Introduction

Attention has been drawn to the burden of poor health in vulnerable and marginalised social groups and the need for effective policies and services that will reduce social and health inequalities [1]. Prisoners are one such group. They often have a background of social and economic deprivation and early life adversity, factors that are associated with health problems and antisocial behaviour including in juveniles [2]. Deficits in self-control are central to many theories of crime. Neurobiological reward systems that are not well controlled can cause a failure to assess risk and behaviours that are associated with sensation seeking and rule breaking. This risk is heightened in adolescence where there is an imbalance between well developed reward systems that respond to emotion and self-regulatory systems that are cognitively mediated and that are continuing to mature [3–5].

Arguably, neurodevelopmental challenges or environmental adversity could exaggerate an imbalance of this kind in adolescence or prolong it into adulthood. Consistent with this, adolescents who offend are at greater risk of having neurodevelopmental, mental health and substance misuse problems with associated distress, depression, self-harm and PTSD, and are characterised by poorer neuropsychological test performance, poorer self-control, negative emotions and risk taking behaviour [2, 3]. These characteristics of juvenile offenders are common after head injury, as also is poor self-monitoring, poor judgement and a lack of empathy, and they can result in repeated risk taking, rule breaking, antisocial behaviour and violence [4]. Indeed, head injuries are more often found in offenders than in the general public [6] and there is growing concern about the long term impact of head injury in young offenders [7, 8]. It is known that the brain continues to mature into the mid 20s and the effects of damage earlier in life may become increasingly apparent in adolescence, when there are complex demands associated with social judgement. Head injury early in life is associated with poorer adult outcomes including in educational attainment and health [9]. More specifically head injury has the potential to affect normal brain development and have neurobehavioural effects that include poorer impulse control [7, 10]. Given this a causative link between head injury, crime and recidivism is plausible, particularly given the high prevalence of head injury in juvenile offenders [11] and indeed there is some evidence to support this [12, 13]. There is a need however, to account for the contributions of co-morbidities that are associated with deprived backgrounds and not only to investigate whether there is a history of head injury, but if the severity is likely to have caused brain damage and persisting disability. There is little published evidence in this area and none on disability, except in adult female prisoners [14, 15]. An understanding of the potential contribution of head injury to offending is important when developing services and determining what assessments and inputs are required and when [14, 16]. Hence, this study investigates relationships between a history of significant head injury (SHI) and other current and historical factors, and the outcomes of cognitive function, disability and offending. We use the term offender in this study for clarity and in accordance with current usage in the criminal justice and forensic mental health systems, and it is not intended to be pejorative.

Methods

The study took place at HM Young Offenders Institute (HMYOI) Polmont which is Scotland’s national holding facility for male young offenders aged 16–21 years. There are two other institutions in Scotland, both of which take adults and juveniles: HMP and YOI Grampian and HMP and YOI Cornton Vale. The last takes women only. As the number of juvenile females in prison in Scotland is small and was about 15 during the study period [17], the study only included males, given potential gender differences in causes of head injury and comorbidity [14]. The average number of juvenile males in prison in Scotland during the recruitment period was 310 of which 98% were resident in HMYOI Polmont [17]. Juvenile males were recruited from Polmont between January 17, 2019 and August 30, 2019. To be included, participants had to be aged 16 years or older, fluent in English, not have a severe acute disorder of cognition or communication and give informed consent. Parental consent was not a requirement of the Ethics Committee and was not obtained. They were recruited through word of mouth and posters placed in prison halls. To avoid bias towards recruiting participants with head injury, the project was advertised as a study on prisoner health. Those with interest were to ask a prison officer for an information sheet and if wishing to take part, a prison officer informed the researchers. Potential prisoner participants then met with a researcher (JM or HA) to find out more about the study and provide written consent. It was made clear to potential participants that they could withdraw from the study at any time and did not need to give an explanation. It was also made clear to them that taking part in the study would not affect any treatment they may be undergoing or their custodial sentence. If a researcher became aware of a significant health issue for a participant, this information was passed to prison health care staff with the participant’s consent. If a researcher became concerned about a participant’s risk to self or others this information would be passed on to the relevant prison officer who could signpost relevant agencies and if deemed necessary, this action would be taken without the participant’s consent.

Prisoners are allocated a personal officer who has a case management role and meets them regularly. They encourage prisoner engagement including in rehabilitation or reintegration initiatives, and update case records and reports. The personal officer for each recruited prisoner was asked to complete a subset of independent observer forms about the prisoner, and commented on level of functioning and behavioural/management challenges.

Procedure

Assessments took 1–2 hours with breaks if necessary and were carried out by a final year doctorate in clinical psychology trainee (JMcV) or by an experienced research worker (HA). Details of assessor training are given here [S1 File]. If the participant was fatigued or seemed unwell, as indicated by self-report or by observation of the assessor, the assessment was completed in a second session.

Results

The sample of n = 103 represents approximately a third of the juvenile male prison population in Scotland [13]. The average age of the sample was 19.0 years (range 16–22), with the majority aged 18 and over (84; 82%). Almost all (96/102; 94%) self-described as ethnically white. Four self-described as mixed-race or multi-race, one as Asian and one as Black. Three quarters (73/96; 76%) were from the two most socially deprived quintiles [19]. Scottish prison statistics for 2019–2020 [17] reveal very similar demographics to that of the sample for age (83% aged over 17 years) ethnicity (94% white) and social deprivation quintiles (45% most deprived quintile) suggesting that the sample was representative of the juvenile prison population (Table 1 and S2 File). More than half had received 1:1 support in school or special schooling (63/102; 62%). Almost all had truanted (94/98; 96%) with 74/98 (76%) doing so weekly and schooling being further disrupted by suspension or exclusion in the majority (84/96; 88%). Many were unemployed before imprisonment (44/102; 43%). Differences between SHI (n = 82) and NoSHI groups (n = 21) were slight and not statistically significant for age, ethnicity, deprivation, years of education or school attendance (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Baseline characteristics.

https://doi.org/10.1371/journal.pone.0287312.t001

Physical and mental health

Physical problems not associated with recent drug use were rarely reported (4/103; 4%). These were migraine (1/103), pain (2/103) and nausea and dizziness not associated with head injury (1/103).

On the HADS, scores for anxiety and depression were both significantly higher in the SHI than in the NoSHI group (Table 2) and clinical anxiety more common in the SHI group (33/82;40%) than in the NoSHI group (0/21; 0%); p = 0.0001). No one in the NoSHI group was clinically anxious or depressed (Table 2). PCL-5 scores were significantly higher in the SHI group but differences became marginally non-significant (p = 0.070) when scores were converted using criteria for PTSD diagnosis, although the magnitude of difference in proportions was large (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Mental health.

https://doi.org/10.1371/journal.pone.0287312.t002

Drug and alcohol misuse

Most participants self-reported a history of problematic alcohol or drug use (80/102; 78%). Although rates for both were higher in the SHI group, differences were statistically non-significant (alcohol OR = 2.5, 95% CI 0.8–8.8, p = 0.088; drugs OR = 2.9, 95% CI 0.96–9.0, p = 0.055; Table 3). Alcohol use was above the clinical cut-off on the AUDIT-C in 96/103 (93%) and drug use above the cut-off on the DAST in 43/103 (42%). Both were significantly more common in the SHI group (alcohol OR = 6.0, 95% CI 1.2–45.1, p = 0.030; drugs OR = 5.6, 95% CI 1.5–32.1, p = 0.006), although imprecisely estimated (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Alcohol and drug use.

https://doi.org/10.1371/journal.pone.0287312.t003

Disability

Disability that was attributed to SHI using the GODS was found in 11/82 (13%). Cause of disability was uncertain but possibly attributable to HI, in a further 1/82 (1%). Disability associated with SHI was moderate in 9/11 and severe in 2/11 (Table 4). Difficulties underpinning SHI-associated disability were most commonly mental health (10/11; 91%) and anger control (4/11; 36%). Two of the eleven juveniles who were disabled by SHI did not have disability from other causes.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Disability outcome on the Glasgow outcome at discharge scale.

https://doi.org/10.1371/journal.pone.0287312.t004

In a multivariable logistic regression model, juveniles with current clinical anxiety were significantly more likely to have SHI-attributed disability (vs good recovery) than those without (OR 12.2, 95% CI 2.2–67.7), although this was imprecisely estimated due to the small number with SHI-attributed disability. SHI attributed disability was not significantly associated with current clinical depression, current PTSD, history of CNS diagnosis, abuse as a child or when aged 16 or over or alcohol or drug abuse assessed on the AUDIT-C or DAST [S3 File].

Disability from any cause was found in about half of participants on the GODS (53/103; 52%). SHI was associated with any cause disability in a univariable model (OR 3.5, 95% CI 1.2–10.0). When other potential risk factors were included as a multivariable model, SHI became non-significant when adjusting for current (OR 1.4; 95% CI 0.4–4.3) or historical factors (OR 2.6; 95% CI 0.9–7.8). In the multivariable model, current clinical anxiety was associated with any cause disability (OR 8.7, 95% CI 2.8–27.3). Current PTSD or clinical depression, a history of abuse, CNS diagnosis or alcohol or drug abuse, while all having positive ORs, did not show statistically significant associations with any cause disability [S3 File].

Disability was most commonly reported as resulting from mental health problems (73/103; 71%), particularly anxiety (28/103; 27%) and low mood (12/103; 12%). Problems resulting from anger and temper control were also often reported as disabling (36/103, 35%). Psychotic (largely paranoid) symptoms were reported in the context of effects of drug misuse (5/103; 5%). Group differences were not significant for any of these [S2 File].

Offending

SHI and NoSHI groups did not differ in the number of convictions or longest length of sentence and the age at first offence was similar between groups (Table 5). There was a weak but statistically significant correlation between age at first SHI and age at first offence (r = 0.268; 95% CI 0.051–0.460; p = 0.016). Multivariable analysis indicated that a CNS diagnosis was marginally associated with having 1.9 times more convictions as those without (95% CI 1.0–3.7), while there was no evidence that SHI, any abuse, problematic alcohol or drug misuse, or current clinical anxiety, depression or PTSD was associated with number of convictions. Longest length of sentence or number of convictions was not associated with any current or historical factor in the multivariable analysis [S3 File].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Offending history.

https://doi.org/10.1371/journal.pone.0287312.t005

Differences between SHI and NoSHI groups by type of offence, including violence, were not statistically significant (Table 5), A history of violent offences was very common (86/103; 84%). Differences in the most serious violent offence did not differ between groups (p = 0.435) [S2 File]. Multivariable analysis indicated that self-report of problematic alcohol or drug misuse was a risk factor for violent offending (OR 6.2, 95% CI 1.8–21.4), whereas a history of SHI, abuse, CNS diagnosis or current clinical anxiety, depression, PTSD or alcohol/drug misuse showed no evidence of association [S3 File]. More in the SHI group had recorded prison incidents (49/61; 80%) than in the NoSHI group (9/16; 56%; p = 0.047).

Discussion

This cross-sectional study on one third of male juveniles in prison in Scotland, revealed a history of SHI in four out of five participants and in two thirds, mild head injury that had occurred repeatedly over lengthy periods of time. Despite the high prevalence, disability was associated with SHI only in one in eight, and cognitive test scores or characteristics of offending did not differ between groups with and without SHI. Those with SHI did more often have a history of other CNS diagnoses, clinical anxiety or depression and problematic alcohol or drug use. Of these factors, only clinical anxiety was associated with disability in the SHI group. There was an indication that the SHI group had greater difficulty with self-regulation and control of impulsive behaviour. This was evidenced by higher self-report scores on the DEX questionnaire and indications of behavioural expression of this in prison, where recorded prison incidents were more frequent in those with SHI. However, this effect is nuanced, given the absence of group differences in offending history, including violence.

Our sample was demographically similar to the juvenile prison population in Scotland suggesting it is representative. Juveniles in Scottish prisons most often have backgrounds of deprivation, disrupted schooling and unemployment or unskilled work and almost all identify as white [6, 17, 36]. As SHI and NoSHI groups did not differ on these factors, their backgrounds seem similar and group effects not likely to be artefacts of background differences.

Four in five juveniles in the present study had a history of SHI and this is much higher than the 30% reported in the meta-analysis by Farrer et al [11]. The disparity may partly be due to different definitions of head injury across studies. When studies on juvenile offenders define head injury as loss of consciousness and/or being dazed after head injury, prevalences are 67–72% and more consistent with the 80% found in the present study [37, 38]. When the definition is restricted to loss of consciousness, studies typically report a much lower prevalence (31–41%) [18, 39–41]. We believe it is important to include those with repeat head injury without loss of consciousness and symptomatic effects, given increasing evidence for cumulative and persisting effects of mild head injury [20, 35, 42].

Male juveniles with SHI in prison in Scotland differ from the general population with a history of head injury. Fighting and assault were by far the most common lifetime causes of SHI in juvenile prisoners (four in five), with a history of episodes of repeat head injury also typical. In contrast, a fall is the most common cause of hospitalised head injury in the general population and a history of single incident accidental head injury more typical [43, 44]. In the general population, head injury resulting from an assault is more likely in adolescents than in adults, although the prevalence is much lower than found in juvenile prisoners here [39, 40], with one review of eighteen studies indicating a median prevalence of 22% [45].

Studies on juvenile offenders [41, 46, 47] including our own, do not find poorer performance on cognitive tests in those with head injury. This may be because of widespread effects of deprivation and multiple morbidity on cognition in offenders [3, 40, 48]. Consistent with this, the overall sample reported here had significantly poorer scores on a test of mental flexibility than expected from general population norms [S2 File]. There is however evidence to suggest that cognitive function is poorer in adult male prisoners with head injury than in those without [49]. This could imply that juvenile offenders with SHI who have poorer cognitive function more often continue offending into adult life or that juvenile offenders with SHI who continue to offend, are at greater risk of becoming cognitively impaired. Follow-up work is needed to elucidate this, given that intervention strategies would differ by aiming either to reduce the impact of existing cognitive impairment or reduce risk of future SHI.

Juveniles in prison often have neurodevelopmental difficulties, problems associated with mental health and substance misuse and previous trauma [2, 50]. Indeed, this was the case in our sample. However, few other studies have looked at relationships between these factors and head injury in juveniles. Similar to our own findings, those that have, report a greater risk of mental health problems including anxiety, trauma and substance misuse in those with head injury [39, 41, 46, 51]. In our study, group differences in those fulfilling diagnostic criteria for PTSD were statistically non-significant, although overall PCL-5 scores were significantly higher in the SHI group. This may reflect sample size limitations or alternatively, greater vulnerability to PTSD later in life in those with SHI. The higher scores on the TLEQ in the SHI group may support the latter interpretation. Clearly though, the higher PCL-5 and HADS-anxiety scores indicate greater distress in the SHI group and this is often associated with prisoners being more difficult to manage [52], as found in the SHI group who more often had recorded prison incidents.

There is very little evidence about disability after head injury in offenders [14] and none previously on disability after head injury in juvenile prisoners [S1 File]. The prevalence of disability after SHI found here (13%) was low compared to the very high prevalence of SHI and the 35–40% prevalence of SHI-attributed disability found in adult prisoners with SHI in Scotland [15, 53]. This difference between juvenile and adult prisoners might be explained by a continuing risk of developing SHI-associated disability in juveniles who continue to offend into adulthood. Cumulative damage from repeated mild SHI is known to occur in non-offender populations [20, 35, 42] and is a risk for juvenile prisoners if released into the same environment that previously exposed them to SHI. Education about the effects of head injury and support in the community might reduce this risk. Knowledge about head injury and its impact on behaviour is poor in adult male prisoners, but can be improved by brief education [54].

Other CNS disorders were not associated with disability in the overall sample, or in the SHI group. Clearly, diagnosis of CNS disorder is not synonymous with persisting disability. However for some, disability may be situation specific, occurring when higher demands are placed on functions that are weak or impaired, such as communication or memory during stressful legal processes. These difficulties may not be evident in normal daily routines [55], and the absence of an association between other CNS disorders and disability assessed on the GODS might reflect this.

Evidence for violent crime as a distinguishing characteristic in juvenile offenders with head injury is not strong, with most studies reporting no group effects between offenders with and without head injury [39, 40, 41, 56, 57]. This seems surprising, given evidence for greater impulsivity and reactive aggression after head injury in the general population, including in studies on exposed and non-exposed juvenile twins [58] and because theories of offending emphasise aggressiveness and poor self-control as being criminogenic [4, 58]. A few studies report greater risk of severe violence in juvenile detainees with head injury [18, 40] and Kenny et al [40] note the potential for head injury and drugs to act synergistically and reduce self-control. However other studies, including our own, find no greater risk of severe violence (or violence) in juvenile prisoners with head injury [38, 41]. Differences in environment, deprivation and exposure to head injury might explain these inconsistencies across studies. Indeed, evidence for head injury as a risk factor for violent offending is stronger in adult offenders [13] where exposure to trauma, problematic alcohol/drug use and repeat head injury is likely to have been longer, increasing the risk of poorer self-control and impulsive behaviour. In contrast, it may be that juveniles differ by being at a transitional phase in maturation. Many juveniles desist from offending in their mid to late twenties, but some become life-course persistent offenders [3, 59]. This difference in outcome has been ascribed to developmental history, including schooling and family interactions and ongoing factors such as drug misuse, neuropsychological deficits, risk taking behaviour, antisocial attitudes, having peers who are offenders and unemployment [4, 59]. If so, juvenile prisoners may be at a ‘crossroads’ in their life-course, with those who have a history of SHI being at risk of further head injury and at greater risk of transition to persistent adult offending including crimes of severe violence. Our findings are consistent with this given that those with SHI have poorer self-control and higher levels of distress, but with group effects for violent offending not yet evident.

This study is limited by the absence of follow-up data which could further explore the hypothesis of juveniles with SHI being at a greater risk of recidivism. The high prevalence of SHI and problematic substance use contributed to multivariable models estimating some effects with low precision and correspondingly wide 95% CIs. The data on other CNS history relied on participants having knowledge of early life events or diagnoses. More generally, much data was collected by self-report introducing a possibility of error in recall. The Glasgow Outcome Scales have been extensively used in retrospective assessment of disability including in studies on late outcome [60], with work substantially based on single incident head injury. There might be greater risk of error in assessing disability when assessing change in function following repeated mild head injuries that occur over extended periods of time and that might have cumulative effects. Techniques to minimise potential sources of error were adopted [S1 File] and with respect to assessment of head injury, offenders do not always attend hospital when injured [61] and medical records may not reflect the high frequency of repeated, mild injuries found. Hence, self-report using a validated method is necessary. Compliance with the assessment was good. The assessment was completed in one session in about 90% of participants. The remainder required two sessions; around half of these were because of prison administrative reasons such as lock up, classes or lunch and half for participant reasons such as fatigue or restlessness. None who required a second session refused to attend.

As one in three juvenile prisoners in the UK are reconvicted within 12 months of release [62, 63] there is an imperative to offer intervention and support to reduce risks of recidivism and potentially of a transition to lifelong offending. This study indicates that juvenile prisoners with SHI show greater psychological distress and signs of poorer behavioural control and suggests they are at greater risk of further SHI, associated disability and continuing to offend. Arguably, therefore there is a need for holistic intervention that includes education about brain injury [12], support in the community and that takes account of the multiple morbidity found in this vulnerable group to facilitate lifestyle changes and provide opportunities for sustaining productive employment.

Supporting information