|
Items in cart [0] |
Questions? Call 888-624-8373 |
| Copyright © 2009. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 172
1 ~
Paying for
Evaluative Research
ALAN M. GARBER AND DOUGLAS K. OWENS
The sponsors of medical technology assessment, the principal form of evalu-
ative research concerning medical interventions, are as diverse as the goals and
methods of the research itself. In this chapter we describe the sponsors of such
research in the United States and discuss why sponsorship matters and why some
sponsors may be more suitable than others. The identity of the sponsors of
technology assessment is important insofar as most of them have an interest in
the results of the assessment, so their ability to produce evaluative research of use
to parties whose interests do not coincide with their own may be limited.
We first define medical technology assessment and describe the institutions
in the United States that sponsor or perform medical technology assessment.
Often, the responsibility for evaluating technologies is divided among several
institutions, particularly the federal government and private institutions. Each
institution may in turn play multiple roles, giving it a set of incentives to perform
research that varies with the particular technology and context.
We then discuss mechanisms for sponsoring technology assessment. Specif-
ically, when is it most appropriate for the government to fund such activities and
when do such activities properly belong in the private sector? To what extent
should sponsorship of technology assessment come from payers and to what
extent should it come from consumer groups or the producers of the technolo-
gies? We close by drawing some general guidelines about how and by whom
technology assessment should be performed.
172
OCR for page 173
PAYING FOR EVALUATIVE RESEARCH
WHAT IS TECHNOLOGY ASSESSMENT?
173
One of the most widely used definitions of technology assessment comes
from an Institute of Medicine publication. According to that definition, technol-
ogy assessment is
any process of examining and reporting properties of a medical technology used
in health care, such as safety, efficacy, feasibility, and indications for use, cost,
and cost-effectiveness, as well as social, economic, and ethical consequences,
whether intended or unintended (Institute of Medicine, 1985, p. 2~.
Medical technology itself can be construed in broad terms; according to the con-
gressional Office of Technology Assessment, medical technology consists of the
techniques, drugs, equipment, and procedures used by health-care professionals
in delivering medical care to individuals, and the systems within which such
care is delivered (cited in Institute of Medicine, 1985, pp. 1-2~.
These broad and-seemingly all-encompassing definitions of medical technol-
ogies and formal evaluations of those technologies can be given greater specifici-
ty by distinguishing the stages) of technology assessment. Fuchs and Garber
(1990) defined three stages of technology assessment (stages I to III), in which
each stage encompasses progressively broader dimensions of evaluation. Stage
I, according to this classification, focuses on technical characteristics. This might
include measures of the ability of an antibiotic to inhibit the growth of bacteria in
a test tube or measures of the resolution, speed, and electromagnetic hazards of a
new computed tomography scanner.
Stage II investigates clinical efficacy, in which the clinical outcomes may be
surrogate endpoints. For example, a stage II clinical trial of an antihypertensive
medication might measure its efficacy at lowering blood pressure, but not its
effects on mortality or on the incidence of stroke. Stage II includes most tradi-
tional randomized clinical trials as well as nonrandomized evaluations of the
efficacy of treatment.
The purview of stage III trials encompasses comprehensive clinical, eco-
nomic, and social endpoints. Cost-effectiveness analyses of medical technolo-
gies are stage III studies, at least when they attempt to incorporate comprehen-
sive measures of health outcomes in addition to costs. Although stage III studies
are the most comprehensive, to the extent that they can build upon the results of
stage II studies they need not be expensive. A traditional randomized clinical
trial, a stage II assessment, is likely to be very expensive; a stage III assessment
that combines information from such a trial with data from other sources may be
far less costly. Many technologies never receive a stage III assessment.2
~ These stages should not be confused with the phases in pharmaceutical evaluation used by the
Food and Drug Administration, which have a precise regulatory interpretation.
2 Recently, clinical trials that assess a broad range of endpoints, including the impact of the inter-
vention on economic endpoints and quality of life, have been initiated. Such trials are likely to
OCR for page 174
174
ALAN M. GARBER AND DOUGLAS K. OWENS
The distinctions between stages of technology assessment are useful because
the purpose of early-stage assessments (stages I and II) differs from that of later-
stage assessments (stage III). Stage III assessments deal most directly with the
policy, reimbursement, and clinical issues that determine when a technology
should be used, how it should be implemented, and whether the benefits the
technology offers can be obtained at an acceptable cost. These issues are not
addressed in stage I or II assessments, which are valuable for different reasons.
Stage I and II assessments answer precise technical and clinical questions using
well-established methods.
The sponsors of technology assessment also have different incentives for
doing studies at different stages. That is why they often concentrate their efforts
on a particular stage. The incentives facing the sponsors influence the way they
select technologies for study and the endpoints that they use for the evaluation.
Seldom will an insurer devote substantial resources toward evaluating an aspect
of technical performance the resolution of a scanner, for example just as the
manufacturer of the scanner may be unwilling to devote resources to determining
the impact of its scanner's use on the functional recovery of individuals whose
metastatic cancer is found through use of the device.
WlIO PERFORMS TECHNOLOGY ASSESSMENTS
IN THE UNITED STATES?
The United States has long been a leader in performing medical technology
assessment, even though countries with national health insurance would seem to
have stronger reasons to support evaluative research. In the United States, as-
sessment activities have not been tightly linked to reimbursement; much of the
funding has come from organizations that have a loose connection, if any at all,
to major insurers. In this section, we give an overview of medical technology
assessment in the United States and describe the groups that perform it. We also
briefly describe the amounts that they spend and their methods for choosing
which technologies to assess. The available expenditure information does not
allow us to distinguish between the stages of technology assessment. However,
it seems clear that if stage I assessments were included, the largest sponsors of
technology assessment would be drug and device manufacturers, much of whose
research and development expenditures can be considered forms of technology
assessment. Our discussion below focuses on stage II and III evaluations.
become the most convincing stage III assessments; by designing trials that encompass these compre-
hensive endpoints at the outset, the investigators obviate the need to link data from disparate sources
and to model phenomena that were not measured as part of the trial.
OCR for page 175
PAYING FOR EVALUATIVE RESEARCH
175
Identifying the Sponsors
The sponsors of evaluative research in the United States are remarkably
diverse and numerous. Although no comprehensive inventory of sponsors has
been published, two reports prepared by the Institute of Medicine provide helpful
overviews (Institute of Medicine, 1985, 1988~. In particular, the Medical Tech-
nology Assessment Directory (Institute of Medicine, 1988) catalogues many of
the organizations that sponsor technology assessment.
Most sponsors of evaluative research can be classified broadly into three
categories. They are producers of technology, consumers, and payers. The
reasons for sponsoring evaluative research differ markedly among the categories
(discussed later in the section "Sponsoring Evaluations of Medical Technolo-
gy"~. Because their goals differ, to the extent that the studies are subject to biases
and that there is selectivity about the outcomes to be investigated and the tech-
nologies to be studied, the sponsor of the research can have a major impact on the
nature of the findings. As we discuss below, many of the sponsors of technology
assessment play different roles in different contexts, so they do not face the same
incentives or bring the same focus to each one of their activities.
Producers
Producers of medical technology include the pharmaceutical industry, de-
vice and equipment manufacturers, professional societies and associations, and
some government agencies. Professional societies often evaluate the services
and products that their members supply. For example, the Diagnostic and Thera-
peutic Technology Assessment Program of the American Medical Association
produces reports that aim to disseminate state-of-the-art information about the
effectiveness of medical treatments to the practicing medical community and to
the public at large.
Some government agencies, such as the National Institutes of Health (NIH),
act in part as producers of technology. NIH is responsible for conducting or
sponsoring the research that leads to many important medical innovations. Its
function as the major government sponsor of biomedical research in the United
States, combined with its large and impressive intramural research program,
means that it has contributed in some capacity to most of the major biomedical
advances of the years after World War II. In some cases, NIH holds patents or
otherwise controls the production of biomedical technology. For example, al-
glucerase injection, a treatment for the inherited disorder known as Gaucher's
disease, is produced in the private sector under a licensing agreement with NIH.
NIH, which discovered the drug, holds one patent and has applied for another
patent concerning its production (Gerber et al., 1992~. NIH also has a very large
budget for evaluative research, including small clinical studies, small and large
randomized clinical trials and, to a limited extent, some stage III evaluations,
including cost-effectiveness analyses.
OCR for page 176
176
Consumers
ALAN M. GARBER AND DOUGLAS K. OWENS
Consumers of medical technology are those people who use medical care
directly (that is, patients) or those organizations that represent patients' interests.
Several agencies of the federal government have the responsibility of safeguard-
ing the interests of patients, such as the Food and Drug Administration in its role
as the regulatory agency for medical devices and drugs. The Agency for Health
Care Policy and Research (AHCPR) is perhaps the largest federal sponsor of
stage III technology assessments. For all three stages of medical technology
assessment, however, NIH spends substantially more than AHCPR because it
sponsors many clinical trials, which are often large and expensive. A major
focus of the research sponsored by AHCPR is determination of the effectiveness
and the cost-effectiveness of medical technologies. Medical professional societ-
ies also often sponsor evaluations of technologies that they do not produce. For
example, the American College of Physicians (ACP), an organization of intern-
ists, has studied and issued guidelines regarding the use of mammography; the
ACP and its members neither sell the equipment used for mammography nor,
typically, do they collect the professional fees for performing mammography (a
task that usually falls to radiologists). Rather, internists act as agents for their
patients. Other private, nonprofit groups, such as research organizations (for
example, the Battelle Memorial Institute, the Hastings Center, and other universi-
ty-based programs) and foundations (for example, Project HOPE, the Center for
Health Affairs, the Hartford Foundation), may sponsor or perform evaluative
research that serves to protect the interests of consumers.
Payers
Payers, both public and private, also sponsor technology assessment. Feder-
al payers that participate in evaluative research include the Health Care Financ-
ing Administration (HCFA), which is responsible for administering the Medicare
program and the federal participation in Medicaid, and the Department of Veter-
ans Affairs (VA). The VA Health Services Research and Development Service,
in particular, sponsors a broad range of evaluative research activities. Private-
sector payers that sponsor technology assessments include insurance companies
(for example, the Blue Cross and Blue Shield Association Technology Evalua-
tion and Coverage Program), health maintenance organizations (McGuire, 1990),
and self-insured employers.
Expenditures for Evaluative Research
The diversity of organizations and the varied natures of their activities make
precise estimations of total expenditures for evaluative research exceedingly dif-
ficult. Nevertheless, the following figures are included to give a broad sense of
OCR for page 177
PAYING FOR EVALUATIVE RESEARCH
177
expenditures for evaluative research. According to the Institute of Medicine,
U.S. expenditures for technology assessment in 1984 were almost $1.3 billion, or
0.3 percent of total health care expenditures. This figure includes roughly $750
million spent by the pharmaceutical industry on clinical trials, $235 million spent
by NIH on various evaluative activities, and approximately $35 million spent by
the medical device industry. The VA Cooperative Studies Program also sponsors
clinical trials. Expenditures for these trials may reach $18 million to $20 million
when the cost of associated clinical care is included (Janet Gold, Department of
Veterans Affairs, personal communication, September 10, 1992~. In addition to
clinical trials, NIH funds the Consensus Development Program, which sponsors
six to eight consensus conferences per year at a cost of $130,000 to $140,000
each (William Hall, Office of Medical Applications Research, NIH, personal
communication, September 8, 19921. HCFA sponsors approximately 10 technol-
ogy assessments per year via contracts from the Bureau of Eligibility, Reim-
bursement and Coverage to the Office of Health Technology Assessment at AH-
CPR. These activities cost approximately one million dollars each year (Samuel
Dellavecchia, Health Care Financing Administration, personal communication,
September 11, 1992~.
Table 13-1 indicates the technology assessment expenditures by a number of
other organizations. These expenditures are approximate and, because defini-
tions may vary or programs that are responsible for technology assessment often
include other activities, they may overstate the magnitude of assessment activi-
ties. The sources of funding for technology assessment vary among the types of
organizations. Professional societies and trade associations fund technology as-
sessment primarily from the dues of members. The primary funding sources for
technology assessment by private, nonprofit organizations are grants and con-
tracts from federal government agencies, industry, and private sources, including
foundations. Academic organizations similarly fund technology assessments pri-
marily from grants and contracts from federal and private sources.
Selection of Technologies for Evaluative Research
Although the process of selecting which technologies to assess varies sub-
stantially, some common themes emerge. In general, the selection process is
influenced strongly by whether an organization is acting primarily as a producer,
as a representative of consumers, or as a payer. As discussed below in detail,
producers naturally focus on the technologies that they have developed. In con-
trast, organizations that act on the behalf of consumers have various selection
processes. AHCPR has focused many of its large-scale technology assessments
on conditions and treatments for which there is significant geographic variation
in their use (a potential indicator of inappropriate utilization or uncertainty about
efficacy) and that are expected to account for substantial Medicare expenditures.
OCR for page 178
OCR for page 179
OCR for page 180
OCR for page 181
OCR for page 182
OCR for page 183
OCR for page 184
OCR for page 185
OCR for page 186
OCR for page 187
OCR for page 188
OCR for page 189
OCR for page 190
OCR for page 191
OCR for page 192
Representative terms from entire chapter:
medical technology
c~
cq
c~
?
o
~o
o
c~
s~
o
c~
o
v)
1
1 oJ)
~3
~ -
-
b4
m
C~
._
Ct
C)
~=
C)
O
m
U~,
C~
O
OZ¢ O O O Oo
OO O O O
OO O ~ O
O ~O
__I ~t^
ff}
Go . _
ct . U ~G
,,, £ £
~ £
a
~ ~ 0 0
Ct C ~o =.
. ~ ¢
·- c' ~ C) L~
°C P~ £ Z
o ~ Ct o
~,, - ~ C
¢= ¢ ~ ~ V)
) o ¢ V V
x ~ ~ m a:
O ¢ O ~ O O O O O O O
O O ~ O O O O O O O
O Z O ~ O O O O O O O
O O O O O O 0 0 0
O ~O O O O O O O
O ~ aN ~ ~O O
~ ~ ~ ~ ~O ~
(f3 C ~CM ~3
_
04
o O
C~
:O
C~ Ct Ct ~ ~ .o
a a L ~ a ° ~
O O O O U, ~ C)
C) ~ ~E_ ~ ~i
:~ :: ~ ,= a :~ ~ ~ ~ ¢
E¢- ¢ ¢ ¢ ~ o ~ ·° ~ E_
0 ~ ~ ~ z z z z ;> ~ 0
¢ O ¢ O O O O ¢ O ~ O ~ O O O O
Z 0= Z g Oo O g Z Oo Z Oo °° g~ g~ O~
M0 0 ~ O O ~ O
- 00 0 - O
_ ~ ~ ~ ~3 0
~3 ~ (f3 (*
E e ~Y ~ ° ° 3 ~ ,~ ~' ~
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ c
~ ~ 0 ~ c' ~ <' c' ce ~ 2 G S 1- O ·- Ct · t
C) ~ O ~ O ~ O O O C) · ~ ~ O O · C,) O ~i ~ ~
¢ ¢ £ ~ V 0 V ~) V ~ ~ V ~ ~ ~ G e~ °J) ~ £
. Ct 3 . C~ ¢
c~ c~ ct
PAYING FOR EVALUATION RESEARCH
179
In addition, AHCPR sometimes performs or sponsors evaluations requested by
HCFA, as mentioned above.
Professional associations generally select technologies on the basis of re-
quests from members, standing committees, third-party payers, and government
organizations. The most notable assessment efforts by a professional society are
those of the Clinical Efficacy Assessment Project (CEAP) of ACP. CEAP identi-
fies potential projects on the basis of a review of the current literature, academic
opinion, policy needs, as well as requests from other ACP committees and out-
side organizations including payers. CEAP has had a long and successful collab-
oration with the Blue Cross and Blue Shield Medical Necessity Project, which
commissioned a series of papers from the project. A large number of CEAP
assessments of diagnostic procedures and screening tests have been published in
the Annals of Internal Medicine. These papers were written by ACP members
who had particular expertise in evaluative research and are reprinted in two books
(Eddy, 1991; Sox, 1987~. On the part of ACP, the final selection of topics to be
assessed by CEAP depends on, among other factors, the anticipated level of
interest to practitioners of internal medicine, the potential for wide application
and benefit of the technology, and the risks associated with wide application of
the technology. As is true in this example, in general, professional associations
assess technologies that are relevant to their members, either because of ques-
tions about efficacy or because of controversy about coverage policy.
Payers usually focus on technologies that are likely to result in significant
claims. They emphasize technologies that are new, expensive, or of uncertain
efficacy. For its collaboration with the ACP, the Blue Cross and Blue Shield
Association Medical Necessity Program prioritizes projects on the basis of re-
quests from member plans. HCFA sponsors technology assessments to deter-
mine if medical technologies are or continue to be eligible for Medicare cover-
age. More specifically, for a technology to be assessed it must (1) represent a
significant advance in medical science or be potentially outmoded, (2) have the
potential for a significant financial impact on the Medicare program, (3) have the
potential for rapid diffusion, and (4) have generated controversy regarding its
efficacy or appropriateness of coverage (Samuel Dellavecchia, Health Care Fi-
nancing Administration, personal communication, September 11, 1992~.
In summary, a diverse group of people and organizations sponsors medical
technology assessments. Although it is difficult to identify all of the participants,
it is helpful to categorize the sponsors by the roles they play. Those sponsors that
have formulated explicit policies for choosing technologies that should be as-
sessed usually adopt criteria that reflect the interests of the people and the institu
tions they represent.
WHO SHOULD SPONSOR TECHNOLOGY ASSESSMENT?
As long as the process of technology assessment is open, the quality of the
research underlying it is high, and the selection of technologies for study is
180
ALAN M. GARBER AND DOUGLAS K. OWENS
appropriate, the value of the studies will depend little on the sponsor of the
funding. Because these conditions are not always met, the identities of the spon-
sors can matter a great deal. Another reason for concern about the sponsors of
the research is that current funding of technology assessment may well fall short
of what is needed. Who might bear the cost of any incremental funding dedicat-
ed to this activity?
Of the three main groups of sponsors that we have discussed producers,
consumers, and payers-each has something different to gain from an assess-
ment. Insofar as their goals differ, they will not always agree in their rankings of
the importance of specific technologies for study or in the outcomes that matter.
The producers of a health care technology have an interest in showing that the
technology is valuable; the consumers of the technology have a direct interest in
learning about its efficacy, but perhaps not its cost; payers have a direct interest
in its cost and a less direct interest in its efficacy.
Producers of Medical Technology
Ordinarily, the producers of a technology sponsor or perform an evaluation
of the technology in the early stages of its development. One of the costs of
developing a product or technology is establishing its technical characteristics.
For some technologies, like medical devices, early evaluations may lead to
changes in the design of or manufacturing process for the product. Early assess-
ments may be conducted before the product is publicly announced or when many
of its characteristics are known only to the producer. Thus, the producer will be
uniquely aware of the limitations and advantages of the technology and might
justifiably refuse to cooperate in an external evaluation. The producer also has the
incentives to perform the evaluation, insofar as the return to expenditures for evalua-
tion will accrue to the producer. In the United States, a great deal of producer-
sponsored evaluation is mandated by regulation, particularly for drug approval.
Producers might not perform certain kinds of evaluative research that would
be useful to consumers. The most important impediment arises if the producer
cannot expect to gain from performing the research. Consider a manufacturer of
acetaminophen, a drug that has been available in generic form for many years.
No company has a monopoly on the sale or distribution of acetaminophen, and
any research showing its superiority to alternative analgesics would benefit all
current and potential producers, not only the sponsor of the research. If the
research instead concerned a newly approved nonsteroidal anti-inflammatory
drug, the company that had exclusive marketing rights would have strong incen-
tives to fund a study likely to demonstrate its superiority to other nonsteroidal
anti-inflammatory drugs. Only when a producer can capture returns that exceed
the costs can it be expected to sponsor evaluative research; this situation is un-
likely to arise when there are multiple current or potential competitors in the
production and sale of the drug or device.
PAYING FOR EVALUATIVE RESEARCH
181
Similarly, a producer cannot be expected to sponsor a study that is unlikely
to show that its product is superior to an alternative therapy. If the study is
devoid of marketing value and is not required for approval, there is little reason
to proceed. Nor will a producer sponsor a study whose results will not be avail-
able during the lifetime of a drug or device patent. The financial benefit to the
manufacturer of a study that will not produce results until the expiration of the
patent or period of monopoly might not exceed its costs. That is one of many
reasons why long-term trials of the effects of preventive care on survival are
seldom sponsored by producers; how can a manufacturer of an antihypertensive
drug conduct a trial that is expected to show a mortality benefit in 15 years if the
patent will expire by then?
Even in circumstances in which producers are willing to sponsor technology
assessments, however, the studies that result may not have the credibility of
studies sponsored independently. For example, if a company can suppress studies
with negative findings and selectively release studies favorable to its products,
published studies will give a misleadingly positive impression of the therapies.3
The design of a study may similarly reflect the interests of the producer
rather than those of the consumer. Because the producer of a technology will be
motivated to perform a study to meet regulatory requirements or to increase
revenues, it may select endpoints and study designs that show the technology to
its greatest advantage. The resulting endpoints may not be the most relevant to
clinical or policy decisions. For example, a producer-sponsored study might
emphasize the advantages of a drug in terms of its side effect profile. The
producer may be less willing to study the long-term efficacy of the drug, particu-
larly if there are significant doubts that it would be better than older and less
expensive alternatives. Therefore, it is critical that standards for the conduct and
review of producer-sponsored technology assessments be developed and imple-
mented.
In summary, when a producer is not a monopolist, when an evaluative study
is unlikely to demonstrate the superiority of the product, or if the producer will
no longer be a monopolist by the time the research is completed, the producer
will have little incentive to sponsor a technology assessment. Even when the
proper incentives are in place for the producer, however, there are additional
reasons why the public, as consumers or payers, might not place much credence
in producer-sponsored technology assessments, unless there is clear evidence
that the study adheres to rigorous methodologic standards.
3 This bias is similar to, but distinct from, "publication bias." Publication bias refers to the tenden-
cy of peer-reviewed medical journals to favor publication of positive findings over negative ones.
Although one might presume that readers could adjust for such a bias, it is hard to do so without any
information about rejected studies. The same would hold true for company-sponsored negative
studies that are never submitted for publication. As a practical matter, a producer of technology
would find it difficult to suppress a multi-institutional randomized trial, but it is probably feasible to
repress dissemination of the results of a small study.
182
ALAN M. GARBER AND DOUGLAS K. OWENS
Consumers of Medical Technology
When producers of a technology are unsuitable as sponsors of an assess-
ment, the consumers of the technology are potential alternative sponsors. No
party has a greater direct stake in the efficacy of a health care technology than the
patients (consumers) who hope to benefit from it. Unlike the producers of the
technology, its consumers do not have an incentive to overstate its benefits, un-
less it is an issue for negotiation with payers.
There is ample precedent in other contexts for consumers sponsoring assess-
ments. For example, car manufacturers are consumers of steel and other inputs
into automobile production; they expend significant time, effort, and money on
evaluations of these components. Consumers Union, a nonprofit consumer test-
ing and advocacy group, provides extensive testing of consumer products. Yet,
as the preceding section suggests, despite the exceptions of federal agencies and
occasional provider groups that may represent consumers, consumers and con-
sumer groups infrequently sponsor formal evaluations of medical technology.
The reasons are not hard to appreciate.
Consumers or patients may not have the expertise to evaluate technologies.
Although their lack of expertise would make it difficult for them to perform
evaluative research, it should not compromise their ability to sponsor it. Of
greater importance is the set of obstacles that economists label "transaction
costs." Consumers of health care are many and decentralized; producers are
fewer and in some cases unique. For any one consumer, the cost of sponsoring
an evaluation is high and the information acquired is of relatively little value to
that individual, although for consumers collectively, the information may be quite
valuable. If consumers could easily cooperate and pool their resources to spon-
sor such studies, they would often be an appropriate source of funding. Thus,
consumer interests are most likely to be represented by agencies of the federal
government, such as AHCPR, acting on their behalf.
Consumers may also be ill-suited to sponsor such studies, at least as individ-
uals, because insurance coverage may narrow their interests. For consumers
whose health care costs are completely covered by insurance, there is little inter-
est in knowing the price of the technology; their sole concern is its safety and
effectiveness. If they bear a copayment, they will have a direct stake in the costs
of the technology, but to the extent that they are insured, their sensitivity to cost
will be blunted. Hence, consumers often have little incentive to sponsor research
that evaluates the costs as well as the benefits of medical technologies. Consum-
ers or their representatives can play a role in sponsoring evaluations of medical
technologies, but they are unlikely to take the lead in such efforts.
Payers for Use of Medical Technology
Unlike consumers, payers have a direct interest in the costs of the technolo-
gy, and at least some payers have a large enough share of the insurance market to
PAYING FOR EVALUATIVE RESEARCH
183
justify making the expenditures to evaluate medical technologies. Unlike pro-
ducers, payers may perform evaluations of drugs and devices that are not patent-
ed (or that compete with several similar products). The large representation of
payers among the entities sponsoring technology assessment attests to the impor-
tance of this activity to payers and to their suitability as sponsors of evaluative
research.
Although payers are often large organizations with thousands of enrollees,
none can be considered a monopsonist (sole buyer) in the market for health
services or a monopolist in the sale of insurance. This is one of the reasons that
payers have limited incentives to perform evaluative research. Even a large
organization may not have sufficient economic interest to sponsor a costly evalu-
ation of a medical technology, particularly if the information will be available to
other payers (some of them competitors). Although payers collectively might
find it worthwhile to learn who should receive coronary angioplasty, no single
payer may be responsible for a large enough share of the market to justify spend-
ing the money required for a full evaluation of that procedure. Thus, payers tend
to underinvest in such activities, although in many respects they are appropriate
sponsors of evaluative research.
Payer-sponsored studies are subject to concerns about payer motivation.
Payers are criticized from time to time for inappropriately concentrating on the
bottom line of expenditures rather than on the well-being of their enrollees. Un-
doubtedly, some insurers focus inappropriately on the short-term payouts, but
such an attitude cannot be a viable long-term strategy for a payer. Payers, whether
they are insurers selling policies to enrollees or companies financing health care
for their employees, cannot long retain subscribers or employees if the coverage
package they offer does not seem worth the cost. If current insurers are systemat-
ically attempting to deny valuable health technologies to enrollees, they are not
very successful; it is probably easier for privately insured Americans than for the
citizens of any other country to gain access to expensive medical technologies.
Other Sponsors
The federal government (and to a lesser extent, state governments) is per-
haps the most important sponsor of evaluative research. It plays multiple roles;
hence, it does not fit completely into one of the categories of sponsors of technol-
ogy assessment discussed above. Some agencies, such as AHCPR, reflect the
interests of many parties and take seriously their role as a protector of consumer
interests. The FDA similarly protects consumers in its efforts to keep unsafe or
ineffective drugs and devices off the market.
In summary, even though many of the entities that sponsor evaluative re-
search do not fit neatly into a single category, the classification is useful because
it helps to provide an understanding of the influences and biases of each sponsor
(see Table 13-2~. Furthermore, a sponsor's willingness and ability to undertake
184
ALAN M. GARBER AND DOUGLAS K. OWENS
TABLE 13-2 Role of Health Care Organizations
Consumers/
Organization Producer Representatives Payers
Health maintenance organizations X X
Providers X X
Government agencies X X X
Hospitals X X
Research organizations, foundations, IOM X
Pharmaceutical industry X
Device manufacturers X
Health Care Financing Administration X
evaluative research depend directly on its role. Payers might be willing to spon-
sor a study of bypass surgery, but producers-the surgeons who perform the
procedure, and the hospitals where the surgery is performed would not. The
producers would be reluctant because they are too dispersed and would benefit
too little from the results of such an evaluation to devote resources to it. Con-
sumer groups and government agencies might sponsor a study of the use of
aspirin for the prevention of heart attacks, but producers would not, because it is
a generic product. Payers would not sponsor such a study, because their expendi-
tures for aspirin are insignificant, and unless the use of aspirin was likely to
markedly reduce the expenditures associated with heart disease, the treatment
would not have a large financial impact. Any one sponsor payers, producers,
or consumers would bear significant costs, but the information would be avail-
able to all. Hence, no one sponsor is likely to be able to fund studies of all
technologies with which it is concerned. Other methods for sponsoring evalua-
tive research may overcome this problem.
SPONSORING EVALUATIONS OF MEDICAL TECHNOLOGY
Previous Proposals for Sponsoring Technology Assessment
Each of the parties identified above can benefit from evaluations of technol-
ogy. Their interests and capabilities differ, and they may not benefit equally
from evaluative research, so they may not be equally suitable as sponsors of
evaluative research. Although it is not clear that any one party should bear the
responsibility for sponsoring all such research, the responsibility is now shared in
a haphazard fashion. Too little is probably spent on such research because the
benefits that all groups share undoubtedly exceed the total amount that produc-
ers, consumers, and payers spend on evaluative research as individual entities.
Furthermore, the small percentage of health care expenditures devoted to formal
PAYING FOR EVALUATIVE RESEARCH
185
technology assessments seems reasonable only for technologies that are stable,
unchanging, and of well-established effectiveness; it is hard to imagine a less apt
description of modern medical technologies. The recognition that too little tech-
nology assessment is done has periodically motivated individuals to propose new
. . ~ . . .
mechanisms for sponsoring Such activities.
Most of the published proposals for changing the funding of technology
assessment suggest that payers take a prominent role (Bunker et al., 1982; Mc-
Givney, 1992; Relman, 1980; Yarbro and Mortenson, 1985~. Of the many indi-
viduals and organizations that have an interest in the evaluation of medical tech-
nology, payers have the most to gain from accurate assessments. The simple
reason is that payers can directly use the results from evaluative research to assist
in coverage decisions. A fair process for studying technologies and for using the
results of evaluative studies to determine coverage can enable payers to facilitate
the delivery of cost-effective care; to the extent that they are successful, and their
success is known to employers and potential subscribers, they will attract addi-
tional subscribers.
One of the most comprehensive plans was offered by Bunker and colleagues
(1982), who proposed the formation of an Institute of Health Care Evaluation.
The institute would include representation from the public (employers, employ-
ees, and consumers), the medical profession (researchers, academicians, clinical
practitioners, and professional societies), payers, and the federal government.
Funding would come from a per capita levy on public and private payers, grants
and contracts from the federal government, and a fee from all other members.
The major goal of the institute would be to perform outcomes-based evaluative
research, including cost-effectiveness analysis, on technologies chosen by the
members. The institute's role would not include policy decisions; rather, respon-
sibility for policy and coverage decision would rest with the members. Although
agencies of the federal government and professional societies would contribute
to the institute, much if not most of the funding would come from payers. The
proposal does not mention the producers of medical technology explicitly.
A proposal formulated by the Jackson Hole Group, an organization pro-
pounding a specific model of managed competition, addresses the issue of spon-
soring technology assessment within the framework of a comprehensive plan to
reform the health care system (Ellwood, 1992~. The plan of the Jackson Hole
Group calls for the establishment of several new oversight agencies that together
create a public-private health partnership. An independent government agency,
the National Health Board, would oversee the transition to a health care system
characterized by universal health insurance coverage, private-sector competition,
and ongoing technology assessment. The National Health Board would recom-
mend which health benefits should be covered on the basis of an evaluation of
their effectiveness. Evaluations of effectiveness depend on the activity of two of
three proposed private-sector agencies. A Health Standards Board would assess
medical technologies and medical practice variations. This activity would be
86
ALAN M. GARBER AND DOUGLAS K. OWENS
facilitated by the creation of an Outcomes Management Standards Board, which
would set standards for data collection and health outcomes reporting and would
oversee the collection of outcomes data on all delivered health care. Both boards
would be sponsored by payers, employers, consumers, and health provider
groups.
The aspects of the Jackson Hole Group's plan that are relevant to evaluative
research are as follows. First, evaluative activities would be made an integral
part of health care by requiring that outcomes data be systematically collected,
reported, and analyzed. Second, technology assessment would be sponsored
primarily by payers (insurance companies and employers), with contributions
from consumer and health provider groups. The plan of the Jackson Hole Group
is similar to the proposal of Bunker and colleagues (1982) in several important
respects. Both plans highlight the need for systematic, increased funding of
technology assessment, with participation of public and private entities, and both
emphasize the role of payers as sponsors. The Jackson Hole Group model ex-
tends to many other aspects of health care, however, and goes considerably fur-
ther in integrating collection of the data needed for technology assessment into
routine health care delivery.
Several other proposals recommend that payers serve as the primary spon-
sors of technology assessments, perhaps in combination with government payers
(Friedman and McCabe, 1992; McGivney and Hendee, 1988; Yarbro and
Mortenson, 1985~. For example, the payers might cover the medical care costs of
patients enrolled in clinical trials but not the incremental research costs attribut-
able to the trial itself.
Sponsorship of Stage III Assessments: The Leading Role of Payers
The chief reason for focusing on payers is that they have both the ability and
the need to sponsor evaluative research, particularly stage III studies. Assuming
that the research itself is valuable, an assumption that we do not explore here, too
little is done for the sole reason that it is a public good. Use of the information by
one person does not make it less available to others; in economist's terminology,
the information is "nonrival." In contrast, the bread that one person consumes is
bread that is not available to another. It is also difficult to charge for access to
some kinds of information; economists describe such information as "nonexclud-
able." Because the information is valuable to many payers, consumers, and other
interested parties, its total value may be great, but if access to the information
cannot be controlled, no single entity that produces it will be able to recover its
costs by selling the information to others.
These conditions characterize a form of market failure, which means that the
standard argument that competitive markets produce optimal outcomes does not
apply here. For an imperfect market such as this, Ronald Coase (a recent winner
of the Nobel prize in economics) provided a useful framework for thinking about
PAYING FOR EVALUATIVE RESEARCH
187
regulatory solutions. If there were a costless way to negotiate detailed contracts,
to monitor them, and to enforce compliance with them so that everyone with
access to the information had to share in the cost, this source of market failure
could be overcome. These ideal conditions, called zero transaction costs, cannot
be met, so the next best solution is to determine who would have the responsibili-
ty for producing the information if the ideal conditions were met. Coase's argu-
ments had to do with property rights; he claimed that the legal system should
(and did) assign rights to the party that would have purchased them if there were
zero transaction costs. The best sponsor of evaluative research can be discovered
by analogous means; it is the party or parties that would have the responsibility
for sponsoring or producing the research if efficient markets for information
could be established.
We believe that payers would have the primary responsibility for sponsoring
evaluative research if such market conditions existed, because payers have the
most to gain and have natural advantages in sponsoring the relevant research.
Insurers can serve their enrollees better if they base their coverage decisions on
good information about what medical technologies work best for a given expen-
diture. Furthermore, payers have unique access to the claims databases that can
be used as part of the assessment of the costs and cost-effectiveness of medical
technologies. These arguments have far less force for stage I assessments than
for stage III assessments because, as we have noted, producers usually have the
best access to the information relevant to stage I assessments and also have
greater incentives to sponsor them.
Alternatives to Payer Sponsorship
Assigning the sponsorship responsibility to insurers is only one solution to
raising the overall financial commitment to evaluative research. An obvious
alternative is an increased level of direct government involvement. Why not
increase the already substantial federal activity in technology assessment? Inas-
much as the federal government represents all interested parties and has an exten-
sive and successful record of sponsoring such research, it too would seem to have
strong reasons to support such work and the competence to do so. For many
purposes, involvement of the federal government is desirable. But the means of
financing the evaluative research is the most important reason for preferring
sponsorship by payers (which includes some federal agencies and programs) to
sponsorship by the federal government generally.
Government funding may lack one of the desirable features of sponsorship
by payers. Namely, if payers are the sponsors perhaps contributing to a fund in
proportion to their premiums-the "tax" that funds the research falls most heavi-
ly on those who benefit most. The people with the most comprehensive insur-
ance coverage would pay the most for the research, whereas those with minimal
insurance would pay less. Government-sponsored research could be funded by a
188
ALAN M. GARBER AND DOUGLAS K. OWENS
variety of mechanisms, but if it were funded by general tax revenues rather than a
premium tax, the people who benefit the most from the research would not neces-
sarily pay the most. Funding on the basis of premiums does not eliminate "free
rider" problems, however, because people who lack insurance would not contrib-
ute to the fund. Nonetheless, the expenditure for technology assessment would
be very small in relation to the value of the uncompensated health care the unin-
sured receive.
Tax Incidence for Evaluative Research
Tax incidence refers to the parties who ultimately bear the burden of a tax. If
a tax is levied on groceries and the price of groceries increases by the same
amount as the tax, the incidence of the tax is on the consumers. We suspect that
the incidence of a tax on health insurance premiums will be on the people who
pay the premiums, not the insurers. Similarly, it may matter little whether the
money used to sponsor an evaluation comes in the form of a tax on health insur-
ance premiums or as a tax on health care expenditures. In both cases, the cost is
likely to be passed on to those consumers who either buy health insurance or use
health care heavily. Suppose that the tax is imposed on health care expenditures
rather than on the insurance premium. Then the insured person's health care
expenditures might be, say, one percent higher than they would be otherwise. If
insurance covered the tax as well as the health care expenditures, the added cost
would be passed through as an increase in the insurance premium, so it would be
nearly equivalent to an insurance premium tax. An important difference between
the insurance premium tax and the health care expenditure tax is the liability of the
uninsured; the uninsured would have to pay the latter but not the former tax.
However, if many of the uninsured are unable to pay their health care expenses,
they might also be unable to pay the tax. Furthermore, because many (but not all) of
the uninsured and underinsured have less wealth and income than the insured,
sparing them the tax might help redistribute income toward the less well off.
If insurers did not cover the health care expenditure tax, the tax would look
to the insured individual like an increase in the copayment rate. Since most
insured patients prefer insurance coverage for essentially all health expenses
(about 80 percent of Medicare enrollees have "Medigap" plans that pay a sub-
stantial fraction of Medicare copayments and deductibles), they are likely to
prefer to have insurance coverage for the technology assessment tax along with
their other health care expenditures. Thus, they would prefer to contribute to the
fund in the form of increased premiums rather than to pay an additional charge
that would be added to copayments.
We believe that the contribution to stage III evaluative research should equal
approximately the level Relman recommended over a decade ago for evaluative
research, that is, 0.2 percent of health expenditures (Relman, 19801; at a very
minimum, the level of expenditure should be 0.1 percent of health insurance
PAYING FOR EVALUATIVE RESEARCH
189
premiums, which would amount to about $200 million (Fuchs and Garber, 1990~.
A fixed percentage contribution would enable the funding for technology assess-
ments to increase if the adoption of a new technology led to increases in health
care expenditures. If the pace of innovation or of the development of new appli-
cations of existing technologies increased, a higher level of funding might be
needed, whereas if innovation slowed, a smaller contribution might be appropri-
ate. These funds would be used primarily for stage III and, to a lesser extent,
stage II assessments and would complement rather than replace the evaluative
research sponsored by producers, as we discuss below.
Sponsorship for Early-Stage Technology Assessment
The published recommendations for financing and sponsoring technology
assessment typically do not distinguish the stages of research, but most of the
recommendations seem to focus on what we call stage II or stage III assessments.
Payers have much less to contribute to stage I assessments. Because stage I
evaluation is often a key component of the development (or approval process) of
a new technology, the producers would usually be best situated to sponsor such
evaluations. They also have incentives to perform stage II assessments when
they are required for approval. Furthermore, stage II assessments may be neces-
sary for discovering how a technology might best be used; for a drug, a stage II
assessment may be required to determine the best dose and dosing interval. The
producer will usually have a strong interest in performing such assessments and,
to the extent that the assessments are designed to refine the technology, the
producer has the incentives to perform a study that consumers, payers, and health
care providers will find useful.
Producers would have a diminished role in stage III assessments, however.
Drug and device manufacturers are increasingly interested in sponsoring cost-
effectiveness evaluations, but they are unlikely to do so if the results of the
evaluation might be unfavorable to their product.
Standards for Evaluative Research
Successful expansion of technology assessment activities requires not only
adequate funding from appropriate sponsors but also the development of stan-
dards for the research. At least as important as the issue of who should sponsor
technology assessment is the question of how it can be made as impartial and as
useful as possible. Evaluative research can be complex, the data can be over-
whelming, and the analytic methods can be difficult; it will not always be amena-
ble to replication without great effort. Consequently, the process itself must be
open and credible. One way to ensure this is to concentrate decisionmaking in a
group that is insulated from short-term political and economic exigencies but that
is responsive to the need to contain health care spending.
190
ALAN M. GARBER AND DOUGLAS K. OWENS
Centralization of the sponsorship of technology assessment activities can
improve the quality and usefulness by such assessments by helping- to aid in the
implementation of uniform standards. The cost-effectiveness techniques that are
prominent in stage III assessments are fairly well developed, despite a number of
areas of disagreement. One of the most imposing obstacles to using cost-effec-
tiveness results to guide resource allocation in health care is the lack of uniformi-
ty in the assumptions used by different groups of investigators. Although any
two studies might adhere to the same set of principles in computing cost-effec-
tiveness estimates, one might use a rate-of-time discount of five percent, whereas
the other might use two percent, one might adopt a societal perspective, whereas
the other might adopt the perspective of payers. Although we do not advocate
strict adherence to a uniform standard for all cost-effectiveness studies, the estab-
lishment of a set of minimum standards can improve comparability of assess-
ments. Thus, for example, the institutions that sponsor or coordinate technology
assessments can define a set of basic assumptions that all studies should include,
although they might also explore the consequences of following an alternative set
of assumptions. This simple step would make the studies much more useful and
would not limit the ability of researchers to explore alternative approaches.
The drawbacks of current approaches to technology assessment are not lim-
ited to the varied and often contradictory assumptions used by different research-
ers. A second problem is the selection of technologies to be studied. As we
discussed above, sponsors now select the technologies to study on the basis of
their own interests. Although they can hardly be faulted for doing so, the process
would be more useful to the public if the method for selecting the topics for study
were explicit. The major factors that need to be considered in studying technolo-
gies are well known: the expense of the technology, the degree of uncertainty
about its effectiveness, and the importance of the health condition that is being
prevented or treated.
The voices of consumers, providers, producers, and payers need to be heard
to help ensure that the assessments address their concerns fairly and as complete-
ly as possible. Most proposals for funding technology assessments and most
international approaches to sponsoring technology assessments, whether they are
part of coverage decisions or not, incorporate representation from the many inter-
ested parties.
CONCLUSIONS
Evaluative research has a critical role to play in determination of the efficacy
and effectiveness of medical technologies. If we can ensure that evaluative re-
search is of high quality and accessible, it also should have a central role in
decisionmaking regarding the coverage of medical technologies and reimburse-
ment policies. Because the U.S. health care system is pluralistic and decentral-
ized, no single entity is suitable as a sponsor of the broad range of medical
PAYING FOR EVALUATIVE RESEARCH
191
technology assessment that is needed. Producers have an important role to play
in stage I and II technology assessments. Payers have the most direct incentives
to sponsor stage III assessments, because they need to anticipate the economic,
health, and to a lesser extent, social consequences of covering a technology. We
believe that they should be the primary sponsors of stage III technology assess-
ments. Other proposals have contained similar recommendations, and all urge
that the assessors possess some independence from short-term economic incen-
tives and direct political control.
This area of research is likely to remain underfunded until cooperative ar-
rangements result in more centralized activities than are currently possible. The
continuing and unsustainable growth in health care expenditures will inevitably
lead to further attempts to reform health care financing mechanisms. If these
reforms are to result in the rational use of U.S. health care dollars, they must be
guided by research that determines what works, and at what cost. Powerful
economic disincentives currently deter the growth in technology assessment ac-
tivities that is needed. A system that generates adequate funding for evaluative
research will need to be designed to overcome these disincentives. Only then
will this country have the opportunity to control expenditures while identifying
those technologies that are of the greatest value to consumers of health care.
REFERENCES
Bunker, J. P., Fowles, J., and Schaffarzick, R. 1982. Evaluation of medical-technology
strategies: Effects of coverage and reimbursements. New England Journal of Medi-
cine 306:687~92.
Eddy, D. M., ed. 1991. Common Screening Tests. Philadelphia: American College of
Physicians.
Ellwood, P. M., Enthoven, A. C., and Etheredge, L. 1992. The Jackson Hole initiatives
for a twenty-first century American health care system. Health Economics 1:149-
168.
Friedman, M. A., and McCabe, M. S. 1992. Assigning care costs associated with thera-
peutic oncology research: A modest proposal. Journal of the National Cancer Insti-
tute 84:76~763.
Fuchs, V. R., and Garber, A. M. 1990. The new technology assessment. New England
Journal of Medicine 323:673~77.
Garber, A. M., Clarke, A. E., Goldman, D. P., and Cluck, M. E. 1992. Federal and
Private Roles in the Development and Provision of Alglucerase Therapy for Gaucher
Disease. Washington, D.C.: U.S. Government Printing Office.
Institute of Medicine. 1985. Assessing Medical Technologies. Washington, D.C.: Na-
tional Academy Press.
Institute of Medicine. 1988. Medical Technology Assessment Directory. Washington,
D.C.: National Academy Press.
McGivney, W. T. 1992. Proposal for assuring technology competency and leadership in
medicine. Journal of the National Cancer Institute 84:742-744.
McGivney, W. T., and Hendee, W. R. 1988. Technology assessment in medicine: The
92
ALAN M. GARBER AND DOUGLAS K. OWENS
role of the American Medical Association. Journal of the American Medical Asso-
ciation 112:1181-1185.
McGuire, P. 1990. Kaiser Permanente's New Technologies Committee: An approach to
assessing technology. Quality Review Bulletin 16:240-242.
Relman, A. S. 1980. Assessment of medical practices. New England Journal of Medi-
cine 303: 153-154.
Sox, H. C., Jr., ed. 1987. Common Diagnostic Tests: Use and Interpretation. Philadel-
phia: American College of Physicians.
Yarbro, J. W., and Mortenson, L. E. 1985. The need for diagnosis-related group 471.
Journal of the American Medical Association 253:684~85.
