A children’s home that restrained often, where children stopped complaining

An inspection of a Welsh local-authority children's home found restraint was frequent, staff were uneasy about how often they used it, and some children had stopped believing complaint was worth making.

Children's homes · 1990s · Method: Rate and complaints

What happened

In 1998 the Social Services Inspectorate for Wales reported on Cartref-y-Gelli, a local-authority children’s home in Carmarthen. The report was obtained by BBC Wales in 2015.

It recorded that “the number of occasions when staff had been required to physically restrain children was high”. Staff were not happy with the necessity of using restraint, and some children felt staff had over-reacted in some situations. Not all the children were confident “of being listened to”, and some said it was not worth complaining.

The report made 14 recommendations and contained no major criticism of the home. Staff told inspectors they were committed to listening and dealing with complaints, though young people had sometimes telephoned the police instead. The inspectors said they could not judge the worth of past complaints.

The outcome

No enforcement action is recorded, and no legal outcome. The home closed in the 1998-99 financial year; Carmarthenshire council would not say why.

Cartref-y-Gelli was later at the centre of Dyfed-Powys Police Operation Almond, covering alleged physical and sexual abuse at the home between 1986 and 1990. Six people had been arrested by March 2015. That investigation is separate from the inspection and establishes nothing about the restraint practice recorded in 1998.

What went wrong

The finding is a rate, not a single hold. No incident was found unlawful. The inspectors put on record that staff were restraining children often, and that staff themselves were uneasy about how often it was necessary.

A team that does not believe in what it is doing, and is doing it at a high rate, has lost what makes restraint defensible: a decision taken deliberately, by people who think it is the last option.

Staff said they were committed to listening. Children said complaint was not worth the effort, and some rang the police. Both statements were recorded and neither was tested. The home was left with a restraint count and a complaints route, and no evidence that either worked.

What it teaches

A high restraint count is a warning sign in its own right, and here it was the finding itself.

Ask for the number first. A home that cannot say how often it restrains cannot say whether the figure is rising, or whether one child accounts for most of it.

Then ask the children. When a child says it is not worth complaining, that is information about the route, not the child. Some here went to the police. That is what a closed complaints system looks like from inside.

And listen to the staff. Their discomfort was written into the report and left there. It is the signal that de-escalation has stopped working, and cheapest to act on while it is still said out loud.

Questions for the room

Evidence and limits

Confidence: probable. The account rests on one publisher, BBC News, reporting an inspection report it obtained itself. Neither that report nor any later regulator record was opened for this entry. The 14 recommendations are not described in the material held, so this page does not say what they were.

The Operation Almond arrests establish nothing about restraint at the home, and no charges or outcomes in that investigation are recorded here.

Our restraint reduction training covers how to read a restraint rate before it becomes a serious case review.

Sources

Reviewed 20 October 2026.

All children's homes cases  ·  Restraint Casebook

Book in a 1-2-1 Meeting

Find Your Solution

What do you need? Use our simple tool to find the most relevant training and support options for your role, sector and challenges.