Twelve-Week Trials, Fifteen-Year Prescriptions

Antidepressant prescriptions have surged since the pandemic ended, climbing alongside diagnoses of anxiety and depression that peaked during lockdown.

We are medicating more Americans than ever before. The question we're circling now — whether this represents genuine need or diagnostic creep, better access or pharmaceutical momentum — is the question we circled in 1980. We got the answer wrong because we never properly looked.

By 1980, roughly one in ten Americans was taking benzodiazepines, with Valium as the most-prescribed drug in the country. Doctors had solid evidence these pills worked for acute anxiety over two to four weeks. The pattern looked like success. Patients filled prescriptions, took them as directed, reported feeling better, and came back for refills.

What nobody measured was what happened over years. The follow-up studies came later. In the 1990s and 2000s, well after the epidemic had peaked. And revealed dependence rates double and triple what anyone expected, cognitive effects that persisted long after stopping, withdrawal syndromes so severe they sometimes required hospitalization. The prescriptions had not matched the evidence. They had simply outrun it. The benzodiazepine literature never corrected because someone finally asked the harder question. What do people look like after years?

The gap between what we test and what we do is where the real question lives.

Antidepressant research has lengthened since then, with better long-term data and cleaner randomized trials than we had for benzodiazepines. And yet the population-level studies. The ones that follow actual people in actual practice over decades, measuring not just symptom reduction but dependence, cognitive effects, tolerance, withdrawal. These remain fragmented and small. We measure efficacy in twelve-week trials and six-month follow-ups, then write prescriptions for fifteen years. The gap between what we test and what we do is where the real question lives.

You see this pattern outside medicine constantly. It is the distance between the controlled measure and the real deployment, where success gets assumed once the short-term proof arrives. It's in how a marketing campaign is judged by clicks rather than conversions, how a new workflow looks perfect in a two-week pilot, how a therapeutic technique proves itself in the clinic and fails at scale. The structure isn't unique to psychiatry. The blindness is universal. We are built to stop looking once the first evidence confirms what we wanted to believe.