Do Antipsychotics Actually Work? Historian Andrew Scull's Deep Dive

In a conversation on the Lex Fridman podcast, psychiatric history researcher Andrew Scull raised a series of sharp questions about the efficacy and costs of antipsychotic medications. Rather than simply dismissing the value of these drugs, the discussion revealed a far more complex reality: these medications work for some people, fail for others, and the side effects they cause have long been deliberately downplayed by the industry.

Effective — But Only for Some

Scull was direct from the outset: antipsychotics are not a universal solution. "The drugs work for some people and don't work for others. A significant number of psychotic patients don't respond to medication, just as a significant number of depressed patients don't respond to antidepressants."

This cuts against a common assumption about psychiatric medication — that diagnosis plus prescription equals improvement. The reality is that drug response varies enormously between individuals, and clinicians still lack reliable tools to identify in advance who will be a "non-responder." That means a substantial number of patients may be enduring side effects while deriving no benefit whatsoever.

Even more troubling is how the industry responded early on. Scull noted that in the initial wave of enthusiasm for these drugs, nearly everyone overlooked their "serious side effects." Some even argued that the side effects were a necessary component of treatment — that patients simply had to "put up with them."

The Side Effects That Were Ignored for Decades

Scull catalogued the harms these drugs can cause in some detail. First is akathisia — an uncontrollable, incurable restlessness that leaves patients in constant motion, unable to keep still, a torment for both the patient and everyone around them.

Second are Parkinson's-like symptoms — what Scull describes as a "dreadful affliction." While L-DOPA can slow its progression, the effect eventually wears off, and there is still no cure.

Most serious is tardive dyskinesia (TD). "Tardive" means it doesn't appear immediately; "dyskinesia" refers to involuntary movements — particularly in the face and tongue: thrusting of the tongue, strange vocalizations, stumbling gait. Scull put it bluntly: when you see someone on the street exhibiting these symptoms, you think "that's a mentally ill person" — and they may indeed have a psychiatric condition, but what you're actually observing is an iatrogenic consequence of their medication.

This problem was ignored by the industry for roughly twenty years. It wasn't until Maryland psychiatrist George Crane published a paper in Science stating plainly that "our profession has been ignoring this terrible problem" that attitudes began to shift. By the 1980s, the American Psychiatric Association was deeply concerned — and while pharmaceutical companies shared that anxiety, they had no clear answers.

---

Akathisia is a subjective experience of suffering that is frequently misdiagnosed or overlooked. Patients feel an irresistible inner agitation and a compulsive urge to move — not simply uncontrolled limb movements. Many describe it as "torment beneath the skin," an inner anguish that is distinct from visible motor disturbances and difficult to articulate. Clinically, akathisia is often misread by physicians as a worsening of underlying psychiatric symptoms, which can lead to increased dosage — a vicious cycle. Research has linked severe akathisia to medication discontinuation, aggressive behavior, and elevated suicide risk, yet because the symptom is difficult to quantify, it has long been underappreciated.

Tardive dyskinesia (TD)'s "tardive" quality makes it especially dangerous — symptoms typically emerge months or even years after starting medication, making it hard for patients and doctors to connect them to drug use. It is estimated that approximately 20–30% of patients on long-term first-generation antipsychotics develop TD. Even after discontinuation, symptoms persist — sometimes irreversibly — in many patients. The FDA approved the first medication specifically targeting TD (valbenazine) in 2017, more than half a century after the condition was first systematically documented.

---

Clozapine and the Tortured Arrival of "Second-Generation" Drugs

A turning point came in the late 1980s. Scull recounted the unusual history of clozapine. The drug had been developed by a small company during the era of Thorazine generics, later acquired by a larger firm and introduced in Europe.

It never made it to the U.S. market — for two reasons. First, it typically did not produce tardive dyskinesia, and at the time, many in the field actually believed that the absence of these symptoms meant the drug wasn't "hitting the right target." Scull called this logic "absurd." Second, it damaged white blood cells in a significant number of patients, causing deaths, and was quickly withdrawn from markets.

As the tardive dyskinesia crisis deepened, clozapine was revived — it proved more effective for treatment-resistant cases, meaning patients who hadn't responded to other antipsychotics. The tradeoff was weekly blood tests; if white cell counts dropped, the drug had to be stopped immediately. When it returned to market in the late 1980s, other pharmaceutical companies rapidly developed drugs with different chemical structures but grouped into the same category — the "second-generation antipsychotics," including risperidone (Risperdal) and olanzapine (Zyprexa).

Scull emphasized that while the public and professionals treat these drugs as a single class, they are chemically quite diverse, with different side effect profiles, mechanisms of action, and efficacy. And there was then — as there remains now — no real scientific framework to guide a physician when a patient walks through the door: which drug should they actually prescribe?

---

Clozapine's mechanism of action differs fundamentally from first-generation antipsychotics. First-generation drugs (like chlorpromazine/Thorazine) act primarily by powerfully blocking dopamine D2 receptors — the same mechanism that produces Parkinson's-like symptoms and tardive dyskinesia, through excessive suppression of dopamine signaling in the basal ganglia. Clozapine has a relatively low affinity for D2 receptors, instead exerting broad effects on multiple receptor types (including serotonin, norepinephrine, and histamine). This "atypical" mechanism is thought to explain its lower rate of motor side effects and laid the conceptual groundwork for subsequent second-generation drug development.

However, clozapine's rate of agranulocytosis (a dangerous drop in granulocytes, a type of white blood cell) is approximately 1–2%. Left undetected, this can be fatal — and in the 1970s, multiple patient deaths directly led to its withdrawal in several countries. Strict blood monitoring protocols (in the U.S., patients must enroll in a national registry and undergo regular blood tests) enabled its controlled return to market, but the burdensome monitoring requirements remain a real-world barrier to broader use, meaning many treatment-resistant patients who could benefit from the drug don't receive it in time.

---

The CATIE Study: Expensive New Drugs Are No Better

The critical evidence in this discussion comes from a 2005 independent study. Scull noted that since the "drug revolution," an increasing share of research has been funded by pharmaceutical companies, which control the data and publish only what flatters them. To obtain FDA approval for efficacy and safety, a drug only needs two successful trials — even if fifteen were conducted and thirteen failed, two successes are sufficient.

In 2005, the National Institute of Mental Health (NIMH) funded an independent study — the CATIE trial, published in the New England Journal of Medicine. It compared one first-generation drug (now off-patent and inexpensive) against four second-generation drugs (still under patent, sometimes costing ten times as much).

The findings were striking. Were the new drugs more effective than the old? The answer was no. And what Scull believes was significantly underappreciated at the time: depending on which drug patients were assigned to, 67% to 82% dropped out of the trial — either because the drug wasn't working or because they couldn't tolerate the side effects. "These are our best drugs, and you can see — there's real complexity here."

Looking more closely, second-generation drugs did show a modest reduction in tardive dyskinesia — but brought an entirely new set of problems: most patients gained significant weight, 10, 20, or even 50 pounds, leading to metabolic syndrome, diabetes, and cardiac complications.

---

CATIE (Clinical Antipsychotic Trials of Intervention Effectiveness) was one of the largest independent clinical trials in the history of psychopharmacology, enrolling approximately 1,500 patients with schizophrenia, randomly assigned to one of five antipsychotic medications, with follow-up extending to 18 months. Crucially, the primary endpoint was not symptom scores but whether patients continued taking their medication — a design that deliberately prioritized "real-world effectiveness" over "efficacy under ideal conditions."

NIMH chose "all-cause discontinuation" as its core metric because the researchers recognized that no matter how impressive a drug's symptom-reduction data looks in a trial, if patients cannot sustain it due to side effects or lack of efficacy, its clinical value is sharply limited. The 67–82% dropout rate thus exposed a reality obscured by industry-sponsored research: in everyday clinical settings, these drugs perform far more modestly than published data suggests. When CATIE's results appeared, they sparked considerable controversy in the field — some psychiatrists questioned the study design — but its independence and sample size have kept it among the most authoritative evidence in the field.

---

No Free Lunch: An Unpredictable Gamble

Scull is careful to say this is not a wholesale rejection of medication. "There's no free lunch in medicine." Even aspirin can cause stomach bleeding severe enough to be life-threatening. Every drug that reaches the market has both a primary effect and side effects.

The genuine dilemma is that this is a cost-benefit calculation patients must make without knowing in advance where they will fall on the probability curve. "You get some relief from psychotic symptoms, but you're also inviting these other dangers that may or may not materialize. And at the outset, we don't know where you're going to land."

No biomarker currently exists to predict response in advance. Scull mentioned colleagues working in psychiatric genetics who hope to one day distinguish responders from non-responders — but he was candid that this remains "a promissory note," a hope rather than a reality, and shouldn't be treated with too much confidence.

Finally, he touched on what may be the hardest dimension to articulate: the side effects aren't only measurable physical changes like weight gain or tremors — they include effects on personality and mind. "Some patients dropped out of the trials because they found that their rich mental life had completely disappeared, and they preferred to endure some hallucinations." It's a painful trade-off — when the cost of treatment is the erasure of one's inner world, the boundary between recovery and selfhood becomes blurred and heavy.