health economics for medical devices - a very good read

8/6/2019 Health Economics for Medical Devices - A Very Good Read

1/7

T

alk about a crisis in U.S.healthcarea crisis of cost,access, and qualityiswidespread and common.

In such an environment ofconcern, it is incumbent upon med-ical technology companies to con-centrate their efforts on developingproducts that will improve the qual-ity and value of healthcare. Medtechmanufacturers will, of course, haveto provide evidence of this benefit topayers. Success in doing so will belikely to expedite reimbursement,which in turn can increase the rate ofadoption of innovative products.Savvy manufacturers, those longattuned to the need for includingclinical outcomes data in regulatorysubmissions,now provide health eco-nomic data for payers too.

Health economics is the study ofhow resources are allocated to andwithin the health-related economy.Conducting health economic analysishelps medtech companies determinewhere the optimal value of a device

might be foundfor example, with-in which patient subgroups or caresettingsor where the product fits

within the disease-treatment contin-uum. This discipline is increasinglyintegrated into the reimbursementprocess.

Reimbursement comprises theareas of coverage, coding, and pay-

ment. Health economics is becomingmore important particularly for thearea of coverage. Coverage is a con-

cept encompassing the process andcriteria used to determine whether aproduct, service, or procedure will bepaid for through a government orprivate insurance program.

This article discusses the ques-

tion of how the value of new medicaltechnologies can be measured anddemonstrated. The full subject

Demonstrating Valueto PayersWhen making coverage decisions, payers want data thatdemonstrate the value of new medical technologies.Melissa Martinson and Rosemary Brekke

Illustrationb

y

JeniferTracy

Business Strategies for Medical Technology Executives


2/7

B U S I N E S S P L A N N I N G & T E C H N O L O G Y D E V E L O P M E N T

includes consideration of how devel-opers and manufacturers of newmedical technologies can garner the

appropriate information to use insuch a demonstration.Also, the arti-cle examines the role that health eco-nomics plays in decision making withrespect to coverage and payment.

Establishing Valuein Healthcare

Value in the purchase of health-care is no different from what a buyer

seeks in the purchase of any othergoods or services. Good value is get-ting much relative to the amountpaid. Maximizing the value of health-care ensures that, regardless of thelevel of funding for healthcareinvolved5% of the budget or50%the care that is delivered pro-vides the most benefit for the money.

It may be emotionally difficultfor some people to talk about health-

care in financial terms, but that focusis necessary in order to quantifyvalue. In healthcare, the term usual-ly employed to signify value is cost-effectiveness, and a measure of thishas been the cost-effectiveness ratio,that is, the dollar cost of a unit ofeffectiveness (see sidebar).

Healthcare effectiveness, ofcourse, can be manifested in a lower-ing of blood pressure, an increase in

joint mobility, a reduction in pain

level, or in other types of physiolog-ical improvement as different asapples and oranges. A commondenominator is needed to make pos-sible value comparisons of healthcareservices with different beneficial out-comes. The usual denominator forthis purpose is the quality-adjustedlife year (QALY), a measure that com-bines quality of life with quantity oflife. Quality of life is measured on a

scale of 0 (death) to 1 (perfecthealth). One year in perfect health

yields 1 QALY. When a health state is

assigned a quality of 0.5, it is becausepeople have been judged to value two

years living in that state only as muchas one year in perfect health (0.5 2= 1 QALY).

This academic-seeming system isused in various places around theworld. The British National HealthService (NHS) has a research arm, theNational Institute for Clinical Excel-lence (NICE), which makes explicitcoverage decisions based on such cri-teria. Currently, NICE uses a cutoff

of approximately $53,000 per QALYbecause, given the NHS budget andthe volume of services demanded,

this is effectively what the countryhas chosen to pay for.1

The recent implantable car-dioverter-defibrillator (ICD) cover-age decisions by the Centers forMedicare and Medicaid Services(CMS; Baltimore) offer a U.S. com-parison. The value of this technologyis about $51,000 per QALY, accordingto a Blue Cross Blue Shield Technol-ogy Evaluation Center (TEC) assess-ment.2 This figure applies to a patientfor whom the probability of arrhyth-

mic death is 10% with conventionaltherapy; other types of patients arecalculated to cost significantly more.

Quantifying the Healthcare Value of New Technology

One technique that has been employed to determine the economic valueof a medical therapy is the calculation of its cost-effectiveness ratio, whichis the dollar cost of a defined unit of effectiveness.

For example, if joint-repair surgery costs $20,000 and improves pain-free mobility in that joint by 20 of motion, then the cost-effectiveness ratiois $20,000 divided by 20, or $1000 per degree of additional pain-freemovement. The chief alternative, joint-replacement surgery, might cost$45,000 and improve pain-free mobility by 70. The incremental cost-effectiveness ratio for this therapy is calculated by dividing the extra$25,000 in cost by the 50 of greater freedom of movement that results.The additional range of mobility costs $500 per degree. Thus, the more

expensive procedure is the better buy in terms of cost per unit of improve-ment in outcome.Most new technologies provide a less clear-cut benefit than in this

example because the incremental cost for each additional unit of effec-tiveness is greater than that of the cheaper existing technology. If the50 mobility improvement with joint replacement costs an additional$75,000, then each additional degree of pain-free motion would have a$1500 price tag.

When the healthcare budget is fixed, choosing among therapies is astraightforward matter with this means of measuring value. If there isenough money in the budget to pay $1500 each for additional degrees ina range of painless motion, then the more expensive therapy can be pro-vided. If not, then either more money must be budgeted for healthcare orcoverage must be limited to the less expensive therapy.

However, there are problems with this method. Not all healthcarebenefits are measured in terms of range of motion. There has to be a wayto compare the value of a 10-mmHg reduction in blood pressure with thevalue of a 20 improvement in mobility range. One good solution is thequality-adjusted life year (QALY), a measure of healthcare benefit used insome industrialized nations to change apples and oranges into fruit, eachtype of which bears a QALY label that allows meaningful comparison.


3/7


There is much discussion in the UKabout raising revenues for the NHSso that the cutoff cost-effectiveness

ratio can be higher. This would allowmore types of patients to be treatedwith ICDs and other technologies.

The degree of rationality broughtto the making of coverage decisionsin Great Britain, Europe,and Canadahas been impossible for the UnitedStates so far. CMS has tried twice inthe past decade to define coveragecriteria, without success. A top offi-cial of the agencys coverage division

believes that cost-effectiveness hasbeen one of the most contentiousissues thwarting these efforts, andclaims that competing interestsamong all of the stakeholdersCMS,Medicare beneficiaries, physicians,hospitals, product manufacturers,and othershave made it impossi-ble to reach consensus.3

CMS has been inching toward adefinition,however. In a recent guid-

ance document, the agency says thatit may issue a coverage decision withthe qualification that more data becollected. This is the language it willuse in such a case: Assessment ofimportant outcomes has not beenevaluated in the available clinicalstudies.These outcomes may include,but are not restricted to, long-termrisks and benefits, quality of life, uti-lization, costs, and other real-worldoutcomes.4 Note the inclusion of

quality of life, costs, and utilizationamong the important outcomes.

Collecting HealthEconomic Data

Until now, CMS has coveredsome services without conditions.However the agency has stated thatit does not intend to continue thispractice. For medtech companies to

ensure that their products receive thebroadest coverage possible, they willhave to provide evidence of the value

of those products in the form ofhealth economic data. Health eco-nomic data are any data pertaining

to the resourceslabor,supplies andmaterials, facilities, money, and soonused in producing healthcare.

The Piggybacked Clinical Trial.

Medtech manufacturers have some-times been reluctant to gather healtheconomic data because of the costthat this adds to product develop-ment.An economic study can be justas expensive as a clinical trial con-ducted to achieve FDA approval.

However, with a little planning, it is

usually possible to piggyback the eco-nomic study onto the clinical trial foronly modest additional cost. Piggy-backing an economic study may add1025% to the clinical trial expense,while a stand-alone economic studycould cost $500,000 to $1 million oreven more.

A medical device manufacturer

will need to make five major deci-sions in designing a piggybacked clin-ical trial. These involve considera-tions of:

Whose perspective the cost-effectiveness analysis shouldassume.

Which comparison groupsshould be included in the trial.

The time horizon of the new

medical technology in terms ofrealization of benefits.

How much detail to pursue in

collecting cost data. What healthcare effectiveness

data to collect, and when to col-

lect it.

Analytical Perspective. Whosecosts and whose benefits should beincluded in the cost-effectivenessanalysis is the first question to bedecided. Health economists oftenrecommend adopting a societal per-spective.5 This entails estimating thecosts borne by everyone affected,from patients through employers

(who lose the productivity of theemployee who is a patient).

Because of the way the U.S.healthcare system is segmented,pay-ers have an incentive to control theirown costs rather than societys. Pay-ers thus are generally more interestedin analyses undertaken from theirown perspective, not a societal per-spective. If the payer is concernedonly with clinical outcomes, then the

trial sponsor (the manufacturer)should collect data relating only todirect medical coststhat is, thoseborne by the healthcare payer. If thepayer is a self-insured large employ-er or workers compensation system,then back-to-work data may becomean important factor.

Following the relevant payersperspective marks a path towarddetermining which economic data toinclude in a study. Data gathering

may still be a challenge, but thiscourse of action is much simpler thantrying to collect all costs.

Comparison Groups. WhileFDA frequently approves clinicalstudies with a placebo or inactivecontrol, CMS is interested in trialsthat include comparisons to the bestavailable therapy or standard of care.This usually means that any controlgroup used in the study must repre-

sent active therapy. FDA generallywill approve trials of this design aswell. The sponsor should be aware

Value in the purchase

of healthcare is no

different from what a

buyer seeks in the

purchase of any other

goods or services

getting much relative to

the amount paid.


4/7

that, whereas comparison with aplacebo requires a demonstration ofsuperior efficacy, a comparison withactive controls will often require only

a demonstration of outcome equiva-lence. The statistical test is different inthe latter case, as those familiar withbioequivalence tests in drug trialsalready know.

Time Horizon. Many new med-ica l technologies , e spec ia lly implantable devices, require consid-erable up-front expenditure.Benefitsfrom the technology, such asimproved quality of life or future

healthcare cost savings, are realizedlater. The trial sponsor should deter-mine how long these benefits willhave to accumulate to justify the up-front costs. Sometimes this will belonger than the duration of a tradi-tional clinical trial. However,becauseFDA often requires postmarketingstudies that are really a continuationof the original trial, the longer timehorizon will not add substantially to

a clinical trials cost in many cases.Cost-Data Detail. If a new ther-apy is expected to reduce hospitalreadmissions substantially, and if it isa credible assumption that outpatientcosts, pharmacy costs, and otherpostrelease costs will not be increased,then hospital readmission rate andhospital charge data may be all thetrial sponsor has to collect. Hospitalsare sometimes reluctant to providethis information; therefore, it is a

good idea for the sponsor to include astipulation in the study contract.Hos-pital charges do not reflect hospital costsvery closely, however, so it is usually alsoa good idea to get the hospitals CMScost-to-charge ratio as well.

On the other hand, if outpatientcosts are important, extra care mayneed to be taken in designing thestudy. This is because follow-up costsare often part of the structure of the

trial. There are two commonapproaches to calculating these costs.One is the utilization approach: vis-

its, special procedures, prescriptiondrugs, and lab tests are simply count-ed, with government and commer-cial databases then being used to pro-

vide a typical cost per unit ofhealthcare delivered. The other is theaccounting approach, in which thekind and amount of each type ofhealthcare labor, material, supply,facility, and so on, used during thetrial is tallied. Here again, the cost ofeach unit is determined from health-care payer prices. The units involvedin the accounting approach are muchsmaller than those costed in the uti-

lization approach.What Data to Collect, and

When. Payers sometimes make a dis-tinction between efficacy and effec-tiveness. Efficacy is the maximumpotential benefit from a device that isrealized in patients who are verycompliant, closely monitored, andintensively treated in a clinical trial.Effectiveness is a measure of whathappens in the real world.Some ther-

apies are more inclined toward a dif-ference in efficacy and effectivenessthan others.If the sponsor is develop-ing a therapy that has a potentiallylarge efficacy effectiveness diver-gence, then it may need to include inthe trial patients who are treatedunder real-world conditions.

If effectiveness is to be reportedin QALYs, patients can completequality-of-life instruments or otherassessments at baseline and then

periodically during the follow-up.The frequency will depend on thenature of the condition and of theintervention that are the focus of thestudy. However, a score on a quality-of-life instrument often cannot betranslated into quality, or utility, ona 0-to-1 scale. It is important tomeasure quality either by a methodthat yields utility directly or byusing an instrument for which a

crosswalk between the instrumentscore and a degree of utility hasbeen developed.

Benefits of Piggybacking.

Gathering health economic data aspart of an FDA clinical trial results,asmentioned, in a cost savings due to

avoiding a stand-alone outcomesstudy. A second benefit is that piggy-backing maximizes the chances of thetechnology achieving coverage for atleast some patients. Clinical trialsgenerally include very complete dataon patient health characteristics, so itis possible to determine which patientsbenefit most,which cost the least, andfor which the technology is the mostcost-effective.CMS is interested in this

information. If some category ofpatients seems to benefit significantlymore than others, that fact can behelpful in making the case for coverage.

A good example is the story ofMedicares coverage of ICDs (see Fig-ure 1). Initially, coverage was limitedto the treatment of life-threateningventricular tachyarrhythmias. Then,in 1999, coverage was expanded tocomprise additional types of cardiac

patients. Coverage was expandedagain in 2003 and in 2005, each timeto encompass indications for whichclinical evidencebalanced withcostsdemonstrated value.

The lesson to be taken away isthat CMS did not automatically allowblanket coverage for ICDs. Instead, itprodded manufacturers into provid-ing clinical evidence for various indi-cations. Once clinical evidence wasavailable, cost-effectiveness analysis

informed the coverage determina-tion; coverage was allowed for thosesubgroups of patients who wouldbenefit most. Those are usually thepatients with whom the therapy is themost cost-effective.

Some product developers mayfind that a piggybacked trial is nolonger an option for them becausethe safety and efficacy trial is alreadywell under way or has been complet-

ed. These companies could considerdeveloping an economic model thatbrings together data from many



5/7

sources, including the clinical trial.Cost and effectiveness data on alter-

native therapies could be gatheredfrom the literature, from commercialand government databases, and fromother sources. The thing such manu-facturers should not do is ignore thepayers need for cost-effectivenessdata simply because they are not ableto collect them during the FDA trial.

Economics andDecision Making

The use of economic incentivesand disincentives to influence behav-ior is commonplace, in healthcare aselsewhere. In 1984, the HospitalInpatient Prospective Payment Sys-tem was implemented to pay hospi-tals for inpatient care of Medicarebeneficiaries. The hospital received afixed payment for each dischargeaccording to its designated diagno-sis-related group (DRG) regardless of

the actual costs it incurred. The pay-ment methodology was based on theeconomic law of averages, which

might be put simply as you winsome, you lose some, it all averages

out in the end. Increased efficiencywas the provider behavior this sys-tem was intended to elicit. However,faced with a fixed payment, the hospitalfocused on cost.Medical device manu-facturers found themselves on thereceiving end of new pricing pressure.

Cost Minimization. In somecircumstances, the use of healtheconomic data to develop a cost-minimization argument has been asuccessful strategy for mitigating

pricing pressure. Vascular closuredevices provide an illustration. Thecost of a closure device was offset bya decrease in consumption of suchresources as nursing time and post-catheterization recovery time.Although the hospital did not receivea higher DRG payment for proce-dures in which the closure deviceswere used, device companies, usingeconomic data,were able to demon-

strate a derivative cost benefit with-in the parameters of the fixed pay-ment. This removed a sales barrier

to the devices.Such a cost-minimization argu-

ment is not always possible,especial-ly for devices costing thousands ofdollars. But Congress did pass legis-lation, the Benefits ImprovementProtection Act of 2000,allowing hos-pitals to be paid additionally for qual-ifying technology.7 The applicationfor a new technology add-on pay-ment does not explicitly require cost-effectiveness data, but it does requestcost data and proof of significantclinical improvement (that is, effec-

tiveness).Coverage and Payment. The

coverage component of reimburse-ment is also favorably influenced byhealth economic data. A survey of228 managed-care plans nationwiderevealed that 90% of the plans con-sider costs in some form when eval-uating new interventions.8 The sur-vey was conducted in 2001. Even atthat time, 40% of the plans were using

formal cost-effectiveness analysis toinform coverage decision making.

Health economics often is a route


Figure 1. As more clinical evidence has become available, the Centers for Medicare and Medicaid Services (CMS) has gradually expanded its cov-

erage of implantable cardioverter-defibrillators (ICDs) to additional types of cardiac patients. Initially, Medicare coverage was limited to the treat-

ment of life-threatening ventricular tachyarrhythmias. In 1999, coverage was expanded to include patients with documented instances of sus-

tained ventricular tachycardia, adding approximately 25,000 beneficiaries. In 2003, CMS approved the prophylactic use of ICDs in patients at high

risk of sudden death due to ischemic cardiomyopathy, expanding coverage to approximately 100,000 patients. And earlier this year CMS expand-

ed coverage again, bringing the total number of eligible patients to approximately 500,000. Each expansion of coverage was made to encompass

treatment indications for which clinical evidencebalanced with costsdemonstrated value. Sources: Medtronic Inc. (Minneapolis); McClellan

and Tunis.6

Year

Eligible Medicarebeneficiaries (est.)

Covered condition

25,000 50,000 100,000 500,000

Life-threateningventricular

tachyarrhythmia

Documentedspontaneous or induced

sustained ventriculartachycardia

Prophylactic use inpatients at high risk ofsudden death due to

ischemic cardiomyopathy

Ischemic and nonischemiccardiomyopathy with

LV fraction 35%and normalQRS complex

1986 1999 2003 2005

health economics for medical devices - a very good read

Documents