A vast amount of valuable biochemical know-how is embedded in genes and in the complex biochemical webs that they create and control. Does the fact that nature invented it mean that it’s up for grabs? Lots of people are eager to grab it, Washington’s aspiring managers of our health-care economy now prominent among them.
At issue in the Myriad Genetics case decided by the Supreme Court in June were patent claims involving two genes, BRCA1 and BRCA2. Myriad has spent $500 million analyzing thousands of samples submitted by doctors and patients in search of BRCA mutations that sharply increase a woman’s risk of developing breast or ovarian cancer. Whenever the company found a mutation that it hadn’t seen before, it offered free testing to the patient’s relatives in exchange for information about their cancer history. As a result, Myriad can now predict the likely effects of about 97 percent of the BRCA mutations that it receives for analysis, up from about 60 percent 17 years ago.
In Myriad, all nine justices agreed that merely being the first to isolate a “naturally occurring” gene or other “product of nature” doesn’t entitle you to a patent. That came as no surprise. A year earlier, in Mayo v. Prometheus Labs, the Court had rejected a patent for a way to prescribe appropriate doses of certain drugs by tracking their metabolites in the patient’s blood, reaffirming long-standing rules that patents may not lay claim to a “law of nature,” “natural phenomenon,” or “abstract idea.” Previous rulings by the federal circuit court that decides patent appeals had reached similar conclusions in addressing attempts to patent all drugs that might be designed to control specific biochemical pathways. A description of a biological “mechanism of action” without a description of a new device or method to exploit it in some useful way is merely a “hunting license” for inventions not yet developed. A patent must describe “a complete and final invention”—a drug, for example, together with a description of the disorder that it can cure.
But very often, much of the cost of inventing the patentable cure is incurred working out the non-patentable molecular mechanics of the disease because all innovation in molecular medicine must in some way mimic or mirror molecular mechanisms of action already invented by nature. And spurred by the enormous promise of precisely targeted molecular medicine, a substantial part of our health-care economy is now engaged in working out the molecular mechanics of diseases. Washington is funding genomic research projects. Drug companies are heavily involved, joined by a rapidly growing number of diagnostic service companies. Doctors are gathering reams of new molecular data, patient by patient. Hospitals are mining their records for internal use and for sale to outsiders. Device manufacturers are racing to provide molecular diagnostic capabilities directly to consumers. Private insurance companies are mining the information they receive when claims are filed. And there are many signs that Washington intends to take charge of all of the above, as it tightens its grip on diagnostic devices and tests, what doctors diagnose, which diagnoses insurers cover at what price, and how the information acquired is distributed and used.
The patentability of genes is thus only one piece of a much broader debate about who will own and control the torrents of information that we have recently begun to extract from the most free, fecund, competitive, dynamic, intelligent, and valuable repository of know-how on the planet—life itself. That the private sector is already actively engaged in the extraction and analysis is a promising sign, but getting it fully engaged will require robust intellectual property rights, framed for a unique environment in which every fundamentally new invention must be anchored in a new understanding of some aspect of molecular biology. Individual gene patents are out of the picture now, but other forms of intellectual property already provide some protection. We should reaffirm and expand them. And we should view Washington’s plans to take charge instead for what they are: the most ambitious attempt to control the flow of information that the world has ever seen.
Drug companies began systematically exploring the molecular mechanics of diseases more than 30 years ago. Sometimes a gene itself is the essence of the cure. The insulin used by diabetics was extracted from pigs and cows until Genentech and Eli Lilly inserted the human insulin gene into a bacterium and brought “humulin” to market in 1982. Other therapies use viruses to insert into the patient’s cells a gene that codes for a healthy form of a flawed or missing protein. Recent “cancer immunotherapy” trials have shown great promise: the patient is treated with his or her own immune-system cells, genetically modified to induce them to attack their cancerous siblings. More often, a drug is designed to target a protein associated with a specific gene. “Structure-based” drug design and the biochemical wizardry used to produce monoclonal antibodies allow biochemists to craft molecules precisely matched to a target protein that plays a key role in, say, replicating HIV or a cancer cell. An FDA official recently estimated that 10 percent to 50 percent of drugs in pharmaceutical companies’ pipelines involve targeted therapies, and about one-third of new drugs approved by the FDA last year included genetic patient-selection criteria.
The development and use of all such therapies hinge, however, on understanding the roles that genes and proteins play in causing medically significant clinical effects. And as Myriad’s huge database illustrates, what looks like a single disorder to the clinician can often be caused by many different variations in genes that may interact in complex ways and that sometimes change on the fly, as they do in fast-mutating cancer cells or viruses like HIV. The molecular-to-clinical links are still more complex when drugs are added to modulate one or more of the patient-side molecules and unintended side effects enter the picture.
The databases and sophisticated analytical engines that must be developed to unravel these causal connections usually go far beyond “abstract ideas” or what any scientist would call a “law of nature.” To guide the prescription of HIV drug cocktails, Europe’s EuResist Network draws on data from tens of thousands of patients involving more than 100,000 treatment regimens associated with more than a million records of viral genetic sequences, viral loads, and white blood cell counts. Oncologists now speak of treatment “algorithms”—sets of rules for selecting and combining multiple drugs in cocktails that must often be adjusted during the course of treatment, as mutating cancer cells become resistant to some drugs and susceptible to others. IBM recently announced the arrival of a system to guide the prescription of cancer drugs, developed in partnership with WellPoint and Memorial Sloan-Kettering and powered by the supercomputer that won the engine-versus-experts challenge on Jeopardy. It has the power to sift through more than a million patient records representing decades of cancer-treatment history, and it will continue to add records and learn on the job.
Other companies are vying to combine efficient DNA-sequencing systems with interpretive engines that can analyze large numbers of genes and the clinical data needed to reveal their medical implications at prices that rival what Myriad is charging to analyze just two genes. Last year, 23andMe, a company founded to provide consumer genetic-sequencing services, announced that it would let other providers develop applications that would interact with data entrusted to 23andMe by its customers. Hundreds soon did. Their interests, Wired reported, included “integrating genetic data with electronic health records for studies at major research centers and . . . building consumer-health applications focused on diet, nutrition and sleep.” For individuals, 23andMe’s platform will, in the words of the company’s director of engineering, serve as “an operating system for your genome, a way that you can authorize what happens with your genome online.” Numerous websites are already coordinating ad hoc “crowd-sourced” studies of how patients respond to treatments for various diseases. The not-for-profit Cancer Commons is pursuing an “open science initiative linking cancer patients, physicians, and scientists in rapid learning communities.”
How much protection, if any, these databases, online services, and analytical engines will receive from our intellectual property laws remains to be seen. Indeed, where Myriad leaves the thousands of gene patents already granted is unclear—it will take years of further litigation to find out. Genes themselves or their biochemical logic are routinely incorporated into a wide variety of medical diagnostic tests and therapies; other sectors of the economy use genetically engineered cells, plants, and live organisms. Most such applications of genetic know-how will probably remain patentable because the constituent parts of an innovative product or process need not be patentable in themselves. Innovative methods for sequencing genes or analyzing their clinical implications will also remain patentable. What does seem clear is that in Myriad, the Court went out of its way to reaffirm and even broaden the scope of its earlier Mayo decision: a claim involving nothing more than the mechanistic science or an empirical correlation that links molecular causes to clinical effects isn’t patentable.
From the innovator’s perspective, however, patents that cover biological know-how only insofar as it is incorporated into an innovative drug or a diagnostic device provide little, if any, practical protection for what is often a large component of the ingenuity and cost of the invention. The successful development of a pioneering drug reveals key information about the molecular mechanics of a disease and a good strategy for controlling it. Armed with that knowledge, competitors can then modify the drug’s chemistry just enough to dodge the pioneer’s patent and rush in with slightly different drugs developed at much lower cost. In the end, the pioneer can easily be the only player that fails to profit from its own pathbreaking work. A Japanese researcher extracted the first statin from a fungus and discovered that it inhibits an enzyme that plays a key role in cholesterol synthesis; a colleague established its efficacy in limited human trials in the late 1970s. Following that lead, others quickly found or synthesized slightly different statins that worked better and had fewer side effects. Pfizer’s Lipitor, which was licensed two decades later, became the most lucrative drug in history.
“Repurposing” presents the flip side of the same problem. Nature often uses the same or very similar molecules to perform different functions at different points in our bodies. These molecules may then be involved in what used to be viewed as different diseases, and the same drug can then be used to treat them all. But when independent researchers or practicing doctors discover the new use, they usually lack the resources to conduct the clinical trials needed to get the drug’s license amended to cover it and can’t, in any event, market the drug for the new use so long as the patent on its chemistry lasts. And for both independent researchers and drug companies themselves, there is no profit in searching for new uses for an old drug unless there is a profitable market ahead that will cover the cost of the search. The new use may be entitled to what is called a “method” patent, but the patent will easily be dodged if the original patent has expired and cheap generic copies of the drug are readily available for doctors to prescribe.
Similar valuable spillovers occur as drug companies and others catalog the genes that determine how drugs interact with molecular bystanders. Countless patients owe a large debt to those who established that genetic variations in one group of enzymes cause some people to metabolize antidepressants, anticoagulants, and about 30 other types of drugs too quickly, before the drugs have a chance to work, and cause others to metabolize them too slowly, allowing the drugs to accumulate to toxic levels. The discovery of a new drug-modulating gene will often help improve the prescription of other drugs, too. Here again, only a fraction of the value of working out the link between variations in a particular gene and a drug’s performance will be captured by the company that discovers the link and incorporates that information into a first drug’s label.
Innovative diagnostic services are particularly vulnerable to free riders because their sole purpose is to convey information. Soon after it became clear that the Supreme Court would probably invalidate Myriad’s gene patents, two academic researchers teamed up with a gene-testing company and a venture-capital firm to buy clinical records from doctors who have treated patients using reports supplied by Myriad. A website set up on the day that the Supreme Court released its Myriad ruling invites patients to give away the same data instead. Neither scheme will create any new know-how; to the extent that they succeed, both will simply replicate Myriad’s database.
Washington is now hatching plans to free all the biological information that (in Washington’s view) deserves to be free. It’s also trying to make sure that we don’t misunderstand or even bother collecting the information that doesn’t.
To facilitate the development of “a new taxonomy of human diseases based on molecular biology,” a 2011 report commissioned by the National Institutes of Health recommends creation of a broadly accessible “Knowledge Network” that will aggregate data spanning all the molecular, clinical, and environmental factors that can affect our health. Last June, the NIH expressed strong support for a plan to standardize the collection, analysis, and sharing of genomic and clinical data across 41 countries. The 2011 report endorses government-funded pilot programs; the curtailment of at least some existing intellectual property rights (though “guidelines for intellectual property need to be clarified and concerns about loss of intellectual property rights addressed”); and “strong incentives” to promote participation by payers and providers. Indeed, the government may “ultimately need to require participation in such Knowledge Networks for reimbursement of health care expenses.” And data-sharing standards must be framed to discourage “proprietary databases for commercial intent.”
Somewhat paradoxically, the report hastens to add that its proposals extend only to the “pre-competitive” phase of research, presumably to leave some room for intellectual property rights directly tied to diagnostic devices and drugs. But there is no such phase—in pursuit of better products and services, the private sector is already competing to do most everything that the report describes. And a new molecular taxonomy of disease is already emerging: a steadily growing number of yesterday’s diseases now come with prefixes or suffixes that designate some biochemical detail to distinguish different forms—for example, “ER+,” for “estrogen-receptor-positive,” in front of “breast cancer.”
Meanwhile, Washington’s paymasters are busy solidifying their control of much of the data gathering and sharing. The Preventive Services Task Force decides which screening tests must be fully covered for which classes of patients by all private insurance policies, and which should be skipped for either medical or cost reasons. What insurers may charge is regulated, too, so more money spent complying with screening mandates will inevitably mean less spent on disfavored tests and associated treatments. Washington also has ambitious plans to take charge of pooling and analyzing the data that emerge. A new national network, overseen by a national coordinator for health information technology, will funnel medical data from providers and patients to designated scientists, statisticians, and other public and private providers and insurers.
For its part, the FDA has made clear that it intends to maintain tight control of the diagnostic services and devices that enable individuals to read their own molecular scripts. In 2010, Walgreens abruptly canceled plans to sell a test kit called Insight when informed by the FDA that the kit lacked a license. The mail-in saliva-collection kit would have told you what your genes might have to say about dozens of things, among them Alzheimer’s, breast cancer, diabetes, blood disorders, kidney disease, heart attacks, high blood pressure, leukemia, lung cancer, multiple sclerosis, obesity, psoriasis, cystic fibrosis, Tay-Sachs, and going blind—and also how your body might respond to caffeine, cholesterol drugs, blood thinners, and other prescription drugs. Invoking its authority to license every diagnostic “contrivance,” “in vitro agent,” or “other similar or related article,” the FDA announced a crackdown on all companies that attempted to sell such things to the public without the agency’s permission. The agency is determined to protect consumers from what it considers to be “unsupported clinical interpretations” supplied by providers or medically inappropriate responses to diagnostic reports by consumers themselves.
There is much reason to doubt, however, that Washington is qualified to teach the rest of the country how to gather or analyze biological know-how. For the last 50 years, the FDA, by scripting the clinical trials required to get drugs licensed, has controlled how those trials investigate the molecular factors that determine how well different patients respond to the same drug. The trials have, in fact, investigated appallingly little, because the agency still clings to testing protocols developed long before the advent of modern molecular medicine (see “Curing Diversity,” Autumn 2008). A report released in September 2012 by President Obama’s Council of Advisors on Science and Technology (PCAST) urges the FDA to adopt “modern statistical designs” to handle new types of trials that would gather far more information about the molecular biology that controls the development of the disease and its response to drugs. In a speech given last May, Janet Woodcock, head of the FDA’s Center for Drug Evaluation and Research, acknowledged the need to “turn the clinical trial paradigm on its head.”
More generally, Washington has been a persistent (and often inept) laggard in moving its supervisory, transactional, and information-gathering and dissemination operations into the digital age. The FDA’s “incompatible” and “outdated” information-technology systems, the PCAST report notes, are “woefully inadequate.” The agency lacks the “ability to integrate, manage, and analyze data . . . across offices and divisions,” and the processing of a new drug submission may thus involve “significant manual data manipulation.” Other parts of Washington launched plans to push digital technology into the rest of the health-care arena around 2005, encouraged by a RAND Corporation estimate that doing so would save the United States at least $81 billion a year. A follow-up report released early this year concluded, in the words of one of its authors, that “we’ve not achieved the productivity and quality benefits that are unquestionably there for the taking.”
Washington launched the modern era of comprehensive genetic mapping when it started funding the Human Genome Project in the late 1980s. Rarely has the federal government found a better way to spend $3 billion of taxpayer money. But far more mapping has been done since then with private funds, and finishing the job is going to cost trillions, not billions—far more than Washington can pay. In a welcoming environment, Wall Street, venture capitalists, monster drug companies, small biotechs, research hospitals, and many others would be pouring intellect and money into this process. The biosphere offers unlimited opportunity for valuable innovation. The technology is new, fantastically powerful, and constantly improving; the demand for what it can supply is insatiable. But Washington’s heavy-handed control of both the science and the economics of the data acquisition, analysis, and distribution will drive much of the private money out of the market. We should instead give the market the property rights and pricing flexibility that will get the private money fully engaged and leading the way.
As the Supreme Court has recognized, a property right that curtails the use of basic science may foreclose too much innovation by others later on. The discovery of a promising target, for example, shouldn’t be allowed to halt all competitive development of molecules that can detect or modulate it. Instead, we need rights that will simultaneously promote private investment in the expensive process of reverse-engineering nature and the broad distribution of the know-how thus acquired, so as to launch a broad range of follow-up research and innovation and allow front-end costs to be spread efficiently across future beneficiaries.
Models for property rights framed to strike such a balance already exist. As it happens, federal law already grants drug companies a copyright of sorts: a “data exclusivity” right that, for periods up to 12 years, bars the manufacturers of generic drugs from hitching a free ride through the FDA by citing the successful clinical trials already conducted by the pioneer. Data exclusivity rights could easily be extended to cover the development of biological know-how as well, enabling the market to spread the cost of developing such knowledge across the broad base of future beneficiaries. The rights should be narrowly tailored to specific diseases, but they should also last a long time.
The agency that issues data exclusivity rights should also have the authority to oversee a compulsory licensing process analogous to the one that allows singers to perform songs composed by others—or to the government’s “march-in” right to license patents for inventions developed with government funding to “reasonable applicants” if the patent holder has failed to take “effective steps to achieve practical application of the subject invention” or, more broadly, as needed to address the public’s “health and safety needs.” Licensing fees should be modest, and total licensing fees collected might be capped at a level commensurate with the cost of developing the biological know-how and its role in creating new value thereafter.
In addition to mobilizing private capital to develop know-how, well-crafted property rights promote its broad and economically efficient distribution. To begin with, they make the know-how immediately available—though not exploitable for profit—to researchers and competitors. Copyrights require publication of the protected content; patents require a description complete enough to enable others in the field to exploit the “full scope” of the claimed invention. Going forward, Myriad and companies like it will instead treat their discoveries and databases as trade secrets and frame contracts to forbid disclosure of their data by the doctors and patients who use their services.
As entertainment and digital software and hardware markets have also demonstrated, markets protected by the right forms of intellectual property usually find ways to develop tiered pricing schemes that allow widespread distribution of information at economically efficient prices while the rights last. By contrast, current policies in the health-care arena load most of the cost of developing biological know-how onto the shoulders of the patients who buy the pioneering drug or device while its patent lasts. The market will develop more pioneer drugs—and government paymasters will be more willing to accept them—if one important component of the know-how costs can be spread efficiently across many more products, services, and patients.
The emergence of a robust market for molecular and clinical information will also draw an important new competitor into drug markets: the independent researcher or company that develops the biological science needed to select the most promising targets and to design successful clinical trials for new drugs. Drug companies will inevitably continue to play a large role in the search for patient-side information. But for much the same reason that software companies welcome advances in digital hardware, drug companies should welcome the rise of independent companies with expertise in connecting molecular effects to clinical ones in human bodies. Independent companies are permitted to interact much more freely with doctors and patients. And by accelerating the development and wider distribution of information about the molecular roots of health and disease, companies that specialize in predictive molecular diagnosis will help mobilize demand for the expeditious delivery of drugs.
Finally, private markets have established that we do not need a government takeover to maximize the value of information by consolidating it efficiently in large databases. For well over a century, competing companies have been pooling their patents and interconnecting their telegraph, phone, and data networks, online reservation systems, and countless other information conduits and repositories. Aggregators buy up portfolios of patents and licenses for songs, movies, TV shows, and electronic books, and then offer access in many different, economically efficient, ways to all comers. The global financial network depends on instant data sharing among cooperating competitors. Many online commercial services today hinge on the market’s willingness to interconnect and exchange information stored in secure, proprietary databases.
With appropriate property rights in place, the economics of the market for clinical data will, in all likelihood, be largely self-regulating. The accumulation of new data points steadily dilutes the value of the old, though their aggregate value continues to grow. New data can be acquired every time a patient is diagnosed or treated. The best way to accelerate the process that makes information ubiquitously and cheaply available is to begin with economic incentives that reward those who develop new information faster and come up with ways to distribute it widely at economically efficient prices.
Conspicuous by its frequent absence in debates about who owns biological know-how is the patient’s right not to disclose information about his or her innards to anyone—the corollary of which should be a right to give it away, share it, sell it outright, or license it selectively, at whatever price the market will bear. The individual patient won’t often be interested in haggling over what a clinical record might be worth to a hospital or an insurance company. But emphatically reaffirming the private ownership of private information is the first, essential step in creating markets for those who would help collect and analyze the data. It would also help set the stage for some serious constitutional challenges to Washington’s attempts to displace them.
Washington assures us that patient privacy will be protected when it pools medical records for analysis by Washington-approved researchers and statisticians—but its policies reflect the conviction that the patient’s privacy rights will be bought by whoever is paying for the care, whether Washington itself or a private insurance company operating under Washington’s thumb. The Health Insurance Portability and Accountability Act of 1996 does indeed require that records, before they are shared, must be redacted to ensure that they can’t be linked back to individual patients. But Washington’s guidelines on “de-identification” of private health-care data are naively optimistic about what that would require. In a paper published in Science last January, an MIT research team described how easily it had used a genealogy website and publicly available records to extract patients’ identities from ostensibly anonymous DNA data. Genetic profiles reveal gender, race, family connections, ethnic identity, and facial features; they correlate quite well with surnames. DNA fingerprinting following an arrest is now routine, and a close match with a relative’s DNA can provide an excellent lead to your identity. Further, state-run clinics and hospitals are exempt from the federal requirements, and some are already selling patient records to outsiders, sometimes neglecting to remove information such as age, zip codes, and admission and discharge dates.
Even if privacy were fully protected, a promise to “de-identify” data doesn’t give the government the right to seize and distribute private information in the first place. Washington, it appears, now aspires to create what is, for all practical purposes, a federal mandate to participate in an insurance system in which Washington minutely regulates which diagnostic tests are administered and which treatments are provided—and then requires the sharing of the data acquired in the course of treatment. It seems unlikely that such a system can withstand constitutional scrutiny. A government grab for information, whether direct or through private intermediaries, is subject to the Fourth Amendment, which places strict limits on such invasions of our privacy.
At the opposite pole, the FDA’s decision to protect the masses from unapproved clinical interpretations of molecular data, or to limit or bar access to the information on the ground that the masses aren’t smart enough to handle it wisely, also looks constitutionally vulnerable. Home-use diagnostic devices and services offer patients the absolute medical privacy that they are entitled to. And the Supreme Court has repeatedly concluded that freedom of speech includes a private right to listen, read, and study that is even broader than the right to speak, write, and teach. Enabling tools and technology—the printing press and ink that produce the newspaper, for example—enjoy the same constitutional protection. Lower courts have recently also affirmed the First Amendment right to discuss off-label drug prescriptions. A constitutional right to engage in officially unapproved discussion about what a drug might do to a patient’s body must surely also cover unapproved discussion about what the patient’s own genes and proteins might do.
If we let them, markets for biological know-how will do exactly what markets do best and deliver what molecular medicine most needs. Propelled as they are by dispersed initiative and private choice, free markets are uniquely good at extracting and synthesizing information that’s widely dispersed among innovators, investors, workers, and customers—“personal knowledge,” in the words of Michael Polanyi, the brilliant Hungarian-British chemist, economist, and philosopher. Nowhere could the free market’s information-extracting genius be more important than in a market for products whose value depends on their ability to mirror biochemical information inside the people who use them.
The Supreme Court’s Myriad ruling is a reasonable construction of the patent law as currently written. But that we now find ourselves pondering whether Washington should be taking full control of the private genetic and clinical data that individuals supply—and that private companies like Myriad are eager to collect and analyze—provides a chilling reminder of how rapidly we are slouching down the road to medical serfdom.
Photo: Myriad Genetics spent half a billion dollars developing a database of mutations of the BRCA1 and BRCA2 genes. When proteins coded in the genes (as shown in BRCA1 above) fail to do their genetic repair work properly, cancer risk is increased. (DR. MARK J. WINTER/SCIENCE SOURCE)