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 25
4
IMPACTS ON BIOMEDICAL AND HEALTH
RESEARCH
The impacts of research on the health of people in the United
States and around the world may not be measured by economic analyses,
but historically these impacts have been among the most important
benefits of research. Five speakers with diverse backgrounds addressed
this topic at the workshop. They found evidence of substantial benefits
while also identifying areas where benefits may be overlooked by current
approaches. In addition, they called attention to problems with the
funding of federal research, such as the damage up-and-down funding
can do to the careers of researchers and the difficulty of allocating
limited funds across categories of research.
REVIEWING THE LITERATURE ON HEALTH IMPACTS
Bhaven Sampat, Assistant Professor of Public Health at Columbia
University, presented a brief summary of a commissioned paper
(Appendix D) that discusses representative studies of the effects of
publicly funded biomedical research on a range of outcomes. Public
funding accounts for about one-third of all biomedical and health
research, with NIH-sponsored research accounting for most of the federal
component along with additional investments by NSF, DOE, DOD,
USDA, and other agencies. In 2007, funding for biomedical research
totaled slightly more than $100 billion.
Sampat showed a stylized albeit simplified view of the innovation
system in which publicly funded R and D leads to improvements and
efficiencies in the private sector, to new drugs and devices, and ideally to
improved health outcomes (see Appendix D, Figure D-1). This flow of
knowledge occurs through many channels. One channel encompasses
25
OCR for page 26
26 MEASURING THE IMPACTS OF FEDERAL INVESTMENTS IN RESEARCH
publications, conference presentations, markets, and informal networks.
A second channel is through the creation of prototypes for drugs and
devices. Since the Bayh-Dole Act of 1980, these prototypes have tended
to be developed in universities and licensed out to firms to turn them into
successful products. A third channel includes funding for clinical trials
and clinical research that informs clinical practice— such as the
knowledge that doctors should give people an aspirin after a heart attack
— along with funding of other applications-oriented work, such as
contracts to fund the development of technologies and to conduct
consensus conferences.
Sampat called attention to another impact of new biomedical
technologies that is being discussed among health policy researchers.
Most economists believe that biomedical technologies are the biggest
source of long-run increases in health care costs. The clinical value from
these technologies may exceed their costs, but technology-driven cost
increases may be unsustainable, Sampat observed.
The Case of Cardiovascular Disease
Sampat described some of the literature on improvements in health
outcomes that can be traced to research. Cutler and Kadiyala (2007)
looked at improvements in cardiovascular disease mortality over the five
decades beginning in 1950, when mortality fell by two-thirds. They
concluded that about one-third of the advance is attributable to new high-
technology treatments, one-third to new drugs, and one-third to
behavioral changes such as not smoking and not eating salty or fatty
foods. Using a standard evaluation of $100,000 per year of life used by
health economists, they then computed the rate of return on investments
in treatments. New treatments provided a 4-to-1 rate of return, while new
behavioral knowledge produced a 30-to-1 rate of return. According to
this paper, Sampat said, “the publicly funded R in R and D has been
worth it.”
This paper makes little mention of NIH or public research except for
NIH’s sponsorship of large epidemiological trials and conferences,
which makes it hard to trace outcomes back to basic research. Another
issue, said Sampat, is the counterfactual: What would have happened in
cardiovascular disease absent any public funding in that area?
A paper by Heidenreich and McClellan (2007) focused on
improvements in heart attack care. These authors go farther than Cutler
and Kadiyala in relating changes in clinical practice to specific outputs of
OCR for page 27
27
IMPACTS ON BIOMEDICAL AND HEALTH RESEARCH
R and D. The authors concluded that the medical treatments studied in
clinical trials accounted for much of the improvement in heart attack
outcomes. The challenges with this paper include the fact that the authors
generally did not trace changes in clinical practice back to basic research.
Also, clinical practice often leads to publicly funded R and D because
informal learning by clinicians generates important research questions.
This learning is often subsidized by Medicare payments to teaching
hospitals and other non-research sources. Finally, clinical trials can lead
to negative results and lead clinicians to stop doing things they were
doing, which can be an unmeasured benefit to research.
Other Disease Categories
A statistical study by Manton et al. (2009) related mortality rates in
four disease areas to lagged NIH funding for the relevant institutes from
1954 to 2004. For two of the diseases studied—heart disease and stroke
— the authors found a relationship between funding and outcomes. For
the two other diseases— cancer and diabetes - the evidence was weaker.
But relying on funding aggregated by institute is difficult, as an institute
can fund widely varying research. Also the counterfactual is hard to
demonstrate since many factors could be driving changes in disease
rates.
Over this time period, competing risks changed. One reason for the
absence of a decline in cancer mortality— and maybe even an increase—
is that fewer people are dying of heart disease, so they live longer and are
more likely to develop cancer.
Relationship of Public and Private R and D
Papers by Toole (2007) and by Ward and Dranove (1995) sought to
relate public sector R and D to private sector R and D and found strong
evidence that they are complements rather than substitutes, in that public
research tends to spur private research. Private and public R and D in a
given area could be driven by scientific opportunity. For both forms of
research, there are challenges linking R and D to health outcomes,
Sampat observed.
Another line of research regarding private sector R and D is whether
proximity to public sector scientists makes firms more productive. A
range of studies have indicated that the answer is probably “yes,” Sampat
OCR for page 28
28 MEASURING THE IMPACTS OF FEDERAL INVESTMENTS IN RESEARCH
said, especially survey research asking firm R and D managers how
much they rely on public sector R and D.
Sampat noted that in surveys the drug industry reports greater
reliance on public sector R and D than do other industries. In contrast,
the device industry tends to be at or below the mean in terms of reliance
on public sector R and D. The drug industry relies mostly on medicine,
biology and chemistry. The device industry relies on medicine and
biology and, third, on materials science, which tends to be funded by
NSF and DOD.
Sampat then turned to drug and device innovation. Very recent
studies have used accounting methodologies to look at, for example, the
impact of public sector R and D in producing drugs that are then
marketed. In a study of drugs in FDA’s Orange Book, about 10 percent
of marketed drugs come from universities or public laboratories,
meaning that these institutions hold key patents (Sampat and
Lichtenberg, 2011). The number is higher, about 20 percent, for
clinically important drugs.
The Case of HIV Drugs and Vaccines
HIV is a special case, Sampat observed. The role of the public
sector in directly generating new drugs is much higher in HIV than in
other arenas; nearly one third of drugs in this area rely on public sector
research. Also, nearly all commercially and therapeutically important
vaccines over the last 25 years have come from the public sector,
according to Stevens et al. (2011). Surprisingly, efforts to relate funding
by disease area to later drug innovation tend not to show much of an
effect.
Device Development
In the areas of devices, Sampat described a case study by Morlacchi
and Nelson (2011) on the development of the left-ventricular assist
device (LVAD). The scientific understanding of heart failure remained
quite weak throughout the period that the LVAD was developed. But
NIH was holding consensus conferences to diffuse best practice and
contracting with firms for device development and clinical trials. “The
more applied side of the activities seems to be important” in this case,
Sampat concluded.
OCR for page 29
29
IMPACTS ON BIOMEDICAL AND HEALTH RESEARCH
Conclusion
The literature shows “consistent evidence of public sector funding
on private sector innovative effort,” Sampat concluded. The literature
also shows that public sector R and D has been important in the
generation of a non-trivial number of important drugs. However, it shows
less impact on other innovative outputs.
“There is surprisingly little research on the health benefits of public
sector biomedical R and D,” Sampat observed. Most of the evidence to
date is from the cardiovascular area. In addition, case studies point to the
importance of public clinical research, applied research, and diffusion
activities. Devices have important differences from drugs. And despite a
good deal of discussion, there has not been much study of the effects of
public sector research on health costs.
THE VOLATILITY OF FEDERAL R AND D SUPPORT
Richard Freeman, Herbert Ascherman Chair in Economics at
Harvard University, addressed the unintended effects of variability in
federal government funding for R and D. Using changes in the budgets
of the National Institutes of Health as an example, Freeman said that
chief among these effects is the damage done to people’s careers by
changes in grant rejection rates and increased uncertainty about future
career prospects. Scientific careers “looked dicey” even after the Wall
Street implosion and lay-offs in banking and consulting made finance
less attractive.
Funding variability may also affect the productivity of scientific
research. His study of the recent doubling of the NIH budget found that
before the doubling period more papers were produced per dollar of
grant than when more money was available. This decline in marginal
productivity may make it easier to cut future funding due to “failing to
meet ‘promises,’” Freeman stated. By contrast, the private sector has not
been as variable in its R and D support.
Gaps in Monitoring Science
Finally, Freeman suggested that the scientific community needs to
do a better job of monitoring the state of science. For example, non-
traditional measures of the supply of jobs might include real-time data
from Internet job boards, searches for information about science and
OCR for page 30
30 MEASURING THE IMPACTS OF FEDERAL INVESTMENTS IN RESEARCH
engineering jobs, and databases on Ph.D. dissertations. Downloads of
working papers could indicate hot areas of research. Online science and
social discussion groups and web-based communications from meetings
and conferences could contain information useful to the policy
community. Companies and other institutions should be accessing these
databases regularly, he said.
Information on what industry is doing is weak. The aggregate
amounts of money spent do little to map the steps to innovation. Even the
NSF BRDIS survey provides little data beyond the amounts of money
spent. Further, basic and applied research tend to be artificially divided,
but anything that is an innovation is going to go back and forth between
the two categories of research, Freeman observed.
MEDICAL DEVICE INNOVATION
Many people assume that the development of biomedical devices is
similar to drugs, said Paul Citron, retired Vice-President at Medtronic,
Inc., and now at the University of California, San Diego; but in fact “they
have very different characteristics as they traverse the pathway from
bench to bedside.” Drugs tend to be more discovery-based and derived
from in-house activity. Devices are engineering-based. A specification is
generated, along with an idea of how to realize that specification.
Moreover, devices evolve over time. The first device is very different
from subsequent generations, whereas a drug tends to be static for its
lifetime.
For devices, the timelines are longer and the markets are smaller
than in the pharmaceutical industry, Citron explained. It is very rare for a
medical device to have a billion dollar market, unlike pharmaceuticals.
The evolution of a device can be heavily influenced by federally
funded research, according to Citron. For example, research can enable
an industry to bring a device from concept to clinic. Federal funding can
build the underlying knowledge needed to make a technology safe and
effective. Federal research also can yield new materials, whereas the
complexity and cost of coming up with a new biomaterial to be
implanted in the human body can be beyond the ability of any one
company. Clinical trials may be crucial in improving a device.
A successful outcome for interventions using a medical device
depends on rigorous manufacturing, which can be improved through R
and D cycles involving federal research. Most medical product recalls
OCR for page 31
31
IMPACTS ON BIOMEDICAL AND HEALTH RESEARCH
are due to manufacturing issues that arise after approval, which can be
reduced through R and D.
Vivariums at academic centers are another crucial investment
underwritten by federal support. Prototype products are often tested at
these institutions, and even large companies may need to use academic
centers for access to animals.
Finally, “probably the most important output of federal inputs,” said
Citron, is students. “We hire the products of the campus” because “that is
where the intellectual horsepower for tomorrow resides.”
Citron listed four criteria an industry uses to decide whether to
pursue a project. (1) Does the technology fit with a company’s internal
capabilities? (2) Is the fit with the customer good? (3) What is the market
opportunity, including the number of customers, price, and the details of
application? (4) Finally, what is the time to market, including the time
needed to satisfy the regulatory process?
MAKING DECISIONS IN THE PHARMACEUTICAL INDUSTRY
The pharmaceutical industry and regulatory bodies need to evaluate
drugs thoroughly and expeditiously as they go through years of clinical
development before gaining approval for use in the treatment of a
particular disease state, observed Dr. Garry Neil, Corporate Vice
President for Science and Technology at Johnson and Johnson. Dr. Neil’s
company discovers and develops therapeutic products and technologies
that are evaluated by regulatory agencies around the world to assess the
efficacy and safety of a product for its intended use. “We have set the bar
very high for ourselves [about] what is expected and what we need to
deliver to our stakeholders, and we take that very seriously.”
Drug discovery spans years of study or phases of study, from
prediscovery to post-marketing surveillance, which allows for continued
follow up in a real-world setting after a therapy had been approved – but
getting to the point of approval can be challenging. Typically thousands
of compounds are synthesized to yield just a few potential candidates
that enter preclinical study, and for every five thousand to ten thousand
synthesized compounds, one approved drug on average may emerge.
And for many reasons, the costs to develop new drugs have risen
precipitously, which is further complicated by the fact that fewer drugs
are commercially successful “Despite all this, we continue to press very
hard because we recognize that there is unmet need and there are
OCR for page 32
32 MEASURING THE IMPACTS OF FEDERAL INVESTMENTS IN RESEARCH
financial rewards for real innovation that can really help people, even if
it’s the exception rather than the rule,” said Neil.
In recent years, public confidence in the pharmaceutical industry
and in the regulatory system has eroded. This may have the effect, if the
regulatory process is lengthened, of delaying the introduction of
innovative products or adding additional expense to the process and
ultimately the final approved product or medicine. “We can’t sacrifice
rigor, and no one is suggesting that, but we need to recognize the
consequences of raising the regulatory bar.”
To improve both productivity and regulatory certainty, said Neil,
work needs to continue on understanding basic biology. “It’s not easy,”
he said. “This is going to require a lot of collaboration between industry
and academia.” In addition, a new tool set is needed for drug discovery
and development as it relates to translational medicine, and these tools
need to be customized for particular diseases to increase the likelihood of
an efficacious therapeutic agent for a particular disease.
The United States should invest in an infrastructure akin to the
Internet or the interstate highway system in which it would be possible to
enroll patients in clinical trials much more rapidly, whether for drug
trials, observational studies, investigations of medical devices, or other
research. Only 3 percent of cancer patients enroll in clinical trials today.
“We make it inconvenient for them. Do we need an institutional review
board in every university? Why can’t we have national review boards?
Why can’t we have national safety monitoring committees? Why can’t
we bring the cost down and make the efficiency much better? Why can’t
we include patients of color, women, and older people? We’re not
getting those people today.”
The nation needs a more sophisticated and effective safety and
performance monitoring system for drugs once they enter the market.
And, most important, said Neil, health care providers need a system to
provide them with the latest information at the point of care to help them
make the best possible decisions for each individual patient.
FDA regulates 25 percent of the U.S. economy, representing over
$1 trillion worth of spending and a third of all the imports, with just
11,000 people and a $3 billion budget. “They need help,” said Neil,
including contemporary tools and techniques for pre- and post-marketing
evaluation. They also need new risk assessment tools and much better
engagement of patient communities.
OCR for page 33
33
IMPACTS ON BIOMEDICAL AND HEALTH RESEARCH
“The standard way of looking at this is to talk about risk and
benefit,” said Neil. “What is the benefit of the treatment? What is the
risk? . . . I think a better way of looking at this is risk and risk. There is a
risk of not treating a disease. What is that risk? Then there is a risk of
treating the disease. What is that risk, and what does that risk ratio mean
in the minds of the patient?”
RESEARCH AND OUTCOMES CASE STUDY: PEDIATRIC HIV
Laura Guay, Research Professor at the George Washington
University School of Public Health and Health Services, provided a
perspective on research funding and evaluation by a philanthropic
foundation. As vice president for research, she spoke about the work of
the Elizabeth Glaser Pediatric AIDS Foundation, founded in 1988 to
prevent HIV infection and eliminate AIDS among children in the United
States and abroad through research, advocacy, and treatment programs.
Early on, the foundation studied how children are infected, how
many children are infected, and why children are infected, chiefly
through “scientist awards” to encourage young investigators to develop
their careers in this less known field. The awards have provided
$750,000 to individual investigators for capacity building rather than for
specific research questions. Since 2007 the foundation also has made
operations research grants to improve treatment program design and
scale-up. As is often the case with medical research, there are obstacles
to the delivery of science into the field, especially in developing
countries, Guay noted.
Guay said that the foundation chooses innovative studies that are
less likely to be funded through NIH. “Why isn’t this fundable by the
NIH” is one question on its application. For example, while funding for
HIV vaccine-related studies is plentiful, very few of these funds focus on
a vaccine in infants born to breast-feeding mothers. The foundation also
may fund young investigators who do not have sufficient credentials to
compete successfully for NIH grants.
To measure the impact of its research investments, the foundation
needs performance metrics for deciding the impact of that funding, Guay
observed. For example, an important question has been how awards have
leveraged additional funds. Dating from the first funding of scientists in
1996, the foundation identified early leading scientists, which has
OCR for page 34
34 MEASURING THE IMPACTS OF FEDERAL INVESTMENTS IN RESEARCH
generated an “exponential increase” of originally small investments over
time.
Guay described two examples of the foundation’s operations
research, both influential in improving maternal HIV diagnosis and
antiretroviral treatment in African countries. The first involved a
controlled experiment in training nurses in the appropriate follow-up to a
positive diagnosis of the infection. The second involved an experiment in
rapid syphilis testing in connection with rapid HIV testing. Both projects
provided evidence for methods of identifying more infected women,
preventing transmission to their babies, as well as attracting men for
testing and treatment.
Guay concluded her remarks with an illustration of successful
application of research results that was recalled several times later in the
workshop (Figure 4-1). In 1994, when research results showed that
treatment of pregnant women with antiretroviral drugs could prevent
babies from being born infected with HIV, the number of perinatally
acquired AIDS cases dropped from approximately 900 per year in the
United States to virtually zero over the next decade and a half. This
dramatic outcome depended on progress dating to well before that date in
human capacity, laboratory capacity, and clinical capacity, Guay
observed. It is important to consider “all of the pieces that had to be in
place” as “we continue to eliminate pediatric HIV in the rest of the
world.”
Sampat asked how the Glaser Foundation allocates funds among
basic studies, vaccine development, and operations research, and Guay
said that in the early years the foundation considered its funds
unrestricted. But as more work has been funded by NIH and others, the
trend has been to “donor-driven” funding for particular projects or areas.
Because “people believe NIH has a lot of money,” it is harder to raise
foundation funds for basic research. And the biggest challenge, said
Guay, is “we have a lot of science we haven’t figured out how to
deliver.”
OCR for page 35
35
IMPACTS ON BIOMEDICAL AND HEALTH RESEARCH
1,000
800
No. of Cases
600
400
200
0
1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007
Year of Diagnosis
FIGURE 4-1 After it was shown that treatment of pregnant women with
antiretroviral drugs could prevent babies from being born infected with HIV, the
number of perinatally acquired AIDS cases in the United States and dependent
areas dropped precipitously.
SOURCE: Guay, 2011
DISCUSSION
Given the “exceptional return” of HIV-AIDS research in the United
States, “we need more case studies on failures” to figure out why some
research avenues have not been more productive, said Sampat. Also, in
evaluating the outcomes of public sector health research, it is hard to
aggregate across disease areas. Sampat cited the Research, Condition,
and Disease Categorization (RCDC) database started by NIH in 2009,
which reports on 229 diseases and research areas of interest to Congress,
but “these are not necessarily the diseases of historical interest to
economists and policy analysts.”
Kai Lee of the Packard Foundation, who spoke later in the
workshop, asked if the data show “there is a lot more to be gained in the
biomedical field from behavior-focused research?” He noted that
Freeman, Citron, and Guay had all suggested the importance of human
and institutional elements to outcomes. Freeman agreed that institutional
and behavioral factors are important in the environmental area; they
appear frequently on NIH’s list of grand challenges as well. For example,
OCR for page 36
36 MEASURING THE IMPACTS OF FEDERAL INVESTMENTS IN RESEARCH
the biggest success in preventing cancer has been behavioral, with
regulatory, marketing, and other factors all working to reduce smoking.
Citron pointed out that that over time the optimal ratio of
biomedical research to behavioral change could change. For example,
though cigarette smoking has declined, many aspects of diet still need to
change to improve health.
