CAP TODAY publisher
Bob McGonnagle spoke recently about the in vitro diagnostics industry
with Brian R. Buxton, principal and co-founder of Easton Associates, New
York, NY. In the interview, Buxton shared his thoughts on Wall Street's
new interest in the industry, personalized medicine, and more—from
his perspective as a former service provider and now consultant to the
Before establishing Easton Associates in 2000,
he spent four years as a consultant with the Wilkerson Group and two as
an independent consultant. Before consulting, Buxton was a corporate development
and marketing executive for 12 years with two commercial laboratories—Unilab
Corp., the largest clinical laboratory in California, and MetPath Inc.
(now Quest Diagnostics).
He also taught marketing planning to MBA students
as an adjunct professor of marketing at Columbia University's Graduate
School of Business. He holds an MBA with High Distinction (Baker Scholar)
from Harvard Business School.
Following is an abridged version of the interview
as published online by the Diagnostic Marketing Association.
Bob McGonnagle: Many of our readers may not
be familiar with Easton Associates, so tell us a bit about the company
and your position there.
Brian Buxton: Easton Associates is a boutique
consulting firm focusing solely on the medical products arena-pharmaceutical,
biotechnology, diagnostics, and medical device companies and their investors.
Our work includes commercial opportunity assessment, strategic planning,
product and partner searches, and due diligence.
The founders, of which I am one, all came from the
Wilkerson Group, a leading health care consulting firm for more than 20
years. My role in the company is to manage projects and to serve as the
industry expert for diagnostics and clinical laboratory testing.
McGonnagle: Can you comment on the atmosphere
in the diagnostics industry and Wall Street's new interest in IVD companies?
Are IVD companies like a rock band that just got discovered?
Buxton: The diagnostics industry is attracting
investor attention for a number of reasons: the interest of companies
like Siemens and GE in acquisitions, its sustained performance versus
disappointments in other sectors, and the attention around new biomarkers.
Siemens' acquisitions of Diagnostic Products Corp.,
Bayer Diagnostics, and now Dade Behring—despite relatively slow
growth in the core businesses—were driven in my view by a shift
in the source of competitive advantage from best test—all the tests
are accurate now—to hospital preferences for best system performance—lower
transaction costs, fewer vendors, and assured integration with IT and
radiology. The value of a diagnostics vendor now has more to do with its
ability to deliver information from the lab, via the hospital information
system, to clinical pathways and algorithms that facilitate treatment
While diagnostics industry veterans remember fondly
the days of many instrument companies and higher growth rates, current
growth rates even for routine diagnostics are actually quite attractive
to companies like Siemens, Philips, and GE, compared with other sectors
of the economy. And these companies really know how to deliver system
performance to hospitals.
But this is not the whole story. Growth has shifted
to a new set of tests, instruments, and companies, those involved in complex
tests—genomics and molecular diagnostics, anatomic pathology tests,
rapid infectious disease tests—and the eventual promise of tests
for personalized selection of therapeutic drugs.
There are many different types of these new, complex
tests. In less severe diseases and conditions, like hypercholesterolemia
and the risk of MI or stroke, we see a number of advanced lipid tests.
These are not really genomic tests, but they are able with great sensitivity
and specificity to stratify individuals at risk for cardiac events.
How big is the opportunity for these tests? It's useful
to compare them to the "previous solution"—the LDL test. Most primary
care physicians use a simpler approach in managing patients at risk for
MI or stroke: If the patient's LDL is 130 or higher, a statin prescription
and a return visit in six weeks to re-check the LDL. This may not be sophisticated,
but it works fairly well, for a lot of physicians.
But when we look at more severe conditions—cancer
is the most obvious one where new tests are used to match patients with
the appropriate drug therapy—we see much greater advantages over
the previous solution—dramatic improvements in life expectancy,
quality of life, and health care system costs. Add to this a system where
the actors—medical oncologists—are specialists well versed
in the clinical literature and the clinical trial results, and comfortable
working with statistics in making clinical decisions, so they are ready
to adopt and use more complex tests.
Also, payers are becoming much more interested in the
ability of tests to affect health system costs. Paying $25,000 to $50,000
a year for targeted therapies for patients with cancer has focused their
attention on whether these patients are getting the best therapy, and
whether there is a significant cost-to-benefit ratio.
Adjuvant chemotherapy is the standard of care for many
patients with node-negative primary breast cancer. While this involves
older, less expensive drugs, the system cost is still quite high, and
so a company that can identify when this treatment will likely be effective
and when it won't is going to get a lot of payer attention.
All of this illustrates what we believe to be a general
rule: Tests that provide relevant clinical information for severe diseases
and conditions, and for specialists, will have a much stronger value proposition
than those for less severe conditions and for primary care physicians.
One disappointing counter-example is in HIV. The viral
load tests are reasonably reimbursed, but the physician time spent interpreting
the results is not. Physicians who spend this time without reimbursement
get more and more HIV patients. These people are saints to work the hours
they do, often for $70,000 a year.
McGonnagle: What about personalized medicine—what
is your take in this area?
Buxton: It may be useful to split personalized
medicine into two different types of tests; both have seen a lot of activity
recently. The first is tests that identify people likely to have adverse
drug reactions. The other is tests that distinguish responders from nonresponders
to a specific drug or set of drugs. I think the adoption model for these
two types of tests will be very different. When the FDA talks about personalized
medicine, I think it is mainly interested in the adverse reactions side.
In the case of tests to identify people with adverse
reactions to a drug, there is a huge public health issue that has received
inadequate attention. A series of gene sequences and SNPs related to cytochrome
P450 [CYP 450] determine human metabolism of a very high percentage of
drugs in current clinical use.
Testing patients for their CYP 450 SNPs is a once-in-a-lifetime
test that costs maybe $300 to $800, much less if there were large test
volumes. Putting these results into the pharmacy benefit management systems
in pharmacies would identify people likely to have an adverse reaction,
or to get no benefit, at the prescribed dose. This would be an enormous
public health benefit. But so far, no one is willing to "prime the pump."
The pharmaceutical sponsors of these drugs may not
find these tests appealing because they address risks in using drugs that
are being marketed as very safe. And the insurance plans and managed care
organizations aren't sure the consumer will stay in their health plan
long enough to see a return on the investment in the test. But patients
stay a lot longer in Medicare and the VA, so they might be the first to
pursue covering these tests.
Promoting these tests would be a great thing for a
consumer health care company to do to enhance its public image. It might
not make a ton of money doing it, but it would be useful to provide the
test to the entire U.S. population and to have everyone's response profile
stored in some sort of privacy-protected record that would be connected
to all of the PBM computer systems in the U.S. So that when people go
to the pharmacy to get a prescription filled, the pharmacist would not
only catch drug interactions but also drug response information for that
specific person and drug.
McGonnagle: In practical terms, this aspect
of personalized medicine is not getting too much traction in clinical
routine practice at this time.
Buxton: That is an understatement. This is an
old story. We can say that this is a terrible situation, but why would
any company in the U.S. take upon itself the cost burden to champion something
that has no real expectation of financial gain, that may not work out,
and that may cause it embarrassment.
But I expect that some company or organization may
be able to figure out a decent way to do this because the need is compelling.
One naturally thinks that a company like J&J is a prime candidate to spearhead
this type of program. Theoretically, even a smaller company that is intensely
focused on this and is able to mount a campaign for decent reimbursement
should also be able to do this. This is something we might see in the
future of personalized medicine.
Doing this would also clarify the opportunity for the
other type of personalized medicine diagnostic testing—selecting
the best responders for a drug therapy. These tests have also had a disappointing
reception. The average person can see a clear value to providing the best
therapy for each individual, but pharmaceutical marketers have a hard
time seeing a favorable business case—it looks like this type of
test would take a drug that could have gotten 100 percent of the market
and leave it with 30 percent or 40 percent of the market. That's not likely
to be a career-enhancing move for a pharmaceutical marketer.
Yet there are signs of change, for several reasons.
First, the tests—usually molecular—have gotten better, more
accurate. Second, the science of drug development seems increasingly headed
in the direction of developing drugs that need tests like this. Third,
as the cost of clinical trials continues to rise, and the number of FDA—approved
NMEs [new molecular entities] continues to fall, this 30 percent to 40
percent of the market proposition starts to look better than no new drugs
at all, or a newly approved drug that cost a fortune to bring to market.
In addition, we are beginning to see a trend in reimbursement
for drugs, medical devices, and diagnostic tests based on relative efficacy.
So a "me too" drug, device, or test, even FDA cleared, may in the future
get either no coverage or a very low reimbursement rate. Payers are beginning
to ask to see clinical studies of comparative effectiveness against the
current best choices in the market. And here, too, a test that would identify
the best responders, or rule out non-responders or adverse-reactors, could
add considerably to the reimbursement of a new drug, by establishing a
strong comparative effectiveness versus drugs already on the market.
Finally, many are projecting that by January 2009,
the Democrats will control the White House and both houses of Congress,
that health care will be a major issue, and that it will be handled with
more skill than in 1993. In such an environment, diagnostic tests and
drugs that show greater comparative effectiveness may be the best kind
of product to have.
McGonnagle: Diagnostics is almost always in
the orbit of the pharmaceutical economic model. Is that fair to say?
Buxton: Sure. First, total spending on prescription
drugs in the U.S. is well over $200 billion, while spending in the U.S.
on lab tests is under $50 billion to the labs, and maybe $12 billion to
$15 billion is for diagnostics instruments and reagents, so roughly 15
times more money is spent on drugs than on diagnostics.
Second, having a drug to address the condition has
a dramatic impact on the ordering of a lab test. When people ask how good
is a new lab test, the old answer is the best: It depends on what the
physician will do differently with the information it provides. Ordering
or not ordering a drug is one of the easiest things the physician can
Look at what statins did for lipid testing—cholesterol,
HDL, LDL, and trigs, but especially LDL. Before drugs like Mevacor, there
was quite a bit of interest in lipid test results, and concerted promotion
of cholesterol LDL and HDL tests by the labs. But the only action physicians
could take was to urge their patients to make dietary and lifestyle changes.
Most of the time the patient didn't comply and nothing changed. Now patients
take their statins, the LDL level falls, the patient is happy, and the
physician is a hero. And the risks of MI and stroke across the population
But it's not just pharmaceutical therapies. Another
important "do differently" test would be one that, for people getting
coronary stents, could identify those with a high risk of thrombus. These
patients would get the older bare metal stents, and everyone else a drug-eluting
stent, but with less risk of the life-threatening thrombotic events that
have been in the news.
So in a broad way, diagnostic tests are maturing and
carrying more and more clinical and economic value, which brings us back
to your opening question about why the market is showing so much more
interest in diagnostics and is valuing this sector much more highly than
in the past.
McGonnagle: There is keen interest now in the
regulatory environment for IVDMIAs [In Vitro Diagnostic Multivariate Index
Assays]. Care to comment?
Buxton: I understand that ACLA and many other
groups had fairly negative feedback for the FDA at hearings on IVDMIAs.
I am not sure how the FDA is going to respond, but I don't think it will
give up easily. IVDMIA tests are new. They are not classical method tests
(like serum lead), nor are they analyte-specific reagents. Their results
are of major importance; for example, in cancer treatment, they may determine
the treatment a patient receives. And they are beginning to receive high
reimbursement rates. So the FDA feels that all this gives it a basis to
regulate these tests, to be sure the information they deliver is accurate.
The dilemma is how to do this. These are very complex
tests—combinations of gene sequences, single nucleotide polymorphisms,
haplotypes—and their diagnostic value lies not so much in the raw
results but in how these signatures correlate with patient outcomes. The
FDA is saying that it won't allow a test that simply provides a measure
of relative risk but leaves the underlying data in a sort of black box.
While this sounds reasonable, it gets complicated when you open this box.
McGonnagle: What implications would you say
provider consolidation has for the future of lab medicine?
Buxton: Lab consolidation has two aspects. One
is consolidation of routine testing—the chemistry, hematology, and
immunoassay tests that make up about over 60 percent of the testing revenue
at major commercial labs like Quest and LabCorp. The consolidation of
this type of testing has been going on for 30 years or more, and it is
continuing because the economics work. The more volume Quest and LabCorp
get, the more efficiently they can perform pickup, testing, reporting,
and billing. So I expect it to continue until it runs out of businesses
to acquire, and already there are fewer labs with enough routine business
to cover the cost of doing the transaction.
The other side is labs acquiring a more complex mix
of tests—esoteric, genomic, and anatomic pathology tests. This type
of testing is faster growing, higher-priced, and more profitable. And
economies of scale also apply to this type of testing.
We used to say that for every dollar of laboratory
testing costs, one-third was for specimen acquisition—sales, marketing,
and couriers; one-third was for results reporting, billing, IT systems,
and corporate overhead (including top management); and one-third was spent
performing the test—instruments, reagents, staff, and space.
So if that is how the business works, it is clear that
the firm that has economies of scale in activities such as nationwide
marketing, distribution, picking up specimens, and reporting will have
a significant competitive advantage. Most reference labs have these economies;
all that is missing is the lab technologists able to do these sophisticated
genomic and tissue-based tests. They clearly have been able to hire these
technologists and have also been able to attract scientists who have developed
high-quality methods by which they can do these tests at faster throughput,
lower cost, and greater capacity than any academic and hospital lab could
It's an attractive model for pathologists as well.
The career path for a pathologist used to be to join a pathology group
and buy into the partnership. Now a lot of pathologists go into commercial
labs and do quite well.
For all these reasons, the laboratory services sector
continues to do well. We did a study comparing revenue, operating profit,
and share price growth for the major commercial labs versus the leading
diagnostics companies. We found that the labs outstripped the diagnostics
companies on all three parameters.
McGonnagle: Would you say this has led to a
new business strategy?
Buxton: Not long ago, the strategy for a company
with a novel diagnostic test, before it went onto an instrument platform
as an FDA-cleared IVD kit, was not to do it yourself. It was simply too
difficult. There was the need for a nationwide sales force, a courier
network, specimen accessioning, a complex computer system, lab techs,
floor space, instruments, plus results delivery, complicated billing,
getting reimbursement coverage—a deep and complex value chain. So
the only sensible thing was to take it to Quest and LabCorp, who would
do all of this, and keep 70 percent to 80 percent of the revenue.
Then companies like Genomic Health, XDx, and others
challenged this model. They felt that the Quest and LabCorp current sales
forces might not be able to explain a novel, complex test to specialists.
They weren't sure these commercial labs would push that hard to maximize
the test's reimbursement. So they did a radical thing: They built the
entire value chain themselves, essentially from scratch. It seemed crazy
at the time, but the results so far are pretty impressive.
So we are seeing a new business model in its early
stages. It's called the CLIA lab, but the real point is to capture all
the value beyond the test itself—the other two-thirds of the expense
Now it helps when the test is very complex, very useful
clinically, high-priced, and only available from the new CLIA lab. It's
not going to work for simpler, more routine tests, nor will it work that
well for tests where empiric therapy—the statins and the LDL that
we talked about—will suffice. But it's likely to work very well
for tests for cancer and other high-severity conditions.