The documents represented four years of research conducted by Sprague's close friend and onetime protégé Stephen Breuning. A 34-year-old wunderkind, Breuning had recently finished a landmark study revealing the dark side of a family of powerful tranquilizers known as neuroleptics. Though the drugs had been used for years to help calm violent mentally retarded patients, newcomer Breuning had suspected that the treatments might do more harm than good. His justcompleted studies supported his theories, showing, among other things, that IQS miraculously doubled when patients were taken off the drugs. The research won him tips of the hat from countless older colleagues. If Breuning's findings were correct, neuroleptics could someday be discarded. ut something about the uneasy. Many of Breuning's numbers were too goodperfect, in fact. Somewhere in the charts, graphs, dates, and case studies scattered across his desk, Sprague feared, there was fakery. On the third day he found what he was looking for. Sprague had estimated that some of Breuning's experiments would have required him to have spent at least 270 working days in his lab, examining patients; but the year in question had only 260 working days during which the lab would have been available to him, and Sprague knew Breuning was working on other time-consuming studies as well. Breuning could simply not have conducted this work. Sprague spent three months verifying his facts and turning up other incon Fraud is potentially deadly. The findings is playing with lives. sistencies in Breuning's The case of Stephen Abbé Gregor Mendel, the nineteenth-century monk who founded modern gene theory by breeding and crossbreeding pea plants, came up with such suspiciously perfect results that later investigators concluded he had tailored his data to fit predetermined theories. Perhaps the most extraor dinary cover-up of this cen- But times have supposedly Yet bad work still slips past the sentries. Worse, in cases like Breuning's, the work isn't just bad, it's potentially deadly. The astronomer who fakes the discovery of a planet may set back his science, but once he's caught, the damage is reversible. A researcher who fudges his drug findings, however, is playing with more than moons and star maps; he's playing with lives. How then do people like Breuning manage their scientific scams? Where is the breakdown in the system that allows wooden-nickel research into circulation? Perhaps more important, where is the breakdown in the researchers themselves? What are the pressures, the ambitions, that turn scientists into counterfeiters, fact finders into forgers? t was in 1979 that Sprague and Breuning first met at the Coldwater Regional Mental Health Center in Michigan. The two could not have been more different. Sprague was older and understated, and he didn't always publish his ongoing studies of the mentally retarded. Breuning was young and aggressive and preferred a higher profile. They took to each other quickly-the patient mentor and the passionate apprentice. Sprague often visited Breuning's home and drove the young psychologist to conventions. "Breuning was bright, and I wanted him to work for me," Sprague explains in his gentle midwestern accent. For 15 years Sprague had been studying neuroleptics. The drugs calm the kind of severely retarded patients who bloody themselves by beating their head against a wall. Neuroleptics, however, produce a side effect, a spasm called tardive dyskinesia. Most of the time this is harmless-a tongue may wag, an arm may flap. But in rare instances the spasms can be lethal; the chest muscles can contract and suffocate the patient. Sprague and many other researchers believed that in extreme cases the benefits outweighed these risks. In December 1980 Breuning left Coldwater for the research was, however, Sprague began to get a funny feeling about its tidiness. Sprague knew from experience that the process for evaluating the extent of tardive dyskinesia left plenty of room for human error: dozens of patients are examined by nurses, who then rate the severity of 34 different movements-from tapping feet to wagging tongues-on a scale of zero to four. In Breuning's lab the nurses agreed, point by point, 100 percent of the time. Conducting the identical experiment, Sprague's nurses agreed just 80 percent of the time. University of Pittsburgh. Out on his own, he was beginning to doubt Sprague's beliefs, questioning whether neuroleptics were such a good idea. Working independently, he conducted a number of studies, most of which condemned the medications. His work was widely read and respected. In addition to his findings concerning the drug's effects on IQ, he concluded that the severity of the tardive dyskinesia was often too great to warrant the drug's administration. Moreover, he found that if patients were taken off the medication, their withdrawal symptoms would last a mere 16 weeks, far less time "I had a strong hunch that than expected. Between 1980 Breuning fudged his data," and 1983 Breuning wrote 24 Sprague remembers. "But I papers on neuroleptics and had no evidence." He wasn't related topics-a full third of even sure if he should look for all the literature on those sub- any. Never the aggressive jects produced during that sort, Sprague hated conperiod. flicts. Besides, he raConvincing as Breuning's tionalized, snitching on an esteemed colleague could only mean trouble. But Sprague had personal motives pushing him to pursue Breuning: the work with drug treatments and the mentally retarded hit close to home. As a foster parent, Sprague often cared for retarded children. And his wife was slowly dying from a failed kidney-drugs were the only things prolonging her life. "My wife was completely dependent on medication to keep her alive," he says. "I would have been furious-livid-if someone had cheated on the research behind her medication and harmed her." Sprague gathered up as much of his former colleague's research as he could and began his detective work. Breuning's studies, he found to his dismay, crumbled to the touch. In addition to the too perfect tardive dyskinesia scores and the experiments Breuning "conducted" nearly every day of the year, Sprague found that one study Breuning claimed to have performed took place, if at all, in the young researcher's absence. The experiment was alleged to have been conducted at Coldwater, but at the time Breuning was hundreds of miles away at the University of Pittsburgh. ODecember 4, 1983. n the morning of Sprague called Breuning and ILLUSTRATION BY DAVID SUTER eral patient exams in the files were dated after Breuning had left Coldwater, yet he was listed as the examiner. "I asked him, 'How could you have conducted these?"" Sprague recalls. "He said, 'I don't know."" As Sprague stared at the mounting evidence, his heart sank. "To suspect someone you've invested in-friendship, emotions, money-is crushing," he says. Possibly the most courageous act of Sprague's life was to roll a piece of paper into a typewriter on December 20; he sent the NIMH a six-page letter. With his accusation in the mail, Sprague thought the job was over, that he could sit back, relax, and let the federal agency go to work. "Boy, was I naive," he says, laughing. the agency dragged its feet. It waited a full month and a half before even acknowledging that the University of Pittsburgh had not investigated enough of Breuning's work. It would take 14 months after Sprague first went to the NIMH with his suspicions for the agency to round up its own panel and conduct the investigation that the university had neglected. Meanwhile Breuning was practicing and conducting research. Colleagues across the country still trusted him. Frustrated with the NIMH bureaucracy, Sprague wrote an angry letter to Science, which published an article criticizing the agency. Finally, on April 28, 1987close to three and a half years after Sprague sent his letter -the NIMH declared Breuning guilty. Why did it take a man motivated by his wife's impending death and his compassion for retarded children to blow the whistle? After all, science is "self-correcting." The profession's checks and balances should catch bogus research before it has any impact on the public. Yet, according to the Food and Drug Administration, the Breuning case is just the tip of the fraud and misconduct iceberg. Investigators at the FDA run across so much shoddy research that they have quippy terms like "Dr. Schlockmeister" for a bad scientist, and "graphite statistics" for data that flow from the tip of a pencil. Every year, as a qualitycontrol measure, the FDA conducts investigations of key studies or researchers involved in getting new drugs to the agency for approval. "This is the last stop for drugs before they go to the public," explains Alan Lisook, who heads the FDA investigations. "You'd think we'd get some of the cleanest science around." But in 1986, when he analyzed the investigations of the previous ten years, Lisook compiled some shocking numbers: Nearly 200 studies contained so many flaws that the efficacy of the drug itself could be called into question. Some 40 studies exhibited not simple oversights but recklessness or outright fraud. In those ten years the FDA banned more than 60 scientists from testing experimental drugs, after finding that they had falsified data or engaged in inept research. As Sprague says, "something is clearly not working." Only one month after the NIMH announced its verdict in the Breuning investigation, the medical community was shaken by yet another scandal. For 22 years internist Charles Glueck had risen steadily through the hierarchy of science. Since graduating from medical school in 1964, he had published nearly 400 papers at the furious rate of close to 17 a year. For his leading-edge research on cholesterol and heart disease Glueck had won the University of Cincinnati's prestigious Rieveschl Award in 1980. As head of the lipid unit and the General Clinical Research Center at the university, Glueck was one of the most powerful and heavily funded scientists on staff. But last July the National Institutes of Health found that a paper of Glueck's published in the August 1986 issue of the journal Pediatrics case is just the tip of the fraud and misconduct iceberg. was riddled with inconsistencies and errors. As written, the NIH explained, the paper was utterly shoddy science, its conclusions empty. The paper tackled a controversial treatment for children at risk for developing heart disease. Doctors could not agree on whether a lowfat diet, combined with cholesterol-lowering drugs called resins, would hinder a child's physical development. Since the drugs would absorb not only cholesterol from food but also nutrients, there was and continues to be-real concern that the children's growth could be stunted. Glueck's 11-page paper, bearing the cumbersome title "Safety and Efficacy of Long-Term Diet and Diet Plus Bile Acid-Binding Resin Cholesterol-Lowering Therapy in 73 Children Heterozygous for Familial Hypercholesterolemia," declared the treatment perfectly safe. Yet shortly before the study appeared, the National Institutes of Health received two phone calls from sources close to Glueck-they have since requested anonymity -warning that the pending paper was grievously flawed. Glueck, they said, was depicting his research as far more thorough than it actually was. The NIH, however, had no proof that the allegations were true. Moreover, even if they were true, the agency saw no grave medical danger in allowing the flawed study to appear in print; no research suggested that the potential retardation in the children's growth would be dramatic or radical. "This was not like publishing a story about a supposedly safe AIDS vaccine that actually causes cancer," said one NIH spokeswoman. Despite their laissez-faire response, the agency's of ficials were troubled by the phone calls. Over the next year the NIH conducted an investigation to determine the truth of the charges. Three separate committees -two at the University of Cincinnati and one at NIH headquarters - interviewed dozens of doctors, lab assistants, and others close to Glueck and scrutinized hundreds of pages of documents and transcripts. In July the NIH issued a harsh verdict: Glueck's methods were "unacceptable by any scientific standard." His paper contained "serious deficiencies" that amounted to scientific misconduct. Among those "deficiencies" was Glueck's statement that 73 children in the study were examined every year. In an earlier draft he had written that he saw the patients every six months. In reality some patients went as long as six years between exams. Also, one of Glueck's main points was that the children developed normally. Yet hardly any patient charts contained information on growth and development. Glueck told the NIH that he called upon "intimate knowledge" of the children when drafting the paper. In other words, Glueck relied upon one of science's least precise tools: memory. Finally, Glueck purported to have thoroughly tracked cholesterol levels, height, and weight for each patient. Yet almost half the patient charts lacked some of that data. As a result of the NIH findings Glueck was barred from receiving agency funding for three years. He was given the embarrassing task of retracting the article in a letter to Pediatrics. He also resigned his university position. Even the New York Times weighed in with an opinion, censuring the researcher in a scathing editorial. G lueck, who has appealed the NIH decision, is unflustered by the scandal, calling his errors "random" and "unintentional" and adding contritely that he and his assistants "profoundly regret" the incident. He also points out that when the study was repeated with authentic data in place of the ersatz stuff, the results were the same. However, Robert Berliner, former dean of Yale University Medical School and a consultant on the case, is unmoved by all this. Now a professor emeritus, Berliner contends that Glueck's errors were not random; rather, he says, the doctor methodically filled gaps in his research to make the paper more conclusive. Even the University of Cincinnati's internal investigating committee was shocked by the doctor's "lack of recognition of the seriousness of his actions." When asked how he could allow any such flawswhether deliberate or accidental to find their way into his research, Glueck explains simply that he was "overworked." As alibis go, it seems like a thin one. But increasingly both the accused and the accusers in science-fraud scandals are agreeing that overwork may indeed play a big role in leading researchers off the straight and narrow. Science, they point out, is as much a career as a calling, and when career pressures get too great in any field, people do funny things. ILLUSTRATION BY DAVID SUTER |