Prison and probation services are under-researched settings from a health perspective. Previous blogs here have explored the prevalence of mental health problems including self harm in prisons, the psychological impact of incarceration and unmet mental health needs related to reincarceration, but never before a randomised controlled trial in a prison setting. A 2014 review of 14,000 applications to UK Research Ethics Committees over 2 years found that only 100 (0.7%) planned to involve participants in prison or under probation (Charles et al, 2014), of which only three tested treatment interventions.
This blog addresses the findings of just such a study newly published in The Lancet Psychiatry, an RCT of Mentalisation-Based Therapy (MBT) for the common but frequently misunderstood disorder known as Anti-Social Personality Disorder (ASPD).
‘Personality disorders‘ are a group of conditions where the way an individual thinks, perceives, feels or relates to others differs significantly from an average person. Within this, ASPD is typically by a sustained pattern of impulsive, irresponsible and often criminal behaviour. Our best estimate is that ASPD affects 46% of men and 27% of women within the prison population, though prevalence studies have their methodological challenges.
The manifestations of ASPD are both individually distressing and impairing, and relevant at a societal level, yet evidence for effective treatments is limited. There are currently no specific treatments recommended in NICE guidance, though it is recommended to “consider offering group-based cognitive and behavioural interventions” targeted at symptoms such as impulsivity and offending (NICE, 2013).
The M in MBT refers to a person’s ability to understand and consider their own mental states and those of others, an important aspect of healthy relationships and emotion regulation. MBT has previously been shown to be effective as a treatment for borderline personality disorder (BPD) by two of the authors of the current study (Bateman and Fonagy, 1999 and 2001), and has also been evaluated amongst those with comorbid BPD and ASPD (Bateman et al, 2016). However, its effectiveness for ASPD alone has not previously been examined in clinical trials.
Randomised Controlled Trials (RCTs) are very rare in prison settings, so it’s encouraging to be blogging today about the MOAM trial (Mentalization for Offending Adult Males).
Methods
The study included men aged 21 or older who were convicted of an offence and under probation at one of the 13 study sites, who met diagnostic criteria for ASPD and had an elevated baseline score (>15) on an assessor-rated measure of impulsive aggression, the Overt Aggression Scale-Modified (OAS-M).
Those in the intervention group received 12 months of weekly group MBT and monthly individual therapy.
The comparison group were allocated to “probation as usual”, the current standard approach. This involves regular appointments that are mandatory as part of supervision and risk monitoring (i.e. missing appointments may lead to recall to prison) and also involve provision of wider support around community integration.
The primary study outcome was participants’ OAS-M score at 12 months. The OAS-M measures frequency and severity of verbal and physical aggression (including against oneself). Mean scores at baseline were 169 in the control group and 158 in the treatment group.
Participants were excluded from taking part if they been convicted of child sexual offences, had a neurodevelopmental disorder, schizophrenia or bipolar disorder or had “inadequate language/cognitive skills”.
The study was assessor blinded, i.e. therapists were aware they were delivering MBT, but not the researchers doing follow-up assessments.
Recruitment was via the Community Offender Personality Disorder (OPD) Pathway Service (the community part of a set of psychologically informed services, jointly funded by HMPPS and NHS England, aimed at reducing reoffending and improving mental health in this cohort).
The study included men aged 21 or older who were convicted of an offence and under probation at one of the 13 study sites.
Results
313 men with ASPD and high aggression participated in the study. They had an average age of 34.2 years and 79% of them were white ethnicity, 10% black.
Primary outcome, aggression:
- 157 received the MBT treatment.
- By the end of treatment, adjusting for other non-treatment differences between the groups, the therapy group scored on average 73.5 points lower on the OAS-M (95%CI -33.2 to -113.7, highly statistically significant).
- The authors followed participants up for a further 12 months after treatment ended and found that the mean difference in aggression scores between the groups had reduced to 25.4 points (95% CI -60.1 to 9.2, p = 0.16) which was no longer statistically significant.
Secondary outcomes, ASPD and offending
The authors also looked at two secondary outcomes including whether participants still met diagnostic criteria for ASPD and offending.
- On the ASPD outcome there was a large improvement in both treatment and control groups at 12 months. This improvement was more substantial in the treatment group and statistically significant compared to the control group (mean difference -0.71, 95% CI -1.4 to -0.051, p 0.035) though the difference was lower and no longer statistically significant at 24 months.
- Receiving MBT did not have a discernible effect on offending across the three years of follow-up (IRR 1.00, 95% CI 0.83 to 1.20, p = 1.00) though the authors note that offending was lower for the MBT group in the third year of the study compared to control and the difference was statistically significant (IRR 0.54, 95% CI 0.36 to 0.81, p=0.0024).
- Frequent attenders at MBT sessions were less likely to reoffend across first and second years of follow-up. However, the study can’t tell us whether it was the extra MBT sessions attended that led to that difference, or some other unmeasured confounder like greater overall motivation or more effective coping strategies.
Mediation analysis
Further evidence that the treatment was exerting an effect comes from a mediation analysis, where the authors showed that the treatment appeared to work in this group by reducing participants’ uncertainty about (their own and others’) mental states.
Attrition
By 12 months, 15% of the treatment group and 21% of the control group had withdrawn from the study. By the end of 24 months of in-person follow-up, these figures had reached 32% and 38% respectively.
There were also a large proportion of participants who did not withdraw but missed measurements. At the 12-month point, less than half of all original participants were measured (50% in treatment arm versus 44% in the control arm), due to withdrawals and people missing that measurement. At the 24-month point, the proportion with a measurement increased (68% in the treatment group and 61% amongst controls) as those missing but not withdrawn were contacted for a measurement.
This trial suggests that mentalisation-based treatment holds promise as an effective intervention for males with antisocial personality disorder within a forensic setting.
Conclusions
The authors conclude that:
MBT-ASPD holds promise as an effective intervention for individuals with antisocial personality disorder within a forensic population. Future research should explore these findings’ generalisability and the sustainability of treatment gains.
Strengths and limitations
Strengths
This study was a significant logistical undertaking in a setting that is neither set up for nor used to hosting clinical trials. There are some specific methodological strengths. Ensuring balanced (stratified) groups, for example, avoided the pitfalls of unbalanced study arms commonly encountered in RCTs. The large number of participating sites makes the findings more generalisable. Checks of MBT adherence demonstrated that treatment was being delivered similarly across the sites. Adjusting for the effect of different centres (using a so-called multi-level model) improved the validity of results. Involving lived experience researchers in the study may have helped retain participants.
Limitations
An unknown proportion of those referred into the study were excluded due to having low language and cognitive skills (along with those with co-morbidities). Very low skills in these areas could make psychological therapy inappropriate. However, we would have liked to see the authors define this criterion more clearly and justify their decisions, to allow readers to assess who did (and didn’t) participate in the study and therefore judge to whom the evidence supports giving MBT. Poor literacy skills are common in prison populations, making this is an important consideration for the real-world implementation of such an intervention (Creese, 2016).
The missing data issues likely reflect challenges faced by researchers trying to follow up a group who typically experience high levels of social instability. Frequent changes, or lack, of address, difficulty accessing and engaging longitudinally with services and high levels of substance misuse are common in this population, and all add to the fragmentation of care that poses such a clinical challenge. Given how rarely researchers even attempt to study these populations, it could be viewed as an achievement to follow up so many in the control arm working with an overstretched probation service.
Having said this, the high proportion of missing values at that crucial 12-month timepoint, the focus of the authors’ analysis, should temper our excitement about the findings. The authors also acknowledge that imputing missing data would not fully capture the actual population characteristics and reduces the reliability of findings.
We don’t know what the aggression scores would have been in the large group who missed that measurement, but if those missing had higher-than-average aggression scores then it would artificially inflate the reduction in aggression compared with baseline. Coupled with the fact that more people were missing in the control group (56% vs 50%), there is a chance that the headline treatment effect is inflated. Given that most of those participants missing at 12 months nonetheless came back for a measurement at 24 months, it could also help to explain why the reduction in treatment effect (from 73.5 points difference between groups at 12 months to 25.4 at 24) was so large. The effect of time passing – and the loss of mentalising skills amongst the therapy group – could also have contributed.
Also worthy of note, and a specific challenge of a study in this context, is that those in the control group spent more time in prison during follow-up (due to breaching their probation license or reoffending), which would have reduced their time-at-risk of repeat offending and reduced the overall duration of their involvement in the control intervention.
The authors are forthcoming about these limitations and others in the paper.
Understandably there were significant amounts of missing data which may reduce the confidence we can have in outcomes.
Implications for practice
A 2020 Cochrane review of the evidence on effects of psychological therapies for ASPD (Gibbon et al, 2020) concluded:
very limited evidence available on psychological interventions for adults with ASPD… No intervention reported compelling evidence of change in antisocial behaviour. Overall, the certainty of the evidence was low or very low.
This paucity of good evidence for effective ASPD therapies means that the emergence of a promising potential treatment is very welcome. Despite the rigorous study design, the difficulties of following participants up highlight the challenges of researching interventions for forensic populations. The findings are promising, but further work and replication is needed to consider implementing into day-to-day practice. However, the authors should be congratulated for their work and its rigour, in a challenging context for research.
A key issue to examine in future research is whether MBT helps improve real-world outcomes like offending behaviour. The authors plan to publish economic evaluations showing the cost-effectiveness of the intervention. Given the social and economic costs associated with ASPD and how prevalent it is, a treatment that reduces offending even by a small amount has the potential to have an important economic and indeed social impact. In time, examining the intervention in women (who make up a small but nevertheless important sub-group of those diagnosed with ASPD) will also be important.
From a personal clinical perspective of working in prisons, the high prevalence of ASPD – often co-morbid with other psychiatric disorders – and the lack of evidence on effective treatment options can contribute to therapeutic hopelessness. Perhaps as a consequence, pharmacological interventions, such as the “off-licence” use of antipsychotic medication to manage symptoms like emotional dysregulation, are relatively common (Hassan et al, 2016). Whilst there is some evidence for a reduction in violence, there are associated issues with adverse effects and polypharmacy (Herttua et al, 2022 and subject of a previous Elf blog).
This study therefore offers some hope for scalable alternatives and complements other, novel, emerging strands of research in ASPD such as pharmacologically targeting underlying processing deficits (e.g. Tully et al, 2023). It also acts as a timely reminder that the same standards of evidence and methodological rigour can be applied to research in these populations as in the rest of medicine (Tully et al, 2024).
The authors are to be congratulated for completing the study in challenging circumstances with findings that provide clinical optimism for this group.
Statement of interests
Tom receives funding for his own research from NIHR, who also funded this study.
No other interests.
Links
Primary paper
Fonagy P, Simes E, Yirmiya K, et al (2025) Mentalisation-based treatment for antisocial personality disorder in males convicted of an offence on community probation in England and Wales (Mentalization for Offending Adult Males, MOAM): a multicentre, assessor-blinded, randomised controlled trial. Lancet Psychiatry; 12,3: 208–19.
Other references
Bateman A and Fonagy P (1999) The effectiveness of partial hospitalization in the treatment of borderline personality disorder – a randomised controlled trial. Am J Psychiatry; 156:1563–1569. https://doi.org/10.1176/ajp.156.10.1563.
Bateman A and Fonagy P (2001) Treatment of borderline personality disorder with psychoanalytically oriented partial hospitalisation: an 18-month followup. Am J Psychiatry; 158:36–42. https://doi.org/10.1176/appi.ajp.158.1.36.
Bateman A, O’Connell J, Lorenzini N, et al (2016) A randomised controlled trial of mentalization-based treatment versus structured clinical management for patients with comorbid borderline personality disorder and antisocial personality disorder. BMC Psychiatry; 16: 304 https://doi.org/10.1186/s12888-016-1000-9.
Charles A, Rid A, Davies H et al (2014) Prisoners as research participants: current practice and attitudes in the UK, Journal of Medical Ethics, 42:246–252 https://doi.org/10.1136/medethics-2012-101059.
Creese, B (2016) An assessment of the English and maths skills levels of prisoners in England. London Review of Education, 14 (3), 13-30 https://doi.org/10.18546/LRE.14.3.02.
Favril L, Rich JD, Hard J et al (2024) Mental and physical health morbidity among people in prisons: an umbrella review. Lancet Public Health. 9(4):e250–e260 https://doi.org/10.1016/S2468-2667(24)00023-9.
Gibbon S, Khalifa NR, Cheung NH, et al (2020) Psychological interventions for antisocial personality disorder. Cochrane Database Systematic Reviews. Sep 3;9(9):CD007668 https://doi.org/10.1002/14651858.CD007668.pub3. PMID: 32880104; PMCID: PMC8094166.
Hassan L, Senior J, Webb RT, et al (2016) Prevalence and appropriateness of psychotropic medication prescribing in a nationally representative cross-sectional survey of male and female prisoners in England. BMC Psychiatry. 16(1):346 https://doi.org/10.1186/s12888-016-1055-7.
Herttua K, Crawford M, Paljarvi T et al (2022) Associations between antipsychotics and risk of violent crimes and suicidal behaviour in personality disorder. Evid Based Ment Health. 25(e1):e58-e64 https://doi.org/10.1136/ebmental-2022-300493.
NICE (2013) Antisocial personality disorder: Prevention and management [CG77], available online at https://www.nice.org.uk/guidance/cg77.
Tully J, Sethi A, Griem J, et al (2023) Oxytocin normalizes the implicit processing of fearful faces in psychopathy: a randomized crossover study using fMRI. Nat Ment Health. 1(6):420-427 https://doi.org/10.1038/s44220-023-00067-3.
Tully J, Hafferty J, Whiting D, et al (2024) Forensic mental health: envisioning a more empirical future. Lancet Psychiatry. Nov;11(11):934-942 https://doi.org/10.1016/S2215-0366(24)00164-0.
