During the Internet boom, Marty Tenenbaum made a fortune creating online marketplaces. In 1992, his company Enterprise Integration Technologies carried out the first commercial transaction on the Internet: an email in which Tenenbaum bought a reprint of an academic paper. Two years later, he founded CommerceNet, an industry association that worked to speed development of technologies that could link, for example, an online bookseller to a credit card company to a freight shipper. By 1999, Tenenbaum was chief scientist at Commerce One, a stock market darling that focused on business-to-business networks through which a General Motors or an IBM could network with its thousands of suppliers. Along the way, he learned a lot about forging digital connections between people.
Eleven years ago, when he was 55 and enjoying his success, Tenenbaum found a lump under his arm. The tumor was from metastatic melanoma and Tenenbaum’s prognosis was dire: perhaps 12 months. Desperate, he trekked from one prominent oncologist to another. Each offered a different opinion on how to treat his cancer. None offered hope. But Tenenbaum turned out to be extraordinarily lucky. Professional connections got him in touch with some of the country’s top melanoma experts at the National Cancer Institute, and he found his way to a clinical trial of a melanoma vaccine. The trial failed when most of the other participants showed no improvement. But the vaccine seemed to work for Tenenbaum. He survived.
The ordeal left him happy to be alive, of course, but also appalled. Inside doctors’ offices, all he heard was that his options were limited. Outside, he believed he was witnessing a scientific revolution. At the turn of the millennium, just one year after Tenenbaum got his terrifying diagnosis, scientists had announced the first draft of a complete human genome. Medicine—and business—seemed poised to capitalize on a new understanding of the genetic code. There was new energy, new research, new diagnostics. But wherever Tenenbaum looked, physicians were not connected to scientists, patients were not connected to experts, and researchers were walled up in corporate and institutional silos. The interlinked global village had become a cliché, yet desperate cancer patients could not seem to find a connection anywhere. The genomics revolution seemed to be doing them little good.
In 2008, motivated by his ordeal, Tenenbaum did what a business leader knows how to do. He launched CollabRx, a company that takes disparate aspects of medical research and the health care industry—from raw data coming out of academic labs, to drug trial results from pharmaceutical companies, to discussions between doctors and patients—and creates virtual medical communities that match patients with the best available treatments while encouraging private companies to work together on new therapies. “We’re trying to position ourselves as a clearinghouse between the industry, which consists of hundreds or thousands of silos, and the doctors and patients who need these things,” Tenenbaum says.
The rationale behind Tenenbaum’s company is part of a broader trend in the business world, called Open Innovation, first enunciated by business professor Henry Chesbrough of the University of California, Berkeley in 2003. Chesbrough realized that to seize on the best ideas, and maximize profits, companies need to look for technological innovations both within and beyond their own walls. For example, a fast-growing number of “crowdsourcing” websites—Atizo, InnoCentive, and Amazon’s Mechanical Turk, to name a few—enable companies to post a particular problem and offer a reward for whoever comes up with the best solution. Corporate giants like BMW, Dell, Kraft Foods, and Starbucks have launched websites that allow customers to post and vote on product suggestions. For medical companies, the old model of secretive research, exclusive patents, and proprietary data made sense only a few years ago. But in this new era of genomic medicine, such proprietary models may hinder business, not further it.
By applying new thinking, business leaders in the genomics industry might do more than discover effective treatments. They might create a better position from which to navigate reimbursement and regulatory hurdles. And they might realize not just cures for people like Marty Tenenbaum, but profits.
When people talk about the genomics revolution, they're usually referring to the rise of "whole-genome sequencing"--the ability to read the 6-billion-letter "book" of our DNA.
It’s difficult to grasp the dimensions, elegance, and complexity of the human genome—the 6-billion-letter “book of life.” Were your two DNA strands to be stretched out in a line, they would be longer than six feet, yet they’re packed into the microscopic nucleus of each of your trillions of cells. The government-led Human Genome Project finished its sequence of a human genome in 2003, at a cost of some $3 billion. Dozens of companies have since joined the sequencing game, and now, for a small blood sample and less than $20,000, you can buy the full readout of DNA letters that defines your personal genome. Experts predict the price will drop to $1,000 within a couple of years.
But reading the sequence is not the same as knowing what to do with it. The difference is something like learning an alphabet versus fully understanding a language’s words, grammar, syntax, and idioms. Are irregularities in your personal sequence part of normal human variation, or telltale signs of disease? In most cases, researchers don’t know yet. Even before machines that could read every letter of an individual’s genetic book, researchers could detect one-letter “typos,” dubbed “SNPs,” that appear in small segments of the general population. By this approach, they identified lots of risk variants—SNPs that crop up more often in people with a certain disease. The trouble is, carrying these variants does not always result in illness. Paul Billings, director and co-founder of Omicia, a genome interpretation company based in Emeryville, California, drolly notes that were risk variants always indicative of disease, “we’d all be dead.” Now research companies have the technology to amass a staggering amount of data compared to the old methods. Yet they’ve learned that they still might not have nearly enough to generate knowledge useful to patients, doctors, or the company’s revenue stream.
Therein lies a business opportunity for anyone who can aggregate ever-larger datasets and mine them for useful patterns and correlations. Omicia does this by using large databases of genetic information built slowly over the past decade by government, academic, and industry scientists. Once the company’s researchers identify a significant pattern—such as a combination of mutations or genetic rearrangements that, when carried together, usually lead to a specific disease—they can look for those markers in a customer’s genome and make educated guesses as to how that variation will affect health. But even these massive databases rarely include detailed information about participants’ medical histories. This is important because the health consequences of a genetic glitch almost always depend on environmental factors such as fetal conditions, diet, medications, and even social interactions. The biggest business opportunity in personalized medicine, Billings says, lies in accelerating the creation and lowering the costs of more comprehensive data. “People who can drive that accumulation of data are going to be in control,” Billings says.
How long such an accumulation will take depends on the willingness of various players to share their data. The track record of the industry with the most experience in analyzing genetic information, cancer diagnostics, does not bode well. Ten years ago, cancer diagnosis was done only one way: Doctors sent biopsy samples to a pathologist, who examined the cells’ shape for telltale signs of disease. Today, doctors can opt for more precise analysis of certain cancers by sending blood or tissue samples out for specialized molecular profiling. The companies selling these tests usually specialize in only one kind of technology. They don’t share information with one another. Yet the development of useful, and therefore profitable, diagnostic tools may be impossible without pooling knowledge. For example, researchers are learning that cancer-causing genetic variations are much more diverse—and less predictable—than previously thought. In a recent study in Nature, scientists screened 441 tumors from breast, lung, ovarian, prostate, and pancreatic tissues and found nearly 2,600 mutations—2,400 more than had been previously noted. About 65 percent of the tumors contained more than one mutation, and 14 percent carried more than 10. Companies focusing on a narrow range of indicators won’t have enough information to make meaningful medical interpretations for most patients.
In an ideal world, private and public databases could be combined so that everyone could benefit from patterns that emerge from the larger data set. But Felix Frueh, vice president of personalized medicine research at Medco, a pharmacy benefit manager with 70 million U.S. customers, says, “There’s so much work and so much money going into the generation of these data sets, that if you’re relatively big, and you have large data sets, you probably look at them as a competitive advantage.” Nevertheless, many genetic testing companies are coming around to the idea of sharing data and collaborating with academic researchers, notes Garry Cutting, medical director of the DNA Diagnostic Laboratory at Johns Hopkins Hospital. “When money gets involved, it always gets a bit more sticky,” Cutting says. “But more often than not, working together benefits everyone, including those in the commercial sector.”
Academic medical centers will have to embrace the sharing ethic, too. Johns Hopkins’ various hospitals have used electronic medical records, which contain important data, for more than 20 years, but always on a closed internal network. “We’ve never been willing to allow it to be shared outside of the Hopkins environment,” says Stephanie Reel, vice provost for information technology for Johns Hopkins University. But to make use of medical records in the fractured health care system, the data has to be shared, she says. Hopkins has been working with the state of Maryland on an initiative—the Chesapeake Regional Information System for Patients, or CRISP—to securely exchange patient data from all medical providers in the state. The first phase of the project is set to go live soon.
Neil De Crescenzo, senior vice president and general manager of software giant Oracle’s Health Sciences Global Business Unit, says that academia and industry should work together to implement electronic medical records on a wide scale. “That’s the challenge we have going forward—to figure out how to get these innovations not just in the large academic medical centers but into the community hospitals and into an outpatient setting,” De Crescenzo says. According to some estimates, less than 20 percent of U.S. doctors use electronic medical records. That number is gradually rising as the issue receives more attention; the federal stimulus package authorized $27 billion over the next decade for doctors and hospitals to install electronic systems for keeping medical records.
Meanwhile, a growing number of nonprofit advocacy organizations and academic groups also are pooling what they know, to try to build a more useful knowledge base. The Personal Genome Project, for example, aims to collect whole-genome readouts and extensive medical information from 100,000 volunteers, and will make all of the data freely accessible on the Internet. Some for-profit concerns have begun similar projects. Twenty pharmaceutical companies, two nonprofits, and several government institutes have contributed to the Alzheimer’s Disease Neuroimaging Initiative, which makes pooled data from brain scans, genetic tests, and drug trials available to anyone. This summer, Omicia joined biotech giant Life Technologies and several large academic centers to create the Genomic Cancer Care Alliance, which will investigate how whole-genome sequences can help guide medical decisions for cancer patients who do not respond well to initial therapy.
Tenenbaum, whose melanoma has been in remission for six years, has spearheaded something called Cancer Commons to unite thousands of cancer patients with clinicians and researchers to find tailored treatments for their specific tumors. Whatever is learned from an individual patient will be fed back into the commons so that it might help the next patient. “The goal is to learn as much as possible from every patient, because a patient is a very precious resource,” he says.
According to some estimates, less than 20 percent of U.S. doctors use electronic medical records. That number is gradually rising as the issue receives more attention; the federal stimulus package authorized $27 billion over the next decade for doctors and hospitals to install electronic systems for keeping medical records.
A more united, collaborative commercial genomics industry not only could spur scientific progress, but better equip itself to tackle challenges in insurance reimbursement and government regulation. No matter how effective a particular test, it won’t sell if insurance companies won’t pay for it. “Most entrepreneurs are totally naive in thinking that the secret sauce, the magic formula, is in the technology,” says Steve Burrill, whose San Francisco venture capital firm, Burrill and Company, specializes in personalized medicine. “God, in this world, is the payer— Medicare or your insurance company or your employer will determine what you get.” Working alone, companies have a tough time convincing payers to reimburse for specific technologies. The companies must demonstrate, to each individual payer, that a test is reliable, leads to better outcomes, and saves money in the long run. The studies needed to prove the technology’s effectiveness are expensive, which can be problematic for most diagnostic tests because they will have relatively small markets and slim profit margins.
Companies that manage to clear this hurdle do so partly through large and politically savvy collaborations. For instance, Genomic Health, a 10-year-old company out of Redwood, California, sells a test that measures gene expression in breast cancer tumors to predict a woman’s likely response to chemotherapy and her risk of recurrence. When Genomic Health first released the test, most payers didn’t cover it. But the company had strong relationships with patient advocates and prominent oncologists, groups that lobbied hard for the insurance industry to approve the test. Within three years, it was considered “standard of care” and covered by most insurers.
More than 200 companies, nonprofits, and research groups belong to an education and advocacy organization called the Personalized Medicine Coalition. The Washington, D.C.–based outfit convenes meetings, writes policy statements, and lobbies Congress not only about insurance reimbursement, but about the equally weighty issue of government regulation. Currently, the legal landscape concerning the genetic testing and diagnostic industries “is a muddled mess,” says the coalition’s president, Edward Abrahams. The Food and Drug Administration must approve genetic tests, like Genomic Health’s, that are sold to clinicians, yet the agency does not oversee the same types of tests when they are developed and used by labs in-house. The fledgling direct-to-consumer industry is largely unregulated. But in the past few months, the FDA and congressional committees have indicated that they will be ramping up restrictions. Because the potential payoff is so high, companies are working together to navigate these policy issues, Abrahams says. “The challenges are not insurmountable, they’re just challenges.”
That’s how Tenenbaum sees it, based on his experience. Health care, he believes, is ready to undergo the same transformation experienced by commerce once it began to understand the potential of an open-source Internet. He says, “The only thing that’s missing is the infrastructure for bringing everyone together.”
