In a recent opinion piece in STAT, researchers Stephen Soumerai and Ross Koppel argue that the black box warning that the Food and Drug Administration (FDA) placed on antidepressants in 2005 led to a decline in youth access to mental health care and increased suicide attempts.
Here’s some background on the black box warning. Starting from the time that selective sertonin reuptake inhibitors (SSRIs) such as Prozac, Zoloft, and Paxil came on the market in the late 1980s, patients have reported disturbing side effects like mania, aggression, suicidal thoughts, and akathasia (extreme psychological and physical agitation). Evidence linking antidepressants to suicidal thoughts in youth led to the FDA issuing advisories in 2003-2004, followed by a “black box” warning implemented in 2005 (the strictest warning used by the FDA) and an expansion of that warning in 2007.
Since the black box warning was issued, researchers have debated whether the warning goes too far in reducing access to treatment for depression. In this latest piece, Soumerai and Koppel claim that trends in diagnoses of depression, antidepressant use, and rates of psychotropic drug poisonings show that the FDA warning has made depression outcomes worse for youth. However, their arguments contain severe flaws and jumps in logic that cannot go without mention.
Here’s why their claims are profoundly wrong:
Soumerai and Koppel argue that the FDA warning led to a decrease in antidepressant prescriptions for children and young adults, which they show in their first graph. However, the graph only shows a decrease in prescriptions among children and adolescents from 2003-2005. What happened to prescribing after 2005? Prescribing for adolescents began to increase again starting in 2008, while antidepressant prescribing for children with psychological impairment has returned to pre-2004 levels. In fact, antidepressant prescribing in general is very high, with about one in ten Americans age 12 and over taking an antidepressant.
To show an association between two events, the timing of those events must be clearly defined. However, as FDA Deputy Director for Safety Dr. Marc Stone points out in a thorough and scathing criticism of their study, there were many different actions the FDA took to increase recognition of the suicidal risk of antidepressants for children. “After which action would researchers recognize that a ‘policy’ has been implemented and begin surveillance?” asks Stone.
And how long after the “policy” happens can a trend be attributed without this being confounded by other factors? “The rationale by which the authors justify their period of observation places no restriction on when an effect of the policy could be expected to occur,” Stone argues.
In their op-ed, Soumerai and Koppel attribute several trends to the FDA warning, event though these trends started at different times. They point to a decline in diagnosing depression and antidepressant prescriptions in youth starting in 2004, after the FDA advisory but before the black box warning was put into place. Then they cite an increase in suicide attempts starting in 2005, after the black box warning was implemented.
They also attempt to show that the warning decreased talk therapy treatment as well as antidepressant use, by showing a decline in psychotherapy in 2003. However, this trend was already in decline, as shown by their own data from previous years (see image below).
Most egregiously, they point to an increase in youth suicides starting in 2003, but draw their trend line before the upward trend begins (see image below)! Journalist and activist Robert Whitaker highlights this misdirection on the Mad in America blog: “This is a kind of visual misdirection that magicians use: The reader’s eye sees the trendline and thus misses the drop in suicides—as marked by the four yearly dots—from 2004 to 2007.”
The authors’ conclusion is that the warning is preventing children and young adults from receiving needed treatment, and therefore, “It’s appropriate to think about removing this warning.” But even if the trends matched up chronologically, this would not show that the FDA warning was the cause. As Stone writes, “FDA actions were preceded and accompanied by public controversy involving news reports, journal articles, lawsuits, etc. How would [the authors] distinguish the effect of an FDA advisory from ongoing public controversy?”
The authors write that they are concerned not only about fewer antidepressant prescriptions, but also less access to talk therapy, which as they show, steadily declined throughout the 2000s. So why is their primary policy recommendation to remove the warning? Why not make different kinds of therapy more available?
Antidepressants have helped many people struggling from depression and other mental health disorders. However, the risk of suicidal ideation, among other side effects, cannot and should not be denied – especially after drug makers covered up these risks for so many years. The FDA warning should be just that – a warning, so that patients and there families can consider the risks along with the benefits, and know what side effects they should be looking for, in the unfortunate possibility that side effects may occur.
It is difficult enough to provide patients and the public information about the risks of drugs and medical devices, against the flood of information provided by drug companies. We don’t need misinformation from researchers that further skews the balance toward more pharmacological interventions.