In a world of retail research, who can trust the results?
On a frigid January morning, a group of young scientists gathered in a small meeting room at Massachusetts General Hospital to share their latest findings. Every year, the Massachusetts Alzheimer's Disease Research Center, a network of Harvard Medical School faculty members, holds this 'poster session,' a time-honored form of scientific show-and-tell. At the 1999 meeting, scientist Dennis Selkoe moved through the room like a celebrity, looking very Brooks Brothers in a sea of rumpled L.L. Bean. In this small community of national researchers, Selkoe is a star, the man responsible for shedding light on the biochemical breakdowns that destroy the brains of Alzheimer's patients.
Selkoe is also one of a new breed of university-based entrepreneurs who have found ways to move their work out of the lab and onto Wall Street. As a founder of California-based Athena Neurosciences, he's earned millions by linking his scientific fortunes with the pharmaceutical industry.
This spring, a Harvard review panel investigated whether Selkoe let his corporate interests interfere with his scientific judgment. Last year, he added his name to a National Institutes of Health (NIH) report endorsing a controversial blood test for Alzheimer's. What he neglected to mention in the report is that Athena manufactures the test and stands to profit from NIH endorsement.
An anonymous complaint filed last December with the Harvard Medical School dean's office charged Selkoe and another Harvard neurologist with conflict of interest and suggested that the report played down research critical of Athena's test. Last May, a Harvard panel decided that Selkoe did not violate the university's conflict-of-interest rules. But the university didn't completely let Selkoe off the hook. He agreed to publish an unusual after-the-fact disclosure statement in The Neurobiology of Aging, the journal that published the original report in April 1998. In addition, the university said it plans to clarify its conflict-of-interest regulations to ensure that researchers disclose their industry links whenever they report on related research, said medical school spokesman Don Gibbons.
Selkoe regrets not noting his relationship with Athena Neurosciences in the NIH report. 'Otherwise,' he says, 'I stand by the accuracy and validity of the statements in the report.'
But not everyone agrees that this is a simple case of poor judgment. Susan de la Monte, a Harvard pathologist who helped develop a competing Alzheimer's test, says 'it's about time' Harvard dealt with Selkoe's dual roles as academic scientist and drug-company consultant. 'I have conflicts too,' she says. 'But I don't sit on a committee telling people to use the test.'
It's not just a Harvard problem. Increasingly, academic science overlaps with pharmaceutical marketing. Scientists are not required to disclose their corporate ties; individual schools have rules for dealing with the issue, but even if a scientist's work turns out to be compromised, there's no guarantee that doctors or patients will ever know.
If you think science is neatly split between research in independent university labs and in companies that produce products for profit, you're about 20 years behind. In the past, most campus scientists paid their bills with grants from the federally funded NIH, but a growing percentage of campus labs are now linked to pharmaceutical or biotech companies. Some fund a single study, while others fund an entire lab in exchange for access to its findings. Scientists often sell the rights to their discoveries--from disease-linked genes to potential new drugs--and share the spoils with the university. And, increasingly, university-based scientists are starting their own companies to cash in on their findings.
Some of these links have allowed key discoveries to move off the lab shelf and into the medicine chest: The technology for the CAT scan was discovered at Stanford University and licensed to General Electric. Harvard's artificial skin for burn victims is now made by Cambridge-based Genzyme. But consumer groups and health ethicists worry that without proper safeguards, bias may creep into the work of scientists who stand to profit, or that scientists concerned about patent rights may be less willing to share findings with others who could challenge or build on their work.
Bias happens. The University of California at Irvine shut down a cancer-research lab in 1997 after finding that scientists there had invested in a company that hoped to sell the drugs they were testing--and then had failed to report side effects and used experimental drugs without Food and Drug Administration approval. Researchers at the University of Toronto looked at a series of heart-drug studies in 1998 and found that researchers funded by companies that make the drugs were far more likely than independent scientists to produce studies supporting their use.
Sheldon Krimsky, a Tufts University urban studies professor and research watchdog, is disgusted by a recent $25 million deal struck between the University of California and a Swiss drug company that now has first dibs on Berkeley's plant genetics discoveries. 'An entire department at Berkeley is being bought off by Novartis,' he says.
Krimsky predicts that a private company will probably buy an entire university some day; in the meantime, it's one professor at a time. 'There's no question in any reasonable person's mind that who you get the funding from affects your work,' he says. He sees a pressing need for a system that would require scientists to acknowledge their corporate affiliations whenever they publish a study, speak at a conference, or sit on a review panel.
'Readers should have the opportunity to form their own opinions on whether conflict of interest exists,' he adds. But they rarely do. In 1997 Krimsky found that in 272 of 800 scientific papers, the authors owned stock, served as consultants, or had some other financial stake in the findings. Only one-half of 1 percent of the 62,000 articles he checked included disclosure statements.
The Athena Alzheimer's test is a classic case of how the pharmaceutical industry can dress up marketing and call it academic research. Alzheimer's affects about 4 million people in the United States; they and their families are desperate for a cure. Testing at least offers hope to patients suffering from memory loss, whose symptoms could be due to drug interactions or minor strokes. But so far there is no blood test or brain scan to diagnose Alzheimer's; doctors rely instead on a combination of MRIs, memory tests, and gut instinct.
In 1993, Duke University scientists discovered an Alzheimer's-related gene. Two years later, Athena bought the right to use the discovery to develop ADmark, a promising but imperfect diagnostic test that hit the market in 1996. In 1997, Quebec-based Nymox Corporation began marketing a different test based on Harvard pathologist Susan de la Monte's discovery of a protein found in high levels in the blood of Alzheimer's patients.
Since both tests are new and imprecise (both claim 80 to 90 percent accuracy), insurance companies won't pay for them. Athena thus found itself with a hard sale, new competition, and skepticism among doctors. So--following the 'nine-out-of-ten-doctors-recommend' approach--it elicited expert endorsements by granting the nonprofit Alzheimer's Association $100,000 to explore the usefulness of Alzheimer's tests. The nonprofit group then asked the highly respected NIH to host a panel and endorse the final report. Athena's role was thus veiled behind two layers of nonprofit credibility. The Alzheimer's Association rounded up experts to conduct the research; among them were Harvard's Dennis Selkoe and Alan Roses, the Duke scientist who discovered the gene.
The final report, published in the April 1998 issue of The Neurobiology of Aging, endorsed the Athena test. The article noted Athena Neurosciences' 'sponsorship,' but nowhere did it mention that Selkoe co-founded Athena, or that Roses held the patent.
The endorsement was subtle. The scientists evaluated not products but the relevance of 'biomarkers'--physical signs that the tests look for. But it did call the technology behind the Athena test the only one that 'can add confidence' to conventional diagnostic techniques. They called de la Monte's protein test 'promising' but said it would 'require further study.'
Then, in December 1998, The Wall Street Journal pointed out both Selkoe's and Roses' links to Athena. A week later, an anonymous complaint charged Selkoe and another Harvard Alzheimer's researcher--John Growdon, who chaired the panel--with violating the university's conflict-of-interest rules.
Selkoe offered a written statement saying he suggested disclosing his corporate affiliation, but Growdon had thought it unnecessary. 'However,' Selkoe wrote, 'it should be considered that my affiliation with Athena was already very widely known in the Alzheimer's disease field and was included in published articles in the past.'
Well, not always. Science magazine reported in 1992 that Selkoe routinely wrote about Alzheimer's without mentioning his corporate affiliation. And a random check of eight studies Selkoe authored and published in 1996 and 1997 found no mention of his relationship with Athena--a relationship that has paid off richly. When Athena went public in 1993, Selkoe owned 255,000 shares worth $3 million, according to Science. When Elan Corp PLC, an Irish biotech company, bought Athena in 1996, Selkoe had a $50,000-per-year consulting contract and sat on the company's board, according to Securities and Exchange Commission documents. He now sits on Elan's board of directors.
So who's minding the store? With so much money at stake, the scientific community has been unable to reach a consensus. Congress paved the way by passing a law in 1980 that made it easier for federally funded scientists to patent and profit from their findings; many became stockholders, board members, and consultants. The NIH, which dispenses $15 billion annually in research grants, tried cracking down in 1989 by asking researchers to report their corporate links when applying for funding, but angry respondents called the plan an attack on academic freedom.
Instead, the NIH asked universities to set their own rules. Unfortunately, the process remains secret, which defeats the purpose, says Mildred K. Cho, a professor at Stanford University's Center for Biomedical Ethics: 'The disclosure is not a true disclosure if it is not public,' she says.
Because schools collect a share of royalties, they have their own conflicts, she adds. These days, universities encourage 'technology transfer': research that is sold or licensed to private companies. MIT earned $21.2 million in royalties and from the sale of stock in startup companies in 1997--up from $10.2 million in 1996. Harvard filed 61 patent applications, signed 67 licensing agreements, and collected $16.5 million from existing licensing contracts in 1997, up from $7.6 million in 1996.
The Dennis Selkoe case is interesting because Harvard created a particularly strict conflict-of-interest policy in 1990, when it was still smarting from the case of a Massachusetts Eye and Ear Hospital doctor who tested a questionable eye drug on hundreds of patients. The drug turned out to be useless, but not before the doctor earned $1 million by selling stock in the manufacturer. So Harvard forbids faculty members to test their own companies' drugs. But publishing rules are vague: a faculty committee must grant researchers permission to present results without disclosing relevant financial interests.
On a national level, this sort of policy means that neither patients nor their doctors find out about potential conflicts of interest unless someone blows the whistle.
Drummond Rennie, West Coast editor for the Journal of the American Medical Association, is demanding change. In a recent speech at MIT, Rennie asked scientists and universities to support efforts to make disclosure of corporate financing routine and public. Meanwhile, in July, scientists for what was once Athena Neurosciences and is now part of Elan Corporation, announced that they had discovered a substance that seems to break down fatty substances, or amyloid plaques, in the brains of mice. The company's president said they plan to file an application with the Food and Drug Administration by the end of the year for permission to begin human tests.
When Selkoe's team published results in the April 8, 1999, issue of Nature, there was no disclosure. The journal doesn't require it, says Krimsky. 'My guess is, if the drug does not pan out--that's when people will begin looking at the conflicts of interest. If it's successful, the attitude will be that the scientists deserve the windfall.'
Still, the drug is unlikely to be perfect. And even if Athena is first, at least 20 other Alzheimer's drugs are currently under study. Once they hit the market, scientists will be asked to help patients decide whether they're any good. And unless those researchers disclose their industry links, no one will know whether they speak for the science or for the company.
'The patient in me,' says JAMA's Rennie, 'sees that as a real threat.'
Tinker Ready is a Boston-based freelance journalist. From the Boston Phoenix (April 29, 1999). Subscriptions: $80/yr. (52 issues) from 126 Brookline Av., Boston, MA 02215.