Jumping to Conclusions
Why we should be careful not to discount traditional suicide prevention methods.
May is National Mental Health Month, so it was unsurprising to see a flurry of articles about psychiatry, psychology, depression, and suicide begin to appear in late April. Usually, these articles remind us that mental illness is serious but treatable and call on our leaders to ensure better access to mental healthcare.
But one article by psychiatrist Amy Barnhorst was quite different and garnered significant online attention (almost 600 comments). The article, appearing in the New York Times on April 6, was titled “The Empty Promise of Suicide Prevention: Many of the problems that lead people to kill themselves cannot be fixed with a little extra serotonin.” The major theme of the article is indisputable: social factors like poverty, homelessness, isolation, and lack of education play a significant role in all illnesses and require at least as much attention as a patient’s physiological problems. Indeed, social determinants of health (SDOH) contribute about three quarters of the risk for all human illness.
Yet Barnhorst’s article also contains serious shortcomings that strike at the heart of perhaps the most common obstacle to understanding the validity of a health claim: the tendency to be swayed by a single dramatic story. In the process of making her point, Barnhorst also manages to seriously trivialize the role of antidepressant medication in the treatment of depression and to imply that given societal woes, there isn’t much we can do to try to prevent suicides—aside from limiting access to lethal means.
Two Different Stories
Barnhorst begins her piece by describing a patient of hers who presented with severe depression, a clear suicide plan, and a likely diagnosis of bipolar depression. The patient “responded well to lithium, one of only two psychiatric medications shown to reduce suicide.”
But Barnhorst goes on to tell us that this clear-cut case is an exception in her practice. She then offers a second story, of a “middle-aged woman with no psychiatric history” who was homeless and overdosed on the nonsteroidal anti-inflammatory drug (NSAID) ibuprofen (Motrin). Further details of the story are that this woman had “seven years of sobriety” but relapsed by taking methamphetamine “to stay awake at night after she was sexually assaulted in the park where she had been sleeping.” This unfortunate woman obviously had multiple adverse social determinants of health, including extreme poverty, homelessness, lack of social support, low educational level, and lack of health insurance.
When a resident tells Barnhorst that his plan is to admit the patient to the hospital and start her on an antidepressant, she writes, “I could tell he knew how ridiculous it sounded.” Antidepressants, she says, have not been proven to reduce suicide and only work in “40 to 60 percent of patients who take them.” Suicide, she correctly notes, is extremely hard to predict—and therefore to prevent—because substantial numbers of people who make suicide attempts do so impulsively, giving no warning. “This is because many of the problems that lead to suicide can’t be fixed with a little extra serotonin. Antidepressants can’t supply employment or affordable housing, repair relationships with family members, or bring on sobriety.”
Why criticize a plea for more attention to what Barnhorst believes to be “the root causes of our nation’s suicide problem—poverty, homelessness and the accompanying exposure to trauma, crime, and drugs?” Of course, we all agree that we have done a woeful job in this country dealing with these devastating social problems.
The problem, however, is that Barnhorst bases her claims on a story, the story of one person. And that story raises many questions. It isn’t necessarily true that the woman had no psychiatric history; from the article, she clearly had a history of substance abuse. Did she also have a history of other psychiatric illnesses, like depression or post-traumatic stress disorder? It would have been nearly impossible for clinicians to answer that question in the few hours and even days after a woman who has been sleeping in the park overdoses on over-the-counter medication. Understanding how she got to be homeless in the first place and what her actual medical and psychiatric history is would take an extended period of careful inquiry.
It is also by no means certain whether this woman would benefit from antidepressant medication. That is not to neglect, of course, that she desperately needs help with housing, employment, and income, or that counseling and psychotherapy are likely to be of some help to her. But it is also possible that medication might help her get better. After all, her most immediate problem is that she no longer wanted to live.
The Case for Antidepressant Medication
Antidepressant medications do more than offer a “little more serotonin.” The antidepressants we call selective serotonin reuptake inhibitors (SSRIs), with brand names like Prozac, Paxil, and Lexapro, do in fact rapidly increase the concentration of the brain neurotransmitter serotonin. But this effect is now understood to be insufficient to explain how the drugs actually work. The biology of antidepressants is far more complex than “a little more serotonin.”
The issue of whether antidepressants reduce the suicide rate is also far more complicated than Barnhorst implies. Almost all gold-standard, randomized clinical trials (RCTs) that have evaluated the effectiveness of antidepressants deliberately excluded people at risk for suicide. Drug companies who make these drugs have traditionally been unwilling to take the risk that a suicide might occur during a clinical trial of an investigational antidepressant and cause delays in getting FDA approval. So in fact we have very little data on whether suicides are prevented by taking antidepressant medication.
Epidemiological data also show that antidepressants do seem to have a protective effect against suicide. We also know that approximately 90 percent of people who die by suicide had a diagnosable mental illness at the time of their deaths. Taken together, it is reasonable to recommend that most people with depression severe enough to include thoughts of ending one’s life should consider an antidepressant medication as part of their treatment plan. An effective way to prevent suicide is to diagnose and treat psychiatric illnesses like depression before they get to that point.
There is no question that suicide rates are higher among homeless people than the general population, but so are rates of depression and other mental illnesses. We also know that most homeless people do not kill themselves. It is imperative, therefore, that when a homeless person presents with suicidal ideation and other risk factors for self-harm, that clinicians consider multiple possible causes and consider the range of interventions available.
A careful reader of Barnhorst’s article in the New York Times will easily agree that her point regarding social factors is important—but after reading her two patient stories, the reader must ask the perhaps pedestrian but nevertheless essential question: What do the data show across larger samples of people? Stories, no matter how gripping, cannot substitute for carefully collected evidence. A good story can illustrate what the data show us, but we must be prepared to push aside our emotions when we read one and inquire whether the story really represents more than the person in it.