How to get away with academic misconduct at the University of Minnesota

In early 2009, antipsychotic fraud was making headlines.  Eli Lilly had announced in January that it would plead guilty to charges that it had illegally marketed Zyprexa. The company agreed to pay a record-breaking $1.42 billion in penalties. Meanwhile, AstraZeneca was defending itself in federal actions over its own antipsychotic, Seroquel, in which damning internal documents showed that the company had buried and manipulated data to boost sales. In March 2009, articles in the St Paul Pioneer Press and the Minneapolis Star Tribune reported that Charles Schulz, the Chairman of the Department of Psychiatry at the University of Minnesota and a consultant both to Lilly and AstraZeneca, was  implicated in AstraZeneca’s deception.  AstraZeneca would eventually settle the federal investigations for $520 million.

            By December 2009, the University of Minnesota’s Board of Regents had received a complaint about Schulz. That complaint came from Mike Howard, a friend of Mary Weiss, whose son, Dan Markingson, had committed suicide in an AstraZeneca-sponsored trial of antipsychotic drugs at the university. Howard’s complaint was not about that trial, however. Rather, it concerned a wide range of questionable activities by Schulz, ranging from his weirdly evasive responses to questions about akathisia in a deposition, to the use of a placebo in a clinical trial involving acutely psychotic patients, to his dubious relationships with the pharmaceutical industry, including his paid presentations about Seroquel. When the university finally responded to Howard’s complaint over a year later, the response came from Mark Rotenberg, the General Counsel for the university, who dismissed the complaint in its entirety. “After careful examination of each allegation,” Rotenberg wrote, “we have found no University policy violations or other improper conduct by Dr. Schulz warranting any University actions against him.”

            Given the time it took to prepare, Rotenberg’s defense of Schulz is remarkable not just for its brevity but for the absence of any evidence or argument. For each of the nine allegations that he identifies, Rotenberg simply refers to the findings of the “RIO,” or “Research Integrity Officer.”  The Research Integrity Officer does not speak for herself in the document, nor is she named, but in fact, she is Dr. Frances Lawrenz, the university’s Associate Vice-President for Research.  While all of the issues in Howard’s complaint deserve to be aired, Schulz’s work on behalf of AstraZeneca deserves special scrutiny.

            One of the most notorious trials in the AstraZeneca deception was called Study 15, which compared Seroquel to Haldol, an older antipsychotic developed in the 1960s.  Haldol is often chosen as a comparator drug in antipsychotic studies because it is easy to beat. Yet Study 15 had returned some unexpected results. For one thing, it showed that patients on Seroquel gained a significant amount of weight – a fact that AstraZeneca went on to hide for years by burying and spinning its data.  In a 1997 email, for example, AstraZeneca physician Lisa Arvanitis is congratulated on the “great smoke-and-mirrors job” she has done on Study 15. But the weight gain was only one problem.  Study 15 also showed that Seroquel was no more effective than Haldol. In fact, on some measures, it performed worse. This fact had to be hidden as well, which the company accomplished by cherry-picking the results that would be published and those that would be buried. In one email, a company employee named John Tumas identified the buried studies and asked, “How are we going to face the outside world when they accuse us of suppressing data?”

            By the time AstraZeneca had recruited Charles Schulz to present a “meta-analysis” of Seroquel studies at the 2000 meeting of the American Psychiatric Association, the company had conducted an internal analysis showing that Seroquel was no more effective than Haldol. (“The data don’t look good,” wrote one company analyst. “In fact, I don’t know how we can get a paper out of this.”) Yet this was not what Schulz told the APA. In his presentation and in press releases, Schulz claimed that Seroquel was “significantly superior” to Haldol. He spoke about the “dramatic benefits” of the drug. Yet in March 2009, when Schulz was caught, the University of Minnesota backed him fully. As the St. Paul Pioneer Press reported, “A U spokesman said that the dean of the medical school, Dr. Deborah Powell, is aware of the controversy over Schulz's research and has offered him her full support.”

            In the university’s reply to Mike Howard, Rotenberg addresses Schulz’s meta-analysis and APA presentation. But Rotenberg’s defense of the meta-analysis makes no sense. According to Rotenberg, the “Research Integrity Officer” – Frances Lawrenz -- determined that the presentation did not violate the university academic misconduct policy because it was “based on different data sets and different measures being analyzed and reported.”  But of course, this is exactly what cherry-picking involves. Schulz used the positive data and excluded the negative in order to make Seroquel appear better than Haldol.

            Later, in a trial known as Study 41, AstraZeneca compared Seroquel XR – the extended release version of Seroquel – to placebo in patients experiencing acute schizophrenia. The University of Minnesota was one of the study sites, and Charles Schulz was the site investigator. Remarkably, Study 41 showed Seroquel XR performing no better than placebo for schizophrenia.  In emails, AstraZeneca officials refer to Study 41 as a “failed study,” or “code red.” Advisors (such as Schulz) and employees were instructed not to discuss the study. So what did AstraZeneca do? They buried it and tried it again, this time in India, Bulgaria, Romania, the Philippines, Russia, Greece and South Africa. That trial was called Study 132, and it turned out positively. So in 2007 AstraZeneca published that study in the Journal of Clinical Psychiatry. The second author of that study was Charles Schulz, who presented the data at the annual meeting of the American Psychiatric Association and served as the academic point man in company press releases.

            Again, in his response to Mike Howard, Rotenberg evades the real issue, which is the role Schulz played in the AstraZeneca spin machine.  Instead, Rotenberg does some spinning of his own.  Schulz wasn’t actually an investigator in Study 41, Rotenberg says – because Schulz never enrolled any subjects.  (Even if this is true -- Rotenberg provides no evidence -- it would not have prevented Schulz from seeing the negative results, of course.)  And when Schulz cited his own discredited meta-analysis as evidence of Seroquel’s efficacy?  Not a problem, says Rotenberg, because the discredited analysis still showed Seroquel was better than placebo. The charge that AstraZeneca buried the failed Study 41?  Not true, says Rotenberg; AstraZeneca published an abstract. (He does not say where or when that abstract was published.) And the fact that Study 132 produced much better results than Study 41, Rotenberg says, was based on “scientific reasons, such as different study designs.”  (In reality, the study designs are almost exactly the same.)

            In fact, Rotenberg’s review doesn’t read anything like a legitimate review. It reads like a contrived defense of Schulz, complete with evasive statements, cherry-picked facts, and technical justifications of academic misconduct.  Actually, not only does it read like a defense of Schulz; it reads like a defense of AstraZeneca, which had already agreed to pay over half a billion dollars to settle federal charges addressing these very issues. All of which raises a number of disturbing questions about officials at the University of Minnesota.  Why is the General Counsel defending the actions of a pharmaceutical company? In fact, why is the General Counsel involved in an academic matter at all?  Why didn’t Frances Lawrenz, the university’s Research Integrity Officer, speak for herself?  Where is the supporting evidence for Rotenberg's statements? And why isn't anyone at the university looking into such a clearly deceptive review?

-- Carl Elliott