The number is worth sitting with before the policy language begins. Sixty-one million Americans currently take a psychiatric medication. One in eight adults. One in three children in foster care. The prescribing rate has tripled since 1990. The suicide rate has not declined.

On May 4, the federal Department of Health and Human Services issued a formal directive — a signed letter from the heads of SAMHSA, CMS, HRSA, and the Administration for Children and Families — telling physicians across the country to reconsider that prescribing, to provide patients with information about withdrawal before starting these drugs, and to seek reimbursement for non-drug treatments that the evidence has supported for years. The letter cites exercise, dietary improvement, psychotherapy, and family support as evidence-based alternatives. It establishes billing codes for medical nutrition therapy. It names foster children explicitly as a population whose psychiatric drug exposure has been inadequately scrutinized.

None of this should be news. The fact that it is news tells the real story.

The real story is not about any particular medication. It is about what happens to a medical culture when a framework becomes too institutionally entrenched to be questioned from inside the institution — and what it costs the people inside that framework, most of whom are patients, while the institution takes thirty years to arrive at a correction it could have made sooner.

The serotonin theory of depression — the idea that depression results from a chemical imbalance correctable by drugs that increase serotonin availability — was never as well-established as patients were told. A 2022 umbrella review in Molecular Psychiatry, examining the accumulated evidence across seventeen studies, found no consistent support for the hypothesis. That finding was contested, as major findings should be. What was not meaningfully contested was the separate, older observation that the clinical trial evidence supporting antidepressant efficacy in mild to moderate depression — when unpublished negative trials submitted to the FDA were included alongside published positive ones — showed effect sizes that were modest and in many cases below the threshold considered clinically significant. That analysis was published in 2008 by Irving Kirsch and colleagues in PLOS Medicine. It was conducted using data obtained through Freedom of Information requests because the negative trials had not been published. The positive ones had.

The prescribing rate continued to climb after 2008. The framework did not change. The patients were not told.

There is a human cost to that silence that aggregate statistics do not capture well, but that anyone who has tried to stop an antidepressant understands immediately. Discontinuation syndrome — the withdrawal response that occurs when serotonergic and noradrenergic medications are stopped — produces symptoms that are nearly identical to the conditions those medications were prescribed to treat: anxiety, insomnia, irritability, low mood, sensory disturbances, cognitive fog. A patient who experiences these symptoms on stopping a medication looks, to the prescriber, like a patient whose underlying condition has returned. The clinical default — in the absence of explicit training in withdrawal recognition and structured tapering — is to restart the medication or increase the dose.

That loop has played out millions of times. It is not primarily a story of bad intentions. It is a story of a framework that was not designed to see what it was doing to people, because the framework was not designed with the possibility of its own error as a starting premise. A clinical culture built around the hypothesis that the problem is a chemical imbalance and the solution is a drug does not easily generate the clinical observation that the drug itself is now a problem. The observational apparatus was calibrated away from that finding.

The HHS letter released this week now requires, as a documented element of informed consent before any psychiatric medication is initiated, that patients be told about potential discontinuation symptoms and the risks of abrupt cessation. That requirement is not, in itself, a policy revolution. It is the correction of an omission that should not have persisted for thirty years.

The deeper question — the one that the policy document approaches but does not fully name — is what it means that the framework held this long.

Medicine does not fail to evolve because the people inside it are indifferent to patients. Most of the physicians who prescribed these medications over the past three decades were doing what they were trained to do, using tools the system made available and reimbursed, within a framework that organized psychiatry defended with increasing institutional investment as the questions about that framework mounted. The problem was structural, not individual. The billing system made the 15-minute medication management visit the economically rational choice. The research funding system made pharmaceutical-sponsored trials the dominant source of evidence. The training system passed the framework forward without passing forward the contested evidence alongside it.

This is what happens when the institution that generates the evidence is also the institution that benefits from one interpretation of that evidence. The correction, when it comes, comes slowly and from outside — from patients who could not stop their medications, from researchers who obtained unpublished trial data through FOIA requests, from advocates who testified before Congress, from a federal agency that finally wrote it down.

The alternatives that the HHS letter now endorses — exercise, dietary improvement, sleep intervention, psychotherapy, family support — have not been waiting for discovery. They have been in the literature. The evidence that regular physical activity produces clinically meaningful reductions in depressive symptoms was available years before this directive. The evidence connecting dietary pattern to mood and cognitive function has accumulated across multiple well-designed trials. The evidence that sleep disruption is not merely a symptom of psychiatric conditions but a primary physiological driver of them has been growing for a decade.

These interventions did not fail to reach standard practice because the evidence was insufficient. They failed to reach standard practice because there was no financial architecture to deliver them. You cannot build a clinical system around exercise and nutrition and family therapy using the reimbursement structures that American medicine inherited from a pharmaceutical-centered model of disease. The billing codes for medical nutrition therapy existed. Practitioners were not being told they could use them. They are being told now.

Whether this directive produces real change in what happens in the examination room will depend on forces that a Dear Colleague Letter cannot fully control — on training programs, on how the Technical Expert Panel that HHS convenes in July is composed, on whether the institutional machinery that built the current prescribing culture finds ways to absorb and neutralize this correction before it reaches the patient.

The patients who have been waiting for this correction have been waiting a long time. The letter acknowledges the debt without quite saying so. That is how institutional admissions work. The acknowledgment matters. What comes next matters more.

For rigorous analysis of the primary literature underlying this policy shift, including the evidentiary record on exercise, dietary intervention, and the contested SSRI trial data, visit Popular Rationalism at popularrationalism.substack.com. For practical, evidence-based approaches to reducing toxic burden and supporting neurological and metabolic health, the Detox America program is available at detoxamerica.net.