Care Not Custody: training on reducing criminalisation of children in careA broken window is not a crime scene.

Story. Nathan, 15, is told he can’t go out. He punches a hole in his bedroom door. Staff call 999, he spends the night in a cell and is charged with criminal damage. At home with his own family, a parent would probably have made him pay for the door.

Theory. The DfE’s national protocol (2018) was written because children in care are more likely than other children to end up in the justice system, and it asks “would this be good enough for my child?”. Howard Becker’s labelling theory (1963) explains why it matters: once a child is labelled an offender, adults treat them as one, and they can start to see themselves that way.

Action. In mixed groups of home staff and police, people plot Nathan’s evening minute by minute and mark every point where a different decision could have kept it in the home. Each group then does the same for an anonymised incident from its own area, and the room agrees the points where a different decision matters most.

By the end of this part people can describe how an incident in a home becomes a record, and name the points where it could have stopped.

You leave with A timeline of where incidents in your area could stop, built with the people who make those calls.

Story. Chloe, 14, pushes a member of staff during an argument about her phone, then locks herself in the bathroom. The staff member is shaken but not hurt. It is 11pm and the manager is not on site.

Theory. The protocol sorts incidents into three levels. Level I is managed in the home. Level II is reported to the manager and the child’s social worker, who decide whether police are needed. Level III needs an immediate police response because of a risk of serious harm or significant damage. Where a child may be a victim of trafficking or modern slavery, section 45 of the Modern Slavery Act 2015 can protect them from prosecution.

Action. In mixed groups, people sort a set of incident cards drawn from your setting into the three levels, decide what happens next and write down why. One card carries signs of criminal exploitation, and the group works out what changes when it spots them. Groups then swap cards and test each other’s reasons, so home staff and police hear how the other side would call it.

By the end of this part people can apply the three levels to an incident, justify the decision in writing, and recognise signs that a child may be criminally exploited.

You leave with A one-page decision guide to use before picking up the phone.

Story. Jordan, 16, smashes a window after a phone call with his mum. Staff keep everyone safe, don’t call the police, and don’t try to talk it through that night. The next morning he is expecting a row.

Theory. Haim Omer’s Non-Violent Resistance (2004) teaches adults to delay: respond when everyone is calm, not in the heat of the moment, and bring in other trusted adults. Howard Zehr’s work on restorative justice (Changing Lenses, 1990) sees harm as damage to people and relationships, so the answer is putting it right. The protocol expects restorative approaches to be used.

Action. In threes, people practise the next-morning conversation with Jordan: what happened, who was affected, and what he will do to put it right. Everyone takes a turn as the adult, as Jordan and as the observer giving feedback. They also practise the calm, short words staff use in the moment to bring things down.

By the end of this part people can plan a delayed response to harm and hold a restorative conversation that holds a child to account without a police report.

You leave with De-escalation steps for use on shift, and a restorative conversation script.

Story. Maya, 17, lives in supported accommodation 80 miles from home. When police are called to an argument there, the officers don’t know she is in care and nobody tells them.

Theory. The protocol asks every area to have a local protocol that covers named leads, information sharing, joint decisions on whether to prosecute, children placed from other areas, care leavers up to 25, and data each agency checks. It lists options short of charge, such as community resolution and Outcome 21, where a crime is recorded but formal action is not in the public interest.

Action. Homes and police check your local protocol against the national model element by element, mark what is working and what is missing, and agree actions with a named owner and date for each. The group then tests the agreed actions against Nathan’s evening from part 1.

By the end of this part people can describe the police options short of arrest and charge, and leave with agreed actions to strengthen their local protocol.

You leave with A joint protocol checklist, with actions agreed by homes and police.