The most serious findings on overprescribing do not come from critics of psychiatry. They come from drug regulators and government-commissioned reviews.
REGULATORFDA 2005 & 2008
A death warning on every antipsychotic
In 2005 the FDA imposed a boxed warning — the strongest it has — on all atypical antipsychotics, after an analysis of randomised trials showed increased mortality in elderly patients with dementia. In 2008 the warning was extended to conventional antipsychotics as well.
Note the chronology: the promotion for which Lilly, AstraZeneca and Janssen were penalised targeted precisely this population.
[5] FDA boxed warning 2005 / 2008 · impact review: JAMA Network Open 2020
REGULATORBANERJEE REVIEW UK DEPT OF HEALTH 2009
1,800 extra deaths a year, in one country
Professor Sube Banerjee, in a report commissioned by the UK Department of Health, estimated that roughly 180,000 people with dementia in England were receiving antipsychotics. His estimate of the cost, verbatim: «an additional 1,620 cerebrovascular adverse events, around half of which may be severe, and an additional 1,800 deaths per year on top of those that would be expected in this frail population». That last clause matters: deaths in addition to the expected number, not total deaths.
This is by some distance the most serious documented finding on psychiatric overprescribing. And it is not about antidepressants in young people — it is about antipsychotics in elderly people with dementia.
[6] Banerjee S. "The use of antipsychotic medication for people with dementia: time for action", 2009
REGULATORFDA SEPT 2020 AUG 2016
Benzodiazepines, decades late
In September 2020 the FDA updated the boxed warning on all benzodiazepines — alprazolam (Xanax), bromazepam (Lexotanil), diazepam, lorazepam — acknowledging that physical dependence can develop after a few days to a few weeks of steady use — at the prescribed dose, without any misuse.
Earlier, in August 2016, it had imposed a boxed warning on roughly 400 products for the combination of an opioid with a benzodiazepine: respiratory depression and death.
[7] FDA Drug Safety Communication 23.09.2020 · [8] FDA DSC 31.08.2016
REGULATORMHRA 8 JANUARY 2026 UNITED KINGDOM
The newest: an exit plan before treatment even starts
On 8 January 2026 the UK medicines regulator (MHRA), following a review of pharmacovigilance data, clinical databases and patient reports, concluded that existing labelling did not adequately communicate the risk of dependence. Packaging and patient information leaflets will now carry the statement that the medicine "may cause addiction, dependence and withdrawal reactions".
The decision covers three classes: benzodiazepines, Z-drugs (zolpidem — Ambien/Stilnox, zopiclone, zaleplon) and gabapentinoids — pregabalin (Lyrica) and gabapentin (Neurontin), which are widely prescribed for neuropathic pain and anxiety and which patients often do not think of as psychiatric drugs at all.
The most consequential part is not the label but the instruction to prescribers: before treatment begins, a strategy for reducing and stopping should be agreed with the patient. The exit plan is part of the prescription, not a problem to be discovered later.
[19] MHRA Drug Safety Update, "Improving information supplied with gabapentinoids, benzodiazepines and z-drugs", 08.01.2026 · advice of the Commission on Human Medicines
SHARE
Sources for this section
[5]FDA boxed warning, increased mortality in elderly patients with dementia: atypical antipsychotics 2005, extended 2008. Review: JAMA Network Open 2020
[6]Banerjee S. "The use of antipsychotic medication for people with dementia: time for action". Report for the UK Department of Health, 2009.
[7]FDA Drug Safety Communication — Benzodiazepines, boxed warning for dependence and withdrawal, 23.09.2020.
[8]FDA Drug Safety Communication — Opioid/benzodiazepine co-prescribing boxed warning, 31.08.2016.
[19]MHRA Drug Safety Update — "Improving information supplied with gabapentinoids (pregabalin/gabapentin), benzodiazepines and z-drugs", 08.01.2026. gov.uk
Written byPetros Chatzianastasiou I am not a doctor. Every claim cites its primary source. Any step you take with your own treatment, always in consultation with your treating doctor and under their monitoring and guidance.