I recently helped a woman in her early 60s reduce the dose of fluoxetine (commonly known by its brand name Prozac) that she had been taking for over 35 years, and subsequently the dose of bupropion (also known as Wellbutrin) that she had been taking for over 10 years.
But Kennedy’s effort confuses true clinical need with claims that are unsupported by evidence, some of which are clearly dangerous.
When she came to see me in 2024, she had been trying to taper off fluoxetine twice before, many years apart. But each time her depression returned, she had to go back to a higher dose. We gradually tapered her fluoxetine down by 5 milligrams every 3 months, continuing for an additional month if she felt unstable, with a mood diary, regular treatment visits, and close monitoring. The fluoxetine taper took 18 months overall. I then waited another 3 months and started tapering off the bupropion, reducing it by half over several months.
She tolerated fluoxetine taper well. But when he cut his bupropion dose in half while under intense occupational stress, his anxiety spiked and his previously well-controlled blood sugar levels spiked. (Bupropion has moderate hypoglycemic effects, and halving the dose may have removed that protection.) We temporarily returned to the previous dose and resumed tapering once the stressor resolved.
This process has been going on for a long time not because discontinuing these drugs is dangerous, but because deprescribing is not a one-time decision but a dynamic, individual process that requires ongoing clinical judgment.
There are legitimate clinical questions at the heart of Secretary of Health Robert F. Kennedy Jr.’s efforts to help Americans stop taking antidepressants. Deprescribing is poorly researched, poorly educated, and poorly covered. For 40 years, drug companies have funded clinical trials to show that their drugs work. Few are funding trials that show when and how to stop.
Stopping antidepressants may not be as dangerous as widely believed, review suggests
But Kennedy’s effort confuses true clinical need with claims that are unsupported by evidence, some of which are clearly dangerous.
The field itself has been seriously addressing this issue. In 2025, the American Society of Clinical Psychopharmacology (ASCP), a leading professional organization for psychiatrists and researchers, held a summit on antidepressant deprescription, which brought together 45 international experts and reached agreement on 44 of 50 consensus statements.
Their recommendations contain real clinical value. In other words, consider discontinuing the prescription if the drugs have redundant functions, minimal improvement, the drug stops working, or the side effects outweigh the benefits, and make only one change at a time.
Importantly, the task force also recognized that stopping prescriptions may not be right for everyone. For patients who have had three or more episodes of depression in their lifetime, the consensus supports continuing the medication indefinitely.
These are sound, evidence-based principles, but their safe implementation requires individual clinical judgment, gradual tapering, psychological support, and available alternatives. These are elements that broad federal efforts have not yet delivered at scale.
Weaning a patient from drug therapy is only clinically responsible if an alternative is available. it’s not.
The 2025 JAMA Psychiatry Survey found that recent increases in the use of psychotherapy have been concentrated almost entirely among urban adults who are young, affluent, college-educated, and have private insurance. No significant increases were seen among adults who were older, less educated, uninsured, or lived in rural areas. This is one reason why approximately 1 in 6 American adults now take SSRIs. You can access your prescriptions in a different way than weekly treatment appointments. The gap widens further when we move clinicians away from prescribing without investing in treatment infrastructure.
The largest meta-analysis to date on discontinuation symptoms found that patients who discontinued antidepressants experienced an average of one additional symptom compared with placebo, below the threshold for a clinically significant discontinuation syndrome. A related Lancet Psychiatry analysis found that about 15% experienced symptoms stemming from the suspension, and about 3% had severe symptoms.
These are real numbers. But they do not support Kennedy’s claim that SSRIs are harder to quit than heroin. Keith Humphreys, an addiction researcher at Stanford University, says antidepressants and heroin are “worlds apart” when it comes to addiction risk. As stated in the DSM-5, antidepressant discontinuation syndrome does not include drug craving and is unrelated to the reinforcing effects of substance dependence. And Kennedy’s claim that SSRIs are involved in school shootings lacks any causal support in the research literature. This statement by federal authorities would stigmatize treatment for depression and discourage people from seeking treatment.
The ASCP consensus also notes that SSRIs with long half-lives, such as fluoxetine, rarely cause significant discontinuation symptoms, and pharmacokinetic data support this. However, the FDA label still recommends gradual tapering rather than abrupt discontinuation. And a 2026 Lancet Psychiatry Network meta-analysis of 76 trials found that abruptly stopping antidepressants was associated with a significantly higher risk of relapse than slowly tapering off with psychological support. Fluoxetine’s long half-life reduces discontinuation symptoms. This does not eliminate the risk of depression recurring. These are two different clinical endpoints, and confusing them makes the framework dangerous.
Stopping antidepressants almost doubles the rate of relapse of depressive symptoms, but a 2021 Cochrane review found that many trials failed to distinguish between true relapses and withdrawal symptoms incorrectly classified as relapses. This warning strengthens the case for structured deprescribing. Slow tapering with psychological support was equivalent to continuation in preventing relapse, but slow tapering alone was not significantly more effective than abrupt discontinuation. Taper is necessary. Psychological support is not optional.
What I learned after quitting SSRIs after almost 10 years
For patients with three or more lifetime depressive episodes, ASCP itself supports indefinite maintenance therapy. Federal deprescribing efforts applied to this population without individualization represent a real clinical risk.
In contrast, the Kennedy event featured speakers advocating phasing out mental health screenings in schools and tobacco-style warnings on antidepressant packages, arguments that are ideological and not evidence-based.
This does not mean that Kennedy’s initiative does not contain anything of value. A CMS reimbursement mechanism to lift time restrictions is the most evidence-aligned proposal, as inadequate reimbursement and distorted quality standards incentives have been documented as structural barriers to lifting adoption restrictions. However, reimbursement only works if clinicians have evidence-based protocols to follow and patients have access to psychological support.
Deeper requirements belong to the FDA. Although existing guidelines provide general tapering strategies, what is absent are drug-specific tapering protocols derived from randomized discontinuation trials for each approved drug in this class. The FDA has never required manufacturers to do them.
This conversation also rarely brings up workforce issues. The majority of antidepressant prescriptions in the United States are written not by psychiatrists but by primary care physicians, clinicians who see large numbers of patients, come in for short periods of time, and have limited psychiatric training. Careful tapering over months requires a skill set and time commitment, which most primary care settings are not set up to provide.
Deprescribing on a large scale would require either a dramatic expansion of psychiatric capacity or an entirely new clinical infrastructure, neither of which this initiative considers.
And the FDA issue deserves even more scrutiny than usual. Some claim that the FDA never required manufacturers to study ways to stop the medication. But the question is whether it is truly unprecedented, or whether regulatory mechanisms exist and have simply never been invoked for this drug class. If precedent exists elsewhere in the FDA’s history or in other countries’ regulatory frameworks, the barrier is institutional rather than legal. Clinical pharmacologists and regulatory historians may have more to say on this point than is still needed in the field.
Even though these drugs have been prescribed for 40 years and one in six adults take them, we still do not know how to optimally discontinue them for all patients. The ASCP Special Committee is in a position to formally make that request. Psychiatrists should urge them to do so.
Deprescribing requires gradual tapering, psychological support, careful risk stratification, and available alternatives, none of which Kennedy’s efforts provide at scale. Patients taking antidepressants deserve both an honest conversation about whether they still need the medication and a system in place to safely discontinue the medication if it is not needed. Right now, we don’t have the data or the infrastructure to provide that.
Dr. Jonathan Slater is a clinical professor of psychiatry at Columbia University Irving Medical Center.
