Another advertisement, this one in 2005, for the drug paroxetine, said, “With continued treatment, Paxil can help restore the balance of serotonin,” a neurotransmitter.

“[T]he serotonin hypothesis is typically presented as a collective scientific belief,” write Lacasse and Leo, though, as they note: “There is not a single peer-reviewed article that can be accurately cited to directly support claims of serotonin deficiency in any mental disorder, while there are many articles that present counterevidence.”

Despite the lack of evidence, the theory has saturated society. In their 2007 paper, Lacasse and Leo point to dozens of articles in mainstream publications that refer to chemical imbalances as the unquestioned cause of depression. One New York Times article on Joseph Schildkraut, the psychiatrist who first put forward the theory in 1965, states that his hypothesis “proved to be right.” When Lacasse and Leo asked the reporter for evidence to support this unfounded claim, they did not get a response. A decade on, there are still dozens of articles published every month in which depression is unquestionably described as the result of a chemical imbalance, and many people explain their own symptoms by referring to the myth.

Meanwhile, 30 years after Prozac was released, rates of depression are higher than ever.

* * *

Hyman responds succinctly when I ask him to discuss the causes of depression: “No one has a clue,” he says.

There’s not “an iota of direct evidence” for the theory that a chemical imbalance causes depression, Hyman adds. Early papers that put forward the chemical imbalance theory did so only tentatively, but, “the world quickly forgot their cautions,” he says.

Depression, according to current studies, has an estimated heritability of around 37%, so genetics and biology certainly play a significant role. Brain activity corresponds with experiences of depression, just as it corresponds with all mental experiences. This, says Horwitz, “has been known for thousands of years.” Beyond that, knowledge is precarious. “Neuroscientists don’t have a good way of separating when brains are functioning normally or abnormally,” says Horwitz.

If depression was a simple matter of adjusting serotonin levels, SSRIs should work immediately, rather than taking weeks to have an effect.  Reducing serotonin levels in the brain should create a state of depression, when research has found that this isn’t the case. One drug, tianeptine (a non-SSRI sold under the brand names Stablon and Coaxil across Europe, South America, and Asia, though not the UK or US), has the opposite effect of most antidepressants and decreases levels of serotonin.

This doesn’t mean that antidepressants that affect levels of serotonin definitively don’t work—it simply means that we don’t know if they’re affecting the root cause of depression. A drug’s effect on serotonin could be a relatively inconsequential side effect, rather than the crucial treatment.

History is filled with treatments that work but fundamentally misunderstand the causes of the illness. In the 19th century, for example, miasma theory held that infectious diseases such as cholera were caused by noxious smells contributing “bad air.” To get rid of these smells, cleaning up waste became a priority—which was ultimately beneficial, but because waste feeds the microorganisms that actually transmit infectious disease, rather than because of the smells.

* * *

It’s possible our current medical categorization and inaccurate cultural perception of “depression” is actually causing more and more people to suffer from depression. There are plenty of historical examples of mental health symptoms that shift alongside cultural expectations: Hysteria has declined as women’s agency has increased, for example, while symptoms of anorexia in Hong Kong changed as the region became more aware of western notions of the illness.

At its core, severe depression has likely retained the same symptoms over the centuries. “When it’s severe, whether you read the ancient Greeks, Shakespeare, [Robert] Burton on [The Anatomy of] Melancholy, it looks just like today,” says Hyman. “The condition is the same; it’s part of being human.” John Stuart Mill’s 19th century description of his mental breakdown is eminently familiar to a contemporary reader.

But less severe cases, in the past, may have been chalked up to simply being “justifiably sad,” even by those experiencing them, whereas they’d be considered a health condition today. And so, psychiatry “reframes ordinary distress as mental illness,” says Horwitz. This framework doesn’t simply label sadness depression, but could lead people to experience depressive symptoms where they would have previously been simply unhappy. The impact of this shift is impossible to track: Mental illness is now recognized as a legitimate health issue, and so many more people are comfortable admitting to their symptoms than ever before. How many more people are truly experiencing depression for the first time, versus those who are acknowledging their symptoms once kept secret? “The prevalence is difficult to determine,” acknowledges Hyman.

* * *

Perhaps unraveling the true causes of depression and exactly how antidepressants treat the symptoms would be a less pressing concern if we knew, with confidence, that antidepressants worked well for the majority of patients. Unfortunately, we don’t.

The work of Irving Kirsch, associate director of the Program in Placebo Studies at Harvard Medical School, including several meta-analyses of the trials of all approved antidepressants, makes a compelling case that there’s very little difference between antidepressants and placebos. “They’re slightly more effective than placebo. The difference is so small, it’s not of any clinical importance,” he says. Kirsch advocates non-drug-based treatments for depression. Studies show that while drugs and therapy are similarly effective in the short-term, in the long-term those who don’t take medication seem to do better and have a lower risk of relapse.

Others like Peter Kramer, a professor at Brown University’s medical school, are strongly in favor of leaning on the drugs. Kramer is skeptical about the quality of many studies on alternative therapies for depression; people with debilitating depression are unlikely to sign up for anything that require them to do frequent exercise or therapy, for example, and so are often excluded from studies that eventually purport to show exercise is as effective a treatment as drugs. And, as he writes in an email, antidepressants “are as effective as most treatments doctors rely on, in the middle range overall, about as likely to work as Excedrin” for a headache.

Others are more circumspect. Hyman acknowledges that, when taken in aggregate, all the trials for approved antidepressants show little difference between the drugs and placebo. But that, he says, obscures individual differences in responses to antidepressants. “Some people really respond, some don’t respond at all, and everything in between,” Hyman adds.

There are currently no known biomarkers to definitely show who will respond to what antidepressants. Severely depressed patients who don’t have the energy or interest to go to therapy should certainly be prescribed drugs. For those who are healthy enough to make it to therapy—well, opinions differ. Some psychiatrists believe in a combination of drugs and therapy; some believe antidepressants can be effective for all levels of depression and no therapy is needed; and others believe therapy alone is the best treatment option for all but the most severely depressed. Unfortunately, says Hyman, there’s little evidence on the best treatment plan for each patient.

Clearly, many people respond well to antidepressants. The drugs became so popular in large part because many patients benefited from the treatment and experienced significantly reduced depressive symptoms. Such patients needn’t question why their symptoms have improved or whether they should seek alternative forms of treatment.

On the other hand, the drugs simply do not work for others. Further, there’s evidence to suggest framing depression as a biological disease reduces agency, and makes people feel less capable of overcoming their symptoms. It effectively divorces depression from a sense of self. “It’s not me as a person experiencing depression. It’s my neurochemicals or my brain experiencing depression. It’s a way of othering the experience,” says Horwitz.

It’s nearly impossible to get good data to explain why depression treatments work for some and not others. Psychiatrists largely evaluate the effects of drugs by subjective self-reports; clinical trials usually include only patients that meet a rarefied set of criteria; and it’s hard to know whether those who respond well to therapy benefitted from another, unmeasured factor, such as mood resilience. And when it comes to the subjective experience of mental health, there’s no meaningful difference between what feels like effective treatment and what is effective treatment.

There are also no clear data on whether, when antidepressants work, they actually cause symptoms to fully dissipate long-term. Do antidepressants cure depression, or simply make it more bearable? We don’t know.

* * *

Depression is now a global health epidemic, affecting one in four people worldwide. Treating it as an individual medical disorder, primarily with drugs, and failing to consider the environmental factors that underlie the epidemic—such as isolation and poverty, bereavement, job loss, long-term unemployment, and sexual abuse—is comparable to asking citizens to live in a smog-ridden city and using medication to treat the diseases that result instead of regulating pollution.

Investing in substantive societal changes could help prevent the onset of widespread mental illness; we could attempt to prevent the depressive health epidemic, rather than treating it once it’s already prevalent. The conditions that engender a higher quality of life—safe and affordable housing, counsellors in schools, meaningful employment, strong local communities to combat loneliness—are not necessarily easy or cheap to create. But all would lead to a population that has fewer mental health issues, and would be, ultimately, far more productive for society.

Similarly, though therapy may be a more expensive treatment plan than drugs, evidence suggests that cognitive behavioral therapy (CBT) is at least as effective as antidepressants, and so deserves considerable investment. Much as physical therapy can strengthen the body’s muscles, some patients effectively use CBT to build coping mechanisms and healthy thought habits that prevent further depressive episodes.

In the current context, where psychiatry’s system of diagnosing mental health mimics other medical fields, the role of medicine in treating mental illness is often presented as evidence to skeptics that depression is indeed a real disease. Some might worry that a mental health condition treated partly with therapy, exercise, and societal changes could be seen as less serious or less legitimate. Though this line of thinking reflects a well-meaning attempt to reduce stigma around mental health, it panders to faulty logic. After all, many bodily illnesses are massively affected by lifestyle. “It doesn’t make heart attacks less real that we want to do exercise and see a dietician,” says Hyman.

No illness needs to be entirely dependent on biological malfunctions for it to be considered “real.” Depression is real. The theory that it’s caused by chemical imbalances is false. Three decades since the antidepressants that helped spread this theory arrived on the market, we need to remodel both our understanding and treatment of depression.

📬 Sign up for the Daily Brief

Our free, fast, and fun briefing on the global economy, delivered every weekday morning.