When someone leaves prison, who is managing the domestic abuse risk?

Much of the recent debate about changes to prisoner release has understandably focused on which offences should be excluded. Domestic abuse perpetrators are not excluded as a category from the changes due to take effect in October, and the Domestic Abuse Commissioner has raised serious concerns about pre-release risk assessment, probation capacity and whether victims will receive enough information and support.

Those concerns are well founded, but there is a wider issue here which I think risks being lost in the argument about offence lists. The offence someone is serving a sentence for may tell us why they are in prison, but it does not necessarily tell us the domestic abuse risk they pose when they leave.

A perpetrator may be serving a sentence for assault, criminal damage, breach, drugs, weapons or another offence altogether, while their wider domestic abuse history sits elsewhere. Police intelligence, previous incidents, MARAC information, probation records, restraining orders, stalking behaviour, safeguarding referrals and information from specialist services may all tell part of that story. If the focus stays on the index offence, there is a real danger that the wider pattern is missed.

The index offence is not the whole picture

This is not a new problem. The Domestic Abuse Commissioner has previously highlighted evidence that around a third of people managed by probation were known domestic abusers even where domestic abuse was not their index offence.

That gets us to the heart of the issue. The criminal justice system necessarily deals in offences and convictions, but risk management must look beyond that to behaviour, history, and context. A conviction records what was proved in court and what somebody was sentenced for. It may tell us very little about previous partners, stalking, coercive control, threats, strangulation, repeated breaches, escalating behaviour or the level of fear experienced by a victim.

Domestic abuse is particularly difficult to understand through isolated offences because apparently separate events can take on a very different meaning once the wider pattern is visible. Criminal damage may be part of intimidation. A breach may show persistence and disregard for protective measures. Contact that appears relatively innocuous when read on its own may look very different when placed alongside stalking or coercive control. Offending involving someone else may also be relevant to what professionals know about the person’s capacity to cause serious harm.

Before release, then, the question cannot simply be what offence someone is serving a sentence for. It has to include what is known about their behaviour, their history and the risk they pose now.

MAPPA already recognises the problem

I have chaired both MAPPA and MARAC during my career. They are different processes serving different purposes, and it would be wrong to blur the two.

MARAC is focused on reducing risk to high-risk victims of domestic abuse and their children through coordinated multi-agency action. MAPPA provides the statutory framework through which police, probation and prisons, together with other agencies where appropriate, assess and manage eligible offenders who may pose a risk of serious harm.

What is relevant to this debate is that MAPPA guidance already recognises that an offence label does not always tell professionals enough. A domestic abuse perpetrator can potentially be referred into MAPPA arrangements even where the conviction itself is not for a violent offence, and somebody convicted of a relatively minor offence, or one which appears unrelated to domestic abuse, can in the right circumstances be considered for Category 3 management where their behaviour indicates a capacity to cause serious harm and formal multi-agency management is required.

The guidance also expects professionals to remain alert to domestic abuse throughout the sentence rather than treating risk identification as something completed at the start. That is the correct approach. Risk assessment should move with the circumstances, not become fixed because a risk assessment was completed or a label applied months earlier.

The planned national expansion of the Domestic Abuse Perpetrators on Licence scheme appears to recognise the same principle. Electronic monitoring can be used because an offender is assessed as presenting domestic abuse risk, rather than simply because the conviction itself is labelled as domestic abuse. If that approach is accepted for monitoring, the same thinking should shape the wider management of risk on release.

Release changes the circumstances

One of the points I regularly make in risk assessment and MARAC training is that risk is dynamic. An assessment does not remain reliable simply because it was once completed, and release from prison is a significant change in circumstances.

While somebody is in custody, their ability to approach, monitor, follow or physically reach a victim is restricted. Once they are released, that changes. The victim may have moved, started a new relationship or made decisions about children, housing or work based on the perpetrator being in custody. The perpetrator may also have formed a new relationship. There may have been contact from prison, attempts to communicate through third parties or behaviour suggesting that fixation has continued.

Equally, the absence of recent incidents may tell us very little if the person has simply been unable to reach the victim. A period of apparent calm should not be read automatically as a reduction in risk.

For some victims, imprisonment offers a period in which they can begin rebuilding their lives. Release can alter that position very quickly, which is why it should trigger an active reassessment of risk rather than a simple check that a release plan already exists.

Sharing information is not the same as managing risk

Multi-agency working is often talked about as though fragmented risk can be solved simply by sharing more information. Information sharing is essential, but it is only useful if somebody then makes sense of what has been shared.

Police may hold previous domestic abuse incidents and intelligence. Probation may know about offending, compliance and behaviour under supervision. Prison staff may know about conduct in custody. MARAC may have considered risk to a particular victim, while an IDVA or specialist domestic abuse service may understand that victim’s current circumstances and level of fear. Children’s services, housing, health and substance misuse services may each hold something else.

The issue is not simply whether all of that information exists or has been exchanged. It is whether somebody has pulled it together and asked what it means when seen as a whole.

Has the behaviour escalated? Are there previous partners? Has the perpetrator repeatedly breached orders or conditions? Is there stalking or fixation? Have the tactics changed? Are there threats of suicide, homicide or serious violence? Is the proposed address appropriate? Do the licence conditions reflect the behaviour causing concern? Has the victim been told about the release and had an opportunity to provide information relevant to their safety?

Those are not administrative questions. They are risk management questions, and somebody needs to be responsible for answering them.

MARAC and MAPPA may see different sides of the same risk

A victim may have been discussed at MARAC while the perpetrator is being managed through prison, probation or MAPPA arrangements. Those systems may therefore be looking at the same relationship from different directions.

That can be a strength, but only if the information connects. MARAC should not be assumed to be managing the perpetrator simply because their behaviour is part of the discussion, and offender management should not assume that it has a complete understanding of victim risk simply because the perpetrator has been assessed.

Where both processes are relevant, the question is whether those different perspectives are informing one another. MAPPA guidance recognises the need to draw information from police, children’s services, domestic abuse services and other agencies, while ensuring that the victim’s perspective informs risk management.

That is what effective multi-agency work should achieve: not more process, but a fuller understanding of the risk and a clearer plan for dealing with it.

What happens when someone does not meet a formal multi-agency threshold?

MAPPA is an important mechanism, but it is not the answer for every domestic abuse perpetrator leaving prison, nor should it be. Not every offender will require management at MAPPA Level 2 or 3, but falling outside a formal multi-agency process does not mean that the risk has disappeared.

A decision that somebody does not require that level of MAPPA management does not remove the need for agencies to work together where the circumstances require it. This is where professional judgement becomes particularly important. Thresholds and processes help organise work, but they should not replace thinking.

If the available information shows stalking, repeat domestic abuse, previous partners, escalating behaviour, repeated breaches or other indications of serious harm, professionals still need to understand what those factors mean and what response is required, regardless of which meeting the individual qualifies for.

There still needs to be a risk management plan, appropriate information sharing and clarity about responsibility. Somebody still needs to own the risk.

Domestic abuse-related death reviews often take us back to the same problem

My work chairing and authoring Domestic Homicide Reviews and Domestic Abuse Related Death Reviews has reinforced something I had already seen through policing, MARAC and MAPPA: the most significant information is rarely contained in one incident or held by one organisation.

Reviews reconstruct what happened after a death and, once the chronology is assembled, information which previously sat in different places can take on a very different significance. That does not mean every death was predictable or preventable, and reviews must guard carefully against hindsight, but they repeatedly show why professionals need to understand patterns, cumulative harm, changing circumstances, and the relationships between pieces of information that may have looked unconnected at the time.

We should not need a review after somebody has died before that wider picture is assembled. The purpose of risk management is to do that work beforehand.

Better identification will help, but it will not solve everything

The Sentencing Act 2026 introduces a formal judicial finding of domestic abuse where a court determines that an offence involved domestic abuse. That should improve visibility within the criminal justice system and address a longstanding difficulty, because domestic abuse does not sit within one offence. The underlying conviction may be for assault, criminal damage, coercive control or something else.

Better recording should help professionals see that history more easily, but it cannot capture everything. Previous relationships, unreported abuse, police intelligence, stalking behaviour, previous victimisation and behaviour which was never part of a prosecution may still be highly relevant when assessing future risk.

A flag can make information easier to find. It cannot replace professional curiosity.

The real test comes when the prison gate opens

The debate about earlier release will continue, including legitimate arguments about which offences should or should not be excluded. Whatever policy is finally implemented, offenders will continue to leave prison and be managed in the community. For domestic abuse, the test cannot simply be whether the correct release date was calculated.

The real question is whether the system recognised the risk the person left prison with. Was their domestic abuse history known? Were previous partners considered? Were stalking, coercive control, threats and breaches understood as part of a pattern? Was the victim’s current situation known? Did the change in circumstances lead to a fresh assessment? Were MARAC, MAPPA and other multi-agency arrangements used where appropriate rather than assumed to be somebody else’s responsibility? Did licence conditions reflect the behaviour causing concern, and was somebody clearly responsible for making sure the plan worked?

The offence tells us why somebody entered prison. Managing domestic abuse safely requires us to understand considerably more before they leave it.

About the author

John Trott is an independent domestic abuse and safeguarding specialist, former Detective Chief Inspector and former force lead for domestic abuse, stalking and public protection. During his police career, he chaired both MARAC and MAPPA and chaired Domestic Abuse Serial and Serious Perpetrator meetings. He has chaired more than 300 MARACs and continues to deliver specialist training on domestic abuse, stalking, risk assessment, MARAC, professional judgement and multi-agency safeguarding. John also undertakes independent Domestic Homicide Review and Domestic Abuse Related Death Review Chair and Author work.

© 2026 John Trott, AbuseFreeLife. All rights reserved.