Aaron E. Carroll, a pediatrician, is the president and chief executive of AcademyHealth.
Health Secretary Robert F. Kennedy Jr. recently announced plans to begin significantly reducing the most-used class of antidepressants in America. Under his proposal, the medications would be subject to federal deprescribing guidelines, and Medicare and Medicaid would adopt new reimbursement rules to encourage tapering them off.
The problem here isn’t that Kennedy wants to reexamine the widespread use of these drugs; it’s how he’s going about it. The plan appears to start with a conclusion, and the government is now assembling evidence to support it. That’s no way to change policies on drugs that millions of Americans rely on.
In 2018, I argued that the benefits of antidepressants, known as selective serotonin reuptake inhibitors (SSRIs), were overstated. Research showed that the effects were mostly modest, and too little was known about their use over the long term or for mild depression. Not much has been published in the interim to make me rethink those concerns. The evidence, both for and against the drugs, remains a mess.
A decade earlier, the New England Journal of Medicine published a study examining all the FDA-registered trials of antidepressants from 1987 through 2004, covering 12 different drugs. That included 74 different trials, of which 38 were seemed to have a clinical effect. Of those, 37 were published in the medical literature.
Of the 36 that didn’t show such an effect, 22 were never published. An additional 11 were written in a way that made them sound positive. Only three were published in a way that made it clear the trial was a failure.
When I give talks on publication bias, this is the example I point to. Any reading of the peer-review literature would conclude that SSRIs are overwhelmingly useful. In reality, however, only half of all the trials conducted were successful.
Eight years ago, recognizing this limitation, researchers published a new study with more sophisticated methods that included 522 trials of 21 different antidepressants. The good news was that every single one of them was found to “work” in the sense that they outperformed a placebo. But even then, the authors rated the certainty of the evidence as “very low” to “moderate.”
Further, the trials mostly focused on short-term treatment of moderate to severe depression. John Ioannidis, a co-author and famous skeptic of medical research, warned that the headlines announcing the drugs worked went well beyond what the data could support.
That doesn’t mean, however, that any policy to reduce their use is good. That’s because doing so could result in depression going untreated.
Here, too, history can teach us. In the early 2000s, trials found that 4 percent of children using antidepressant treatments showed suicidal thoughts or behaviors, compared with about 2 percent of children receiving a placebo. Even though no child died by suicide in any trial, this prompted the FDA to put a black box warning on antidepressants for children and adolescents, later extended to young adults up to 25 years old.
What happened next is hotly contested. In 2024, a review published in Health Affairs linked these warnings to a reduction in depression diagnosis, treatment and even psychotherapy. In some analyses, they were linked to a rise in suicide deaths. These studies had methodological weaknesses (as I said, the evidence is a mess). But the bottom line is that federal action on these drugs 20 years ago may have harmed more children than it helped. We still don’t know for sure.
Kennedy’s concerns are defensible. Prescriptions for antidepressants have climbed for decades, and some patients could probably taper off them safely. But he is seeking to change federal policy as if he has all the answers, not a bunch of questions.
If his department announced plans to fund research on those questions, I would applaud it. Agencies already exist that can make that happen: The Patient-Centered Outcomes Research Institute, for instance, funds research on questions that matter to patients, and the Agency for Healthcare Research and Quality evaluates existing research and helps doctors use it.
These institutions are perfectly positioned to support studies nobody has run, such as long-term randomized discontinuation trials that will show who can stop using SSRIs safely and who might be at risk of relapsing. HHS can also focus research on mild depression, where prescribing is common but the evidence is thin. Finally, it could fund researchers to synthesize the evidence and study how to safely deprescribe patients.
How well these drugs work is an empirical question; whether relying on one is shameful is not. Depression sits in an unfortunate bucket of conditions that society quietly moralizes. People with the condition are suspected of just not “trying hard enough.” Words like “dependency” and “overmedicalization” imply that leaning on a drug is a failure of will.
I started an SSRI years ago after a bad panic attack, and it has helped me. That’s an anecdote without a counterfactual. It proves nothing about the drug. But neither do the value propositions hiding inside Kennedy’s policy proposals.
The FDA still deems SSRIs safe and effective. The honest answer is that we don’t fully know how much good they do, for whom and for how long. What I’d prescribe in this case is doing the research before a policy change rather than after.
Evidence is supposed to be able to change your mind. If it can’t, it’s just decoration.
The post The problem with RFK Jr.’s crusade against antidepressants appeared first on Washington Post.
