Finding foul play in the laboratory
Brian D. Johnson
In the antiseptic laboratories of pure research, where it’s so often assumed that science handles the truth with rubber-gloved integrity, a strange virus has set in. Known as scientific fraud, it has triggered an escalating series of scandals that have ruined brilliant careers and weakened public faith in the previously inviolate discipline,most often in the competitive
and glamorous field of cancer research.
Fictional science is as old as alchemy—in fact, the father of genetics, Gregor Mendel, who conducted the famous green pea experiments, was suspected of altering data. But it has come a long way, even since 1974, when William Summerlin at New York’s Sloan-Kettering Cancer Center confessed he had painted “skin grafts” on white mice with a black felt pen. Some more recent lab frauds have been as ingenious as the discoveries they claim to substantiate. Their scenarios can feature all the elements of a biochemical potboiler-intrigue, espionage, betrayal, jealousy and coverup.
While Canada has yet to produce a truly spectacular case, and while most scientific research is valid, Maclean's has learned there have been at least four alleged cases of altering scientific data (all unrelated) at the University of Toronto alone. Three of these occurred in the medical genetics department, which conducts much of its research at the Hospital for Sick
Children. Two cases involved technicians, both discreetly fired after being caught falsifying data. Another involved a brilliant post-doctoral student on a $20,000-a-year Centennial fellowship who was asked to leave his lab but had his fellowship renewed. And a fourth with repercussions still in the courts, implicates a U of T biochemist running a lab at the hospital—after being formally cleared of fraud, he was unofficially reprimanded and eventu-
ally promoted to the rank of full professor. Most Canadian scientists consider fraud rare, but a few are sounding the alarm. Dr. Byron Lane, a U of T biochemistry professor who had a gruelling encounter with an incident of fraud early in his career, says cases are hushed up “in a very Canadian fashion.” Scientists have trouble agreeing on an exact definition of scientific fraud— it’s a legally tainted word—but actions that cause ethical concerns in the scien-
tifie community come in a variety of shades: from massaging and constructing data to fudging, cooking, forging or plagiarizing data. Whatever the method, using false findings to “help the truth” breaks a long-standing taboo in the scientific community. Although outright fraud may be rare, sloppy science is known to be quite common, especially in the industrial labs straining to prove conclusions preordained by corporate needs. While one may not expect applied science to remain unsullied by marketplace pressure, basic research—that publicly funded excursion into the realm of pure discovery—belongs to the inner sanctum of science. As one writer put it, finding fraud there is like catching the Pope cheating at poker.
Budget cutbacks in the U.S. and Canada have greatly increased the pressure on scientists scrambling for public funds. They recommend one another for grants via a peer review system, and because the best mark of success in the lab is a paper published in a profes-
sional journal, data can get trampled in the rush to publish or perish. The review system can also bog down in the flood of complex and highly specialized papers. Byron Lane says scientists often bend the truth as they try to live up to such words as “significant” and “exciting ” for grant applications. Although journal material is generally screened by referees in each field before publication, it’s difficult for them to tell whether original lab results have been fudged. In one 1980 case a $750,000 research project at Massachusetts General Hospital collapsed when a scientist discovered his senior colleague, Dr. John Long, had invented data in a paper published on Hodgkin’s disease. Long, who was forced to resign from the hospital, confessed he buckled under pressure from a grant application. And only last August, Dr. Philip Felig, newly appointed chief physician at the Columbia Presbyterian Medical Center in New York, resigned after a junior associate admitted falsifying data in a paper they coauthored at Yale. In Canada, the National Research Council (NRC) puts out 11 of the country’s top journals, and Claude Bishop,
NRC director of biological science, admits, “We have caught one or two things, not so much scientific fraud as plagiarism.”
Fraud is “not a big problem here,” says Bishop, “probably because science in Canada might be a little less competitive than in the United States.”
However, Canadian scientists usually send their most ambitious work to prestigious jour-
nals outside the country. This is
what happened in a highly controversial case that has plagued the Hospital for Sick Children for five years. Dr. Bibudhendra Sarkar, a U of T biochemistry professor who runs a lab at the Hospital for Sick Children, published photographic data that a retouch artist altered with an airbrush. The photos, uncovered by his research assistant, stemmed from two experiments that were part of a project to study how copper accumulates in the body, a condition of a rare illness called Wilson’s Disease.
One experiment, cited in a paper Sarkar coauthored with a prominent Swedish scientist, combined copper with human albumin (a common protein in blood) to show how the metal binds albumin molecules into clumps of several distinct “species.” Photographs taken of the albumin cell as it whizzed around a centrifuge showed a distinct curve with two peaks, each representing a species. An airbrush was used to exaggerate a vague ripple (see photos) that Sarkar claimed indicated a third species of copper albumin. In the other ex-
periment, coauthored with two French scientists, Sarkar combined copper with amino acids and peptides and claimed to have produced a variety of compounds. The solution was run across a chromatograph, a silica blotter that segregates different molecules according to their physical properties. Each original chromatograph showed an indistinct smudge, which was later retouched to display a spectrum of sepa-
rate molecular bands. Sarkar admitted in the paper he had trouble with the technique.
Charges that Sarkar had fudged the data in both experiments came from his research assistant, Theo Kruck, who had been working with him at the hospital since 1964. Their amiable rapport exploded in 1976 when Sarkar accused him of supplying false data, and Kruck retaliated by discovering the photos. Kruck’s brother, Helmut, was the retouch artist who, says he, had altered them according to Sarkar’s specifications. According to Sarkar, Helmut was hired to clean up photo backgrounds for publication. But Helmut said Sarkar also had him darkening blips. “Then he had me improving curves,” he says. “I just did what I was told. I didn’t know what were all these blobs and blips and curves.” Sarkar denies that lie gave such instructions.
Obscure though they may be, the alterations created considerable turmoil. To the hospital’s credit, Dr. Aser Rothstein. director of the hospital’s research
institute, took the allegations against Sarkar “very seriously,” asked the U of T to postpone Sarkar’s promotion to full professor, and promised to set up an outside tribunal to investigate. Meanwhile, Kruck was transferred to the hospital’s service division, which he considered a demotion and refused to accept. By September, 1977, he was out of a job and preparing to sue for wrongful dismissal. The hospital called it resignation, a finding that was upheld by the Supreme Court of Ontario. The
decision is now under appeal. In October, the tribunal was set up: two referees, eminent biochemists from Harvard and the University of Alberta, ruled that Sarkar wasn’t guilty of fraud but that he may have reached his conclusions too quickly. Kruck, still unsatisfied, sought a police investigation, but the Crown attorney refused to lay charges, arguing that the courts are not a suitable forum for “a slim, arcane paper dealing with an obscure disease.” Sarkar’s promotion in 1978 from associate to full professor came at a time when, according to Dr. Rothstein, the biochemist had already suffered enough. His promotion was delayed two years, and his wife suffered a nervous breakdown. Rothstein is fed up with the flak arising from the incident. “There’s fudging and fudging,” he says, and adds, “[Sarkar] was trying a little too hard to make his case.”
The cloak-and-dagger politics surrounding such cases can become Byzantine considering the size of the molecules originally under investigation. Byron Lane says he was an easy-going
guy until he ran into a major scandal at New York’s Rockefeller Institute in 1960. Lane, one of 20 PhDs working in the lab of Nobel laureate Fritz Lipmann, was asked to reproduce an experiment by a highly trusted “hot young prospect” in the lab. The experiment consistently failed until the hot prospect diverted Lane to a new supplier for one of his ingredients. The ingredient had been spiked with a substance to make the experiment appear successful. Later it was discovered that the lab’s entire supply of a chemical under study had been switched with calcium chloride. Lane was horrified to find calcium chloride powder in his own lab-coat pocket: “It began to look like I was being framed.” While eventually vindicated, Lane never quite recovered. “This notion of science based on trust is the biggest crock,” he says. “No scientist should resent being mistrusted.”
A classic case involving joint research with a trusted protégé occurred in a medical genetics lab at U of T, where Dimitrious Spandidos, regarded as an exceptionally gifted young scientist, began work in June, 1976. He spent two years in the lab under the supervision of Dr. Lou Siminovitch, chief geneticist at the Hospital for Sick Children, and pioneered gene-transfer experiments that the international scientific community saw as exciting breakthroughs. Together Spandidos and Siminovitch published five papers in leading scientific journals. But early in 1978, suspicion darkened the teacher’s devotion to his prize student. Siminovitch couldn’t understand how Spandidos was turning out results so fast considering how little time he spent actually performing experiments. And the results always seemed infallible. Eventually Siminovitch discovered there were not enough dishes in the lab’s incubator to handle all the experiments reported. Also worrisome were the journal referee’s doubts about how the work could have been completed.
In April, 1978, Siminovitch finally asked Spandidos to conduct an experiment in the presence of another postdoctoral fellow, Bill Lewis, who had spent four months trying to duplicate Spandidos’ gene-transfer technique without success. “When I couldn’t get it to work,” says Lewis, “I thought I was doing something wrong, or that Spandidos had left important steps out of the published procedures.” The experiment involved transferring genes from a drug-resistant cell to a sensitive cell to see if the sensitive cell would become similarly resistant under the foreign genetic influence. But Lewis says that when he watched Spandidos perform the technique, he transferred entire cells (not just genes), thus rendering the results meaningless.
Siminovitch called his lab personnel together and announced that Spandidos had been asked to leave. Spandidos left, claiming he had been slandered, and asked Ottawa’s Medical Research Council (MRC), which funded the lab (and which awards $100 million in research grants and scholarships annually), to set up an impartial inquiry. The MRC turned down the request, and Spandidos’ fellowship was, for the time being, cancelled. He sent an open letter to members of the scientific community arguing his career had been destroyed “in a cruel and ruthless manner” by
Siminovitch, who “was probably misled by some jealous and incompetent people.” He also claims that “other scientists in the world have been successful in repeating many of my results.”
Cases such as the Spandidos episode illustrate the scientific community’s enormous difficulties in policing itself. Siminovitch now wonders whether he was right in not issuing a public statement. And whilehe cancelled recommendations for four high-profile jobs he had given his student only a few months earlier, he has not retracted most of the papers they had coauthored. Spandidos,
meanwhile, is now working at the prestigious Beatson Institute for Cancer Research in Glasgow. The MRC also restored his Centennial fellowship, a move that Siminovitch, who is on the MRC executive, attributes to “spurious reasoning.” Dr. Francis Rolleston, director of special programs of the MRC, says, “The council felt the individual should be given the benefit of the doubt and not have his scientific career destroyed by the reflex of a central bureaucracy.” Rolleston further suggests the courts are better equipped than the MRC to cope with charges of fraud. “The kind of witch hunts that have gone on in the States have been very destructive,” he says, “because they’ve called into question the whole funding process on the basis of very few cases.” (In the wake of well-publicized frauds in the U.S., a congressional subcommittee on “oversight and investigation” was set up last spring, and the National Institute of Health introduced measures to cut off funds where data is falsified. Stricter controls are also being exercised by the American scientific community on coauthorship.)
Perhaps the most bizarre of the U.S. scandals was the case of Mark Spector, a 24-year-old boy wonder who might well have snagged a Nobel Prize for his insights into the causes of cancer, when a foreign substance was found in his work last year. Studying as a graduate student under the wing of an eminent scientist at Cornell University, 68-yearold Efraim Racker, Spector discovered and purified a chain of cancer-cell kinases (enzymes) in just six months—a task expected to take years. His elegant theory of “kinase cascade” was so seductive that, despite the controversy, many researchers believe it may still stand. And the addition of foreign material, which Spector says was not his doing, involved a substitution of chemicals that was no less ingenious than the theory.
There are striking parallels between Racker and Siminovitch. Each is a distinguished elder of science who developed a bond of trust with a brilliant young protégé. Both Spector and Spandidos were suspect because they produced immaculate results with alarming speed. Both made discoveries that may be valid and both left their mentors perplexed as to how much of the work might be tainted. Racker, unlike Siminovitch, retracted all the papers he had cowritten with his student and locked himself in the lab eight hours a day to see what he could salvage from Spector’s work. Siminovitch says it would take two years to reproduce Spandidos’ work, and the results could still be inconclusive. “I have a lot of sympathy with Racker,” says Siminovitch.
Sometimes falsification starts at the
bottom rung of the lab hierarchy— with the lab technician who is trying too hard to produce the results his boss is seeking. At this level, cases are generally dealt with expeditiously. Dr. Rod Mclnnes, another U of T geneticist working at the Hospital for Sick Children, concedes he once fired a lab technician for falsifying data. “It was traumatic for everybody involved,” he says. While refusing to discuss details, Mclnnes suggested the technician’s motive was perhaps “an inappropriate desire to please.” Dr. Paul Sadowski, an-
other U of T medical genetics professor, also admitted a technician in his lab was fired for “quite blatant faking of data.” Sadowski said he pities the culprits in such cases.
The most susceptible targets for fraud are large labs run by top scientists too busy to oversee the fine details. Dr. John Coulter, head of biochemistry at the University of Alberta, says he has never encountered a case but agrees that people like Siminovitch and himself, “who have fairly extensive responsibilities, can’t be on top of all our people all the time.” Dr. Jack Campbell, who runs the University of British Columbia’s microbiology department, says: “I’ve never heard of any cases here. But I’m not sure it’s something to be so bloody proud of. It’s probably because we’re not doing very high-level research.”
The problem has become endemic among scientists working under industrial or commercial harness. Five years ago, fraudulent data destroyed the credibility of one of the largest private labs in the U.S., Industrial Bio-test Laboratories of Illinois (IBT) in Chicago. Although most IBT operations are now shut down, pesticides and fungicides tested by its labs are still on the market in Canada. The federal department of health and welfare, finding 80 of the iBT’s 83 cancer studies poorly substantiated, has developed a special hazard label for improperly tested products. Drug companies have been most notorious for shoddy product-testing—the classic example is the thalidomide pill, which, after insufficient research, was recommended for pregnant women to prevent nausea and later was found to cause deformities in offspring. Dr. Gordon Atherley, president of the Canadian Centre for OccupationalHealthand Safety (a Crown corporation), goes so far as to assert that half of commercial science is “shoddy.” “The government doesn’t know what science to draw on for its health and safety standards.” While warping data in applied science can inflict direct injury on the public, the impact of fraud in pure research is less obvious. It’s a more puzzling offence, requiring someone who is clever enough to carry it off but stupid enough to think he can get away with it. A fraudulent discovery of any importance will eventually be ferreted out by other scientists, if only after considerable time and money have been wasted pursuing chimerical data. However, there are more far-reaching repercussions. If science is willing to tolerate even a marginal amount of fraud, the entire perspective of objectivity on which the scientific method is based could break down. And science’s claim to be a yardstick unbent by political and economic influence becomes increasingly dubious. &t;$>