Serious flaws and conflicts skewed results of largest antidepressant study ever done

In 2006, researchers first published results from a $35 million NIMH-funded study of antidepressants known as STAR*D, claiming it proved the effectiveness of second-generation antidepressants used alone and in combination with each other. The NIMH chimed in with press releases extolling "new strategies" that help depressed patients become symptom-free, and the findings became the basis for American Psychiatric Association's guidelines calling for the open-ended use of antidepressants in treating depression.

But, as Edward Pigott, a Maryland psychologist, reveals in several published papers and his blog, it was all a big lie. Pigott shows how the STAR*D authors, 10 of whom had financial ties to antidepressant makers, played unethical games with the data to make all of the antidepressants in the study look far more effective than they really were. For instance:
       *The researchers changed the primary outcome measure from the Hamilton rating scale of depression (considered a gold standard in measuring depression) to a proprietary rating scale owned by the principal investigator, Dr. John Rush, a psychiatrist at the University of Texas (who by the way was investigated by Senator Grassley for failing to disclose extensive conflicts of interest). And they made the change even though the secondary rating scale had been used in clinical treatment, thus tainting it as an objective research measure. Furthermore, in the published results, Rush and his co-authors didn't bother to disclose this change (which skewed the results in favor of the drugs).
       * They failed to count patients who had dropped out as treatment failures, thus further skewing results in favor of the drugs.
       * Halfway through the study, they included patients with only mild depressive symptoms who had originally been excluded because they didn't meet the original criteria for being depressed, again making the results look better than they were.
       * They repeatedly rounded up percentages to make it look as if the antidepressants in the study were more effective than they really were. Yet they didn't round up any findings showing the percentage of negative side effects among patients taking these drugs.

After re-analyzing the data from STAR*D, Pigott found that in contrast to STAR*D’s published findings, only 108 of its 4,041 patients (2.7 percent) went into remission in the acute phase of the study. And of those initial patients, only 38 percent obtained remission after being dosed with other medications in three later phases of the trial. In every phase, more patients dropped out than were remitted, and this drop-out rate increased throughout the study. As Pigott says in a paper published this month in the journal Ethical Human Psychology and Psychiatry, "this reality directly counters the study's false claim that about 70 percent of those who did not withdraw from the study became symptom-free."

Three years later, in a review article for the Journal of Clinical Investigation, Dr. Thomas Insel, director of the NIMH in 2006 and today, pretty much acknowledged the inadequacy of second-generation antidepressants in treating depression. To quote Insel:

In 2007, the third and fourth most heavily purchased medications in the United States were antipsychotics and antidepressants, respectively, with a combined market of $25 billion. Remarkably, despite the heavy use of these medications, we have no evidence that the morbidity or mortality of mental disorders has dropped substantially in the past decades.

Instead, as Pigott points out in his most recent paper, "the morbidity and chronicity of mental disorders appears to be increasing with a twofold to threefold increase between 1987 and 2007 in the number of Americans receiving disability payments for such disorders." Robert Whitaker, of course, argues in his new book, Anatomy of an Epidemic, that it is the very overuse of so many psychoactive drugs with severe side effects that has led to this exponential increase in the number of Americans disabled by mental illness.

Whether or not you buy Whitaker's hypothesis -- and it certainly warrants further investigation -- the fact is that the STAR*D study, like Paxil study 329 was flawed in so many ways that its published results should be retracted.
Not only did most of the STAR*D authors including John Rush have financial ties to Forest Labs, the maker of Celexa, and other antidepressant makers who stood to benefit from the study's positive findings. But as Pigott reveals, the very same NIMH officials who were tasked with the oversight of this $35 million multi-center study were also allowed to put their names on STAR*D studies published in the New England Journal of Medicine and the American Journal of Psychiatry, an egregious conflict of interest that should never have been allowed.

Interestingly, while Insel acknowledges the inadequacy of current antidepressants in his 2009 review, there is nothing on NIMH's website that points to this more sober assessment, despite the existence of five meta-analyses that show "modest to no" advantage of antidepressants over placebo in clinical trials. Pigott concludes that it is hard to find any reason for this pro-drug bias "other than convention, ease to prescribe for physicians, and the success of pharmaceutical companies' relentless marketing efforts."

Let's fix the perverse financial incentives in Medicaid before hacking its budget

Medicaid and Medicare are in the news of late, as Congressional Republicans spar with President Obama and the Democrats on how best to rein in the ballooning costs of these entitlement programs, which make up a growing share of federal and state budgets. But what few policy makers seem to be talking about -- at least in public -- are the perverse financial incentives built into the system that allow the companies who manage Medicaid contracts on a statewide basis to make profits at the expense of quality health care for Medicaid recipients. As a result of these perverse incentives, many Medicaid patients, particularly those being treated for mental health problems, are either terribly under-served or in some cases, badly over-treated. The upshot, in many cases, is poorer health outcomes and higher Medicaid costs.

Allow me to explain. In most states, Medicaid contracts for mental health and substance abuse are separated out from medical services and handled by managed care companies that make their money from keeping patients out of expensive hospitals. As a result of the way many contracts have been drawn up, drug treatments for emotional and behavioral problems do not count as costs to the behavioral health managed care entity, but instead fall under the plan's general prescription drug coverage. However, hospitalizations and psychosocial treatments (such as different forms of therapy and family interventions) do fall under the managed care's costs. So to make a profit, managed care in almost every state has done a spectacular job over the last two decades of limiting hospitalizations and access to psychosocial interventions, according to a new Hastings Center Special Report, which focuses primarily on troubled children.

As the report concludes, it's much easier for "patients to obtain referrals for medication management and psychopharmacology." This perverse incentive goes a long way to explaining why so many Medicaid recipients, both children and adults, are prescribed expensive psychotropic drugs, which cause serious side effects, rather than alternative therapies, which might be more effective and less hazardous to their health. Managed care companies make more money that way.

As one top medical official for New York State told me, "If you're a managed care entity and the pharmacy benefit is not part of your budget, you're happy if a symptom can be controlled by someone prescribing a dubious and expensive drug. It's not your problem." Or, as the authors of the Hastings report put it,

"The country's mental health care system makes it difficult for children to access psychosocial care, but relatively straightforward to access medication treatments (even if those treatments are not monitored or reassessed as recommended)."

And the same, of course, holds true for adults.

Not only do individuals with mental health problems suffer as a result, but state governments actually end up spending more when these patients don't receive the most effective treatments and interventions. Some patients will inevitably "end up on the streets or in prison where the costs don't accrue to the managed care companies," the New York state medical official explains. The states, of course, still pick up the costs -- but out of separate budgets for prisons, homeless shelters and programs for troubled teens.

The plot thickens. In many states, large insurance companies such as United Health care, WellPoint and Blue Cross Blue Shield have joined large managed care companies such as Americhoice and Centene in managing Medicaid contracts, according to this article in The Washington Post. Some states have flat-rate contracts with managed care companies that are designed to balance the need to limit costs with providing appropriate mental and physical health care. But in other states such as Colorado, Florida and Tennessee, managed care companies have what is known as "at-risk contracts" -- contracts whereby they make a percentage profit based on how much is spent on each Medicaid patient. And with such contracts comes another perverse financial incentive that, in some cases, make patients sicker.

For example, some managed care companies actually benefit twice-over when patients are over-prescribed anti-psychotic drugs such as Seroquel and Zyprexa, according to Edward Knight, a former executive with a managed care company in Colorado and a recovery consultant. This is how: while the behavioral health arm of the company makes money by referring patients for drug prescriptions rather than more expensive psychosocial treatments, another arm of the company makes money by managing the diseases such as obesity and diabetes caused by the over-use of these drugs. The off-label use of anti-psychotics like Seroquel, Risperdal and Zyprexa is widespread in many states, with disastrous effects on some patients, as many bloggers, myself included, have already noted. To think that some companies managing Medicaid contracts are actually profiting from that over-use is indeed perverse.

Seems to me that until we do something about the conflicting financial incentives that are built into the Medicaid system, fighting over how to cut Medicaid funding to the states is a waste of time. As the New York medical official says, setting up contracts based on a specific dollar amount is fruitless. "If it's just dollars, [the managed care company] will cheat and figure out ways to get other entities to pick up costs, or they'll just deny care," he says. "We have to change the contract incentives so that these companies can earn X number of dollars when they improve health care. We have to get to risk-based contracts based on performance metrics rather than profits." Amen.

A tale of censorship and secrecy starring the American Psychiatric Association

Psychiatry is supposed to be all about disclosure, disclosing the dark secrets of one's past to a professional in an effort to heal or, at the very least, figure out why one is in such psychic pain. But given the recent actions of the American Psychiatric Association, the largest trade group for psychiatrists in the U.S., one might get the impression that the profession is really all about censorship and obfuscation.

Remember when the Project on Government Oversight (POGO), as part of an effort to get the NIH to crack down on ghostwriting, released documents showing that a psychopharmacology handbook for primary care doctors, authored by then psychiatry kingpins Charles Nemeroff and Alan Schatzberg, had actually been ghost-written by a company hired by GlaxoSmithKline, the maker of the blockbuster antidepressant, Paxil? The New York Times broke the story last fall, relying on internal Glaxo documents obtained in the course of a lawsuit against the drug giant, and a number of other journalists, including myself, blogged about it -- see here. The documents showed that Glaxo hired Scientific Therapeutics Information (STI) to prepare a draft of the textbook with the understanding that Glaxo could review the initial drafts before publication. The handbook itself, which was published in 1999 by the APA, essentially promoted Paxil, among other drugs, as a safe and effective treatment for anxiety and depression.

Sound familiar? As Side Effects reveals, that's exactly how the initial drafts of the notorious Paxil study 329 were prepared -- by the same company, STI, again under contract to Glaxo. In a remarkably similar arrangement, the authors of the Paxil study did not object to the draft's positive wording and made only minor changes to the paper, which concluded that Paxil was safe and effective in treating depression in adolescents even though the actual data showed quite the opposite -- see back story here. (In its later review of Paxil and other antidepressants before attaching black box warnings to them, the FDA labeled study 329 a "negative finding" because the study did not show statistical evidence of the drug's effectiveness and did not accurately represent the extent of Paxil's suicidal side effects among adolescents in the trial).

To get back to the story at hand, after the Times article was published, Nemeroff, who was drummed out of Emory for failing to disclose myriad financial ties to the drug industry and is now chair of psychiatry at the University of Miami, and Schatzberg, the dethroned chair of psychiatry at Stanford and a past president of the APA, called in their lawyers who insisted on several corrections to the Times piece, which I append here:

A headline on Nov. 30 with an article about SmithKline Beecham’s role in the publication of a book about treating psychiatric disorders overstated SmithKline’s actions. While documents show that SmithKline (now known as GlaxoSmithKline) hired a writing company for the book, they do not indicate that the company wrote the book for the authors, Dr. Charles B. Nemeroff and Dr. Alan F. Schatzberg. The article also described incorrectly, in some editions, events outlined in a letter from the writing company to Dr. Nemeroff. The correspondence proposed a timeline for the writing company to furnish the doctors and SmithKline with draft text and final page proofs for approval; the letter did not say that the company had already provided those materials for final approval. And the article misstated the context under which Dr. David A. Kessler, the former commissioner of the Food and Drug Administration, commented about the book’s production. The letter and other documents were described to him; he did not personally review the documents.

All well and good, but then the APA, probably in response to outraged queries from its members, wrote a self-serving and inaccurate piece about the whole affair in its news bulletin, Psychiatric News. Its bromide said the corrections "seemed to throw into doubt the central premises of the original article" and "seriously overstated the role of the writing company." Neither of these allegations, of course, are true -- the corrections appended to the Times article do not change the basic fact that the handbook was largely ghost-written by a company hired by Glaxo to promote Paxil and other antidepressants to primary care doctors.

Indeed, two prominent psychiatrists pointed that very truth out in a letter to the editor of Psychiatric News. While Ron McMillen, a spokesman for the APA, argued in the PN article that Nemeroff and Schatzberg were actively involved "in every stage of the book's development," the letter's authors note that McMillen did not venture to say that Nemeroff and Schatzberg "wrote the initial drafts -- or indeed any drafts." The letter notes:

"The released 49-page sample chapter draft provided by Diane Coniglio and Sally Laden, employees of Scientific Therapeutics Information (STI), is largely reproduced verbatim in the Handbook.

In their January 28, 2011 letter to the editor, Dr. Bernard Carroll, professor and chairman emeritus of the department of psychiatry at Duke University Medical Center, Dr. Robert Rubin, professor and vice chair of the department of psychiatry at the University of California, Los Angeles, and Leemon McHenry, a researcher at California State University, also called on the APA to release internal documents that might shed light how much influence Glaxo did have on the context and tone of the book and its role in approving drafts. The letter writers concluded, "This case highlights the need for disclosure if we ever are to understand the scale of corporate influence on academic publishing."

Here's where things really get absurd. After sitting on the letter for two months, Psychiatric News decided not to publish it. In an email to Carroll et al dated March 22, Cathy Brown, executive editor for the publication, wrote that the issue has been covered extensively already and "therefore, we will not be printing additional information about it at this time."

How sad. Can it be that the APA is so fearful of its members' reactions (or more likely, legal threats by Nemeroff and Schatzberg) that it cannot even run a short and well-reasoned letter to the editor? For an organization that represents a profession whose very modus operandi is based on disclosure and transparency, the APA's track record of censorship and secrecy is unacceptable.