Research and analysis

Effectiveness of interventions to reduce the risk of violence-fixated individuals

Published 24 September 2026

This is not a statement of government policy.

Rapid projects support government departments to understand the scientific evidence underpinning a policy issue or area by convening academic, industry and government experts at a single roundtable. These summary meeting notes seek to provide accessible science advice for policymakers. They represent the combined views of roundtable participants at the time of the discussion and are not statements of government policy.

This publication considers the following question, taken from a meeting note of a roundtable chaired by Professor Paul Taylor (Chief Scientific Adviser, Policing), facilitated by the Government Office for Science on 27 May 2025.

What evidence exists, or is required, to understand the efficacy of interventions to reduce the risk of violence-fixated individuals (VFIs)?

For the purpose of this meeting, violence-fixated individuals (VFIs) were defined as: individuals with a fixation on violence, who are possibly showing intent to inflict violence, but who have no clear ideological motivations.

Summary of roundtable findings

There are various ways in which VFIs can come to the attention of public services and agencies. Touch points and pathways vary by individual, such that a single intervention or triage process will unlikely succeed.

Data and information sharing on individuals who come into contact with these touch points only work when they are timely and unobstructed.

There is limited evidence on the efficacy of interventions. A mixture of anecdotal, clinical and emerging empirical insights suggest that existing interventions can be effective (this could be tested more formally), but the system is fragile and dependent on what opportunities can be offered to those who are vulnerable.

This evidence also reveals the detrimental effects of poorly resourced interventions not delivering effectively and confirming a VFI’s view of society.

Interventions must be based on thorough assessments of individuals to understand their motivations, needs and personal context.

Robust evaluations of interventions are critical to target finite resources effectively and avoid unintended outcomes.

Most risk predicting tools consider variables in isolation and have proven ineffective, especially for adolescents.

Where structured interventions are not available, sustained engagement by services with individuals remains important.

Those working with VFIs need training around pertinent issues, including neurodivergence and digital cultures.

VFI touch points with public services

1. The VFI population is extremely heterogenous, including the very small cohort guilty of committing violence. They can easily be unknown to any authorities.

2. There are various behaviours potentially indicative of violence fascination, and various push factors potentially leading individuals towards violence fascination. However, many individuals ostensibly interested in violence will be consuming violent content for status, identity, friendship or for pleasure only. The vast majority of those accurately categorised as VFI will not commit violence.

3. There are a variety of ways in which VFIs can come to the attention of public services and agencies.

4. Touch points and pathways will be different for each VFI, but they include:

• schools, which raise concerns about students (including on the basis of parental concern)
• Prevent, or child adolescent mental health services, which some parents will approach directly
• children’s homes and welfare units
• the criminal justice system: pathways cover individuals charged under terrorism laws with adjacent violence fascination, those involved in mass violence with a counter-terrorism (CT) footprint, and those with violence fascination but no CT links (for example, individuals convicted for murder)
• GPs and social services, who often refer individuals to mental health services, which may escalate to forensic psychiatric care (this is a relatively uncommon pathway)
• faith leaders and community centres, who may receive disclosures about individuals of concern, but they lack formal referral pathways
• digital footprints: the digital age has increased visibility of violence ideation, especially among impulsive or younger individuals, although disciplined VFIs will leave no digital trace

5. Data and information sharing for the VFI cohort coming into contact with these touch points only really works if it is as timely as possible.

6. Prevent is often the default referral pathway for individuals with violence fascination, including those presenting incel or gender-based ideologies. However, Prevent is sometimes pursued in the absence of other viable options, not because it is the most appropriate (see paragraph 10).

Evidence for interventions

7. There is currently a fragmented and limited evidence base for the efficacy of interventions around VFIs, with a mixture of anecdotal, clinical and emerging empirical insights.

8. Prevent is a well-developed referral process, leading to some effective mentoring and identity-building activities. However, Prevent, as it is currently construed, can exacerbate cases by labelling individuals and stigmatising their views. Similarly, Prevent may raise anxiety or reinforce a sense of grievance, particularly for neurodivergent individuals whose needs and motivations are poorly understood.

9. Prevent struggles to address issues such as domestic instability, while contact with Prevent can lead schools and colleges (which serve a protective purpose) to exclude individuals. This compounds grievance and deepens violence fascination.

10. Interventions need to be based on thorough assessments of individuals to understand motivations, needs and personal contexts. This includes identifying the nature of the violence fixation, what it provides (for example, stimulation, control, identity) and what it manages (for example, distress or isolation). Without understanding the “why” behind behaviour, as well as the capacity of individuals to benefit from available help, any intervention is premature, ineffective and potentially exacerbating.

11. There is a large over-representation of neurodivergence among convicted, high-risk VFIs and among at-risk or vulnerable VFIs in the pre-crime space. This can have significant implications for understanding risk, need and responsivity in those individuals. Understanding the many ways that neurodivergence intersects with push and pull factors for violence fixations is critical for intervention design.

12. Children and adolescents may have different risk factors and needs from adults. All young people’s risks unfold in developmental contexts that must be considered in relation to interventions. There is a danger that interventions with children and adolescents which are too aggressive or too fast will do more harm than good.

13. Robust evaluation of interventions is critical to target finite resources effectively and avoid unintended outcomes.

14. For example, some randomised controlled trials have been used in relation to reducing violence. In one, intensive, person-centred support for knife-carrying young people was found to deliver a 60% reduction in offences, while lighter-touch interventions (for example, mentoring in custody) showed no effect.

15. Research in Thames Valley has found that in-depth, person-centred problem-solving approaches, where case workers work with individuals without pre-conceived plans, were more effective than other types of approach. The more in-depth the problem solving and mentoring, the greater the effect. Light-touch approaches had no effect on violent crime (Olphin and others, 2024; Olphin and Reed, 2024; Stevens and Olphin, 2024).

16. Most risk prediction tools consider variables in isolation and have proven ineffective, especially for adolescents. One tool used in policing only correctly identified 7% of the highest-risk cases.

17. Multivariate analysis considers multiple variables and their interactions simultaneously and offers a more accurate and nuanced understanding of risk, and has been attempted for Prevent. However, it requires more extensive data sharing among services than currently occurs (Clemmow and others, 2024).

18. A tiered approach to interventions may help to set priorities. Imagine a pyramid with:

• a large base of vulnerable individuals showing early signs of intent
• a middle tier of established VFIs who have not offended
• a tiny apex of high-risk individuals who have acted or are close to acting

Properly evaluated interventions should be scaled according to these tiers:

• lighter touch, preventative work at the base
• intensive, multi-agency work closer to the top

19. The prison service and other forensic mental health settings have extensive and nuanced understanding of why individuals commit violence (at the top of the pyramid). These insights should be used to develop better risk assessment frameworks towards the bottom of the pyramid.

20. Even where structured interventions are not available, sustained engagement with individuals remains important. For example, VFIs may not meet thresholds for mental health support or social care but still pose risks, particularly when different services do not share data. In some parts of the country, Forensic Child and Adolescent Mental Health Services (FCAMHS) groups have stepped in to provide support while appropriate wider support is found.

21. Those working with VFIs, especially in education, health and community settings, need training around themes such as neurodivergence. Adolescents require distinct support related to their developmental stage (Borum, 2025).

22. The credibility of intervention providers is important, particularly in terms of understanding the digital cultures and subcultures that shape young people’s fixations. Without this, engagement can be difficult.

23. There is a lack of education for young people around safe online behaviours in relation to crime and violence, as opposed to protecting their identity online. Provision in this area could be improved.

Meeting participants

These participants attended the meeting:

• Paul Taylor (CSA Policing, chair)
• Adam Carter (HM Prison and Probation Service)
• Grace Crawford (Barton Moss and Marydale Secure Children’s Homes)
• Hugh Stickland (CSA, Ministry of Justice)
• Lucy Warner (FCAMHS)
• Margarita Boateng (Ministry of Housing, Communities and Local Government)
• Maria Harkiolaki (Policing)
• Mayura Deshpande (National Health Service England)
• Nicki Fowler (Birmingham and Solihull Mental Health NHS)
• Razia Butt (Birmingham City Council)
• Tori Olphin (ResearchCore)

References

Borum R (2025). Developmental considerations for assessing the risk of extremist violence in adolescents. Journal of Threat Assessment and Management. Advance online publication.

Clemmow C, Fowler N, Seaward A and Gill P (2024). ‘Risk of what and why? Disaggregating pathways to extremist behaviours in individuals susceptible to violent extremism’, Behavioural Sciences and the Law (Accessed: 30 May 2025).

Olphin T, Bottomley S, Webb A and Stock M (2024). What works series: focused deterrence 12 month findings.

Olphin T and Reed K (2024). What works series: school navigators project.

Stevens J and Olphin T (2024). What works series: NEET to EET mentoring rapid evidence review.