient population considerations: 

h l– High cost:  Cost per cycle, £3,000

– Variable treatment duration—treat until tumor progression

• Value: Not cost‐effective at 1st submission (£35,000/QALY, substantial uncertainty).

• Market factors: 

– First‐in‐class, no good alternatives at 1st relapse

– Public payer, market access based on cost‐effectiveness

41

Velcade for Multiple Myeloma in UK

• Performance‐based scheme development: 

– Resubmission to NICE with performance‐based scheme

– Conditional treatment continuation (stopping rule after 4 cycles) and outcomes guarantee (rebate for non‐

d )responders) 

– ICER with rebate, stopping rule:  £20,700/QALY  NICE approval

• Handling uncertainty: Mitigate negative consequences of uncertainty about value

42

15

VELCADE® Response Scheme (VRS) for patients with Multiple Myeloma at 1st relapse within the NHS in England, Wales and N.Ireland- Process Flow

Janssen-Cilag/Ortho BiotechCustomer

JC Drug SafetyVRS Administrator

Orders VELCADE

Signs VRS agreement with JC

VRS claim forms are sent to Customer upon receipt of signed VRS agreement

Hospital completes VRS Drug Safety allocates a

Evaluate responseafter a maximum of 4 cycles (16 vials)Response Y/N

Yes1

No2

Claim validated3

Y/NOrder further vials of VELCADE

Replacement product delivered or credit note/cash refund issued to customer

Yes

No JC advises and discusses with customer

Instruction to despatch replacement stock, issue credit note or full cash refund

VRS claim logged (Audit and Healthcare Compliance)

completes VRS claim form for replacement stock, credit or full cash refund for 16 vials. Fax/Post to JC

VRS Claim Form copied to JC Drug Safety Dept.

Drug Safety allocates a specific UK/DS Number and writes to clinician requesting more details on the non-response event

Notes:

1 Response defined as a patient, at first relapse, with a 50% or greater reduction in serum M-protein, within 4 cycles of treatment, compared to baseline level immediately prior to VELCADE treatment

2 Non-response or minor response defined as a patient, at first relapse, having less than a 50% reduction in serum M-protein, within 4 cycles of treatment, compared to baseline level immediately prior to VELCADE treatment

NB: 15% - 20% of patients do not have measurable serum M-protein levels, and for this group Bence-Jones urine protein (urine free light chains) could be used. Response defined according to the standard EBMT (European Group for Blood and Marrow Transplantation) criteria, i.e. at least a 90% reduction, compared to baseline. This would be compared to baseline immediately prior to VELCADE treatment (and within 4 cycles).

3 Claim Validation to check:

• VRS claim form completed in full and signed

43

Source: Anne Hines, CNPF, NHS

Example 2:  Multiple Sclerosis drugs in the UK

• Effectiveness: 

– Clinical trials have shown that all three interferon products reduce relapse frequency and severity in patients with relapse remitting MS and may also influence duration of relapse. 

• The reduction in frequency amounts to about 30% on average, 

• Equivalent to one relapse avoided every 2.5 years 

• This reduction has been demonstrated for the first 2 years of therapy only.

No evidence on the long term progression of patients after cessation of therapy– No evidence on the long‐term progression of patients after cessation of therapy.

• Budget impact/patient population:

– It is estimated that in England and Wales MS affects some 63,000 people.

– The current annual cost per patient for MS drugs in UK is:

• Betaferon: £7,259 

• Avonex: £9,061 

• Rebif: £9,088/£12,068 (low dose/higher dose)

• Glatiramer acetate is £6,650

– Costs of treatment in a full year could be in the order of £50 million

– Potential to continue treatment despite lack of benefit  no good alternatives

45

16

Multiple Sclerosis Drugs in the UK (cont.)

• Cost‐effectiveness 

– NICE reviewed MS drugs for relapsing‐remitting multiple sclerosis in 2001

• 5 year model:  £380,000 to £780,000/QALY

• 10 year model:  £190,000 to £425,000

• 20 year model: £40 000 to £90 000• 20 year model:  £40,000 to £90,000

• “On the balance of costs and benefits, the beta interferons and glatiramer acetate are not cost‐effective”

– Dependant on long‐term outcomes:  Only 2‐years of follow up data

• Used assumptions and historical control to estimate long‐term outcomes

– Estimates sensitive to the impact of a relapse on quality of life and the time horizon over which benefits may be accrued.

46

Multiple Sclerosis Drugs in the UK (cont.)

• Performance‐based scheme development

– Detailed monitoring over 10 years of a cohort of patients to confirm the cost‐effectiveness of the MS treatments:  Avonex, Betaferon, Copaxone and Rebif

• Treatments initiated by specialist MS centres based on ABN guidelines

– No bar to clinicians prescribing for patients outside these guidelines

• Outcome measure is Expanded Disability Status Score. 

• Actual outcomes reviewed every 2 years against standard MS disease (non‐treated) progression model through the EDSS scale.

– Price adjustment to maintain CE threshold at £36,000/QALY

• PLR, Outcomes guarantee

• Handling uncertainty: 

– Resolve residual uncertainty related to long‐term outcomes

– Mitigate negative consequences of uncertainty about value

Health Services Circular 2002/004 http://www.doh.gov.uk/publications/coinh.html 47

UK Multiple Sclerosis Risk Sharing Scheme:

Test Case for Value‐Based Pricing and Reimbursement?

17

Conclusion

49

Performance‐based schemes by country

10

15

20

25

of

Ag

ree

me

nts

50

0

5

10

France Canada Italy Germany Australia US Sweden UK

Nu

mb

er

Country

CED CTC PLR FU

CED: Coverage with evidence development; CTC: Conditional treatment continuation; PLR: Performance linked reimbursement; FU: Financial or utilization based agreements

Coverage with Evidence Development Summary

• Majority of CED in the U.S. involved devices.  

– Less well developed evidence profile than for drugs.  

– Examples identified are part of the CMS CED initiative

• Majority of pharmaceuticals for which CMS pays (except injectables) are Medicare Part D negotiated by individual Part D contracting insurersMedicare Part D negotiated by individual Part D contracting insurers.  

• Outside the US, CED focused on pharmaceuticals, majority found in EU.  

– Increased willingness in the EU, relative to the US, to limit the use of pharmaceuticals in patients and patient subgroups for which the benefits and cost‐effectiveness remain questionable.

51

18

Thanks!  Questions?

lgarrisn@uw.edu

atowse@ohe.org

carlsojj@uw.edu

52

53

54