Part IV

The Chemical Gamble — Medication and Its Limits

Chapter 12: The Medication Landscape

Psychiatric medication is the other pillar of American mental health treatment. The numbers:

America is the most psychopharmacologically treated nation in human history. The question is whether this treatment is working.

Chapter 13: What Medication Can and Cannot Do

Where medication works (relatively well):

Where medication is more contested:

Chapter 14: The Skeptic's View of Medication

We don't understand the mechanism. The "chemical imbalance" theory of depression — that depression is caused by insufficient serotonin — was never established by scientific evidence. It was a marketing narrative that pharmaceutical companies promoted and that the psychiatric profession allowed to propagate. A comprehensive umbrella review by Moncrieff et al. (2022) in Molecular Psychiatry concluded that there is "no convincing evidence that depression is associated with, or caused by, lower serotonin concentrations or activity." This does not mean that SSRIs don't help some patients — but it means that we don't know why they help, which makes the confident prescription of these medications an exercise in empirical guesswork, not targeted treatment.

Side effects are pervasive and under-discussed. Sexual dysfunction affects 40-65% of SSRI users. Weight gain is common with many psychiatric medications. Metabolic syndrome, cognitive dulling, emotional blunting ("I don't feel depressed anymore, but I don't feel anything"), sleep disruption, and GI disturbance are reported at high rates. For many patients, the side effects of medication are themselves a significant source of suffering.

Prescribing has outrun the evidence. Polypharmacy — the simultaneous prescription of multiple psychiatric medications — is common and largely unvalidated. A patient on an SSRI, a benzodiazepine, a mood stabilizer, and a sleep medication is not receiving evidence-based treatment. They are receiving a pharmacological collage assembled through trial and error, with no RCT evidence for the specific combination they are taking. Yet this is routine in American psychiatry.

The 15-minute med check. Psychiatrists in most practice settings see medication patients for 15-20 minutes per visit, often monthly or quarterly. The "evaluation" consists largely of asking the patient whether they feel better and whether they are experiencing side effects. There is no biomarker for depression, no blood test for anxiety, no imaging study for PTSD. The physician is making medication decisions based entirely on patient self-report — the same self-report that, as we established in Chapter 7, is systematically unreliable due to shame, fear, and social desirability.

Chapter 15: The Opioid Lesson

No discussion of psychiatric medication in America is complete without the opioid crisis — not because psychiatric medications are opioids, but because the opioid epidemic demonstrates what happens when a medical system:

  1. Trusts pharmaceutical company claims about safety and efficacy
  2. Prescribes aggressively based on incomplete evidence
  3. Fails to monitor long-term outcomes
  4. Creates dependency at scale
  5. Lacks the infrastructure to manage the consequences

Every one of these dynamics is present, to varying degrees, in psychiatric prescribing. The opioid crisis killed over 500,000 Americans. It began not with street drugs but with confident physicians prescribing FDA-approved medications based on pharmaceutical company representations that turned out to be catastrophically wrong.

The psychiatric medication system has not produced a single catastrophe as concentrated as the opioid crisis. But the cumulative effect of tens of millions of people on medications of uncertain long-term efficacy, with significant side effects, prescribed through 15-minute evaluations based on unreliable self-report, managed by a system with no outcome tracking — this is not a system operating responsibly. It is a system hoping for the best.


← Part III: The Skeptic's Case Part V: The Honest Picture →