Wednesday, August 6, 2008

COMMERCIAL BIAS at STANFORD and CORCEPT

COMMERCIAL BIAS at STANFORD and CORCEPT

Stanford University has replaced Dr. Alan Schatzberg as Principal Investigator on his NIH grant that studies psychotic depression. This step followed persistent questions from Senator Grassley about conflicts of interest and financial disclosure. Dr. Schatzberg’s drug mifepristone has been in NIH-supported clinical trials at Stanford and in FDA-monitored trials sponsored by Corcept, the company Dr. Schatzberg founded.

The University emphasized “We want to put to rest confusion about the integrity of the research involving mifepristone.” Questions did arise following NIH-supported publications from Stanford in 2001 and 2002. Neither one found significant evidence of efficacy, yet both were repeatedly touted by Dr. Schatzberg and his associates as positive. The 2001 report had no significant findings even after attempts to cherry-pick the data. In the 2002 report the data were so weak that Dr. Schatzberg elected to present no statistical analyses of outcome at all. We now know from Corcept’s three negative FDA-monitored Phase III trials that the early challenges to mifepristone’s efficacy had merit (see Abstract #63). As Dr. Schatzberg should know, wishful thinking and hand waving are no substitute for quality data that are rigorously analyzed. Dr. Schatzberg is accountable for the unwarranted positive tone of those early publications. His later promotional claims reveal a commercial bias that has no place in academic medicine.

Scientists are required to present both the positive and the negative aspects of their research “fully and honestly.” Here is one example of this code (see item 2). Here is another (see item 15). The standard does not change when commerce is involved. The standard applies in review articles as much as in original reports. Did the Stanford-Corcept group meet this standard? That is the issue Stanford now confronts.

In a 2003 textbook that he coauthored (Manual of Clinical Psychopharmacology), Dr. Schatzberg made a clearly biased claim. He stated “Two recent studies of … (mifepristone) … by our group have demonstrated efficacy in the treatment of psychotic depression.” He was referring to the 2001 and 2002 Stanford-NIH publications. Dr. Schatzberg went on to state that “The effects were more pronounced on the psychosis than on the depression, but both were significantly improved within 7 days.”

The claim of “demonstrated efficacy” is categorically false. It may be fraudulent. Neither study achieved any accepted evidence of efficacy. The second claim also is categorically false. No quantitative data on significant improvement of psychotic symptoms or depression severity exist in the two studies for drug over placebo. Lack of efficacy is what the investigative journalist Paul Jacobs exposed so well, with the consultation of independent statisticians, in the San Jose Mercury News July 10, 2006.

These false claims in a well-known textbook promoted a positive climate of opinion for Dr. Schatzberg’s drug in the period leading up to Corcept’s planned IPO. Through the textbook, Dr. Schatzberg also promoted positive professional expectations in advance of the marketing of his drug. Moreover, in this American Psychiatric Press publication Dr. Schatzberg did not disclose his large financial interest.

The pattern of exaggerating positive trends in the data while suppressing mention of negative analyses has become a trademark of Stanford-Corcept publications. Dr. Schatzberg made similar unwarranted claims in a 2003 journal article. In 2006 Dr. Schatzberg did likewise in another APA Press publication (The American Psychiatric Publishing Textbook of Mood Disorders). Once again, Dr. Schatzberg used an APA Press textbook to talk up his drug’s prospects. Once again, Dr. Schatzberg did not disclose his own financial interest.

In 2006, Corcept’s CEO (Dr. Joseph Belanoff) and the former Chief Medical Officer (Dr. Charles DeBattista) published a review article on use of mifepristone in neuropsychiatric disorders. In violation of the journal publisher’s policy, the article was sourced only to Stanford University. These current and past Corcept officers did not disclose their financial interest or their relationship to the corporation. They highlighted the use of mifepristone in psychotic depression. Referring to the 2001 Stanford-NIH trial, they claimed “all five patients showed a substantial improvement in depression.” This claim is categorically false. The question of fraud arises again.

These examples illustrate what I have termed commercial promotion and branding through academic outlets. In all, 13 scientific journals, 4 textbooks, and other media have been tainted by these and similar misrepresentations. These biased claims in reference to the Stanford-NIH trials were also featured in press releases and press interviews while Corcept was raising private and public capital.

This pattern of bias raises questions of ethics, and of possible fraud. One could make the case that the misrepresentations described here do not fall under the safe harbor provisions of US securities law. It is one thing for entrepreneurs to speculate giddily about the potential of their product if the available scientific data are presented accurately (“fully and honestly”). In that circumstance, caveat investor applies. But it is an altogether different matter when entrepreneurs misrepresent the available scientific data by emphasizing positive trends while suppressing the inconvenient, negative aspects of the data. That is what Corcept did in discussing the 2002 Stanford-NIH trial in its SEC filing March 19, 2004 leading up to the IPO. That is also what Dr. Schatzberg and his associates did in the examples given above.

Dr. Schatzberg is slated to become president of the American Psychiatric Association in 2009. The APA would do well to reconsider his fitness for that office. We would have expected Dr. Schatzberg, who is also President of the American Psychiatric Press, the publishing arm of the APA, to understand that he and his corporate associates should discuss their scientific data “fully and honestly,” while being transparent about their competing financial interests. As for Stanford University, these academic biases and systemic conflicts of interest signal the need for a fresh look at the academic-corporate boundary.