Autism spectrum disorder in prison and secure care

Dear Editor,

I read with interest, ‘Autism spectrum disorder and Irish prisoners’ (Moloney & Gulati, 2019). Similar to Ireland, the UK prevalence of autism spectrum disorder (ASD) within the prison estate remains poorly understood. Estimation relies largely on extrapolation from general populations (Robinson et al. 2012). Reasons for this have been considered (Chaplin & McCarthy, 2014) and cited as: lack of screening programme; lack of suitable screening assessment tool; limited use of standardised assessment tools (i.e. Autism Diagnostic Observation Schedule or Autism Diagnostic Interview) by in-reach teams and diagnostic difficulties owing to practical limits on attaining corroborating evidence. Furthermore, it has been highlighted, with reference to neurodevelopmental disorders, that prison mental healthcare and research has historically tended to focus on serious mental illness (Underwood et al. 2013).

Complexity, heterogeneity and co-morbidity are all to be borne firmly in mind in the identification of such patients and in adequately meeting their therapeutic and offending behaviour needs. A recent systematic review and meta-analysis examined prevalence of co-morbid psychiatric disorders in adults with ASD and identified attention-deficit hyperactivity disorder as the most common (Lugo-Marín et al. 2019). Co-morbidity with intellectual disability is well established with rates recorded varying from approximately one-fifth to four-fifths (Postorino et al. 2016). Such a wide range suggests issues around caseness, training and methodological variation. Mood disorders are also common and potentially inappropriately treated due to a reliance on evidence drawn from non-ASD populations (Matson & Williams, 2014). Atypical presentations and masking of affective symptoms need to be considered and there have been calls for research into targeted treatments for co-morbid mood disorders in ASD (Giovinazzo et al. 2013).

As people with ASD negotiate the criminal justice system, their heterogeneity is matched by their vulnerability. The features of ASD have been argued to have a significant bearing on a wide array of pre-trial, trial and sentencing issues (Freckelton, 2013; O’Sullivan, 2017). The literature on offending behaviour and ASD has historically suffered from an over-reliance on case reports and series. Indeed, methodological shortcomings within much of the wider literature on ASD and the criminal justice system have been highlighted (King & Murphy, 2014). Whilst it has been stressed no evidence indicated people with ASD had increased offending rates, evidential support is considered to be mounting for the view those with co-morbidities are at increased offending risk (Chaplin et al. 2013). Various offending behaviours have been associated with ASD (Mouridsen, 2012). Woodbury-Smith et al. (2006) identified a significantly increased likelihood of criminal damage-type offending. However, this study was limited by a relatively small community sample. Barkham et al. (2013) found an over-representation of stalking, arson and sexual offences when compared to non-ASD controls in a medium secure unit. They suggested the majority of ASD patients requiring secure care were likely to be high functioning. This area remains understudied, especially so in relation to female forensic populations with ASD. The heightened media attention and frequent associations made with cybercrime cases and lone actor terrorism could be argued to potentially distort clinicians’ expectations of how ASD may typically manifest in such settings. ASD patients undoubtedly represent a challenging cohort for forensic services. Their vulnerability combined with the limited availability of specialist training and forensic ASD units are prominent factors (Murphy, 2010).

Heterogeneous clinical presentations and offending patterns also present challenges for risk formulation in secure care. High risk factor variability (on HCR-20) was identified in, albeit, small samples of ASD patients placed in high secure care (Murphy, 2013). This could be argued to call into question the validity of applying traditional clinical risk assessment instruments to this cohort (Murphy, 2013). Potential barriers to comprehensive risk assessment would include failure to consider the individual’s cognitive, emotional and sensory profile and deficits. Murphy (2013) conceded it is unlikely a single risk assessment tool will adequately address all the relevant considerations. Gunasekaran (2012) had similarly cautioned against too narrow an interpretation of structured risk instruments. Ideally, an individual’s risk characteristics and specific ASD profile are to be formulated in tandem with established risk instruments with multidisciplinary input.

Research on their experiences in prison has been limited (Robertson & McGillivray, 2015). Distress, vulnerability, isolation and relational issues with staff have all been described. ASD patients in secure care in the UK have been shown to have longer stays (Hare et al. 1999; Haw et al. 2013). The limited coverage and availability of appropriate step-down services and community resources are likely factors. In terms of patient experiences of care and quality of life, this has only recently been examined in a high secure setting for the first time (Murphy & Mullens, 2017). Although using a small sample, this qualitative research demonstrated a means to understand experiences of this intensive level of secure care at an individual level. The authors emphasised the role of regular and robust surveys of views and satisfaction ratings to improve patient experiences and therapeutic outcomes.

Allely (2018) recently published a systematic review on ASD in secure care that examined prevalence, treatment, risk assessment and other clinical considerations. Only 12 suitable studies were identified, mostly from the UK. Regarding screening, they highlighted validity issues in applying the Autism Quotient (AQ) (Woodbury-Smith et al. 2005) in the forensic setting due to poor prisoner literacy skills, among other factors (Murphy, 2011). There have been similar reservations made about its utility in light of low specificity found in non-forensic adult populations (Ashwood et al. 2016).

Achieving an accurate prevalence estimation of ASD in prison remains elusive, but may prove crucial in planning and developing prison in-reach and secure services to meet Ireland’s needs. Improved training and awareness will help; however, there are arguments for a dedicated forensic screening instrument. Furthermore, the general constraints of prison mental health assessments coupled with the scope for subtlety and complexity render it difficult to reach an ASD diagnosis in this setting. Risk formulation in secure care is complicated by clinical heterogeneity, co-morbidity and limits on conventional clinical risk assessment instruments. Considering the clear difficulties in the UK in studying this population and the comparatively much smaller population of Ireland, there may be an argument for an additional focus on qualitative research to improve outcomes and care ratings for those with ASD treated within the Irish prison estate and progressing through its National Forensic Mental Health Service.