Summing up the STAR*D Scandal: The Public was Betrayed, Millions were Harmed, and the Mainstream Media Failed Us All

This article was first published by Mad in America on 4/1/25

Last spring, it seemed certain that the story of the STAR*D scandal, which Mad in America has been reporting on for 14 years, would finally attract the attention of the mainstream media. All of the ingredients for a blockbuster article were now clearly visible, including an acknowledgement from inside psychiatry that this story was of profound importance for all of our society.

Yet, the media has remained silent, and now the scandal is fading away. American psychiatry has weathered the crisis; it will not have to confront a public stunned by news of how the oft-cited 67% cumulative remission rate, in the “largest and longest study ever done to evaluate depression treatment,” was born of scientific misconduct. Instead, that finding will remain in the literature, evidence that can be cited by the media and by the field of the effectiveness of antidepressants.

As such, the scandal now serves as a historical verdict on the ethics of American psychiatry, and by extension, the National Institute of Mental Health (NIMH). As for the mainstream media, this is a story of how it utterly failed the public, cowed into silence by a medical discipline that, for decades, has used the media to promote a narrative that privileges its guild interests but is belied by its own science.

Here is a recap of this scandal, and the harm done.

The Story That Was Told to the Public

When the NIMH launched the $35 million STAR*D study, it promised that the results would be rapidly disseminated and used to guide clinical care. It was well known that industry-funded trials of antidepressants used inclusion-exclusion criteria to select patients most likely to respond well to the drug, and that because of such criteria, 60% to 90% of real-world patients didn’t qualify for such studies. The STAR*D study would be conducted in real-world patients, which is why it would be of such prime importance.

“Given the dearth of controlled data (in real-world patients groups), results should have substantial public health and scientific significance, since they are obtained in representative participant groups/settings, using clinical management tools that can easily be applied in daily practice,” the STAR*D investigators wrote.

The STAR*D trial provided patients with up to four tries of finding an antidepressant or combination of antidepressants that worked. If a first antidepressant hadn’t led to a remission at the end of 12 weeks, then a second antidepressant could be tried for another 12 weeks, and so on, through four levels of treatment. Whenever a patient was found to have remitted at the end of one of the treatment periods, the patient was whisked into a year-long follow-up study to see if the remission could be sustained.

In 2006, the American Journal of Psychiatry (AJP) published four reports on STAR*D outcomes. Three told of remission rates after the first and second stages of the study, and then in November of that year it published a summary of outcomes from the trial. The abstract in the summary report told of a therapeutic success:

The NIMH stayed true to its promise to rapidly disseminate the study findings, trumpeting the good news in a press release. “Over the course of all four levels, almost 70 percent of those who didn’t withdraw from the study became symptom free.”

This “nearly 70%” figure was the bottom-line result from this landmark study, and in the years that followed, the STAR*D investigators and other leaders in American psychiatry published scores of papers touting the 67% remission rate, with graphics like the one below summarizing the outcomes.

The clinical protocol was clear: If the first drug didn’t work, try a second, and if a second didn’t work, try a third or a combination of drugs, and if that didn’t work, try yet again with another antidepressant. Nearly 70% of real-world patients could expect to become “symptom free” with this use of antidepressants.

Deconstructing the Study

While the November 2006 report told of a therapeutic success, readers who were familiar with the STAR*D articles published earlier that year, telling of the results from the first and second treatment levels, could see that something was amiss with this chain of publications. Most notably, the stated number of eligible patients had changed with every publication, from a high of 4,041 to a low of 2,876. The authors also switched from using the HAM-D as the principal instrument to assess the stage one remissions to an instrument called QIDS in the November article, without any explanation for why they had done so.

Moreover, while the abstract of the November report told of a cumulative remission rate of 67%, the discussion section of the paper revealed that this was, in part, a made-up number. The researchers had theorized that if those who had dropped out had remained in the study through all four stages of treatment, they would have remitted at the same rate as those who hadn’t dropped out, and they added these imagined remissions to their tally of patients who had gotten well. Without these imaginary remissions, the reported remission rate in the November 2006 paper would have been around 50%.

There was one other red flag. The paper did not disclose the sustained remission rate at the end of the one-year follow-up. How many of those who had remitted had stayed well? All that the STAR*D investigators wrote was that relapse rates during the follow-up were higher for those who had taken multiple tries to remit. They did publish a graphic which purported to tell of the survival rate for the 1,518 patients who entered the follow-up trial, but as there was no discussion of how to read the graphic, it was impossible to make sense of the numbers.

While these red flags were present in the November 2006 article, any lasting questioning of the STAR*D results would likely have disappeared if it weren’t for the dogged investigative work of psychologist Ed Pigott. He could see something wasn’t right, and once he had obtained a copy of the STAR*D protocol through a freedom of information (FOI) request, he and his colleagues, starting with a paper published in 2009, were able to detail the ways that the STAR*D investigators had deviated from the protocol to inflate their published remission rates.

In a 2010 paper, Pigott and colleagues also made sense of the graphic representing the one-year results. Of the 4,041 patients who entered the trial, only 108 had remitted and then stayed well and in the study to its one-year end, a documented stay-well rate of 3%. All of the others had either never remitted, remitted and then relapsed, or dropped out of the study. This was an outcome that was quite at odds with the 67% “cumulative remission rate” being touted to the public as evidence of the effectiveness of antidepressants.

Mad in America’s review of Pigott’s 2010 paper was titled: “The STAR*D Scandal: A New Paper Sums It All Up.” In 2011, Pigott published a paper titled “STAR*D: A Tale and Trail of Bias,” which laid out the protocol violations in detail, and Mad in America posted the documents that he had obtained through his FOI requests. The public now had easy access to source documents that told of how the STAR*D investigators had deviated from the protocol to produce their inflated 67% remission rate.

Pigott and colleagues published articles in 2015 and 2018 on the STAR*D trial, and each time Mad in America reviewed the articles. In our 2015 book Psychiatry Under the Influence, Lisa Cosgrove and I wrote about the STAR*D scandal in depth, as it served as an example of the “institutional corruption” in psychiatry due to pharmaceutical interests and psychiatry’s own guild interests.

Yet, even as Pigott and colleagues published their findings in peer-reviewed journals, the mainstream media remained silent about this challenge to the STAR*D findings. Instead, mainstream media outlets, such as The New Yorker, cited STAR*D as evidence of the efficacy of antidepressants. As far as I can tell, the only exception to this media silence was when Medscape reported on Pigott’s 2010 paper, its report confirming that the 3% stay-well rate was accurate.

Although there were writers, such as Bruce Levine, who wrote about the STAR*D on other online media sites, such as CounterPunch, the mainstream media remained silent. The New York Times, The Washington Post, the Los Angeles Times, and on and on: Pigott’s deconstruction of the study, while well known to Mad in America readers, did not find its way into any major publications.

The RIAT Reanalysis

In 2019, Pigott and his colleague Jay Amsterdam, who is an emeritus professor of psychiatry at UPENN, obtained access to the patient-level data from the STAR*D trial through the Restoring Invisible and Abandoned Trials initiative (RIAT). This initiative had been established in 2013 to enable researchers to reanalyze data sets from publicly funded trials, particularly when there was reason to think that the initial results may have been incorrectly reported. The other members of their team were Thomas Kim, Colin Xu, and Irving Kirsch.

A RIAT reanalysis meant that their published findings would carry a societal stamp of approval for correcting the research record. Pigott and colleagues could now go through the case reports for each of the 4,041 patients enrolled into the study and assess their progress through the study: their baseline HAM-D scores, whether their HAM-D scores showed them to be in remission at the end of one of the stages of treatment, and whether they had dropped out before remitting. They could compile these outcomes based on following the protocol for doing so.

The STAR*D investigators, in their November 2006 report, had stated that 67% of the patients had remitted by the end of all four levels of treatment (2,460/3,671). In an article published in BMJ Open in August of 2023, Pigott and colleagues found that if the protocol had been followed, 1,089 of 3,110 patients would have been reported as having remitted after all four stages of treatment (35%).

Moreover, with access to patient-level data, Pigott and were able to publish a precise count of how each of three protocol violations had been used by the STAR*D investigators to inflate the remission rate.

1. There were 607 patients enrolled into the trial who weren’t depressed enough to be eligible for the trial (the study required a baseline HAM-D score of 14 or higher), and another 324 who lacked a baseline score. In their initial report on remissions after the first stage of treatment, the STAR*D investigators had rightfully noted that these 931 patients didn’t meet eligibility criteria, and thus weren’t “evaluable” patients. However, the STAR*D investigators snuck this group of 931 back into their count of remissions when they reported level two results, and they continued to include this group as “evaluable” patients when they published their cumulative remission rate in November 2006. This added 570 to their tally of remitted patients.
2. The protocol stated that the HAM-D would be used to assess remissions. At clinic visits, the clinicians would also administer an unblinded tool, the Quick Inventory of Depressive Symptomatology (QIDS), to assess how patients were doing, and since the study was meant to mimic real-world care, this instrument would be used to assess whether dosages should be altered and whether other medications—such as sleep medications—should be prescribed. The protocol explicitly stated that “research outcomes are not collected at the clinic,” and thus the QIDS would not be used as an outcomes instrument. However, in their November 2006 report, the STAR*D investigators used the QIDS scores to report remission rates, omitting any mention of HAM-D scores. This switch to an unblinded instrument, one that was not supposed to be used to report research outcomes, added 195 to the count of remitted patients.
3. Although the study protocol was silent on how dropouts would be counted, a 2004 article by the STAR-D investigators stated that patients with missing HAM-D scores at the end of each treatment step were “assumed to not have a remission,” meaning that they were to be chalked up as treatment failures. Instead, as noted above, the STAR*D investigators theorized that if the dropouts had stayed in the study through all four levels of treatment, they would have remitted at the same rate as those who did stay in the study. This thought experiment, which turned “treatment failures” into remitters, added another 606 to their count of remitted patients.

With these three protocol violations (and one other lesser one), the STAR*D investigators changed the remission equation from 1,089/3,110 (35%) into an equation of 2,460/3,671, a remission rate of 67%. The protocol violations added 1,371 to the count of remitters, meaning that 56% of the remissions came from these deceptions.

The Media Remain Silent

In September of 2023, MIA published a long report titled “The STAR*D Scandal: Scientific Misconduct on a Grand Scale.” We also put up a petition on change.org calling for the AJP to retract the 2006 paper. While AJP editor-in-chief Ned Kalin ignored our petition, which had been signed by more than 2000 people, it seemed to strike a nerve, and in December the AJP published a commentary by John Rush, the principal investigator of the STAR*D study, and four of his colleagues, that sought to defend the integrity of their November 2006 article.

Their reply to Pigott and colleagues, arrogant in tone, revealed an utter disregard for truth.

The most egregious violation that Pigott and colleagues had identified was that the STAR*D investigators had counted remissions in the group of 931 patients who lacked a baseline HAM-D score, or who weren’t depressed enough, based on their HAM-D score, to qualify for the study. As the STAR*D investigators had acknowledged in their report on stage one outcomes, this group were not “evaluable” patients. Yet, in their commentary, titled “The STAR*D Data Remain Strong: Reply to Pigott et al.,” the STAR*D investigators accused Pigott and colleagues of applying “post-hoc criteria” to wrongfully remove these patients from their reanalysis of remissions.

They wrote:

And:

And still more:

The AJP, by publishing this commentary, had now added its name to the scandal. Pigott and colleagues had specifically calculated outcomes according to the protocol—this was the whole point of their RIAT reanalysis—and so the AJP was now presenting to its readers an accusation that was obviously false. Rush and colleagues, together with the AJP, apparently were counting on AJP readers to be so lacking in curiosity that they would not bother to read the Pigott paper, or, even if they had, to be so primed to discount challenges to their profession that they would just assume the STAR*D authors—key opinion leaders in their field of depression—must be telling the truth.

At the same time, for those who had read Pigott’s paper and understood its RIAT-defined methodology, it was evident that the STAR*D investigators could not provide an honest explanation to defend the integrity of their November 2006 paper. Instead, they had made up this “post hoc” accusation in an effort to impugn the credibility of Pigott and colleagues. This would seem to meet the legal criteria for libel, given that both the STAR*D authors and the AJP knew that Pigott and colleagues had done no such thing.

However, the fact that Rush and colleagues published their letter was a sign that public disclosure of the scandal was gaining some momentum. Shortly after their commentary appeared in the AJP, the Psychiatric Times published a cover story on Pigott’s paper. It was titled “STAR*D Dethroned?” and subheaded “Since 2006, STAR*D stands out as an icon guiding treatment decisions of major depressive disorder. But what if it is broken?” This was the moment that it seemed surely the media silence would break.

In his essay, Psychiatric Times editor John Miller neatly laid out the impact STAR*D had on psychiatry. “For us in psychiatry, if the BMJ authors are correct, this is a huge setback, as all of the publications and policy decisions based on the STAR*D findings that became clinical dogma since 2006 will need to be reviewed, revisited, and possibly retracted,” he wrote.

At this point, all of the ingredients for a blockbuster story were in place. This was an acknowledgement from within psychiatry that it was urgent that psychiatry investigate Pigott’s reanalysis, and that if it held up, it would be evident that prescribers of antidepressants, and all of American society, had been grievously misled by the STAR*D investigators, with much societal harm done as a result.

With Psychiatric Times having weighed in, multiple people sent emails to editors at major newspapers telling of this unfolding scandal, urging them to cover the story. And if the media needed any additional reason to cover the story, it came in March 2024, when psychiatrists Nicolas Badre and Jason Compton published a letter in Psychiatric Times titled “STAR*D: It’s Time to Atone and Retract.” They laid out in detail the ethical imperative that the field of psychiatry now faced.

And:

The story was all lined up for the media, with documents neatly in a row: the original 2006 report with the red flags that could now be identified; the dogged pursuit of the “truth” by Ed Pigott; the RIAT reanalysis published in a prestigious journal that identified the protocol violations the STAR*D investigators had employed to grossly inflate the remission rate; the letter by Rush and colleagues defending their work that could easily be shown to lack credibility; and finally the Psychiatric Times story that told of the extraordinary impact the STAR* study had on prescribing practices. There was a list of people to interview, and when it came to fleshing out the harm done, reporters could interview patients who had been prescribed antidepressants since 2006 and ask them: would they have made a different decision about their use if they had known the true STAR*D results, including the 3% stay-well rate at the end of one year?

Yet, even though editors at major newspapers were contacted, and Pigott and Amsterdam had interactions with a reporter or two, the months ticked by, and no articles about the STAR*D scandal appeared in the mainstream media. And with each passing month of silence, the possibility that the general public would ever learn about this scandal began to fade away.

This past October, Pigott and colleagues sent a letter to AJP editor Ned Kalin, detailing the “scientific errors” present in the November 2006 article that told of a 67% remission rate, and calling for the article to be retracted. Given that it was coming from the authors of the reanalysis, this retraction request provided the media yet another possible news hook for the story.

Mad in America published a copy of their retraction letter, but the media silence continued. Kalin did not respond to their letter, and so, given that nearly 18 months have passed since Pigott and colleagues published their RIAT reanalysis, it now appears that news of this scandal is never going to break out of the Mad in America “ghetto,” which is to say that while it is well known to our readers, it will remain unknown by the general public.

Pigott, Amsterdam, and colleagues are continuing their analysis of the patient-level data in the STAR*D study, looking at claims made by the STAR*D investigators in other articles they published, and their ongoing work is certain to expose other aspects of this scandal, such as a failure by STAR*D investigators to report serious adverse effects and drug-induced suicidal ideation. However, it appears unlikely that such future reports will stir the mainstream media into action, given that the blatant scientific misconduct revealed by their RIAT reanalysis failed to do so.

As such, it appears that American psychiatry is going to weather the crisis and avoid having to answer to the public for its scientific misconduct and its promotion of a false 67% remission rate to the public. Instead, the 67% remission rate will remain in the research literature, ready to be cited by the media in articles about the merits of antidepressants.

Indeed, on December 10, The New York Times,  in an article titled “7 Things Everyone Should Know About Antidepressants,” told readers yet again that “a large study of multiple antidepressants found that half of the participants had improved after using either the first or second medication that they tried, and nearly 70 percent of people had become symptom-free by the fourth antidepressant.”

Such was the news that TheNew York Times found fit to print.

A Defining Challenge to American Psychiatry

This scandal presented American psychiatry, and by extension the NIMH, an existential challenge, one that would define their relationship to society. Would American psychiatry do the right thing and investigate whether the November 2006 report on the STAR*D outcomes should be retracted?

That was all that John Miller and the Psychiatric Times were demanding of the American Journal of Psychiatry and the “field” of psychiatry. Launch a public investigation into this matter and determine whether the article should be retracted.

The standards for retracting an article in a medical journal are clear. A  2011 article on retraction in a medical journal noted the following:

This case met that standard. The 67% remission rate published in the AJP in November 2006 could no longer be “considered trustworthy due to scientific misconduct.” Retraction of the article was “essential for correcting the literature and maintaining trust in the scientific process.”

Moreover, the World Association of Medical Editors, in its “Professional Code of Conduct,” specifically states that “editors should correct or retract publications as needed to ensure the integrity of the scientific record and pursue any allegations of misconduct relating to the research, the reviewer, or editor until the matter is resolved.”

Finally, it is evident that this is a story of research fraud. There is an obvious intent to deceive that is present: the inclusion of 931 who weren’t eligible for the trial in the tally of remissions isn’t an instance of an unwitting scientific mistake, but rather this group of ineligible patients was snuck back into the list of evaluable patients in order to inflate the cumulative remission rate. The switch from using the HAM-D to assess remission rates to the QIDS instrument, which the protocol explicitly states would not be used to assess outcomes, is another clear example of an intent to deceive.

There are three primary categories of research misconduct—fabrication, falsification, and plagiarism—and this is an instance of “falsification” of results, as opposed to “fabrication” of data. Falsification is defined as “manipulating research materials, equipment, or processes, or changing or omitting data or results such that the research is not accurately represented in the research record.”

The letter by Badre and Compton, titled “STAR*D: It’s Time to Atone and Retract,” best summed up the ethical challenge that was presented to American psychiatry by the RIAT reanalysis. The key word they used was “atone,” which, of course, has a meaning that derives from Biblical scriptures. In the Bible, atonement, as one online definition states, is the “process of fixing relationships that have been broken by wrongdoing. It’s a prominent concept in both Judaism and Christianity, where it refers to the idea that humans must atone for their sins against God.”

In the case of STAR*D, the falsification of results could be seen as a “sin” against the standards of science, which impaired the “relationship” between a medical profession and the public it is supposed to serve. Badre and Compton were asserting that in the absence of a compelling explanation by the STAR*D investigators for their protocol violations, the field, in order to repair its relationship to the public, needed to confess to their falsification of data and retract the study so that it no longer stained the evidence base for antidepressants.

There Will Be No Atonement

While the STAR*D scandal may seem like a “one-off” instance of falsification, it can be best described as a galling chapter in a larger story of American psychiatry failing to fulfill its duty to be a faithful communicator of scientific findings to the American public. This has been particularly true in terms of its communications to the public about depression and the merits of antidepressants.

This ongoing failure can be traced back to the publication of DSM-III in 1980, when American psychiatry adopted a “disease” model for categorizing and treating mental disorders. At that time, the American Psychiatric Association (APA) launched a public relations campaign to sell this new model to the public (an effort that was funded in large part by pharmaceutical companies). The public was informed that research had shown that mental disorders were diseases of the brain, and there was reason to think they were due to chemical imbalances. We were told that psychiatric drugs fixed those chemical imbalances, like insulin for diabetes.

This was a story of a great medical advance. The second-generation SSRI antidepressants, starting with Prozac’s introduction into the market in 1988, were hailed as breakthrough medications, as were the second-generation “atypical” antipsychotics. Meanwhile, pharmaceutical money flowed to the APA to promote this disease model, and flowed to academic psychiatrists who served as their “key opinion leaders” (KOLs).

The American public organized its thinking and use of psychiatric services around this narrative of a great medical advance. The diagnosing of depression and other mental disorders increased dramatically, and so too the prescribing of psychiatric drugs. Spending on psychiatric drugs in the United States rose from around $800 million in 1987 to nearly $40 billion 20 years later.

This was the governing public narrative when the STAR*D trial was conducted. In 2005, a year before the STAR*D reports appeared, the APA published a press release telling how a survey had found that “75 percent of consumers believe that mental illnesses are usually caused by a chemical imbalance in the brain.” This, said APA president Steven Sharfstein, was evidence of “good news for [public] understanding of mental health.” A psychiatrist, the press release noted, was a “specialist specifically trained to diagnose and treat chemical imbalances.”

That same year, the APA published a “Let’s Talk Facts About Depression” brochure, which delivered the same message: “Antidepressants may be prescribed to correct imbalances in the levels of chemicals in the brain.”

However, while American psychiatry was informing the public of the wonders of antidepressants and other psychiatric drugs, the scientific literature was telling a very different story about the merits of the “disease” model of care. Specifically, researchers were failing to find evidence that people diagnosed with major mental disorders suffered from a chemical imbalance; psychiatric drugs, rather than fix known chemical imbalances, were instead found to induce abnormalities in neurotransmitter function; and NIMH studies were telling of poor long-term outcomes.

In the case of antidepressants, a search of the scientific literature—literature published prior to 2006—had this story to tell:

• In 1984, the NIMH reported that a study of depressed patients had failed to find evidence that a lesion in the serotonergic system was a primary cause of depression. This was the first of many such failures, and the third edition of the APA’s Textbook of Psychiatry, published in 1999, acknowledged this fact, writing that decades of research “has not confirmed the monoamine depletion hypothesis” (serotonin is a monoamine). In 2005, Kenneth Kendler, editor-in-chief of Psychological Medicine, wrote that “We have hunted for big simple neurochemical explanations for psychiatric disorders and not found them.”
• Prior to the introduction of antidepressants, depression was understood to run an episodic course. However, epidemiological studies in the 1980s found that depression was now running a more chronic course, which led at least a few to worry that antidepressants were causing a chronification of the disease. In 1994, Italian psychiatrist Giovanni Fava wrote: “Antidepressant drugs in depression might be beneficial in the short term, but worsen the progression of the disease in the long-term, by increasing the biochemical vulnerability to depression . . . Use of antidepressant drugs may propel the illness to a more malignant and treatment unresponsive course.”
• As our society embraced the use of SSRI antidepressants, disability due to affective disorders dramatically increased. One reason for this was that SSRIs were found to increase the risk that a person with unipolar depression would suffer a manic episode and be diagnosed with bipolar disorder, a diagnosis with a much less favorable prognosis. Indeed, a 1995 NIMH study that tracked the outcomes of medicated and unmedicated depressed people over a period of six years found that those who were “treated” for the illness were three times more likely than the untreated group to suffer a “cessation” of their “principal social role” and nearly seven times more likely to become “incapacitated.”

• Prior to the STAR*D study, the NIMH had funded two small studies that provided a comparison of one-year outcomes for medicated and unmedicated “real-world” patients. In a study of 118 outpatients given the best possible clinical care, only 26% responded to the antidepressant during the first year of treatment (meaning that their symptoms decreased by at least 50% on a rating scale), and only about half of that group had a “sustained response.” Only 6% of the patients saw their depression fully remit and stay away during the year-long study. Meanwhile, in a study of 84 patients of unmedicated depression, 85% of the patients had recovered by the end of one year. These two NIMH studies provide results that were polar opposites: a 6% recovery rate for those treated with antidepressants, and an 85% recovery rate for those who eschewed the drugs.

That record of science provided reason to predict that the STAR*D outcomes would be poor and add to the worry that antidepressants, over the longer term, increased the likelihood that medicated depression would run a chronic course. That was indeed what the RIAT reanalysis of the patient-level data showed: only a 35% remission rate after four levels of treatment, and only a small handful of the 4,041 who entered the trial enjoying a sustained remission at the end of one year.

However, the STAR*D investigators were invested in the narrative that told of how antidepressants were an effective treatment. This was the narrative their guild had promoted, and they also had extensive ties to the pharmaceutical companies that sold antidepressants, serving as their KOLs. The 12 STAR*D authors listed a collective total of 151 ties to pharmaceutical companies. Eight of the 12 had ties to Forest, the maker of Celexa, the study drug given in the first stage of the study.

Rather than report findings that would undercut the public narrative of progress, the STAR*D investigators saved that narrative, and even provided a stronger foundation for it, by reporting that nearly 70% had become “symptom free” at the end of the four stages of treatment. They also hid the one-year results, and provided no information about the serious adverse events that the patients may have experienced (although such information was, in fact, collected during the study). As a result, the lone “evidenced-based” soundbite to emerge from the study was the “nearly 70%” remission rate.

As Badre and Compton suggested in their letter, the RIAT reanalysis actually provided psychiatry with an opportunity to “atone” for its past sins, including the falsification of outcomes in the STAR*D study. As they wrote, psychiatry had “an ethical duty to our patients to take an honest look at the evidence derived from our research when making decisions that will impact their mental health,” and if, in response to the RIAT reanalysis, the AJP, the APA, and the NIMH had launched an investigation into this matter, and made their investigation known to the public, that would have served as evidence that the field was, as it moved forward, fulfilling that “ethical duty to take an honest look at the evidence derived from our research.”

But expecting that to happen was like whistling in the wind.

We now can see that American psychiatry has chosen not to grab that opportunity, and instead has chosen to continue on the path it has been on for more than 40 years, privileging its guild interests over its “ethical duty” to the public. This choice be seen in the letter published by the STAR*D investigators, which falsely accused Pigott and colleagues of applying “post-hoc criteria” to reduce their count of cumulative remissions, and in the fact that AJP editor-in-chief Ned Kalin did not even bother to respond to the letter from Pigott and Amsterdam requesting that the November 2006 article—and other STAR*D articles—be retracted.

That is what we can now conclude: The falsified results will remain in the literature, and there will be no atonement for the field’s sins.

The Media Fails Us Too

While it may seem difficult to understand the media’s silence, given the documentation that exists telling of the STAR*D scandal, there is a history that can help explain it. The reticence of the media to report on outcomes that threaten guild interests is not new, and its roots can be found in a strategy adopted by the pharmaceutical industry and its thought leaders more than 30 years ago.

After it published DSM-III, the APA understood that it needed to court the media to promote its disease model. It did so in multiple ways, regularly holding “media days” to tell of the field’s advances in understanding the biology of mental disorders, and handing out awards to media that published stories telling of such advances. In 1988, Eli Lilly brought to Prozac to market, and soon the media was hailing it as a breakthrough medication, telling of how it fixed a chemical imbalance in the brain. The pill even appeared on magazine covers.

However, Eli Lilly—and by extension, American psychiatry—soon faced a public relations crisis. In clinical trials, Prozac had been shown to stir suicidal and violent thoughts in a significant number of patients, and by 1990, so many people had suffered a bad reaction to Prozac that a Prozac Survivors Support Group had formed. Many harmed by the drug had taken their complaints to lawyers, and in the summer of that year, one lawsuit in particular caught the public’s attention. Five weeks after starting Prozac, Joseph Wesbecker walked into a Louisiana printing plant where he worked and opened fire with an AK-47 assault rifle, killing eight and wounding 12. Victims of the shooting and their families sued Eli Lilly, and suddenly the “breakthrough” sheen around Prozac was in jeopardy.

Then Eli Lilly got the break it needed. The Citizens Commission on Human Rights, which was affiliated with Scientology, issued a press release urging Congress to ban this “killer drug,” and Eli Lilly quickly pounced, as they now had a method to discredit this worry. “These lawsuits are being drummed up by the Scientology group, which has a history of criticizing the use of psychiatric drugs,” it said.

Eli Lilly honed a four-point message for the media. This was an issue being raised by Scientologists; extensive clinical trials had shown Prozac to be safe and effective; the suicidal and homicidal events were due to the disease and not the drug; and people who could be helped are being scared away from treatment by such criticisms, and that was the real public menace. Eli Lilly ran media-training sessions for the academic psychiatrists it paid to be their thought leaders, having them practice their delivery of this message.

This strategy immediately paid dividends for Eli Lilly. An article in The Wall Street Journal informed readers that Scientology was a “quasi-religious/business/paramilitary organization” that was “waging war on psychiatry,” and that it had attacked Prozac’s safety “even though doctors unaffiliated with Eli Lilly had found, during the clinical trials, that there was a lower tendency toward suicidal thinking with Prozac than with other antidepressants, or with the starch capsules given to a control group.”

This PR strategy had proven remarkably successful, and it became a model for psychiatry and its KOLs to adopt as they responded in the media to critics. The critics were Scientologists (or else were maligned as “flat earthers” or some other term that delegitimized their criticism), and any media outlet airing such criticisms was being irresponsible by publishing such criticisms, as it could discourage the public from getting helpful treatment. A responsible media would turn to academic psychiatrists as the experts who could be trusted to properly inform the public about the merits of psychiatric drugs.

The KOLs—and American psychiatry—could expect that this argument would protect their guild interests. Reporters who covered “health” were expected to interview the “experts” for their stories, and this meant turning to academic psychiatrists for quotes and explanations about research findings. Health reporters and their editors at mainstream media took to heart the message from the KOLs that to report on stories that questioned the efficacy of psychiatric drugs was irresponsible, as they were giving credence to critics who were biased and motivated by a hostility towards psychiatry.

In short, more than 30 years ago, the pharmaceutical companies and American psychiatry adopted a PR strategy to delegitimize critiques of the narrative of progress that they were telling the public. This was a strategy that, in essence, cowed the mainstream media, making it afraid to do its own investigations of the scientific literature.

The media’s silence about the RIAT reanalysis reflects that institutional timidity. The authors of the STAR*D study are the experts that mainstream media turn to when they report on the effectiveness of psychiatric drugs, and while the Psychiatric Times did open up the possibility that major print media would report on this scandal, it is now apparent that even that opening was not enough to overcome that institutional timidity.

Or to put it another way, American psychiatry has successfully made mainstream media its handmaiden, and in this instance, that has led to a failure of epic dimensions.

Harm Done

The harm done to the American public by the falsification of the STAR*D results is this: Our society has been denied informed consent about the merits of these drugs, and if it were not for the falsification of the STAR*D results, our societal use of these drugs would likely have been much more restrained.

On an individual level, those who decided to take an antidepressant following the publication of the STAR*D reports did so with the understanding that research had shown that two-thirds of patients so treated got completely well, their symptoms gone. While many may tell of how antidepressants improved their lives, there are also thousands who tell of how these drugs ruined their lives.

They tell of having gone manic on an SSRI and being diagnosed with bipolar disorder; of sexual dysfunction; of tardive dysphoria (an enduring form of depression); of a painful form of agitation known as akathisia; of horrible withdrawal symptoms when they try to taper from antidepressants; and of persistent nervous system injury that remains even after they have withdrawn from the medications. And these patients, who have fared poorly on antidepressants, regularly tell of how they were informed they had a chemical imbalance and given little information about the possible hazards of long-term use.

The story of the STAR*D scandal reveals that multiple institutions have failed the American public. Given the importance that the STAR*D study has had on the prescribing of antidepressants, the American Journal of Psychiatry, the American Psychiatric Association, and the NIMH all had a duty to make the RIAT reanalysis known to the public.

So too all of American medicine. The majority of prescriptions for antidepressants are written by general practitioners. BMJ Open is a general interest medical journal, and one would hope that other medical journals, in their editorials, would have commented on the Pigott re-analysis, and urged an investigation into the STAR*D study, given its profound impact on prescribing practices. The medical field as a whole has a duty to protect its evidence base from being soiled by the “falsification” of results, but to my knowledge, there has not been any clamor from other medical journals about this scandal, in the “largest and longest” study of antidepressants ever conducted.

That is the conclusion to be drawn from a review of the STAR*D scandal. American psychiatry, the NIMH, the larger medical community, and mainstream media . . . all can be seen as having betrayed the American public by failing to make this scandal known.