Effectiveness of psychological and psychosocial interventions for forensic mental health inpatients: A meta-analysis

Abstract

Psychological and psychosocial interventions have significant potential to treat the mental health and criminogenic needs of forensic mental health patients. However, due to a dearth of high-quality evaluation studies there is limited evidence on the effectiveness of these interventions with this population. This review aimed to quantitatively summarise the effectiveness of psychological therapies delivered in forensic hospitals. The literature was systematically searched to identify controlled evaluations of psychological interventions delivered within inpatient forensic psychiatric settings. Twenty-eight studies were included. Methodological quality was assessed using the SIGN Methodology Checklists. Pooled effect sizes were calculated for fourteen outcome domains. Small effect sizes were found favouring psychological treatment over the comparator condition in increasing insight into mental illness, ameliorating symptoms, improving problem-solving ability, reducing pro-criminal attitudes and improving ward behaviour. A medium effect size was found for treatment increasing patients' knowledge of their mental illness. There were few outcomes for which psychological therapy was associated with improvements beyond that of comparison treatment, and these improvements were generally small. Despite more frequent adoption of a randomised-controlled design, methodological quality remains problematic and more well-designed trials are needed to determine the effectiveness of psychological interventions across outcome domains relevant to forensic patients' recovery.

Introduction

Individuals detained in the care of forensic mental health services often present with entrenched, complex mental health problems alongside a range of offending behaviours. By working with forensic patients, forensic mental health practitioners aim to assess and treat patients' mental health needs and reduce their risk of future reoffending while promoting patients' personal recovery (Barnao & Ward, 2015; Vojt et al., 2011). Treatment in forensic hospitals is often long-term and comes with great financial and personal cost. In England, where secure mental health services account for nearly 20% of all public expenditure on adult mental healthcare (Durcan et al., 2011), the average length of stay in continuous secure care in a sample of high and medium secure patients was over 14 years (Völlm et al., 2018). Considering this significant investment of resources, there is a dearth of empirical evidence for the interventions delivered in forensic mental health settings, particularly psychological interventions (Barnao & Ward, 2015; Mallion, Tyler, & Miles, 2020). With few psychological or psychosocial interventions supported by research evidence in this population, the day-to-day treatment delivered is rarely evidence-based. A recent national survey of UK forensic hospitals found that a majority of the offence-specific group treatments delivered by services including substance misuse, firesetting, sexual offending or violent offending treatment programmes, were developed “in house” and only a minority had ever been subject to even a local service-evaluation of treatment effectiveness (Mallion, Tyler, & Miles, 2020).

Determining the effectiveness of interventions for forensic patients and services to deliver interventions found to be effective is crucial if healthcare providers are to remain committed to delivering evidence-based care. There have been many attempts to summarise and synthesise the research evidence for psychological treatment of forensic mental health patients and mentally disordered offenders (e.g. Duncan et al., 2006; MacInnes & Masino, 2019; Sturgeon et al., 2018). Most commonly this includes systematic reviews followed by narrative synthesis, rather than quantitative synthesis via meta-analysis. Authors of narrative reviews of the literature on psychological treatment for forensic patients reported their initial plans to undertake meta-analysis were thwarted by a high prevalence of small scale studies with lax methodological design (Duncan et al., 2006) and high levels of heterogeneity reflected in the treatments under evaluation and the outcomes being assessed (MacInnes & Masino, 2019).

A small number of meta-analyses with relevance to treatment in secure settings have been conducted (e.g. Martin et al., 2012; Papalia et al., 2019; Yoon et al., 2017) but their broad inclusion of studies on prisoners, forensic community outpatients as well as inpatients reduces the generalisability of review findings to patients in long-term forensic inpatient care. Firstly, factors such as the therapeutic skill-set of the intervention facilitator (e.g. clinician vs. correctional officer) and practical issues arising from the treatment environment may impact the effectiveness of treatments delivered in hospital compared to prison or community settings (for example, see Taylor et al., 2020). Secondly, aside from the obvious differences in violence risk between those detained in institutions and those residing in the community, many distinctions can be made between mentally disordered offenders (MDOs) in prisons and secure hospitals (Thomas et al., 2009) which could conceivably lead to differential effectiveness of the same treatment. For example, Thomas et al. (2009) compared prisoners attending the prison mental health clinic to a sample of forensic inpatients. Forensic inpatients were more likely to have a diagnosis of a psychotic disorder, a history of drug misuse, and a history of previous psychiatric admissions than prisoners, while prisoners tend to have more unmet psychiatric and daily living needs. Other research has linked forensic patients' problems with impulsivity and behavioural control (Cullen et al., 2011) and ongoing psychotic symptoms and cognitive impairment to treatment dropout from interventions originally developed for use in correctional settings. A widely studied cognitive skills programme is Reasoning and Rehabilitation (R&R). There is a substantial evidence base supporting the effectiveness of R&R in reducing recidivism among violent offenders in community and institutional settings across the world (Tong & Farrington, 2006). However, there are striking examples of implementation failures of R&R in forensic psychiatric settings. For example, ⁎Cullen et al., 2012a, ⁎Cullen et al., 2012b found only 50% of forensic patients who started the original 36-session R&R completed treatment. R&R was subject to a significant revision to increase its responsivity to forensic patients' needs, including a reduction in the number of sessions from 36 to 16, the addition of a module on cognitive impairments, and the introduction of individual mentoring between group sessions (Young et al., 2010). Yet, trialling the new “R&R2M”, Young et al. (2010) reported 19% of patients who were referred to the group refused to attend, another 22% of patients who initially agreed subsequently failed to start the programme and a further 21% of patients started the R&R2M but terminated treatment prematurely. For the above reasons reviews which adopt broad inclusion criteria but do not undertake moderator analyses exploring the potential for differential treatment effectiveness dependent on the institutional setting or other study characteristics may be of little relevance for practitioners seeking to identify and introduce effective psychological interventions in forensic hospitals.

According to systematic reviews the most effective psychological treatments in secure hospitals appear to be those with an evidence-base in general psychiatry (Dumont et al., 2018). Narrative syntheses (MacInnes & Masino, 2019; Sturgeon et al., 2018; Tapp et al., 2013) have found that psychoeducational programmes lead to improved insight and understanding of one's mental disorder. Cognitive behavioural therapy (CBT), largely considered the gold standard treatment for schizophrenia, depression and anxiety, as well as personality disorder (Hofmann et al., 2012) also reduces forensic patients' psychiatric symptoms. A range of therapies have shown promise in reducing violence risk and aggressive behaviour, including cognitive skills programmes (notwithstanding the implementation problems already noted), dialectical behavioural therapy (DBT) and CBT (MacInnes & Masino, 2019; Sturgeon et al., 2018). Despite the range of reviews undertaken to summarise this literature and a surge in the number of publications of controlled evaluations of psychological interventions which focus on this patient group, to date there has been no meta-analysis of the effectiveness of psychological interventions in addressing forensic mental health inpatients' needs arising from their mental illness and offending behaviours. A meta-analysis would provide a necessary foundation for objective consideration of psychological treatment effectiveness and would facilitate comparisons based on effectiveness and cost-effectiveness with other treatment modalities targeting similar outcomes, including psychotropic medication. This review therefore addressed this gap by conducting a meta-analytic review of controlled evaluations of psychological and psychosocial treatments delivered in forensic mental health hospitals.