What has not ended

The first large study of memory was done in 2007. The method has been in use since 1938.

The paper opens with its own admission, verbatim: "Despite ongoing controversy, there has never been a large-scale, prospective study of the cognitive effects of electroconvulsive therapy". That was written in 2007. ECT had by then been in use for sixty-nine years.

Of 751 patients referred, 347 took part in at least one post-treatment evaluation, across seven New York facilities. The conclusion, verbatim: "adverse cognitive effects were detected 6 months following the acute treatment course".

Specifically: bilateral electrode placement produced "more severe and persisting retrograde amnesia" than right unilateral; sine wave stimulation produced pronounced slowing of reaction time, both immediately and six months later. And: "advancing age, lower premorbid intellectual function, and female gender were associated with greater cognitive deficits".

And what happened next. The damage had been measured by the field itself and published in a leading journal. The method carried on being used — and being imposed on people who do not consent. No critic is needed to say this: they measured it themselves, and they did not stop.

[64] Sackeim HA, Prudic J, Fuller R, Keilp J, Lavori PW, Olfson M. "The Cognitive Effects of Electroconvulsive Therapy in Community Settings", Neuropsychopharmacology 2007;32:244-254, DOI 10.1038/sj.npp.1301180

In use since 1938. Placebo-controlled evidence: 224 patients.

The only way to show that a treatment works beyond expectation is comparison against a sham — for ECT, anaesthesia without the current. There are eleven such trials. They are all earlier than 1986. Between them they cover 224 patients who received ECT and 187 controls.

The five meta-analyses the literature relies on draw on between one and seven of those eleven. The 2019 quality review concluded that the research as a whole "barely supports short-term benefit" and contains no indication of long-term benefit, of suicide reduction, or of greater effectiveness in older people. Its authors called for the method's immediate suspension until properly designed trials are done.

That concession is what counts. It does not come from the critics; it comes from the man defending the method. Eighty-seven years after its introduction, both sides agree that there is almost no evidence about what remains once the treatment stops. They disagree only on whether that is enough to carry on.

This page's position, without hedging. A method in use since 1938; whose entire body of controlled evidence is 224 patients from trials that stopped in 1986; whose own advocates measured persisting memory damage at six months; and whose defender concedes that we do not know what remains once it stops — has not proved what it claims. And it continues to be imposed on people who do not consent.

This is not an open scientific question. It is a practice unjustified by its own evidence.

[65] Read J, Kirsch I, McGrath L. "Electroconvulsive Therapy for Depression: A Review of the Quality of ECT versus Sham ECT Trials and Meta-Analyses", Ethical Human Psychology and Psychiatry 2019;21(2):64-103 · [66] Anderson IM. "Electroconvulsive therapy (ECT) versus sham ECT for depression: do study limitations invalidate the evidence?", BJPsych Advances 2021;27(5):285-291

There is no "UK" framework. There are three.

England and Wales: under s.58A of the Mental Health Act 1983, ECT may not be given to a patient aged 18 or over who is capable of consenting and does not consent — except where s.62 applies, that is, treatment immediately necessary to save life or, not being irreversible, to prevent serious deterioration.

Scotland: a separate and stricter regime applies under the Mental Health (Care and Treatment) (Scotland) Act 2003, where a capacitous refusal is binding and cannot be overridden. Northern Ireland differs again.

Greece: we located no ECT-specific statute; the general provisions of Articles 95–99 of Law 2071/1992 on involuntary hospitalisation apply.

And the finding that is not a number: we searched for official statistics on how often ECT is administered without consent. We found none, in any of the three countries. That a practice of this weight is not systematically recorded is itself the finding.

[48] Mental Health Act 1983, ss. 58A and 62 · [49] Mental Health (Care and Treatment) (Scotland) Act 2003 · Greek Law 2071/1992, arts 95–99 · [50] Kaliora SC et al., J ECT 2013;29(3):219-224

Chemical restraint: drugs for behaviour, not for illness

In 2007 the US HHS Office of Inspector General reviewed six months of Medicare claims and found that 14% of elderly nursing home residents had a claim for an atypical antipsychotic — 304,983 people out of 2.1 million. That 51% of those claims were erroneous, amounting to $116 million. And that 22% of the drugs claimed were not administered in accordance with the standards on unnecessary drug use.

In October 2009 a report to the responsible minister at the UK Department of Health estimated that around 180,000 people with dementia were being treated with antipsychotics each year, and that up to 36,000 would derive some benefit. 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: these are deaths in addition to the expected number, not total deaths.

And the trend line, with equal weight: both reports preceded national reform programmes, and prescribing has since fallen substantially in both countries. 2007 is not 2026.

The distinction from the sister site: psychdrugs.org covers the FDA's mortality warnings. Here the angle is restraint — a drug given to control behaviour rather than to treat illness.

[51] HHS Office of Inspector General, OEI-07-08-00150, May 2011 (review period 01.01–30.06.2007) · [52] Banerjee S, "The use of antipsychotic medication for people with dementia: Time for action", 2009

Residential programs for "troubled teens" — and why the title says non-medical

In October 2007 the Government Accountability Office testified to the House Committee on Education and Labor that it had found thousands of allegations of abuse, some involving death, at residential programs for troubled youth in the United States and at American-owned facilities abroad, between 1990 and 2007. It examined 10 closed civil or criminal cases from 1990 to 2004 in which a teenager died while enrolled in a private program.

GAO states itself that it did not verify the allegations, and that allegations should not be confused with proof of abuse. It also states that no federal agency collected comprehensive nationwide data.

The programs were wilderness therapy programs, boot camps and academies. Contributing factors GAO identified: the hiring of untrained staff and a lack of adequate oversight.

Why it appears here with an asterisk: these are largely residential settings, not psychiatric clinics, and it is a purely American matter. It is given as one exhibit, not as a pattern.

[53] US GAO, GAO-08-146T, "Residential Treatment Programs: Concerns Regarding Abuse and Death in Certain Programs for Troubled Youth", testimony 10.10.2007

Six years for a ban that has still not been issued

In March 2020 the US Food and Drug Administration banned electrical stimulation devices that deliver skin shocks to stop self-injurious or aggressive behaviour. They were used almost exclusively at one facility, the Judge Rotenberg Educational Center in Canton, Massachusetts.

On 6 July 2021 the Court of Appeals for the D.C. Circuit vacated the ban, two votes to one. The reason was not safety: the Court held the FDA had no power to ban a device for a particular use, because another provision of the statute forbids it from regulating the practice of medicine. The Chief Judge dissented.

What the Court did not decide, in its own words: "we do not address petitioners' other arguments, including whether the ban was arbitrary and capricious or whether substantial evidence supported the FDA's factual determinations". The ruling did not say the devices are safe. It said the FDA had taken the wrong route.

In December 2022 Congress closed the gap. Section 3306 of the Food and Drug Omnibus Reform Act amended the statute so that the FDA may expressly ban a device "for one or more intended uses" — a provision written in answer to this very ruling.

On 25 March 2024 the FDA proposed a new ban, finding psychological risks of "depression, anxiety, worsening of underlying symptoms, development of post-traumatic stress disorder" and physical risks of "pain, burns, and tissue damage".

The final rule has not been issued. The deadline has moved repeatedly; in June 2026 it slipped again, to November. The legal basis has been repaired since 2022 and the proposal was filed in 2024. Meanwhile, on the FDA's own finding, one facility still uses them.

[68] Judge Rotenberg Educational Center v. FDA, U.S. Court of Appeals, D.C. Circuit, No. 20-1087 (consolidated with 20-1088), 06.07.2021 · [69] 21 U.S.C. §360f as amended by FDORA §3306, December 2022 · [70] FDA, proposed ban, 25.03.2024

Allegation, not finding: the financial incentive to admit

In July 2020 Universal Health Services, among the largest private providers of inpatient psychiatric care in the United States, agreed to pay $122 million to close a federal investigation. The Department of Justice alleged that across roughly thirty units between 2006 and 2018, people were admitted who did not need inpatient care, were not discharged when they no longer needed it, and that physical and chemical restraint were misused. The investigation arose from nineteen separate whistleblower suits.

And now the part that makes this an allegation and not a finding. The settlement is not an admission of liability. The company stated expressly that it "unequivocally disputes any allegation that it engaged in wrongdoing of any kind" and that the agreement "is not an admission of liability but merely a resolution of a civil claim". No court found that any patient was harmed. No clinician or executive was named.

It appears here for one reason only: it describes a mechanism that appears nowhere else on this page — the financial incentive to keep someone admitted. As an allegation it is documented; as a finding it does not exist.

[71] U.S. Department of Justice and HHS Office of Inspector General, Universal Health Services settlement and Corporate Integrity Agreement, 10.07.2020

Deinstitutionalisation as an accounting exercise: 144 dead, and to date no conviction

In 2015 the government of South Africa's Gauteng province terminated, citing cost, its contract with the private provider Life Esidimeni, which had housed chronic psychiatric patients for years. In the single quarter 1 April – 30 June 2016 about 1,371 patients were hurriedly transferred to hospitals and to 27 non-governmental organisations — 457 a month, up from 13 a month in 2015. The country's Health Ombud, Malegapuru Makgoba, published on 1 February 2017 a report finding that "all the 27 NGOs … operated under invalid licenses" and that "all patients who died in these NGOs died under unlawful circumstances"; it recorded 94 deaths at the time of writing — against the 36 the province had announced — and that "no feasibility studies or costing exercises were performed" on the alternatives. In the sample of deaths it examined, the most common causes were community-acquired pneumonia and uncontrolled seizures — the latter "raises concern that these patients did not receive medication for their epilepsy".

The arbitration. The State conceded liability and referred compensation to arbitration before the former Deputy Chief Justice of the Constitutional Court, Dikgang Moseneke. The award of 19 March 2018 opens: "It is now undisputed that … 144 mental health care users died and 1418 were exposed to trauma", while the whereabouts of 44 remained unknown. It called the reasons given for the move "false, disingenuous and advanced in order to conceal the true reasons" — the provincial Treasury, the province's own finance minister Barbara Creecy testified, "has never demanded that any department cut core services" — and the licensing of the NGOs "unlawful and knowingly fraudulent". It awarded R20,000 for funeral expenses to the families of the deceased, and to each claimant R180,000 for shock and trauma and R1,000,000 in constitutional damages for the "unjustifiable and reckless" breaches of rights.

What the court found, and what it did not. The prosecuting authority had initially declined to prosecute on the 141 death dockets. The inquest of the High Court in Pretoria, in a judgment of 10 July 2024, found that nine of the deaths "were negligently caused by the conduct" of the then provincial health minister Qedani Mahlangu and the then director of mental health Makgabo Manamela, who "created circumstances in which the deaths … were inevitable"; for most other deaths, without autopsy or with insufficient evidence, the court said it was unable to make a finding, and for the then head of department it did not establish causation. It is not a conviction: an inquest finds, it does not try. On 21 April 2026 the National Prosecuting Authority announced that it had decided to prosecute, with charges "expected to include culpable homicide, among others", naming neither accused nor a trial date. As of our last check (September 2026) no one has been convicted. The persons named argued before the arbitration that they could not reasonably have foreseen the deaths; if they are prosecuted, foreseeability is what will be tried.

Why it is here. This is not a case of bad psychiatric treatment — it is a case of an administrative decision taken over people who could not react. Deinstitutionalisation, whose delay Part A records, also has a fully documented failure, when done as an accounting exercise; it belongs here by the same measure. The "about 1,700 patients" in circulation has a source, though not the one usually given: the arbitration writes that the end of the contract "precipitated the move of 1711 mental health care users" between October 2015 and the end of June 2016 (§24); the Health Ombud's 1,371 covers only the final quarter, with about 160 more in the preceding year.

[82] Health Ombud, report into the circumstances surrounding the deaths of mentally ill patients in Gauteng, 01.02.2017 · [83] Life Esidimeni Arbitration, Award, Moseneke, 19.03.2018, §§1–2, 47, 226 · [84] High Court of South Africa, Gauteng Division, Pretoria, inquest I001/21, 10.07.2024, §§3, 570–574 · [85] National Prosecuting Authority, statement 21.04.2026