Utility of the Brain Injury Screening

Index in Identifying Female Prisoners

with a Traumatic Brain Injury and

Associated Cognitive Impairment

Michelle O’ Sullivan

Submitted for the Degree of

Doctor of Psychology(Clinical Psychology)

School of PsychologyFaculty of Arts and Human Sciences

University of SurreyGuildford, SurreyUnited Kingdom

October 2015

1

Abstract

An estimated 60.25% of offenders have a history of traumatic brain injury (TBI).

There is currently no established valid or reliable screening tool for identifying

female prisoners with a TBI and associated cognitive impairment available in the

UK. Using a cross-sectional design, this study aimed to investigate the retest

reliability and construct validity of the Brain Injury Screening Index (BISI).

Convergent validity was explored using self-report measures of mood and

neurodisability, as well as a battery of neuropsychological assessments of

cognitive functioning. Of a planned sample of 73 participants, preliminary data

from 23 participants has been analysed. 69.56% of participants were identified as

having a history of TBI, with a mean of 2.09 TBIs. Intraclass correlation

coefficients reached statistical significance for six of 10 identified key clinical

indicators on the BISI. The BISI variables did not reach statistically significant

convergence with most of the test battery. Two of the four BISI summary

variables demonstrated correlations in the hypothesised directions across the full

assessment battery, however only one BISI variable reached statistical

significance with one subscale in the battery. Analyses provide support for

further investigation into the construct validity and retest reliability of the BISI

with a larger sample. The implications of these findings, particularly in refining

the BISI, and future research and practice are considered.

2

Acknowledgements

Thank you cannot begin to express the gratitude I feel towards the copious amounts of

people who got me to the end of this three years, but it’s probably not a bad place to

start.

Firstly, I would like to thank each person who I have had the pleasure to work with

clinically and took the time to participate in my research. I have witnessed so much

strength and determination in the face of adversity in each of you. I count myself

honoured to have had the chance to learn from you and be part of your stories of

recovery.

I have had the opportunity to learn from some of the most amazing clinical supervisors

over the last three years. A few in particular have shaped my personal and professional

identity so much. Dr Chris Hall taught me patience (albeit a lesson I am still learning!)

and how the most powerful respect is commanded with a quiet confidence. Dr Julie

Nixon taught me the clinical power of personality; how authenticity is one of our most

valuable assets; and how supporting people to be the best they can be extends beyond

our clients, to our team and communities. Dr Oliver Sindall taught me that it is alright to

make mistakes, and that vulnerability is not only not a bad thing, but an asset in the

right context. Dr Manveer Kaur taught me self-compassion and rebuilt my sense of

confidence and competence when my self-belief as a clinical psychologist was at its

lowest. You have all been so much more than supervisors to me and given me so much

to aspire to.

I would like to thank the team at the Disabilities Trust: my field supervisor Prof Mike

Oddy, Dr Sara da Silva Ramos, and Deborah Fortescue, for their unwavering support

throughout my Major Research Project. They enabled me to turn a small idea into a

3

rewarding project, that despite three years of hard work and many pitfalls, I remain

passionate about. Ideas are easy to come by, it is the instigators and facilitators that are

invaluable. I would like to thank my University research supervisors Dr Emily Glorney

and Prof Annette Sterr for their support throughout the three years. I am particularly

indebted to Emily who continued to supervise the project after she left the University.

To have such continued support and commitment during the three years meant the

world to me. I would like to thank Dr Laura Simonds, who continued to act as my

research tutor and a link to the clinical team, even when the role no longer existed. Not

only did she provide me with containment and guidance at times of difficulty, but she

modelled for me what I can only hope to provide to future clinical trainees in my career.

I also owe so much thanks to Prof Chris Fife-Schaw who never seemed phased by the

barrage of stats questions I sent his way. I would like to thank all the team at the

participating prison, particularly all the prison officers who made the day to day running

of the project possible – I would still be wandering the corridors looking for a room

right now if it was not for them. I would also like to extend my appreciation to Elaine

Cameron and Sarah Disspain, who took the risk of supporting the project and enabling

its setup, and gave so much time to thinking about how the research design could fit the

prison regime, and balancing the ethics and the practicalities of the procedures.

I would like to thank my clinical tutor, Louise Deacon, and Honorary Clinical Tutor,

Fiona Goodwin. You have been breaths of reflective fresh air amidst administrative

smog.

Without my family I would never have made it this far. You instilled in me core values

of education, compassion for others, and responsibility. The bedtime stories my father

always read me gave me the power of language and knowledge. Witnessing the

4

influence of my mothers’ warmth on others made me believe in the power of kindness.

Finally, Cohort 41, you turned England from just a place I was living, to a home. You

have given me some of the best memories of my life, made my belly ache with laughter,

and brought me back to life with hugs. You are all wonderful clinical psychologists, so

much so I’m expecting to get an awful bill in the post any day now.

5

Contents

MRP Empirical Paper (including Abstract) p7

MRP Empirical Paper Appendices p77

MRP Proposal (without Appendices) p203

MRP Literature Review (with Appendices) p215

Clinical Experience Précis p264

Summary of Assessments Table p268

6

Utility of the Brain Injury Screening Index in Identifying Female

Prisoners with a Traumatic Brain Injury and Associated Cognitive

Impairment

An estimated 60.25% of offenders have a history of traumatic brain injury (TBI).

There is currently no established valid or reliable screening tool for identifying

female prisoners with a TBI and associated cognitive impairment available in the

UK. Using a cross-sectional design, this study aimed to investigate the retest

reliability and construct validity of the Brain Injury Screening Index (BISI).

Convergent validity was explored using self-report measures of mood and

neurodisability, as well as a battery of neuropsychological assessments of

cognitive functioning. Of a planned sample of 73 participants, preliminary data

from 23 participants has been analysed. 69.56% of participants were identified as

having a history of TBI, with a mean of 2.09 TBIs. Intraclass correlation

coefficients reached statistical significance for six of 10 identified key clinical

indicators on the BISI. The BISI variables did not reach statistically significant

convergence with most of the test battery. Two of the four BISI summary

variables demonstrated correlations in the hypothesised directions across the full

assessment battery, however only one BISI variable reached statistical

significance with one subscale in the battery. Analyses provide support for

further investigation into the construct validity and retest reliability of the BISI

with a larger sample. The implications of these findings, particularly in refining

the BISI, and future research and practice are considered.

Keywords: traumatic brain injury; offenders; prisoners; screening; reliability;

validity

Introduction

Traumatic Brain Injury

Traumatic brain injury (TBI) is defined as “an alteration in brain function, or other

evidence of brain pathology, caused by an external force” (Menon et al., 2010),

capturing the range of presentations which fit under the TBI diagnostic umbrella,

including loss of or decreased consciousness, any loss of memory, neurological deficits,

and any alteration in mental state e.g. confusion (Menon et al., 2010). TBI is the most

common form of acquired brain injury (ABI, see Figure 1; Fleminger & Ponsford,

2005), with an estimated prevalence of 8.5% (Silver et al., 2001) across all levels of

severity; however, rates may overlook milder TBI due to reliance on medical records

(Tennant, 2005) and associated diagnostic and selection biases (Feigin et al., 2013).

Figure 1. Types of Brain Injury

TBI severity traditionally has been classified by scores on the Glasgow Coma Scale

(GCS; World Health Organization, 2006). Other commonly used measures include post-

traumatic amnesia (PTA) and length of loss of consciousness (LOC; Sherer et al.,

2008). Table 1 summarises typical cut-offs used for differentiating mild, moderate and

severe TBI. Absence of a uniform classification system makes comparison across

studies difficult (Corrigan et al., 2010).

Most TBIs are mild (Donnelly et al., 2011). Reports of problematic sequelae following

mild TBI (mTBI) range from only 10% (Albicini & McKinlay, 2014) to 42% (Konrad

et al., 2011). Not all blows to the head result in a TBI, and not all TBIs result in

functional difficulties. However, research suggests that multiple mTBIs can have a

8

Brain injury

Congenital brain injury

Acquired brain injury

Traumatic Brain Injury

Closed head injury

Open head injury

Non Traumatic

Brain Injury

cumulative effect, with similar cognitive and behavioural profiles to more severe

TBI(Collins, Grindel, Lovell, & et al., 1999; Diamond et al., 2007; Miller et al., 2013).

While moderate to severe TBIs tend to be self-evident, deficits from mTBIs can be

easily overlooked (Donnelly et al., 2011).

Table 1. TBI classifications systems

GCS score (World

Health Organization,

2006)

PTA (Lezak,

2004)

LOC (Greenwald et al.,

2003)

Mild TBI 13-15 < 1 hour <30 minutes

Moderate TBI 9-12 1-24 hours ≥ 30 minutes ≤ 6 hours

Severe TBI 3-8 > 24 hours > 6 hours

Up to twice the rate of TBI has been found in males than females in the general

population (Hillbom & Holm, 1986; Hirtz et al., 2007). The apparent protective effect

of female gender appears attenuated in specific populations, including those with

substance use disorder (Felde et al., 2006) and prisoners (Brain Injury Association of

Wyoming, 2008; Ferguson et al., 2012). TBI risk in females appears more pronounced

with mTBI, e.g. Diamond, Harzke, Magalett, Cummins and Frankowski (2007) found

that 54.7% of females in a prison population self-reported TBI with no LOC in

comparison to 40% of males. This fell to 35.6% of females in comparison to 47.8% of

males with LOC of less than one hour. Differences may be attributable to different

gender-related behavioural patterns, e.g. decreased likelihood of reporting mTBIs. TBI

in females may be underestimated due to unreported domestic violence (Valera &

Berenbaum, 2003). Due to the nature of domestic violence it is difficult to get an

accurate prevalence rate of TBI, with rates varying from 30-74% (Kwako et al., 2011).

Females with a TBI are more likely to report elevated psychological distress, past year

9

suicidal ideation, school bullying, drinking, and smoking (Colantonio et al., 2007; Ilie et

al., 2014). Meta-analysis has found that females tend to have worse outcomes after TBI

than males across all TBI types (Farace & Alves, 2000).

TBI in Offenders

A recent meta-analysis (Shiroma et al., 2010a) places TBI prevalence in offender

populations at 60.25% (95% CI: 48-72%), with a male and female prevalence estimate

of 64.41% (95% CI: 53.30-75.53%) and 69.98% (95% CI: 50.18-89.79%) respectively.

Once TBI definition was limited to LOC, excluding milder TBIs, males demonstrated a

higher prevalence than females (59.31% vs. 55.28%). The decrease in prevalence once

limited by LOC supports the prevalence of mTBIs in females. Along with a higher

prevalence than in the general population, prisoners are at higher risk of neurodisability

following TBI, by virtue of reduced cognitive reserve from exposure to factors such

substance use and mental health difficulties (Ropacki & Elias, 2003).

TBI is a complex condition that can result in an array of cognitive, emotional, physical

and behavioural sequelae. The Swedish population registers’ research found that

individuals with TBI have a significantly increased risk of committing a violent crime

(Fazel et al., 2011). Even mTBI in childhood is associated with an array of long-term

negative outcomes, including increased risk of arrest, violent offences, and property

offences (McKinlay, 2014). McKinlay et al. (2014b) found TBI severity and age of

injury to be significant predictors of later offending behaviour; with substance use

mediating the relationship between childhood TBI and later offending (McKinlay,

Corrigan, Horwood, & Fergusson, 2014a). The Repairing Shattered Lives report

emphasised the need for research examining causes and consequences of TBI in female

offenders specifically (Williams, 2012).

10

The field of social neuroscience has begun marrying the previously segregated

understandings of brain development and social behaviour. Damasio’s somatic marker

hypothesis (1994) provides insight into offending behaviour by taking a systems-level

neurological approach with the relationship between motivational and affective

processes, and the coordination of complex reasoning and action. For instance,

individuals with ventromedial prefrontal cortex (vmPFC) injuries demonstrate missing

anticipatory skin conductance responses, which leads to difficulties using the emotional

consequences of past experiences to guide behaviour (Taber-Thomas & Tranel, 2012).

Somatic markers that sit in the vmPFC are essential for normal moral cognition under

conditions of ambiguity, uncertainty and conflict, providing a representation of how the

future self would feel after performing an action (Taber-Thomas & Tranel, 2012). Loss

of somatic markers can manifest itself behaviourally in lack of concern for others,

socially inappropriate behaviour, reduced guilt and shame, and increased aggression

(Anderson et al., 1999; Beer, John, Scabini, & Knight, 2006; Damasio, Tranel, &

Damasio, 1990). The consequences of somatic marker loss may be more pronounced

amongst offenders when contextualised within sociodemographic factors, such as living

in environments with high rates of antisocial behaviour which could increase the

frequency with which individuals are faced with of uncertainty and conflict that are

demanding of moral cognition.

Yeates et al.’s (2012) integrative model of the social outcomes of childhood brain

disorder provides a holistic understanding of the role of social cognition, affect

regulation, factors relating to the neurological insult, and environmental factors, on

social competence. For instance, deficits in social information processing, such as

mentalisation, cue interpretation and outcome evaluation, contribute to poor social

adjustment (Parker, Rubin, Erath, Wojslawowicz, & Buskirk, 2006; Rubin, Bukowski,

11

& Parker, 2006). Consequent antisocial behaviour may arise from choosing

instrumental over prosocial goals; misinterpretation of others’ intent; and generation of

fewer effective responses to social dilemmas (Yeates et al., 2012). Yeates et al. (2012)

emphasise how this can contribute to a negative developmental spiral with social

interactions and relationships, resulting in chronic social problems even if social

information processing improves following TBI, thereby possibly perpetuating

antisocial lifestyles.Without early intervention, individuals with TBI may pose a risk to

others, becoming involved in the criminal justice system (Brower & Price, 2001;

deSouza, 2003; Greve et al., 2001; Kreutzer et al., 1995; Kreutzer et al., 1991; Miller,

1999; Simpson et al., 1999). Once there, they may be more difficult to rehabilitate and

discharge, with services ill-equipped to address their needs. Hawley and Maden’s

(2003) study of TBI in medium secure units indicated that 41.6% of service users had a

history of TBI, and were significantly more difficult to discharge into the community

due to perceived greater risk of violence to others and self-harm. Research

demonstrating increased disciplinary incidents in prisoners with TBI (Merbitz et al.,

1995; Morrell et al., 1998; Shiroma et al., 2010b) suggests they may have increased

difficulty adapting to prison life due to cognitive and behavioural sequelae. For

instance, the somatic marker hypothesis (Damasio, 1994) indicates the vmPFC’s

involvement in the representation and monitoring of reward, highlighting potential

ineffectiveness of traditional behaviour management through contingencies which form

the backbone of the criminal justice system. This has implications for engagement in the

legal process, prison management, and post-discharge and release pathways (Jackson &

Hardy, 2011). Due to inadequate screening and identification of TBI, services are

unable to provide adapted rehabilitation. Under-identification perpetuates inadequate

resources, providing no incentive to fund appropriate interventions. Research in this

12

field is congruent with the Transforming Rehabilitation strategic business priorities

within the UK National Offender Management Service (NOMS; National Offender

Management Service, 2013) and may increase efficiency by informing programs to

reduce recidivism.

Screening for TBI

A valid TBI screen for female offenders is required to facilitate research and clinical

practice, enabling researchers to determine the prevalence of TBI in UK female

offenders, which is currently unknown. Screening is consistent with NOM’s Reducing

Prison Unit Costs strategic business priority by ensuring appropriate services are

commissioned and target the most appropriate offenders (NOMS, 2013), e.g. screening

can provide a cost-effective way of determining who is appropriate for referral to

limited and expensive resources such as neuropsychologists. A valid measure would

enable prisons to add TBI screening to established intake risk assessment procedures.

A clinically useful screen should capture frequency and severity of lifetime TBI, that

can help differentiate uncomplicated head injury from TBI (Diamond et al., 2007). One

difficulty with developing screens is whether the screen purpose is to assess for

exposure to any head injury, which reduces specificity and risks overburdening low

resourced services; or TBI with associated functional sequelae, which is difficult to

capture using self-report. TBI screens in military populations have acknowledged this

conflict, prioritising high sensitivity, compromising on specificity (Donnelly et al.,

2011); however with under-resourced prisons this compromise is less tenable.

There are currently no validated and reliable published screening tools for use with UK

female offenders. Research has begun to support the use of The Brain Injury Screening

Index (BISI; Pitman, Haddlesey, Ramos, Oddy, & Fortescue, 2014) with UK male

13

offenders, but has yet to be extended to females. Much research relies on instruments

developed for individual studies, with little consideration of reliability and validity, or

one to two item methods which underestimate TBI history (Diamond et al., 2007). Two

published screening tools have psychometric data available for use with prisoners,

including females; the Traumatic Brain Injury Questionnaire (TBIQ; Diamond et al.,

2007) and the Ohio State University TBI Identification Method (OSU TBI-ID; Bogner

& Corrigan, 2009). These measures have not been validated within a female UK prison

population. Indices on the OSU TBI-ID which require an estimate of mTBIs, relating to

episodes such as domestic violence, are unreliable (Bogner & Corrigan, 2009).

Therefore, this may be inappropriate for female prison populations. The TBIQ has a

similar layout to the BISI, however includes a symptoms checklist. The TBIQ has only

been validated against short rating scales, with research demonstrating that self-

assessment of cognitive functioning is not representative of cognitive functioning as

measured by objective neuropsychological assessment tools (French, Lange, & Brickell,

2014; Spencer, Drag, Walker, & Bieliauskas, 2010). Reduced insight is well

documented post-TBI (Ham et al., 2013). Self-report neurobehavioural measures

address this using of informant scales to corroborate data (Kreutzer, Marwitz, Seel, &

Devany Serio, 1996), however feasibility of informant scales with offenders is an issue,

e.g. informant scales are usually designed to be completed by family who are not

witness to the daily neurobehavioural presentation of prisoners. Self-report

neuropsychological sequelae tend to correlate highly with psychiatric rather than

cognitive constructs (Spencer et al., 2010). Discriminating between TBI sequelae and

other disorders has proven difficult, e.g., of the 100 symptoms identified in the Brain

Injury Screening Questionnaire, 79 are symptoms of affective or thought disorders

(Walker, Cole, Logan, & Corrigan, 2007). TBI severity and frequency is likely to yield

14

more reliable and valid indicators of TBI sequelae than self-report symptomology

(Donnelly et al., 2011). Screens should have enough questions to facilitate recall of past

TBIs (Diamond et al., 2007), whilst not over-including symptomatic items which have

poor discriminant value. Diamond et al.’s (2007) research indicates that interview

administered screens perform better than self-report checklists.

There is no agreed upon gold standard for the assessment of TBI sequelae by which to

compare screens. Clinical interview is frequently used as the gold standard criterion,

however this is an imperfect method (Albicini & McKinlay, 2014; Donnelly et al.,

2011). Neuropsychological assessments, an objective measure of cognitive functioning,

are valuable additions to post-TBI functioning assessments; however these are rarely

used in assessing the reliability and validity of TBI screens. The gold standard of

clinical interview needs to be bolstered with psychometric measures to evaluate the

most useful screens (Donnelly et al., 2011), particularly when medical records which

would be the strongest criterion measure of TBI, tend to be unavailable in offender

populations.

Aims and Hypotheses

Aims

(1) Explore the reliability and construct validity of the BISI for screening for history

of TBI.

Hypotheses

(1) Presence of TBI; the TBI Severity Index; the TBI Time Index; and the BISI

Global Score will be significantly positively correlated with scores obtained on

the self-report measures of mood and neurodisability.

15

(2) Presence of TBI; the TBI Severity Index; the TBI Time Index; and the BISI

Global Score will be significantly negatively correlated with neuropsychological

measures of cognitive functioning.

(3) The BISI will have statistically significant test-retest reliability intraclass

correlation coefficients (ICC) for all continuous variables examined.

(4) The BISI will have statistically significant test-retest reliability Kappa

coefficients for all binary variables examined.

Methods

Design

A quantitative correlational cross-sectional design using a semi-structured clinical

interview, clinical questionnaires, and neuropsychological measures, was employed.

Participants

The study was conducted at a UK closed women’s training prison, with an operational

capacity of 282. Participants were recruited from new prison receptions. A sample size

calculation for a correlational analysis based on the number of variables being

measured, an anticipated large effect size (Pitman et al., 2014), statistical power of .80

and type I error α of .0004 (Bonferroni correction for multiple comparisons), indicated

that 73 participants were required. Pitman et al.’s (2014) r of -.45 with an N of 189

yielded confidence intervals (CIs) of -.56 to -.33, therefore it was anticipated that for a

similar effect size, an N of 73 would yield CIs of -.62 to -.24. Charter (1999) warns of

such risk of imprecision in the estimation of r in reliability and validity studies with

small Ns, particularly under 400; however due to feasibility related to resource

limitations, and the preliminary nature of this research, a small N was deemed an

appropriate and necessary limitation. Inclusion criteria included prisoners over 18 years

16

of age, with an upper age limit of 80 in line with norms provided in the instruments

used. Exclusion criteria included acute symptoms of physical or mental illness.

Prisoners with a confirmed diagnosis of dyslexia, problems with literacy, inadequate

English fluency, or acquired a TBI in the last six months, were excluded due to measure

validity limitations. Participants with a learning disability (LD) were included unless

queries regarding capacity to consent were raised. Fifty-five prisoners were approached,

eleven of which were ineligible (Figure 2). Twenty-three of the remaining 44

participated, providing a response rate of 52%. Of the 23 participants, 16 self-reported a

TBI.

Participants ranged from 21 to 64 years of age (M = 38.74, SD = 12.73). Premorbid IQ

ranged from 66.50 to 112.30 (M = 88.74, SD = 11.57), while obtained IQ on the

Wechsler Abbreviated Scale of Intelligence-II (WASI-II) ranged from 67 to 126 (M =

95.00, SD = 15.23). The majority of participants identified as White British (69.60%).

Number of years spent in education ranged from 2 to 18 (M = 11.73, SD = 3.57).

Appendix II provides full sample demographics.

No significant correlations were found between the key BISI summary variables and

premorbid IQ, age, educational background, TOMM score, and alcohol use (Table 2).

No participant scored under the cut-off of 45 on the TOMM.

Figure 2. Recruitment flow

17

Table 2. Correlations between sample demographics & BISI summary variables

Correlations (point biserial) with TBI presence (N=23)

Variable Mean (SD) r Sig

Age 38.74 (12.73) -.17a .422

Premorbid IQ

(TOPF raw score)c

33.57 (12.73) .25 a .248

Years in education 11.73 (3.57) .26 a .221

Units of alcohol

consumed daily

26.66 (33.91) -.09 a .680

TOMM Trial 2 49.65 (1.07) .30b .161

Correlation with TBI Global Score (N=23)

Age

Premorbid IQ (TOPF raw score)

Years in education

Units of alcohol consumed daily

.39 a .063

-.11 a .614

-.36 a .083

.24 a .264

-.01 b .956

18

Assessed for eligibility

Ineligibile (n=11)

Acutely unwell (n=3)

Moving prison during assessment

period (n=3)

Insufficient English fluency

(n=4)

Problems with literacy (n=1)

Eligible (n=44)Consented to participating

(n=23)

Self-reported TBI (n=16)

No reported TBI (n=7)

TOMM Trial 2

Correlation with TBI Time Index (n=16)

Age

Premorbid IQ (TOPF raw score)

Years in education

Units of alcohol consumed daily

TOMM Trial 2

.55 b .027

.44 b .084

-.10 b .701

-.10 b .713

.26 b .321

Correlations with TBI Severity Index (n=16)

Age

Premorbid IQ (TOPF raw score)

Years in education

Units of alcohol consumed daily

TOMM Trial 2

.61 a .012

.31 a .234

-.19 a .478

.39 a .126

.26 b .321

aPearson correlation coefficientbSpearman’s rhocFor ease of interpretation M and SD for premorbid IQ once converted to standardised scores are 91.77 (9.21).*Significant at the level of <.0004 based on the Bonferroni correction for multiple comparisons

Measures

The assessment battery used in Pitman et al.’s (2014) study examining TBI in male

offenders was adapted following reflections from the project. Pro formas are only

included for measures freely available due to copyright. Further scale properties are

presented in Appendix III.

Semi-Structured Interview

The semi-structured interview was designed to ascertain history of TBI, offending,

19

mental health, and social history, in prisoners (Appendix IV).

The Brain Injury Screening Inventory (BISI)

The BISI (Appendix V; Pitman et al., 2014) is an eleven item questionnaire designed by

the Disabilities Trust, a UK based charity supporting people with a range of disabilities.

The BISI is designed to enable practitioners of all levels, ranging from health and social

care staff, probation, police, and housing support teams, to screen for TBI using self-

report data. It takes about 10 minutes to complete. Items were constructed on the basis

of items used in previous prevalence studies (Hwang et al., 2008; Williams et al., 2010),

and examines how many times an individual has suffered a serious blow to the head,

and consequent disorientation, PTA and LOC. It includes questions on other potential

ABIs, cognitive sequelae, and neurodevelopmental problems. The BISI provides

qualitative screening data, but efforts have been made to quantify the results using a

TBI Severity Index which is calculated by multiplying the highest rate of

unconsciousness, rated on a 0-3 likert scale, by the number of TBIs (Pitman et al.,

2014). The Disabilities Trust considered use of a TBI Time Index, which is LOC in

minutes multiplied by number of TBIs. The Disabilities Trust also developed a TBI

Global Score based on responses on the first reported TBI, the formula for which is

presented in Appendix V. The BISI Global Score provides an indicator of clinical need,

mapping severity to appropriate treatment pathways. Table 3 summarises the different

scoring systems for the TBI Global Score and TBI Severity Index. There is no

interpretation system developed for the Time Index. There is no published data

comparing these scoring systems. The BISI has been used in research on TBI in a UK

population of homeless people (Oddy et al., 2012) and male prisoners (Pitman et al.,

2014). Pitman et al.’s (2014) research demonstrated preliminary support for the validity

of the BISI with male offenders, with presence of TBI on the BISI and TBI Severity

20

Index correlating with performance on neuropsychological measures. There is no

published data on the reliability of the BISI.

Table 3. Scoring and interpretation of the BISI

The BISI Time Index The BISI Global Score

Severity 1-10: Mild TBI <5: no clinical need

11-30: Moderate TBI 5-8: signposting to healthcare

31-60: severe TBI 9: routine cognitive assessment

61-300: very severe TBI ≥10: routine cognitive assessment

≥300: extremely severe TBI

Beck Depression Inventory-II (BDI-II)

The BDI-II is a 21 item self-report measure of severity of depressive symptoms (Beck

et al., 1996). Items are rated from 0-3. A total score of 0-13 indicates minimal

depression; 14-19 mild; 20-28 moderate; and 29-63 severe depression. The BDI-II has

demonstrated strong validity and reliability across a range of populations, including

offenders (Wang & Gorenstein, 2013). Cronbach’s alpha for the current sample was .91,

indicating excellent internal reliability.

Beck Anxiety Inventory (BAI)

The BAI (Beck et al., 1988) is a 21 item self-report measure of severity of symptoms of

anxiety. Items are rated from 0-3. A total score of 0-7 indicates minimal anxiety; 8-25

mild; 16-25 moderate; and 26-63 severe anxiety. Beck et al. (1988) reported adequate

reliability and validity. Cronbach’s alpha for the current sample was .88, indicating

good internal reliability.

21

The Impact of Events Scale – Revised (IES-R)

The IES-R (Weiss & Marmar, 1997) is a well validated and reliable 22 item self-report

measure of symptoms of PTSD, asking the participant to reference a specific traumatic

event when responding to questions (Appendix VI). Items are rated from 0-4. A total

score over 33 indicates the possible presence of posttraumatic stress disorder (PTSD).

Cronbach’s alpha for the current sample was .91, indicating excellent internal reliability.

The Neurobehavioral Functioning Inventory (NFI)

The NFI (Kreutzer et al., 1999) is a 76 item self-report inventory composed of six

subscales measuring frequency of common difficulties after acquiring a neurodisability,

namely: depression, somatic, memory/attention, communication, aggression, and motor.

Items are rated from 1-5, with higher scores indicating higher levels of neurodisability.

While an informant version of the scale was available, it was only possible to complete

the self-report in this study, as an informant with an appropriate knowledge of the

participants’ functioning would not be available. The NFI has demonstrated adequate

reliability and validity for use in TBI (Kreutzer et al., 1996; Kreutzer et al., 1999).

Cronbach’s alphas for the current sample for the depression, somatic, memory/attention,

communication, aggression and motor subscales are .93, .80, .94, .90, .81, and .82

respectively, indicating good to excellent internal reliability.

The Dysexecutive Questionnaire (DEX)

The DEX (Wilson, Evans, Emslie, Alderman, & Burgess, 1998) is a 20 item self-report

questionnaire assessing frequency of executive-type behavioural problems. Items are

rated from 0-4, with higher scores indicating higher frequency of problems. While an

informant version of the scale was available, it was only possible to complete the self-

report in this study. The DEX has demonstrated adequate reliability and validity for use

22

in TBI (Bennett, Ong, & Ponsford, 2005). Cronbach’s alpha for the current sample

was .95, indicating excellent internal reliability.

The Test of Premorbid Intellectual FunctioningUK (TOPF)

The TOPF (Wechsler, 2009) provides a reliable and valid means of estimating the

premorbid cognitive functioning levels of adults suspected of suffering from intellectual

deterioration. It is designed on the premise that general intelligence is highly correlated

with reading ability, which is relatively resistant to neuropsychological insult (Carlozzi

et al., 2011). It consists of 70 written words which must be read aloud and scored by

pronunciation. The premorbid IQ score can be adjusted for sex and years in education. It

is normed with the Wechsler Adult Intelligence Scale IVUK (WAIS IV; Wechsler, 2008).

The Wechsler Abbreviated Scale of Intelligence II (WASI-II)

The WASI-II (Wechsler & Zhou, 2011) is a reliable and valid abbreviated measure of

cognitive intelligence for individuals aged 6-90 years. It is adapted from the four highest

loading WAIS-IV subtests on g (general intelligence); namely, Vocabulary and

Similarities which constitute the Verbal Comprehension Index (VCI), and Block Design

and Matrix Reasoning which constitute the Perceptual Reasoning Index (PRI). Index

scores are adjusted for age and have a mean of 100 (SD = 15).

The Repeatable Battery for the Assessment of Neuropsychological Status

(RBANS)

The RBANS (Randolph, 1998) is a reliable and valid brief cognitive battery for the

assessment of neuropsychological functioning for individuals aged 12-90 years.

Consisting of twelve subtests (list learning, story memory, figure copy, line orientation,

picture naming, semantic fluency, digit span, coding, list recall, list recognition, story

23

recall, and figure recall), it provides measures of immediate and delayed memory,

visuoconstruction and visuoperception ability, attention, language, as well as a global

summary score. Index scores are adjusted for age and have a mean of 100 (SD = 15).

The Behavioural Assessment of Dysexecutive Syndrome (BADS)

The BADS (Wilson et al., 1996) is designed to be a reliable and valid assessment of

executive functioning for individuals aged 16-87 years. Consisting of six subtests

(temporal judgement, rule shift cards, action program, key search, zoo map, and

modified six elements), the tasks are meant to reflect real life situations. Each task is

scored on a scale of 0-4, which are summed to create a total profile score. Profile scores

can be converted to age adjusted standard scores with a mean of 100 (SD = 15).

The Test of Memory Malingering (TOMM)

The TOMM (Tombaugh, 1996) is a 50 item visual recognition test designed to detect

poor cognitive effort in neuropsychological testing in adults. It consists of two learning

trials and an optional retention trial if a participant scores below the cut-off of 45. In the

learning trials the participants are shown 50 line drawn targets for three seconds. The

test phase is forced choice and each target item is paired with a distractor, requiring the

respondent to indicate the previously shown target. Scores range from 0-50, with five or

more errors on trial two indicating low effort. It has demonstrated its reliability and

validity particularly for use with TBI (Tombaugh, 1997).

Procedure

Participants were recruited through a convenience sample. A list of new receptions from

September 29th 2014 to February 9th 2015 was obtained from the prison. A poster

(Appendix VII) advertising the study was hung in the intake room and reception wing,

24

along with expression of interest slips which could be handed to any prison officer and

returned to the prison psychology office. Recruitment was ongoing during the data

collection period, with one to two participants recruited each week up to three weeks in

advance of assessment. Only two participants returned expression of interest slips. The

remaining participants were recruited in consecutive chronological order by individually

approaching them. Prison officers acted as gatekeepers and were asked if potential

participants met any of the exclusion criteria prior to approaching them. Prisoners in the

healthcare wing were not approached due to exclusion criteria. Copies of the participant

information sheet (Appendix VIII) and consent form (Appendix IX) were dropped into

eligible participants’ cells. A week later prisoners were followed-up face-to-face to

discuss participation and consent. After signing the consent form, a suitable time to

commence assessment was arranged. Participants were allocated a participant number,

which was recorded with their prison number, and stored in a file managed by the

prison psychology team.

Assessment was conducted in a quiet room with a large table, and took place over two

sessions on different days, ideally a week apart unless scheduling conflicts did not

allow. Days between Part One and Part Two of the assessment ranged from three to 42

(M = 12.91, SD = 11.40). During Part One, participants completed the BISI and clinical

interview, taking approximately one hour. They then completed the BDI-II, BAI, and

IES-R. The neuropsychology battery took two to three hours and was administered in

the following order: the TOMM, the TOPF, the RBANS, the WASI-II, the BADS, the

DEX, and the NFI. Participants could choose the Part One end point, to manage fatigue.

Most participants stopped after the RBANS, with a few continuing to the WASI-II.

Participants could request a feedback session at the end of the assessment. Psychometric

data were scored immediately after testing. Feedback was given verbally within one

25

month, at which point they could decide if they wanted clinically significant results to

be passed on to healthcare (see Appendix X for healthcare feedback pro forma). Data

were also extracted from The Offender Assessment System (OASys), a structured

clinical needs assessment tool used throughout NOMS (Home Office, 2006). The

OASys provided brief offence and abuse history (see Appendix XI for list of items

extracted) and was accessed from participant files after the assessment was complete.

Ethics

This study was reviewed and granted favourable ethical opinion by the NOMS National

Research Committee and by the Faculty of Arts & Human Sciences Ethics Committee

at the University of Surrey (see Appendix XII for correspondence with confirmation of

ethical approval). Issues of consent, anonymity and data protection were adhered to

throughout the study according to the requirements set by the above ethical panels.

Data Analysis

While a large dataset was collected, for the purpose of this MRP only data related

directly to these research questions will be discussed. All analyses were done using IBM

SPSS version 20 (IBM, 2011). There were no missing data, as data pertaining to these

research questions came from the interview and neuropsychological assessment. Data

preparation included checking responses, calculating total scores, and assessing

normality of distribution (see Appendix XIII for histograms and z scores). If z scores

were significantly higher than zero (z > 1.96, p <.05) then data were considered to be

abnormally distributed (Field, 2013), in which case non-parametric equivalents of tests

were used where appropriate.

Construct validity of the BISI was explored by examining its convergent validity with a

range of neuropsychological tests. Correlation coefficients were calculated for key BISI

26

summary variables (presence of TBI; the TBI Time Index; the TBI Severity Index; and

the TBI Global Score), the neuropsychological measures, and standardised

questionnaires.

Retest reliability was assessed for key BISI variables, namely: presence of TBI; total

number of TBIs; age of first TBI; presence of any episode of PTA; total number of

episodes of PTA; longest period of reported LOC; TBI Severity Index; TBI Time Index;

and TBI Global Score. For the continuous variables, ICCs using a two-way fixed effect

model for agreement (Rankin & Stokes, 1998), Pearson correlation coefficients, and

Spearman correlation coefficients where appropriate, explored retest-reliability across

the two time points. For the nominal variables, Cohen’s kappa (Cohen, 1960) and Phi

coefficients (Cramer, 1946) assessed retest reliability.

Results

History of TBI

Of the total sample, 69.56% reported a history of any TBI, while 60.86% reported a

history of TBI with LOC. Number of reported TBIs ranged from 0 to 6 (M = 2.09, SD =

1.97). Age of first TBI ranged from 4 to 28 (M = 15.25, SD = 6.85), while age of most

serious TBI ranged from 5 to 33 (M = 19.44, SD = 7.51). Table 4 outlines the reported

causes of the TBIs.

Table 4. Cause of TBIs

27

Time since first TBI ranged from 3 to 50 years (M = 24.94, SD = 18.10), and time since

most recent TBI from 1 to 33 years (M = 8.75, SD = 9.05). While 87.50% of

participants who experienced a TBI reported at least one LOC episode, only 35.41% of

TBIs involved LOC. Most severe LOC reported was over six hours for 18.80% of

participants, between ten minutes and six hours for 35%, and under ten minutes for

43.80%. In 45.65% of cases of TBI, participants did not seek or come to the attention of

medical or professional assistance (Table 5).

A TBI Time Index, consisting of LOC in minutes multiplied by number of TBIs, ranged

between 0 and 11,520 (M = 1,589, SD = 3,522). The TBI Severity Index, consisting of a

likert scale based LOC severity rating multiplied by number of TBIs, ranged from 1 to

12 (M = 4.56, SD = 3.22). The TBI Global Score, which consists of a sum of responses

related to first TBI reported, ranged from 0 to 14 (M = 4.91, SD = 4.36).

28

Cause %

Road traffic accident 17.39

Sporting accident 10.86

Fights 21.73

Intimate partner violence 15.21

Falls (on substances) 13.04

Other (not crime related) 8.69

Falls (sober) 10.86

Childhood abuse 2.17

Table 5. Treatment sought after TBI

Treatment %

Hospital 41.30

Nothing 45.65

Taken to prison 2.17

Paramedic 2.17

Prison/police healthcare 6.52

GP 2.17

Construct Validity

Key summary variables on the BISI were tested for convergence with

neuropsychological measures of cognitive functioning and standardised self-report

questionnaires of mood and neurodisability.

Contrary to hypotheses, the TBI Severity Index demonstrated small to medium

predominantly negative correlations with the self-report mood and cognitive

questionnaires. It was hypothesised that the TBI Severity Index would be negatively

correlated with the neuropsychological measures, however the directions of the

correlations were inconsistent. Across all measures, none of the variables reached

statistical significance (Table 6). Overall, results indicate that as the TBI Severity Index

increased, symptomology on mood and neuropsychological measures did not

consistently increase as expected.

29

Table 6. Correlations between TBI Severity Index (M=4.56 SD=3.22) and variables

from the Neuropsychological Battery (n=16)

Variable Mean (SD) r Sig

Standardised self-report mood and neurodisability questionnaires

BAI 17.56 (11.12) -.15 a .556

BDI-II 24.08 (12.68) -.33 a .211

DEX total score 29.91 (19.05) -.28 a .289

IES-R 45.69 (20.62) -.33 a .208

NFI Depression 40.04 (11.57) -.46 a .067

NFI Somatic 29.73 (8.28) -.39 a .129

NFI Memory 50.30 (17.21) -.09 a .714

NFI Communication 25.95 (8.81) -.14 a .599

NFI Aggression 18.43 (5.78) -.38 a .144

NFI Motor 18.86 (5.45) -.04 a .879

Neuropsychological assessments of cognitive functioning

RBANS immediate memory 90.95 (16.36) .03 a .904

RBANS visuospatial 87.52 (15.03) .11 a .681

RBANS language 95.00 (14.49) .07 a .796

RBANS attention 85.22 (16.66) -.04b .874

RBANS Delayed Memory 93.13 (13.94) .15 b .563

RBANS total score 86.78 (14.54) .02 a .928

BADS Total Score 16.43 (4.13) -.24 b .367

WASI verbal comprehension 99.65 (13.71) .02 a .915

WASI perceptual reasoning 91.13 (15.39) -.17 a .519

WASI full scale 95.00 (15.23) -.08 a .754

WASI-TOPF 6.25 (12.65) -.28 a .292

30

aPearsons’ zero order correlationbSpearman’s rho*Significant at the level of p≤.0004 based on the Bonferroni correction for multiple comparisons

While the TBI Global Score demonstrated small to large correlations in the expected

direction with the self-report mood and neurodisability questionnaires, only the NFI

Motor subscale reached statistical significance (Table 7). The TBI Global Score only

demonstrated small correlations in the expected direction with the neuropsychological

measures of cognitive functioning, and none reached statistical significance (Tables 7).

Overall, results suggest that as the TBI Global Score increased symptomology on mood

and neuropsychological measures increased, however as these were predominantly not

within levels of statistical significance, with such a small sample these results are likely

to be unstable and are interpreted with caution.

Table 7. Correlations between TBI Global Score (M=4.91 SD=4.36) and

Neuropsychological Measures (N=23)

Variable r Sig

Standardised self-report mood and neurodisability questionnaires

BAI .44a .034

BDI-II .29 a .173

DEX total score .37 a .077

IES-R .21 a .328

NFI Depression .25 a .238

NFI Somatic .53 a .009

NFI Memory .58 a .003

NFI Communication .62 a .002

31

NFI Aggression .26 a .220

NFI Motor .68 a <.0004*

Neuropsychological assessments of cognitive functioning

RBANS immediate memory -.27 a .209

RBANS visuospatial -.17 a .417

RBANS language .01 a .933

RBANS attention -.28b .167

RBANS Delayed Memory -.23b .279

RBANS total score -.19 a .382

BADS Total Score -.24 b .255

WASI verbal comprehension -.22 a .300

WASI perceptual reasoning -.25 a .250

WASI full scale -.25 a .240

WASI-TOPF -.07 a .739

aPearson zero order correlationbSpearman’s rho*Significant at the level of p≤.0004 based on the Bonferroni correction for multiple comparisons

Using point biserial correlations, a reported history of TBI on the BISI was correlated

with variables on the neuropsychological test battery. None reached statistical

significance. Similarly to the BISI Global Score, while examination of the correlation

coefficients suggests that the results may be moving in the hypothesised direction, with

some medium and large coefficients (Table 8), lack of statistical significance and a

small sample indicate that these coefficients are likely to be unstable. The strongest

relationships were with the self-report measures.

32

Table 8. Point Biserial Correlations between Presence of TBI and Variables from the

Assessment Battery (N=23)

Variables Sig r

Standardised self-report mood and neurodisability questionnaires

BAI .053 0.40a

BDI-II .284 0.23 a

DEX total score .313 0.21 a

IES-R .172 0.29 a

NFI Depression .228 0.26 a

NFI Somatic .014 0.50 a

NFI Memory .097 0.35 a

NFI Communication .144 0.31 a

NFI Aggression .038 0.43 a

NFI Motor .002 0.70 a

Neuropsychological assessments of cognitive functioning

WASI verbal comprehension .432 -0.17

a

WASI perceptual reasoning .731 -0.07

a

WASI full scale .584 -0.12

a

WASI-TOPF .527 0.13 a

RBANS attention .325 -.21b

RBANS Delayed Memory .474 -.15 b

RBANS immediate memory .496 -0.14

a

RBANS visuospatial .209 -0.27

33

a

RBANS language .231 0.25 a

RBANS total score .661 -0.09

a

BADS Total Score .602 -.11 b

*Significant at the level of p≤.0004 based on the Bonferroni correction for multiple comparisonsaPearson correlation coefficientsbSpearman correlation coefficients

Spearman’s rho was used to assess the correlation between the TBI Time Index and the

neuropsychological battery (Table 9). Contrary to our hypotheses, the TBI Time Index

demonstrated small to medium predominantly negative correlations with the self-report

mood and cognitive questionnaires, as well as neuropsychological measures, none of

which reached statistical significance. Overall, results indicate that as the TBI Time

Index increased, symptomology on mood and neuropsychological measures did not

consistently increase as expected.

Table 9. Correlations between TBI Time Index and Neuropsychological Measures

(n=16)

Variable r Sig

Standardised self-report mood and neurodisability questionnaires

BAI -.018 .947

BDI-II -.187 .487

DEX total score -.294 .269

IES-R -.324 .221

NFI Depression -.496 .051

NFI Somatic -.345 .191

34

NFI Memory -.122 .653

NFI Communication -.170 .528

NFI Aggression -.336 .204

NFI Motor -.038 .888

Neuropsychological assessments of cognitive functioning

RBANS immediate memory -.017 .952

RBANS visuospatial .161 .553

RBANS attention -.037 .892

RBANS delayed memory .271 .310

RBANS language .000 1.000

RBANS total score .058 .830

BADS total score -.218 .418

WASI verbal comprehension -.076 .779

WASI perceptual reasoning -.210 .434

WASI-II full scale -.082 .762

WASI-TOPF -.426 .100

*Significant at the level of p≤.0004 based on the Bonferroni correction for multiple comparisons

Table 10 describes differences between variables tested in this study and Pitman et al.’s

(2014) study, on which our hypotheses were based. Although there was some

consistency in correlation size and direction across studies for Presence of TBI, as

discussed previously these correlations are not statistically significant and likely to be

unstable given the small sample. Table 10 also highlights possible discrepancies in

findings for the TBI Severity Index across the two populations.

35

Table 10. Correlation coefficients and for variables of interest against TBI Presence and TBI Severity Index in the present study and Pitman et

al.’s (2014) study

TBI Presence TBI Severity Index

Correlation coefficients

in this studyc

Correlation coefficients in

Pitman et al.’s sample (2014)c

Correlation

coefficients in this

study

Correlation coefficients in

Pitman et al.’s sample (2014)

NFI Depression .26 .46* -.46 a .49* a

NFI Somatic .50 .46* -.39 a .54* a

NFI Memory .35 .49* -.09 a .51* a

NFI Communication .31 .38* -.14 a .43* a

NFI Aggression .43 .37* -.38 a .40* a

NFI Motor .70 .41* -.04 a .46* a

DEX .21 .40* -.28 a .45* a

BAI .40 .41* -.15 a .47* a

BDI .23 .31* -.33 a .38* a

36

RBANS -.09 -.45* .02 a -.42* a

BADS -.11 -.32* -.24b -.28* a

WASI -.12 -.27* -.08 a -.25* a

WASI-TOPF .13 -.24* -.28 a -.20* a

*Statistically significant at respective p with Bonferroni correctionsaPearson correlation coefficientsbSpearman’s rhocPoint biserial correlation

37

BISI Test-Retest Reliability

Five of the eight continuous variables were consistent with hypothesis three,

demonstrating statistical significance (Table 11). ICC values for six of the eight

continuous variables on the BISI were over .60 (Table 11), meeting minimum criteria

for acceptable retest reliability (Anastasi, 1998; Baumgartner & Chung, 2001; Chinn,

1991), however Total no. of episodes of PTA did not reach statistical significance which

is reflected in its wide confidence intervals. The most reliable variables were Total

Number of TBIs, the TBI Global Score, and the TBI Severity Index, which all had

large positive Pearson coefficients as well as ICCs over .80 (p <.0004), indicating

excellent retest reliability (Landis & Koch, 1977). Total number of episodes of

disorientation, age at first TBI and total number of episodes of PTA had moderate ICC

coefficients and large Pearson coefficients. Longest recorded LOC and the TBI Time

Index did not demonstrate adequate ICC coefficients, despite large Pearson coefficients.

Table 11. Test-Retest Reliability for Key Continuous Variables

Variable Mean (SD) R Sig ICC 95%

CI

Time 1 Time 2

Age at first TBI

(n=16)

15.31 (6.77) 16.94

(7.79)

.814a <.0004

*

.796 .519-.

923

Total no. of episodes

of PTA (n=16)

.69 (.70) .5 (.63) .626b .002 .658 .276-.

864

Total no. of episodes

of disorientation

(n=16)

2.81 (1.37) 2.38 (1.14) .844b <.0004

*

.758 .410-.

910

38

Total no. of TBI

(N=23)

2.04 (1.89) 1.78 (1.56) .989a <.0004

*

.943 .853-.

977

Global TBI Score

(N=23)

4.91 (4.36) 4.65 (4.26) .853a <.0004

*

.857 .693-.

937

Longest LOC (n=16) 479.43

(1003.95)

219.12

(485.48)

.743b .011 .533 .096-.

804

TBI Time Index

Score (n=16)

1589.34

(3522.67)

617.84

(1548.84)

.706b .006 .574 .154-.

824

TBI Severity Index

(n=16)

4.56 (3.22) 4 (2.73) .852b <.0004

*

.830 .588-.

936

aPearson correlation coefficientbSpearman’s rho *Significant at the level of p≤.0004 based on the Bonferroni correction for multiple comparisons

Consistent with hypothesis four, Presence of TBI reached statistical significance (Table

12) and had excellent retest reliability with both Phi and Kappa coefficients of 1

(Landis & Koch, 1977). However, Presence of PTA was inconsistent with our

hypothesis. Although it demonstrated moderate to excellent retest reliability across both

Phi and Kappa (Table 12) it failed to reach statistical significance. Analyses were not

possible for Other ABI as every participant responded negatively to this item.

Table 13 compares reliability coefficients across the BISI, TBIQ and OSU TBI-ID for

variables that were designed to capture the same data. The BISI demonstrated the

highest reliability across three of the four variables.

39

Table 12. Test-Retest Reliability for Key Dichotomous Variables

Variable Consistent

(%)

Inconsistent

(%)

Phi Sig Kappa Sig

Presence of

TBI (N=23)

23 (100%) 0 (0%) 1 <.0004

*

1 <.0004*

Presence of

PTA (n=16)

12 (75%) 4 (25%) .524 .036 .740 <.0004*

Other ABI

(N=23)

23 (100%) 0 (0%) Not

computed

as variable

was

constant

- - -

*Significant at the level of p≤.0004 based on the Bonferroni correction for multiple comparisons

Table 13. Comparison of retest reliability across measures

BISI OSU TBI-ID TBIQ

Presence of TBI 1a N/A 0.56a

No. of TBIs .98b; .94c .87c 0.90b

Age at first TBI .81b; .79c .67c N/A

Longest LOC .74d; .53c .91c N/A

aKappa coefficientbPearson coefficientcICCdSpearman’s rho

Discussion

This study’s results provide some preliminary support for the utility of the BISI in

40

identifying female prisoners with a TBI history and associated cognitive impairment in

the UK, demonstrating the value in further investigation of its use in this population.

History of TBI

The BISI and clinical interview revealed that 69.50% of participants had a history of

any TBI, and 60.86% of TBI with LOC, which is consistent with the female prevalence

of any TBI of 69.98% and 59.31% when restricted to LOC, reported in Shiroma et al.’s

(2010a) TBI in offenders meta-analysis. This female sample experienced slightly fewer

TBIs with LOC than in Pitman et al.’s (2014) UK male sample (87.50% vs 94.20%),

with only 35.41% of TBIs resulting in LOC, and more reporting not seeking help after

the injury (45.65% vs 31.00%). Across both studies, the TBI Severity Index

demonstrated similar means (M=4.06 SD=4.50 in the male sample; M=4.91 SD=4.36 in

the female sample), but with a greater range in the male study (0-18 vs 0-14). The

female sample had a younger mean age of first TBI (15.25 years vs. 17.71 years). These

findings support evidence of gender specific epidemiological pathways in TBI (O'

Sullivan, Glorney, Sterr, Oddy, & Da Silva Ramos, 2015), indicating that while female

offenders have a similar TBI prevalence to males, females experience milder TBIs and

are less likely to seek help. The most frequently reported causes of TBIs were fights,

road traffic accidents, and domestic violence. No studies available studies compare

mechanism of injury across gender (O' Sullivan et al., 2015).

Construct Validity

Results suggest that further investigation of the construct validity of the BISI is

required. Only one of the BISI’s four summary variables demonstrated statistical

significance with one subscale within the assessment battery. Two of the variables

demonstrated correlation coefficients in the hypothesised direction with the battery of

41

neuropsychological measures of cognitive functioning and self-report questionnaires of

mood and neurodisability, however these coefficients are likely to be unstable due to the

small sample. This is the only TBI screening tool in an offender population that has had

its convergent validity investigated against a battery of neuropsychological tests. Our

results emphasise the need to further explore such measures’ convergence with clinical

psychometric assessments.

The TBI Global Score and TBI Presence variables demonstrated correlations in the

expected direction across the measures, however only the NFI Motor subscale reached

statistical significance with the TBI Global Score. There is no prior published research

on the TBI Global Score. Similarly, comparing the scores of those with and without a

self-identified TBI history, the self-report questionnaires demonstrated the strongest

relationship as opposed to the cognitive measures. This mirrors results found in the

male study (Pitman et al., 2014), with the largest effect sizes seen across both studies

being self-report.

Contrary to the hypotheses, both TBI Severity Index and Time Index demonstrated

small to medium correlations predominantly in the opposite direction for the self-report

measures, and in inconsistent directions for the cognitive tests. It is possible that the

TBI Severity Index and Time Index are invalid clinical indicators in this population due

to gender differences in TBI presentation, specifically difficulties in recalling periods of

LOC. Albicini and McKinlay (2014) emphasise the problem with validity that relying

on self-report LOC causes for diagnosis, e.g. individuals confusing PTA with LOC,

which as memory gap is subjectively experienced as the same. Without reliable

corroborating reports, using LOC as an indicator is likely to be misleading. Considering

most women in this study did not seek medical help, corroborating reports are unlikely

42

to exist. Donnelly et al. (2011) found that a binary variable asking whether participants

had experienced LOC in a military population had clinical value, with high specificity

(94%) but lower sensitivity (54%).

Across both studies, the strongest convergence has been with self-report measures of

neurodisability and mood rather than the objective cognitive assessments, highlighting

the complex relationship between mood and subjective, as well as objective, cognitive

functioning (Chamelian & Feinstein, 2006). Chamelian and Feinstein (2006)

demonstrated when mood is controlled for in TBI, subjective cognitive difficulties no

longer predicts most objective cognitive difficulties, with psychological factors

influencing objective recovery. This may be particularly relevant for TBI rehabilitation

considering females report higher levels of somatic depression in particular (Silverstein,

2002). While this may be an artefact of the gender response bias hypothesis (Sigmon et

al., 2005), examining means across self-report measures of mood and cognitive

functioning between this study and Pitman et al.’s (2014) study, the female group did

not consistently report greater pathology across measures, e.g. scores on the BDI-II are

higher across both TBI and non-TBI in the female population, but lower on the NFI

Depression subscale. The only variables which did not demonstrate results in the

hypothesised direction with the TBI Global Score and TBI Presence variables were the

WASI-TOPF and the RBANS Language Index. Premorbid IQ was lower than obtained

IQ across the whole sample, as well as for the TBI and non-TBI groups; with the TOPF

appearing to underestimate premorbid IQ particularly in the TBI group. Although verbal

comprehension was not particularly low in this sample, it is possible that there is lower

literacy proficiency specifically, which would lead to lower scores on the TOPF.

It is notable that on the BISI item which assessed other history of ABI or neurological

43

conditions, no participant provided a positive response. This is unusual, with research

demonstrating frequent comorbidity between TBI and other neurological conditions

(Bazarian, Cernak, Noble-Haeusslein, Potolicchio, & Temkin, 2009), with offenders at

particular risk of hypoxic brain injuries from substance use (Jackson & Hardy, 2011). It

is possible that this item is not a valid indicator of other neurological insults in this

population.

BISI Test-Retest Reliability

These results are the first to provide preliminary support for the retest reliability of the

BISI, with ICCs ranging from .53 to .94 on continuous variables, and Kappa

coefficients ranging from .740 to 1 on binary variables. Six of the 10 variables met

statistical significance as hypothesised. Eight of the 10 variables met minimum criteria

for adequate retest reliability, however two of these did not reach statistical significance

and had wide confidence intervals. Four variables demonstrated excellent retest

reliability. Since the reliability and validity of the OSU TBI-ID has been explored and

refined to its strongest 16 indices, its reliability coefficients range from .63 to .91

(Bogner & Corrigan, 2009). The TBIQ (Diamond et al., 2007) only provides reliability

coefficients for lifetime prevalence of TBI and frequency of TBI. Comparing these

screens, across variables designed to capture the same data, the BISI demonstrated the

highest reliability across three of the four variables. Differences may be attributable to

sample differences: the OSU TBI-ID and TBIQ are American tools; length of retest

period, with the TBIQ reporting approximately 2 to 4 weeks between testing, the OSU

TBI-ID reporting 1 to 2 weeks, while this study had a mean of 12.91 days; or

differences in question phrasing. Phrasing appears to be particularly an issue for the

BISI’s longest LOC item, which asks participants to state length of LOC rather than

providing categories as in the OSU TBI-ID, with poor recall leading to high variability

44

in responses, which is likely to be a factor for those with a TBI history. This is mirrored

in the TBI Time Index which is derived from length of LOC in minutes – this item

could not be reliable if length of LOC was unreliable. In comparison, when LOC was

converted into an ordinal scale for the TBI Severity Index, as in the OSU TBI-ID, the

TBI Severity Index remained reliable. Exploring the reliability of the OSU TBI-ID,

Bogner and Corrigan (2009) also found that items requiring estimation of LOC in

particular had lower reliability.

Comparing types of reliability coefficients both across studies, and within this study,

highlights the variability in statistics used. Pearson coefficients are frequently reported

to demonstrate retest reliability for continuous variables, however as Pearson measures

the strength of linear association rather than agreement, it is possible to have a high

correlation when agreement is low (Rankin & Stokes, 1998). Pearson coefficients do

not take systematic differences into account (Streiner & Norman, 2003). Quality criteria

guidelines specify that because systematic differences are part of measurement error, an

ICC two way models are the most appropriate retest reliability parameter for

continuous measures (Terwee et al., 2007). Pearson coefficients’ overestimation of

reliability is evidenced on a number of variables in this study, e.g. longest LOC, which

has a large Pearson coefficient but only a moderate ICC. The large Pearson coefficient

suggests it is reliable, however the ICC indicates that the frequency of disagreement

would make it unfit for its purpose of providing an estimate of TBI severity. Similarly,

for nominal variables, such as Presence of TBI, Kappa is considered the optimal retest

reliability parameter, as opposed to Phi which is the nominal equivalent of Pearson

(Mokkink et al., 2010; Rankin & Stokes, 1998).

45

Implications for Research and Practice

Providing the first psychometric data on the reliability and validity of a screening tool

for use in UK female offenders, this study has implications for both research and

practice, and is consistent with recommendations of the Repairing Shattered Lives

report (Williams, 2012). Our study demonstrates the value of further investigation into

the utility of the BISI for screening female offenders for TBI and associated cognitive

impairment, which has implications for the provision of specialist services both in

prison and upon release. Although this study was not designed to establish prevalence

specifically, the BISI produced TBI prevalence rates similar to those established in the

literature, for both with and without LOC (Shiroma et al., 2010a), supporting its

sensitivity to TBI in this population. The BISI demonstrated the strongest retest

reliability of comparable screens. However, two of its ten clinical variables, the TBI

Time Index and Longest LOC, which are closely interrelated, did demonstrate

inadequate retest reliability, and a further two did not meet statistical significance which

was reflected in wide confidence intervals. Longest LOC could be rephrased to capture

LOC range, however due to poor validity of self-report LOC, particularly in mTBI,

LOC range may not contribute sufficient clinical value to a screen, and may be best

removed. Similarly, the TBI Time Index demonstrated insufficient retest reliability and

may best be removed.

With regards to data analysis this study demonstrates the value of ICC and Kappa

coefficients over Pearson coefficients; and Phi, Pearson’s binary equivalent, for

establishing retest reliability. Our results highlight how Pearson and Phi coefficients can

substantially overestimate reliability by neglecting to take systematic differences into

account (Streiner & Norman, 2003).

46

Although most of the BISI summary variables did not reach statistically significant

convergence with the assessment battery, results suggest that further data collection is

warranted. No other TBI screen for use with offenders has demonstrated convergent

validity with a battery of neuropsychological measures. Self-report presence of TBI on

the BISI and the TBI Global Score both demonstrated correlations in the hypothesised

directions, although not reaching statistical significance.; Results provide support for

further investigation of the BISI’s utility in identifying those at risk of TBI sequelae,

and extension of Pitman et al.’s (2014) research into the construct validity of the BISI to

a female offender population. This study is the first to explore the new clinical indicator,

the TBI Global Score. As the TBI Global Score is based only on responses to the first

TBI, the BISI instructions should clarify if the first TBI to be reported should be the

first or the most severe; at present it is assumed the first reported will be the most

severe. It could also be considered if obtaining detailed information on other TBIs

makes any further meaningful contribution at a screening level, or if the TBI Global

Score is robust enough to inform clinical pathways.

The relationship between self-identified TBI history and the self-report measures across

both male (Pitman et al., 2014) and female offenders highlights the importance of taking

mood and subjective cognitive functioning into account when researching TBI causes

and consequences. It also demonstrates the difficult negotiation between sensitivity and

specificity when screening for TBI. TBI symptoms and risk factors overlap significantly

with psychiatric constructs. Albicini and McKinlay (2014) highlight the absence of a

gold standard in TBI assessment, emphasising the complex nature and specialist skills

required to diagnose TBI. It is recommended that future TBI research include

neuropsychological cognitive assessments to refine screens and reduce the false

positives, which lead to overburdening and wasting expensive clinical resources. Future

47

research needs to specifically identify the sensitivity and specificity, and diagnostic

odds ratio, for the BISI, whilst taking into account the impact of comorbid mental health

diagnoses on validity.

Both the TBI Severity and Time Indices demonstrated insufficient construct validity,

and the BISI may be strengthened by their removal. Gender differences in TBI

presentation specifically related to LOC, may reduce its validity with female offenders.

The BISI may also be refined by the removal of Question eight, on history of other

neurological insult.

It is advised that these refinements are considered fully once the complete dataset is

available. At present, our study is underpowered, and while some trends have been

detected, no definitive decisions should be made until there is sufficient power to ensure

that non-significant results are statistically sound. Once the BISI is refined it can be

utilised to address under identification of TBI in female prisoners, which has important

implications for allocation of resources, staff training, rehabilitation, and behaviour and

risk management.

Limitations

The study’s primary limitation is its small sample size. As present data are only

preliminary, as discussed, we can see some trends in the data but it cannot be

established whether non-significant hypotheses can be rejected until the full sample size

is reached. The response rate of 52% was also lower than that of the male study, which

had 66% of eligible participants complete the full neuropsychological battery (Pitman et

al., 2014). This difference may be attributable to variation in study design which was

informed by constraints of the prison regime.

Although the proportion of the sample reporting TBI is reported, this is not a prevalence

study, and there is a risk of recruitment bias by virtue of the study’s nature being stated

48

in recruitment materials. It is possible that TBI is overrepresented in our sample,

however difficulties in recruiting participants without a TBI may be a reflection of the

prevalence of TBI in the population. With regards to wider representation of the UK

female prison population, the overall mean age of women in this study was 38 years,

which is consistent with the Ministry of Justice’s (2014) reports that the most common

age range of women in prison is 30-39 years; while 73% of women in prison are

reported to be White British, which is similar to the 69.6% found in this study. Notably,

while the BISI was designed to be administered with minimal training, in this research

the BISI was administered by a trainee clinical psychologist with experience working

with TBI, therefore it is may not be representative of administration in general practice

where staff workloads are high and training in working with TBI is rare. Administration

by a clinician with experience in TBI would likely increase the sensitivity of the BISI,

by virtue of validating mTBI which prisoners are often dismissive of, and which can be

missed when administration is rushed. Furthermore, retesting was always administered

by the same trainee, therefore it was not possible to explore inter-rater reliability of the

BISI, which will need to be investigated in future research.

It is also important to note that as individuals with a diagnosed LD were not excluded

unless they demonstrated difficulties with literacy or were unable to provide informed

consent. Their inclusion could impact some of the results, as individuals with LD tend

to have a different cognitive presentation than those with TBI. Once the full dataset is

gathered this data can be examined separately to assess for any significant impact on the

results.

There are a number of limitations for establishing the retest reliability. Although every

effort was made to ensure a retest interval of seven days, unfortunately due to the

practicalities of prison research this was not always possible, and there was a wide

49

interval range. It could also be argued that knowledge of the BISI results could bias

scoring at the second time point, or scoring of the neuropsychological battery; however

adherence to assessment instructions minimised this risk. Longest LOC, the TBI Index

Score, and the TBI Severity Index, are all not normally distributed, which indicates that

the confidence intervals for these ICCs may be biased. If these variables remain

abnormally distributed once the full sample is collected it is advised that bootstrapping

is used to correct the confidence intervals accordingly. The confidence intervals for the

smaller ICCs are also quite wide. Baumgartner and Chung (2001) advise that a sample

of 50 is necessary to reduce CIs to an acceptable breadth.

Cronbach alphas were not calculated for the established neuropsychological cognitive

assessments as it was beyond the scope of this research; therefore there is a chance that

these measures do not have adequate internal reliability in this specific sample.

However, the internal reliability is well established across a range of populations, and

was considered during the design of this study (Appendix II). Also, the WASI in

particular has had its validity as an assessment tool called into question. Although the

WASI has demonstrated factorial equivalence across standardisation and clinical

samples (Ryan et al., 2003), no studies examining the validity of the WASI-II with TBI

specifically has been conducted, and the Matrix Reasoning subtest has been found to

have no predictive validity for TBI (Ryan et al., 2005).

Conclusions

This study of adult female prisoners in the UK provides support for further

investigation of the retest reliability and construct validity of a short TBI screening tool.

While most results were not statistically significant, two of the four summary variables

demonstrated some correlations in the hypothesised directions with a range of measures

of mood and neurodisability, indicating the value of further research with a larger

50

sample. This is the first data that is available on the Global TBI Score which is a new

summary indicator to help identify level of clinical need and inform appropriate clinical

pathways. Six out of 10 clinical indicators demonstrated statistically significant retest

reliability. These findings have implications for the future refinement of the BISI and

demonstrate value in further investigations with a larger sample. This study is the first

of its kind to explore reliability and validity of a TBI screening tool for female offenders

in the UK, beginning to extend evidence of its utility from male offenders (Pitman et al.,

2014). The development of a reliable and valid screening tool for women with TBI can

enable researchers to address the dearth of research into TBI in female offenders (O'

Sullivan et al., 2015), highlighted by the Repairing Shattered Lives report (Williams,

2012). Adoption of a screening tool by female prisons can inform funding for services,

by ensuring the most efficient use of resources. Identifying this vulnerable population

can help apportion funding into training of prison staff in working with female

offenders with TBI, inform offender rehabilitation plans, promote the populations

engagement with the criminal justice system, and identify who would benefit from

specialist assessment and rehabilitation services. The similar rate of TBI found in this

study with other studies highlights the dangers of the literature’s gender bias.

Differences in presentation such as length of LOC and help seeking behaviours

emphasises the possibility of gender specific epidemiological pathways in TBI, which

require much further research.

51

References

Albicini, M., & McKinlay, A. (2014). Mild traumatic brain injury: A review of

terminology, symptomatology, clinical considerations and future directions. In

F. Sadaka & T. Quinn (Eds.), Traumatic Brain Injury: InTech.

Alderman, N. (2007). Prevalence, characteristics and causes of aggressive behaviour

observed within a neurobehavioural rehabilitation service: Predictors and

implications for management. Brain Injury, 21(9), 891-911. doi:

doi:10.1080/02699050701543560

American Congress of Rehabilitation Medicine. (1993). Definition of mild traumatic

brain injury. The Journal of Head Trauma Rehabilitation, 8(3), 86-87.

Anastasi, A. (1998). Psychological testing (6th ed.). New York: McMillan.

Anderson, S. W., Bechara, A., Damasio, H., Tranel, D., & Damasio, A. R. (1999).

Impairment of social and moral behavior related to early damage in human

prefrontal cortex. Nature neuroscience, 2(11), 1032-1037.

Arnau, R. C., Meagher, M. W., Norris, M. P., & Bramson, R. (2001). Psychometric

evaluation of the beck depression inventory-II with primary care medical

patients. Health Psychol, 20(2), 112-119.

Baguley, I. J., Cooper, J., & Felmingham, K. (2006). Aggressive behavior following

traumatic brain injury: How common is common? The Journal of Head Trauma

Rehabilitation, 21(1), 45-56.

52

Baumgartner, T. A., & Chung, H. (2001). Confidence limits for intraclass reliability

coefficients. Measurement in Physical Education and Exercise Science, 5(3),

179-188. doi: 10.1207/s15327841mpee0503_4

Bazarian, J. J., Cernak, I., Noble-Haeusslein, L., Potolicchio, S., & Temkin, N. (2009).

Long‐term neurologic outcomes after traumatic brain injury. The Journal of

Head Trauma Rehabilitation, 24(6), 439-451. doi:

10.1097/HTR.0b013e3181c15600

Beck, A., Steer, R., & Brown, G. (1996). Beck depression inventory II. San Antonio,

TX: The Psychological Corporation.

Beck, A. T., Epstein, N., Brown, G., & Steer, R. A. (1988). An inventory for measuring

clinical anxiety: Psychometric properties. J Consult Clin Psychol, 56(6), 893-

897.

Beer, J. S., John, O. P., Scabini, D., & Knight, R. T. (2006). Orbitofrontal cortex and

social behavior: Integrating self-monitoring and emotion-cognition interactions.

J Cogn Neurosci, 18(6), 871-879. doi: 10.1162/jocn.2006.18.6.871

Bennett, P., Ong, B., & Ponsford, J. (2005). Measuring executive dysfunction in an

acute rehabilitation setting: Using the dysexecutive questionnaire (DEX).

Journal of the International Neuropsychological Society, 11(04), 376-385. doi:

doi:10.1017/S1355617705050423

Blanchette, K., & Brown, S. L. (2006). The assessment and treatment of women

offenders: An integrative perspective. West Sussex: Wiley.

53

Bogner, J., & Corrigan, J. D. (2009). Reliability and predictive validity of the ohio state

university TBI identification method with prisoners. The Journal of Head

Trauma Rehabilitation, 24(4), 279-291 210.1097/HTR.1090b1013e3181a66356.

Brain Injury Association of Wyoming. (2008). Study of undiagnosed brain injuries in

Wyoming's prison population. Retrieved from

http://www.wybia.org/documents/document_display-DocID-Prison%20study

%20final.pdf.cfm

Brewer-Smyth, K., & Burgess, A. W. (2008). Childhood sexual abuse by a family

member, salivary cortisol, and homicidal behavior of female prison inmates.

Nurs Res, 57(3), 166-174. doi: 10.1097/01.NNR.0000319501.97864.d5

Brewer-Smyth, K., Burgess, A. W., & Shults, J. (2004). Physical and sexual abuse,

salivary cortisol, and neurologic correlates of violent criminal behavior in

female prison inmates. Biological Psychiatry, 55(1), 21-31. doi:

http://dx.doi.org/10.1016/S0006-3223(03)00705-4

Brower, M. C., & Price, B. H. (2001). Neuropsychiatry of frontal lobe dysfunction in

violent and criminal behaviour: A critical review. Journal of Neurology,

Neurosurgery & Psychiatry, 71(6), 720-726. doi: 10.1136/jnnp.71.6.720

Browne, A., Miller, B., & Maguin, E. (1999). Prevalence and severity of lifetime

physical and sexual victimization among incarcerated women. International

Journal of Law and Psychiatry, 22(3–4), 301-322. doi:

http://dx.doi.org/10.1016/S0160-2527(99)00011-4

Bruns, J., & Hauser, W. A. (2003). The epidemiology of traumatic brain injury: A

review. Epilepsia, 44, 2-10. doi: 10.1046/j.1528-1157.44.s10.3.x

54

Butler, T., Allnutt, S., Cain, D., Owens, D., & Muller, C. (2005). Mental disorder in the

new south wales prisoner population. Australian and New Zealand Journal of

Psychiatry, 39(5), 407-413. doi: 10.1111/j.1440-1614.2005.01589.x

Carlozzi, N. E., Stout, J. C., Mills, J. A., Duff, K., Beglinger, L. J., Aylward, E. H., . . .

Paulsen, J. S. (2011). Estimating premorbid IQ in the prodromal phase of a

neurodegenerative disease. The Clinical neuropsychologist, 25(5), 757-777. doi:

10.1080/13854046.2011.577811

Carmines, E. G., & Zeller, R. A. (1979). Reliability and validity assessment (Vol. 17).

Newbury Park, CA: Sage.

Chambers, J. C., Yiend, J., Barrett, B., Burns, T., Doll, H., Fazel, S., . . . Fitzpatrick, R.

(2009). Outcome measures used in forensic mental health research: A structured

review. Criminal Behaviour and Mental Health, 19(1), 9-27. doi:

10.1002/cbm.724

Chamelian, L., & Feinstein, A. (2006). The effect of major depression on subjective and

objective cognitive deficits in mild to moderate traumatic brain injury. The

Journal of Neuropsychiatry and Clinical Neurosciences, 18(1), 33-38. doi:

doi:10.1176/jnp.18.1.33

Charter, R. A. (1999). Sample size requirements for precise estimates of reliability,

generalizability, and validity coefficients. Journal of Clinical and Experimental

Neuropsychology, 21(4), 559-566. doi: 10.1076/jcen.21.4.559.889

Chinn, S. (1991). Statistics in respiratory medicine. 2. Repeatability and method

comparison. Thorax, 46(6), 454-456.

55

Cohen. (1960). A coefficient of agreement for nominal scales. Education and

Psychological Measurement, 20, 37-46.

Colantonio, A., Stamenova, V., Abramowitz, C., Clarke, D., & Christensen, B. (2007).

Brain injury in a forensic psychiatry population. Brain Injury, 21(13-14), 1353-

1360. doi: doi:10.1080/02699050701785054

Cole, W. R., Gerring, J. P., Gray, R. M., Vasa, R. A., Salorio, C. F., Grados, M., . . .

Slomine, B. S. (2008). Prevalence of aggressive behaviour after severe

paediatric traumatic brain injury. Brain Injury, 22(12), 932-939. doi:

doi:10.1080/02699050802454808

Collins, M. W., Grindel, S. H., Lovell, M. R., & et al. (1999). Relationship between

concussion and neuropsychological performance in college football players.

JAMA, 282(10), 964-970. doi: 10.1001/jama.282.10.964

Corrigan, J. D., Selassie, A. W., & Orman, J. A. (2010). The epidemiology of traumatic

brain injury. The Journal of Head Trauma Rehabilitation, 25(2), 72-80

10.1097/HTR.1090b1013e3181ccc1098b1094.

Cramer, H. (1946). Mathematical methods of statistics. Princeton: Princeton University

Press.

Creamer, M., Bell, R., & Failla, S. (2003). Psychometric properties of the impact of

event scale—revised. Behaviour Research and Therapy, 41(12), 1489-1496. doi:

http://dx.doi.org/10.1016/j.brat.2003.07.010

Damasio, A. R. (1994). Descartes' error: Emotion, reason, and the human brain. New

York: Putnam.

56

Damasio, A. R., Tranel, D., & Damasio, H. (1990). Individuals with sociopathic

behavior caused by frontal damage fail to respond autonomically to social

stimuli. Behavioural Brain Research, 41(2), 81-94. doi:

http://dx.doi.org/10.1016/0166-4328(90)90144-4

Daniel, W. W., & Cross, C. L. (2013). Biostatistics: A foundation for analysis in the

health sciences (10th ed.). New York: Wiley.

Daoust, S. W., Loper, A. B., Magaletta, P. R., & Diamond, P. M. (2006).

Neuropsychological dysfunction and aggression among female federal inmates.

Psychological Services, 3(2), 88-96.

deSouza, C. A. (2003). Frequency of brain injury in a forensic psychiatric population.

Rev Bras Psiquiatr, 25(4), 206-211. doi: /S1516-44462003000400005

Diamond, P. M., Harzke, A. J., Magaletta, P. R., Cummins, A. G., & Frankowski, R.

(2007). Screening for traumatic brain injury in an offender sample: A first look

at the reliability and validity of the traumatic brain injury questionnaire. The

Journal of Head Trauma Rehabilitation, 22(6), 330-338

310.1097/1001.HTR.0000300228.0000305867.0000300225c.

Donnelly, K. T., Donnelly, J. P., Dunnam, M., Warner, G. C., Kittleson, C. J.,

Constance, J. E., . . . Alt, M. (2011). Reliability, sensitivity, and specificity of

the va traumatic brain injury screening tool. The Journal of Head Trauma

Rehabilitation, 26(6), 439-453. doi: 10.1097/HTR.0b013e3182005de3

Eslinger, P., Grattan, L., & Geder, L. (1995). Impact of frontal lobe lesions on

rehabilitation and recovery from acute brain injury. NeuroRehabilitation, 5(161),

85.

57

Farace, E., & Alves, W. M. (2000). Do women fare worse: A metaanalysis of gender

differences in traumatic brain injury outcome. Journal of neurosurgery, 93(4),

539-545.

Farrington, D. P., Painter, K., & Britain, G. (2004). Gender differences in risk factors

for offending: Home Office London.

Fazel, S., & Grann, M. (2006). The population impact of severe mental illness on

violent crime. Am J Psychiatry, 163(8), 1397-1403. doi:

10.1176/appi.ajp.163.8.1397

Fazel, S., Lichtenstein, P., Grann, M., & Långström, N. (2011). Risk of violent crime in

individuals with epilepsy and traumatic brain injury: A 35-year swedish

population study. PLoS medicine, 8(12), e1001150.

Fazel, S., Philipson, J., Gardiner, L., Merritt, R., & Grann, M. (2009). Neurological

disorders and violence: A systematic review and meta-analysis with a focus on

epilepsy and traumatic brain injury. Journal of Neurology, 256(10), 1591-1602.

doi: 10.1007/s00415-009-5134-2

Feigin, V. L., Theadom, A., Barker-Collo, S., Starkey, N. J., McPherson, K., Kahan, M.,

. . . Ameratunga, S. (2013). Incidence of traumatic brain injury in new zealand:

A population-based study. The Lancet Neurology, 12(1), 53-64. doi:

http://dx.doi.org/10.1016/S1474-4422(12)70262-4

Felde, A. B., Westermeyer, J., & Thuras, P. (2006). Co-morbid traumatic brain injury

and substance use disorder: Childhood predictors and adult correlates. Brain

Injury, 20(1), 41-49. doi: doi:10.1080/02699050500309718

58

Ferguson, P. L., Pickelsimer, E. E., Corrigan, J. D., Bogner, J. A., & Wald, M. (2012).

Prevalence of traumatic brain injury among prisoners in south carolina. The

Journal of Head Trauma Rehabilitation, 27(3), E11-E20

10.1097/HTR.1090b1013e31824e31825f31847.

Ferguson, P. L., Smith, G. M., Wannamaker, B. B., Thurman, D. J., Pickelsimer, E. E.,

& Selassie, A. W. (2010). A population-based study of risk of epilepsy after

hospitalization for traumatic brain injury. Epilepsia, 51(5), 891-898. doi:

10.1111/j.1528-1167.2009.02384.x

Field, A. (2013). Discovering statistics using IBM SPSS statistics. London: Sage.

Filley, C., Price, B., Nell, V., Antionette, T., Morgan, A., Bresnahan, J., . . . Kelly, J.

(2001). Toward an understanding of violence: Neurobehavioral aspects of

unwarranted physical aggression: Aspen neurobehavioral conference consensus

statement. Neuropsychiatry Neuropsychol Behav Neurol, 14, 1-14.

Fleminger, S., & Ponsford, J. (2005). Long term outcome after traumatic brain injury.

BMJ, 331(7530), 1419. doi: 10.1136/bmj.331.7530.1419

French, L. M., Lange, R. T., & Brickell, T. (2014). Subjective cognitive complaints and

neuropsychological test performance following military-related traumatic brain

injury. Journal of rehabilitation research and development, 51(6), 933-950. doi:

10.1682/jrrd.2013.10.0226

Greenwald, B. D., Burnett, D. M., & Miller, M. A. (2003). Congenital and acquired

brain injury. 1. Brain injury: Epidemiology and pathophysiology. Arch Phys

Med Rehabil, 84(3 Suppl 1), S3-7.

59

Greve, K., Sherwin, E., Stanford, M., Mathias, C., Love, J., & Ramzinski, P. (2001).

Personality and neurocognitive correlates of impulsive aggression in long-term

survivors of severe traumatic brain injury. Brain Injury, 15(3), 255-262. doi:

doi:10.1080/026990501300005695

Ham, T. E., Bonnelle, V., Hellyer, P., Jilka, S., Robertson, I. H., Leech, R., & Sharp, D.

J. (2013). The neural basis of impaired self-awareness after traumatic brain

injury.

Hawley, C. A., & Maden, A. (2003). Mentally disordered offenders with a history of

previous head injury: Are they more difficult to discharge? Brain Injury, 17(9),

743-758. doi: doi:10.1080/0269905031000089341

Haydel, M. J., Preston, C. A., Mills, T. J., Luber, S., Blaudeau, E., & DeBlieux, P. M.

C. (2000). Indications for computed tomography in patients with minor head

injury. New England Journal of Medicine, 343(2), 100-105. doi:

doi:10.1056/NEJM200007133430204

Hayes, S., Shackell, P., Mottram, P., & Lancaster, R. (2007). The prevalence of

intellectual disability in a major uk prison. British Journal of Learning

Disabilities, 35(3), 162-167. doi: 10.1111/j.1468-3156.2007.00461.x

Hibbard, M. R., Uysal, S., Kepler, K., Bogdany, J., & Silver, J. (1998). Axis i

psychopathology in individuals with traumatic brain injury. The Journal of

Head Trauma Rehabilitation, 13(4), 24-39.

Hillbom, M., & Holm, L. (1986). Contribution of traumatic head injury to

neuropsychological deficits in alcoholics. J Neurol Neurosurg Psychiatry,

49(12), 1348-1353.

60

Hirtz, D., Thurman, D. J., Gwinn-Hardy, K., Mohamed, M., Chaudhuri, A. R., &

Zalutsky, R. (2007). How common are the “common” neurologic disorders?

Neurology, 68(5), 326-337. doi: 10.1212/01.wnl.0000252807.38124.a3

Hodgins, S. (1998). Epidemiological investigations of the associations between major

mental disorders and crime: Methodological limitations and validity of the

conclusions. Social Psychiatry and Psychiatric Epidemiology, 33(1), S29-S37.

doi: 10.1007/s001270050207

Home Office. (2006). Offender assessment system OASys user manual. London:

Author.

Honer, M., Ferguson, P., Selassie, A., Labbate, L., Kniele, K., & Corrigan, J. (2005).

Patterns of alcohol use 1 year after traumatic brain injury: A population-based,

epidemiological study. Journal of the International Neuropsychological Society,

11(03), 322-330. doi: doi:10.1017/S135561770505037X

Horwitz, R. I., & Yu, E. C. (1984). Assessing the reliability of epidemiologic data

obtained from medical records. Journal of Chronic Diseases, 37(11), 825-831.

doi: http://dx.doi.org/10.1016/0021-9681(84)90015-8

Houtveen, J. H., & Oei, N. Y. L. (2007). Recall bias in reporting medically unexplained

symptoms comes from semantic memory. Journal of psychosomatic research,

62(3), 277-282.

Hwang, S. W., Colantonio, A., Chiu, S., Tolomiczenko, G., Kiss, A., Cowan, L., . . .

Levinson, W. (2008). The effect of traumatic brain injury on the health of

homeless people. Canadian Medical Association Journal, 179(8), 779-784. doi:

10.1503/cmaj.080341

61

IBM. (2011). IBM SPSS statistics 20. New York: IBM Corp.

Ilie, G., Adlaf, E. M., Mann, R. E., Boak, A., Hamilton, H., Asbridge, M., . . .

Cusimano, M. D. (2014). The moderating effects of sex and age on the

association between traumatic brain injury and harmful psychological correlates

among adolescents. PLoS ONE, 9(9), e108167. doi:

10.1371/journal.pone.0108167

Jackson, M., & Hardy, G. (2011). Acquired brain injury in the victorian prison system

Corrections Research Paper Series. Victoria: Department of Justice.

Jarde, A., Losilla, J.-M., & Vives, J. (2012). Methodological quality assessment tools of

non-experimental studies: A systematic review. Anales de Psicología(2), 617-

628.

Johansson, S. H., Jamora, C. W., Ruff, R. M., & Pack, N. M. (2008). A biopsychosocial

perspective of aggression in the context of traumatic brain injury. Brain Injury,

22(13-14), 999-1006. doi: doi:10.1080/02699050802530573

Kelly, J. C., Amerson, E. H., & Barth, J. T. (2012). Mild traumatic brain injury: Lessons

learned from clinical, sports, and combat concussions. Rehabil Res Pract, 2012,

371970. doi: 10.1155/2012/371970

King, M., Semlyen, J., Tai, S., Killaspy, H., Osborn, D., Popelyuk, D., & Nazareth, I.

(2008). A systematic review of mental disorder, suicide, and deliberate self harm

in lesbian, gay and bisexual people. BMC Psychiatry, 8(1), 70.

62

King, N. (1997). Mild head injury: Neuropathology, sequelae, measurement and

recovery. British Journal of Clinical Psychology, 36(2), 161-184. doi:

10.1111/j.2044-8260.1997.tb01405.x

Konrad, C., Geburek, A. J., Rist, F., Blumenroth, H., Fischer, B., Husstedt, I., . . .

Lohmann, H. (2011). Long-term cognitive and emotional consequences of mild

traumatic brain injury. [10.1017/S0033291710001728]. Psychological

Medicine, 41(06), 1197-1211.

Kreutzer, J. S., Marwitz, J. H., Seel, R., & Devany Serio, C. (1996). Validation of a

neurobehavioral functioning inventory for adults with traumatic brain injury.

Archives of Physical Medicine and Rehabilitation, 77(2), 116-124. doi:

http://dx.doi.org/10.1016/S0003-9993(96)90155-0

Kreutzer, J. S., Marwitz, J. H., & Witol, A. D. (1995). Interrelationships between crime,

substance abuse, and aggressive behaviours among persons with traumatic brain

injury. Brain Injury, 9(8), 757-768. doi: doi:10.3109/02699059509008232

Kreutzer, J. S., Seel, R., & Marwitz, J. H. (1999). Neurobehavioural functioning

inventory. San Antonia, TX: The Psychological Corporation.

Kreutzer, J. S., Wehman, P. H., Harris, J. A., Burns, C. T., & Young, H. F. (1991).

Substance abuse and crime patterns among persons with traumatic brain injury

referred for supported employment. Brain Injury, 5(2), 177-187. doi:

doi:10.3109/02699059109008088

Kwako, L. E., Glass, N., Campbell, J., Melvin, K. C., Barr, T., & Gill, J. M. (2011).

Traumatic brain injury in intimate partner violence: A critical review of

63

outcomes and mechanisms. Trauma, Violence, & Abuse, 12(3), 115-126. doi:

10.1177/1524838011404251

Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for

categorical data. Biometrics, 33(1), 159-174. doi: 10.2307/2529310

León-Carrión, J., & Ramos, F. J. C. (2003). Blows to the head during development can

predispose to violent criminal behaviour: Rehabilitation of consequences of head

injury is a measure for crime prevention. Brain Injury, 17(3), 207-216. doi:

doi:10.1080/0269905021000010249

Lezak, M. D. (2004). Neuropsychological assessment (4th ed.). New York: Oxford

University Press.

Mann, C. J. (2003). Observational research methods. Research design II: Cohort, cross

sectional, and case-control studies. Emergency Medicine Journal, 20(1), 54-60.

doi: 10.1136/emj.20.1.54

Martell, D. A. (1992). Estimating the prevalence of organic brain dysfunction in

maximum-security forensic psychiatric patients. J Forensic Sci, 37(3), 878-893.

McAllister, T. (1998). Traumatic brain injury and psychosis: What is the connection?

Seminars in Clinical Neuropsychiatry, 3, 211-223.

McKay, C., Casey, J. E., Wertheimer, J., & Fichtenberg, N. L. (2007). Reliability and

validity of the rbans in a traumatic brain injured sample. Archives of Clinical

Neuropsychology, 22(1), 91-98. doi: http://dx.doi.org/10.1016/j.acn.2006.11.003

McKinlay, A. (2014). Long-term outcomes of traumatic brain injury in early childhood.

Australian Psychologist, 49(6), 323-327. doi: 10.1111/ap.12084

64

McKinlay, A., Corrigan, J., Horwood, L. J., & Fergusson, D. M. (2014a). Substance

abuse and criminal activities following traumatic brain injury in childhood,

adolescence, and early adulthood. The Journal of Head Trauma Rehabilitation,

29(6), 498-506. doi: 10.1097/htr.0000000000000001

McKinlay, A., Grace, R. C., McLellan, T., Roger, D., Clarbour, J., & MacFarlane, M.

R. (2014b). Predicting adult offending behavior for individuals who experienced

a traumatic brain injury during childhood. The Journal of Head Trauma

Rehabilitation, 29(6), 507-513. doi: 10.1097/htr.0000000000000000

Menon, D. K., Schwab, K., Wright, D. W., & Maas, A. I. (2010). Position statement:

Definition of traumatic brain injury. Archives of Physical Medicine and

Rehabilitation, 91(11), 1637-1640. doi:

http://dx.doi.org/10.1016/j.apmr.2010.05.017

Merbitz, C., Jain, S., Good, G. L., & Jain, A. (1995). Reported head injury and

disciplinary rule infractions in prison. Journal of Offender Rehabilitation, 22(3-

4), 11-19. doi: 10.1300/J076v22n03_02

Miller, E. (1999). The neuropsychology of offending. Psychology, Crime & Law, 5(4),

297-318. doi: 10.1080/10683169908401774

Miller, K. J., Ivins, B. J., & Schwab, K. A. (2013). Self-reported mild TBI and

postconcussive symptoms in a peacetime active duty military population: Effect

of multiple TBI history versus single mild TBI. J Head Trauma Rehabil, 28(1),

31-38. doi: 10.1097/HTR.0b013e318255ceae

Ministry of Justice. (2011). Statistics on women and the criminal justice system 2011.

London: Author.

65

Ministry of Justice. (2013a). Offender management statistics quarterly bulletin.

London: Author.

Ministry of Justice. (2013b). Proven re-offending statistics quarterly bulletin. London:

Author.

Ministry of Justice. (2014). Offender management statistics bulletin: October to

december 2014 and annual 2014. London: Author.

Mokkink, L., Terwee, C., Patrick, D., Alonso, J., Stratford, P., Knol, D., . . . de Vet, H.

W. (2010). The cosmin checklist for assessing the methodological quality of

studies on measurement properties of health status measurement instruments: An

international delphi study. Quality of Life Research, 19(4), 539-549. doi:

10.1007/s11136-010-9606-8

Monahan, J., Steadman, H. J., Silver, E., Appelbaum, P. S., Robbins, P. C., Mulvey, E.

P., . . . Banks, S. (2001). Rethinking risk assessment: The macarthur study of

mental disorder and violence: Oxford University Press, USA.

Morgan, A. B., & Lilienfeld, S. O. (2000). A meta-analytic review of the relation

between antisocial behavior and neuropsychological measures of executive

function. Clinical Psychology Review, 20(1), 113-136. doi:

http://dx.doi.org/10.1016/S0272-7358(98)00096-8

Morrell, R. F., Merbitz, C. T., Jain, S., & Jain, S. (1998). Traumatic brain injury in

prisoners. Journal of Offender Rehabilitation, 27(3-4), 1-8. doi:

10.1300/J076v27n03_01

66

National Offender Management Service. (2013). National offender management

service: Business plan 2013-2014. London: Author.

Norris, G., & Tate, R. (2000). The behavioural assessment of the dysexecutive

syndrome (bads): Ecological, concurrent and construct validity. [Article].

Neuropsychological Rehabilitation, 10(1), 33-45. doi:

10.1080/096020100389282

Nunnally, J. C. (1978). Psychometric theory: 2d ed: McGraw-Hill.

O' Sullivan, M., Glorney, E., Sterr, A., Oddy, M., & Da Silva Ramos, S. (2015).

Traumatic brain injury and violent behaviour in females: A systematic review.

[Advance online publication]. Aggression and Violent Behaviour. doi:

10.1016/j.avb.2015.07.006

Oddy, M., Moir, J. F., Fortescue, D., & Chadwick, S. (2012). The prevalence of

traumatic brain injury in the homeless community in a uk city. Brain Injury,

26(9), 1058-1064. doi: doi:10.3109/02699052.2012.667595

Ogilvie, J. M., Stewart, A. L., Chan, R. C. K., & Shum, D. H. K. (2011).

Neuropsychological measures of executive functiong and antisocial behavior: A

meta-analysis Criminology, 49(4), 1063-1107. doi: 10.1111/j.1745-

9125.2011.00252.x

Parker, J., Rubin, K. H., Erath, S., Wojslawowicz, J. C., & Buskirk, A. A. (2006). Peer

relationships and developmental psychopathology. In D. Cicchetti & D. Cohen

(Eds.), Developmental psychopathology: Risk, disorder, and adaptation (2nd

ed., Vol. 2, pp. 419-493). New York: Wiley.

67

Pitman, I., Haddlesey, C., Ramos, S., Oddy, M., & Fortescue, D. (2013).

Neuropsychological performance markers of self-reported traumatic brain injury

in a sample of adult male prisoners in the uk. Manuscript submitted for

publication.

Pitman, I., Haddlesey, C., Ramos, S. D. S., Oddy, M., & Fortescue, D. (2014). The

association between neuropsychological performance and self-reported

traumatic brain injury in a sample of adult male prisoners in the uk.

Neuropsychological Rehabilitation, 1-17. doi: 10.1080/09602011.2014.973887

Randolph, C. (1998). The repeatable battery for the assessment of neuropsychological

status. San Antonio: The Psychological Corporation.

Rankin, G., & Stokes, M. (1998). Reliability of assessment tools in rehabilitation: An

illustration of appropriate statistical analyses. Clinical Rehabilitation, 12(3),

187-199. doi: 10.1191/026921598672178340

Rao, V., Rosenberg, P., Bertrand, M., Salehinia, S., Spiro, J., Vaishnavi, S., . . . Brandt,

J. (2009). Aggression after traumatic brain injury: Prevalence and correlates.

The Journal of Neuropsychiatry and Clinical Neurosciences, 21(4), 420-429.

Rettinger, L. J., & Andrews, D. A. (2010). General risk and need, gender specificity,

and the recidivism of female offenders. Criminal Justice and Behavior, 37(1),

29-46. doi: 10.1177/0093854809349438

Rogers, J. M., & Read, C. A. (2007). Psychiatric comorbidity following traumatic brain

injury. Brain Injury, 21(13-14), 1321-1333. doi:

doi:10.1080/02699050701765700

68

Ropacki, M. T., & Elias, J. W. (2003). Preliminary examination of cognitive reserve

theory in closed head injury. Archives of Clinical Neuropsychology, 18(6), 643-

654. doi: http://dx.doi.org/10.1016/S0887-6177(02)00153-1

Rosenbaum, A., Hoge, S. K., Adelman, S. A., Warnken, W. J., Fletcher, K. E., & Kane,

R. L. (1994). Head injury in partner-abusive men. Journal of Consulting and

Clinical Psychology, 62(6), 1187.

Rubin, K. H., Bukowski, W., & Parker, J. (2006). Peer interactions, relationships, and

groups. In N. Eisenberg (Ed.), Handbook of child psychology: Social, emotional,

and personality development (6th ed., pp. 571-645). New York: Wiley.

Ruff, R. M., & Hibbard, K. (2003). Ruff neurobehavioral inventory: Professional

manual. Lutz, FL: Psychological Assessment Resources, Inc.

Ryan, J. J., Carruthers, C. A., Miller, L. J., Souheaver, G. T., Gontkovsky, S. T., &

Zehr, M. D. (2003). Exploratory factor analysis of the wechsler abbreviated

scale of intelligence (WASI) in adult standardization and clinical samples.

Applied Neuropsychology, 10(4), 252-256. doi: 10.1207/s15324826an1004_8

Ryan, J. J., Carruthers, C. A., Miller, L. J., Souheaver, G. T., Gontkovsky, S. T., &

Zehr, M. D. (2005). The WASI matrix reasoning subtest. International Journal

of Neuroscience, 115(1), 129-136. doi: 10.1080/00207450490512704

Sander, A. M., Kreutzer, J. S., & Fernandez, C. C. (1997). Neurobehavioral functioning,

substance abuse, and employment after brain injury: Implications for vocational

rehabilitation. The Journal of Head Trauma Rehabilitation, 12(5), 28-41.

69

Setnik, L., & Bazarian, J. J. (2007). The characteristics of patients who do not seek

medical treatment for traumatic brain injury. Brain Injury, 21(1), 1-9. doi:

doi:10.1080/02699050601111419

Sherer, M., Struchen, M. A., Yablon, S. A., Wang, Y., & Nick, T. G. (2008).

Comparison of indices of traumatic brain injury severity: Glasgow coma scale,

length of coma and post-traumatic amnesia. Journal of Neurology,

Neurosurgery & Psychiatry, 79(6), 678-685. doi: 10.1136/jnnp.2006.111187

Shiroma, E. J., Ferguson, P. L., & Pickelsimer, E. E. (2010a). Prevalence of traumatic

brain injury in an offender population: A meta-analysis. Journal of Correctional

Health Care, 16(2), 147-159. doi: 10.1177/1078345809356538

Shiroma, E. J., Pickelsimer, E. E., Ferguson, P. L., Gebregziabher, M., Lattimore, P. K.,

Nicholas, J. S., . . . Hunt, K. J. (2010b). Association of medically attended

traumatic brain injury and in-prison behavioral infractions: A statewide

longitudinal study. Journal of Correctional Health Care, 16(4), 273-286. doi:

10.1177/1078345810378253

Sigmon, S., Pells, J., Boulard, N., Whitcomb-Smith, S., Edenfield, T., Hermann, B., . . .

Kubik, E. (2005). Gender differences in self-reports of depression: The response

bias hypothesis revisited. Sex Roles, 53(5-6), 401-411. doi: 10.1007/s11199-005-

6762-3

Silver, J. M., Kramer, R., Greenwald, S., & Weissman, M. (2001). The association

between head injuries and psychiatric disorders: Findings from the new haven

nimh epidemiologic catchment area study. Brain Injury, 15(11), 935-945. doi:

doi:10.1080/02699050110065295

70

Silverstein, B. (2002). Gender differences in the prevalence of somatic versus pure

depression: A replication. American Journal of Psychiatry, 159(6), 1051-1052.

doi: doi:10.1176/appi.ajp.159.6.1051

Simpson, G., Blaszczynski, A., & Hodgkinson, A. (1999). Sex offending as a

psychosocial sequela of traumatic brain injury. The Journal of Head Trauma

Rehabilitation, 14(6), 567-580.

Spencer, R. J., Drag, L. L., Walker, S. J., & Bieliauskas, L. A. (2010). Self-reported

cognitive symptoms following mild traumatic brain injury are poorly associated

with neuropsychological performance in oif/oef veterans. J Rehabil Res Dev,

47(6), 521-530.

Stoddard, S. A., & Zimmerman, M. A. (2011). Association of interpersonal violence

with self-reported history of head injury. Pediatrics, 127(6), 1074-1079. doi:

10.1542/peds.2010-2453

Strauss, E., Sherman, E. M. S., & Spreen, O. (2006). A compendium of

neuropsychological tests: Administration, norms, and commentary. Oxford:

Oxford University Press.

Streiner, D. L., & Norman, G. R. (2003). Health measurement scales: A practical guide

to their development and use. New York: Oxford University Press.

Taber-Thomas, B. C., & Tranel, D. (2012). Social and moral functioning: A cognitive

neuroscience perspective. In V. Anderson & M. Beauchamp (Eds.),

Developmental social neuroscience and childhood brain insult: Theory and

practice (pp. 65-91). New York: Guildford Press.

71

Tennant, A. (2005). Admission to hospital following head injury in england: Incidence

and socio-economic associations. BMC Public Health, 5(1), 1-8. doi:

10.1186/1471-2458-5-21

Terwee, C. B., Bot, S. D. M., de Boer, M. R., van der Windt, D. A. W. M., Knol, D. L.,

Dekker, J., . . . de Vet, H. C. W. (2007). Quality criteria were proposed for

measurement properties of health status questionnaires. Journal of Clinical

Epidemiology, 60(1), 34-42. doi: 10.1016/j.jclinepi.2006.03.012

The Cochrane Collaboration. (2011). Cochrane handbook for systematic reviews of

interventions J. Higgins & S. Green (Eds.), Retrieved from

http://handbook.cochrane.org/

Timonen, M., Miettunen, J., Hakko, H., Zitting, P., Veijola, J., von Wendt, L., &

Räsänen, P. (2002). The association of preceding traumatic brain injury with

mental disorders, alcoholism and criminality: The northern finland 1966 birth

cohort study. Psychiatry research, 113(3), 217-226.

Tombaugh, T. N. (1996). Test of memory malingering (TOMM). New York: Multi-

Health Systems, Inc.

Tombaugh, T. N. (1997). The test of memory malingering (TOMM): Normative data

from cognitively intact and cognitively impaired individuals. Psychological

Assessment, 9(3), 260.

Valera, E. M., & Berenbaum, H. (2003). Brain injury in battered women. J Consult Clin

Psychol, 71(4), 797-804.

72

Visscher, A. J. M., van Meijel, B., Stolker, J. J., Wiersma, J., & Nijman, H. (2011).

Aggressive behaviour of inpatients with acquired brain injury. Journal of

Clinical Nursing, 20(23-24), 3414-3422. doi: 10.1111/j.1365-

2702.2011.03800.x

Volavka, J. (2002). Neurobiology of violence. Washington, DC: American Psychiatric

Press.

von Elm, E., Altman, D. G., Egger, M., Pocock, S. J., Gøtzsche, P. C., &

Vandenbroucke, J. P. (2007). The strengthening the reporting of observational

studies in epidemiology (strobe) statement: Guidelines for reporting

observational studies. Preventive Medicine, 45(4), 247-251. doi:

http://dx.doi.org/10.1016/j.ypmed.2007.08.012

Walker, R., Cole, J. E., Logan, T. K., & Corrigan, J. D. (2007). Screening substance

abuse treatment clients for traumatic brain injury: Prevalence and characteristics.

The Journal of Head Trauma Rehabilitation, 22(6), 360-367. doi:

10.1097/01.htr.0000300231.90619.50

Walsh, E., Buchanan, A., & Fahy, T. (2002). Violence and schizophrenia: Examining

the evidence. The British Journal of Psychiatry, 180(6), 490-495. doi:

10.1192/bjp.180.6.490

Walter, S. D., Eliasziw, M., & Donner, A. (1998). Sample size and optimal designs for

reliability studies. Statistics in Medicine, 17(1), 101-110. doi:

10.1002/(sici)1097-0258(19980115)17:1<101::aid-sim727>3.0.co;2-e

73

Wang, Y.-P., & Gorenstein, C. (2013). Psychometric properties of the beck depression

inventory-II: A comprehensive review. Revista Brasileira de Psiquiatria, 35,

416-431.

Wechsler, D. (2008). Wechsler adult intelligence scale (4th ed.). San Antonio, TX:

Pearson.

Wechsler, D. (2009). Test of premorbid functioning. San Antonio, TX: The

Psychological Corporation.

Wechsler, D., & Zhou, X. (2011). WASI-II (wechsler abbreviated scale of intelligence)

manual. Bloomington MN: Pearson.

Weiss, D., & Marmar, C. (1997). The impact of events scale - revised. In J. Wilson & T.

Keane (Eds.), Assessing psychological trauma and PTSD (pp. 339-411). New

York: Guildford Press.

Whelan-Goodinson, R., Ponsford, J. L., Schönberger, M., & Johnston, L. (2010).

Predictors of psychiatric disorders following traumatic brain injury. The Journal

of Head Trauma Rehabilitation, 25(5), 320-329

310.1097/HTR.1090b1013e3181c1098f1098e1097.

Williams, H. (2012). Repairing shattered lives. London: Transition to Adulthood

Alliance.

Williams, W. H., Mewse, A. J., Tonks, J., Mills, S., Burgess, C. N. W., & Cordan, G.

(2010). Traumatic brain injury in a prison population: Prevalence and risk for re-

offending. Brain Injury, 24(10), 1184-1188. doi:

doi:10.3109/02699052.2010.495697

74

Wilson, B., Alderman, N., Burgess, P., Emslie, H., & Evans, J. (1996). The behavioural

assessment of the dysexecutive syndrome Bury St Edmunds, UK: Thames Valley

Test Company.

Wilson, B. A., Evans, J. J., Emslie, H., Alderman, N., & Burgess, P. (1998). The

development of an ecologically valid test for assessing patients with a

dysexecutive syndrome. Neuropsychological Rehabilitation, 8(3), 213-228. doi:

10.1080/713755570

Wood, R. L., & Liossi, C. (2006). Neuropsychological and neurobehavioral correlates

of aggression following traumatic brain injury. J Neuropsychiatry Clin

Neurosci, 18(3), 333-341. doi: 10.1176/appi.neuropsych.18.3.333

World Health Organization. (2006). Neurological disorders: Public health challenges.

Geneva: Author.

Yeates, K. O., Bigler, E. D., Gerhardt, C. A., Rubin, K. H., Stancin, T., Taylor, H. G., &

Vannatta, K. (2012). Theoretical approaches to understanding social function in

childhood brain insults: Toward the integration of social neuroscience and

childhood brain insult. In V. Anderson & M. Beauchamp (Eds.), Developmental

social neuroscience and childhood brain insult: Theory and practice (pp. 207-

231). New York: Guildford Press.

Yudofsky, S. C., Silver, J. M., Jackson, W., Endicott, J., & Williams, D. (1986). The

Overt Aggression Scale for the objective rating of verbal and physical

aggression. The American journal of psychiatry.

75

Zappalà, G., Thiebaut de Schotten, M., & Eslinger, P. J. (2012). Traumatic brain injury

and the frontal lobes: What can we gain with diffusion tensor imaging? Cortex,

48(2), 156-165. doi: http://dx.doi.org/10.1016/j.cortex.2011.06.020

Zlotnick, C., Clarke, J. G., Friedmann, P. D., Roberts, M. B., Sacks, S., & Melnick, G.

(2008). Gender differences in comorbid disorders among offenders in prison

substance abuse treatment programs. Behavioral Sciences & the Law, 26(4),

403-412. doi: 10.1002/bsl.831

76

List of Appendices

Appendix I: Neuropsychological Rehabilitation Guidelines for Authors …... 130

Appendix II: Demographic Data ……………………………………………. 140

Appendix III: Additional Psychometric Properties of Measures ……………. 144

Appendix IV: Semi-structured Interview …………………………………… 150

Appendix V: The Brain Injury Screening Inventory ……………..………… 164

Appendix VI: The Impact of Events Scale – Revised ………………………. 168

Appendix VII: Recruitment Poster ………………………………………….. 170

Appendix VIII: Participant Information Sheet ……………………………… 172

Appendix IX: Participant Consent Form ……………………………………. 178

Appendix X: Participant Feedback Pro Forma ……………………………… 180

Appendix XI: OASys Items Extracted ……………………………………… 183

Appendix XII: Correspondence Regarding Ethical Approval ………………. 185

Appendix XIII: Normality Data ……………………………………………... 188

77

Appendix I

Neuropsychological Rehabilitation: An International Journal – Guidelines

for Authors

78

Instructions for authors

This journal uses ScholarOne Manuscripts (previously Manuscript Central) to peer

review manuscript submissions. Please read the guide for ScholarOne authors before

making a submission. Complete guidelines for preparing and submitting your

manuscript to this journal are provided below.

Use these instructions if you are preparing a manuscript to submit to

Neuropsychological Rehabilitation. To explore our journals portfolio, visit

http://www.tandfonline.com/, and for more author resources, visit our Author Services

website.

Neuropsychological Rehabilitation considers all manuscripts on the strict condition that

the manuscript is your own original work, and does not duplicate any other previously

published work, including your own previously published work; the manuscript has

been submitted only to Neuropsychological Rehabilitation; it is not under consideration

or peer review or accepted for publication or in press or published elsewhere; the

manuscript contains nothing that is abusive, defamatory, libellous, obscene, fraudulent,

or illegal.

Please note that Neuropsychological Rehabilitation uses CrossCheck™ software to

screen manuscripts for unoriginal material. By submitting your manuscript to

Neuropsychological Rehabilitation you are agreeing to any necessary originality checks

your manuscript may have to undergo during the peer-review and production processes.

Any author who fails to adhere to the above conditions will be charged with costs which

Neuropsychological Rehabilitation incurs for their manuscript at the discretion of

79

Neuropsychological Rehabilitation’s Editors and Taylor & Francis, and their

manuscript will be rejected.

This journal is compliant with the Research Councils UK OA policy. Please see the

licence options and embargo periods here.

Manuscript preparation

1. Journal-specific guidelines

This journal accepts original (regular) articles, scholarly reviews, and book reviews. The

style and format of the typescripts should conform to the specifications given in the

Publication Manual of the American Psychological Association (6th ed.).There is no

word limit for manuscripts submitted to this journal. Authors should include a word

count with their manuscript. 

2. General guidelines

Manuscripts are accepted in English. Oxford English Dictionary spelling and

punctuation are preferred. Please use double quotation marks, except where “a quotation

is ‘within’ a quotation”. Long quotations of words or more should be indented without

quotation marks.

Manuscripts should be compiled in the following order: title page; abstract; keywords;

main text; acknowledgements; references; appendices (as appropriate); table(s) with

caption(s) (on individual pages); figure caption(s) (as a list).

Abstracts of 150-200 words are required for all manuscripts submitted.

Each manuscript should have up to 5 keywords.

80

Search engine optimization (SEO) is a means of making your article more visible to

anyone who might be looking for it. Please consult our guidance here.

Section headings should be concise.

All authors of a manuscript should include their full names, affiliations, postal

addresses, telephone numbers and email addresses on the cover page of the manuscript.

One author should be identified as the corresponding author. Please give the affiliation

where the research was conducted. If any of the named co-authors moves affiliation

during the peer review process, the new affiliation can be given as a footnote. Please

note that no changes to affiliation can be made after the manuscript is accepted. Please

note that the email address of the corresponding author will normally be displayed in the

article PDF (depending on the journal style) and the online article.

All persons who have a reasonable claim to authorship must be named in the manuscript

as co-authors; the corresponding author must be authorized by all co-authors to act as an

agent on their behalf in all matters pertaining to publication of the manuscript, and the

order of names should be agreed by all authors.

Biographical notes on contributors are not required for this journal.

Please supply all details required by any funding and grant-awarding bodies as an

Acknowledgement on the title page of the manuscript, in a separate paragraph, as

follows:

For single agency grants: "This work was supported by the [Funding Agency] under

Grant [number xxxx]."

For multiple agency grants: "This work was supported by the [Funding Agency 1]

81

under Grant [number xxxx]; [Funding Agency 2] under Grant [number xxxx]; and

[Funding Agency 3] under Grant [number xxxx]."

Authors must also incorporate a Disclosure Statement which will acknowledge any

financial interest or benefit they have arising from the direct applications of their

research.

For all manuscripts non-discriminatory language is mandatory. Sexist or racist terms

must not be used.

Authors must adhere to SI units. Units are not italicised.

When using a word which is or is asserted to be a proprietary term or trade mark,

authors must use the symbol ® or TM.

2. Style guidelines

Description of the Journal’s reference style.

Guide to using mathematical scripts and equations.

Word templates are available for this journal. If you are not able to use the template via

the links or if you have any other template queries, please contact

authortemplate@tandf.co.uk.

Authors must not embed equations or image files within their manuscript

3. Figures

Please provide the highest quality figure format possible. Please be sure that all

imported scanned material is scanned at the appropriate resolution: 1200 dpi for line art,

600 dpi for grayscale and 300 dpi for colour.

82

Figures must be saved separate to text. Please do not embed figures in the manuscript

file.

Files should be saved as one of the following formats: TIFF (tagged image file format),

PostScript or EPS (encapsulated PostScript), and should contain all the necessary font

information and the source file of the application (e.g. CorelDraw/Mac, CorelDraw/PC).

All figures must be numbered in the order in which they appear in the manuscript (e.g.

Figure 1, Figure 2). In multi-part figures, each part should be labelled (e.g. Figure 1(a),

Figure 1(b)).

Figure captions must be saved separately, as part of the file containing the complete text

of the manuscript, and numbered correspondingly.

The filename for a graphic should be descriptive of the graphic, e.g. Figure1, Figure2a.

4. Publication charges

Submission fee

There is no submission fee for Neuropsychological Rehabilitation.

Page charges

There are no page charges for Neuropsychological Rehabilitation.

Colour charges

Colour figures will be reproduced in colour in the online edition of the journal free of

charge. If it is necessary for the figures to be reproduced in colour in the print version, a

charge will apply. Charges for colour figures in print are £250 per figure ($395 US

83

Dollars; $385 Australian Dollars; 315 Euros). For more than 4 colour figures, figures 5

and above will be charged at £50 per figure ($80 US Dollars; $75 Australian Dollars; 63

Euros).

Depending on your location, these charges may be subject to Value Added Tax.

5. Reproduction of copyright material

If you wish to include any material in your manuscript in which you do not hold

copyright, you must obtain written permission from the copyright owner, prior to

submission. Such material may be in the form of text, data, table, illustration,

photograph, line drawing, audio clip, video clip, film still, and screenshot, and any

supplemental material you propose to include. This applies to direct (verbatim or

facsimile) reproduction as well as “derivative reproduction” (where you have created a

new figure or table which derives substantially from a copyrighted source).

You must ensure appropriate acknowledgement is given to the permission granted to

you for reuse by the copyright holder in each figure or table caption. You are solely

responsible for any fees which the copyright holder may charge for reuse.

The reproduction of short extracts of text, excluding poetry and song lyrics, for the

purposes of criticism may be possible without formal permission on the basis that the

quotation is reproduced accurately and full attribution is given.

For further information and FAQs on the reproduction of copyright material, please

consult our Guide.

6. Supplemental online material

Authors are encouraged to submit animations, movie files, sound files or any additional

84

information for online publication.

Information about supplemental online material

Manuscript submission

All submissions should be made online at the Neuropsychological Rehabilitation

Scholar One Manuscripts website. New users should first create an account. Once

logged on to the site, submissions should be made via the Author Centre. Online user

guides and access to a helpdesk are available on this website.

Manuscripts may be submitted in any standard editable format, including Word and

EndNote. These files will be automatically converted into a PDF file for the review

process. LaTeX files should be converted to PDF prior to submission because

ScholarOne Manuscripts is not able to convert LaTeX files into PDFs directly. All

LaTeX source files should be uploaded alongside the PDF.

Click here for information regarding anonymous peer review.

Copyright and authors' rights

To assure the integrity, dissemination, and protection against copyright infringement of

published articles, you will be asked to assign us, via a Publishing Agreement, the

copyright in your article. Your Article is defined as the final, definitive, and citable

Version of Record, and includes: (a) the accepted manuscript in its final form, including

the abstract, text, bibliography, and all accompanying tables, illustrations, data; and (b)

any supplemental material hosted by Taylor & Francis. Our Publishing Agreement with

you will constitute the entire agreement and the sole understanding between you and us;

no amendment, addendum, or other communication will be taken into account when

85

interpreting your and our rights and obligations under this Agreement.

Copyright policy is explained in detail here.

Free article access

As an author, you will receive free access to your article on Taylor & Francis Online.

You will be given access to the My authored works section of Taylor & Francis Online,

which shows you all your published articles. You can easily view, read, and download

your published articles from there. In addition, if someone has cited your article, you

will be able to see this information. We are committed to promoting and increasing the

visibility of your article and have provided guidance on how you can help. Also within

My authored works, author eprints allow you as an author to quickly and easily give

anyone free access to the electronic version of your article so that your friends and

contacts can read and download your published article for free. This applies to all

authors (not just the corresponding author).

Reprints and journal copies

Corresponding authors receive a complimentary copy of the issue containing their

article. Complimentary reprints are available through Rightslink® and additional

reprints can be ordered through Rightslink® when proofs are received. If you have any

queries about reprints, please contact the Taylor & Francis Author Services team at

reprints@tandf.co.uk. To order a copy of the issue containing your article, please

contact our Customer Services team at Adhoc@tandf.co.uk

Open Access

Taylor & Francis Open Select provides authors or their research sponsors and funders

86

with the option of paying a publishing fee and thereby making an article permanently

available for free online access – open access – immediately on publication to anyone,

anywhere, at any time. This option is made available once an article has been accepted

in peer review.

Full details of our Open Access programme

Last updated 11/03/2014

87

Appendix II

Demographic Data

88

Table 14. Frequency Demographic Data

Ethnicity %

White 69.6%

Black or Black British 17.4%

Mixed race 8.7%

Other ethnic group 4.3%

Marital status

Single 69.6%

Married 8.7%

Divorced 8.7%

Separated (but still married) 8.7%

Widowed 4.3%

Highest qualification

No qualifications 39.1%

GCSE/O level 4.3%

A-level/NVQ/Vocational award 39.1%

Diploma 4.3%

Degree 8.7%

Master’s degree 4.3%

Mental health diagnosis

None 34.8%

Depression 26.1%

Anxiety 4.3%

Bipolar 4.3%

89

Personality disorder 4.3%

Comorbid diagnoses (depression;

anxiety; PTSD; bipolar; OCD;

psychosis; personality disorder)

26.1%

Learning disability diagnosis 13%

History of self-harm 26.1%

Accessed mental health

services

65.2%

Experience of childhood abuse (self-reported on OASys assessment) 42.1%

Experience of domestic violence

Self-reported on OASys assessment 50%

OASys and reports from TBI history

combined

59.1%

Category of index offence

Drug offences 21.7%

Indictable motoring offences 4.3%

Violence against the person 26.1%

Robbery 26.1%

Theft & handling stolen goods 4.3%

Sexual offences 4.3%

Fraud & forgery 4.3%

Criminal damage 8.7%

Table 15. Presence of drug use

TBI (%) Non-TBI (%)

90

Heroin 37.50% 14.30%

Prescription drugs 37.50% 14.30%

Cocaine 56.30% 28.60%

Crack cocaine 31.30% 28.60%

Amphetamines 50.00% 57.10%

Ecstasy 37.50% 42.90%

Cannabis 68.80% 42.90%

Hallucinogens 37.50% 0.00%

Solvents 18.80% 0.00%

Sedatives 31.30% 14.3%

Crystal meth 6.30% 0.00%

Table 16. Descriptive Demographic Data

n Range M (SD)

Age of onset of mental health difficulties 15 9 - 55 21 (12.006)

No. of suicide attempts 23 0 - 10 1.43 (2.53)

91

Appendix III

Additional Psychometric Properties of Measures

92

Questionnaire Details Reliability in Relevant Sample Validity in Relevant Sample

Internal consistency Test-retest reliabilities

The Brain Injury Screening

Index

7 item self-report TBI

screen

N/A Demonstrated preliminary

convergent validity in male

prison sample (Pitman et al.,

2014)

Not known Not known

The Test of Memory

Malingering (Tombaugh,

1996)

50 item recognition test

designed to discriminate

between true memory

impairment and

malingering

Adult sample (Strauss, Sherman, & Spreen, 2006)

0.94 Not known Sensitivity: 0.1

Specificity: 0.9

The Wechsler Abbreviated

Scale of Intelligence

(Wechsler & Zhou, 2011)

Standardised measure of

cognitive function,

consisting of 4 subtests, 2

Adult sample (Wechsler & Zhou, 2011)

0.97 0.90-0.96 Clinical group study on

individuals with TBI conducted

93

verbal and 2 performanceby test authors

The Test of Premorbid

Functioning (Wechsler,

2009)

Provides an estimate of

premorbid intellectual

functioning using atypical

grapheme-phoneme

translation to measure

word knowledge through

reading

Adult sample (Wechsler, 2009)

0.95 0.93 Clinical group study on

individuals with TBI conducted

by test authors

The Behavioural

Assessment of the

Dysexecutive Syndrome

(Wilson et al., 1996)

A battery of 6 subtests to

test executive function

Adult sample – general population & neurological (Wilson et al., 1996)

0.70 -0.08-0.71 Construct validity: differentiated

between neurological and

healthy participants on 74% of

occasions (Norris & Tate, 2000;

Wilson et al., 1996)

The Repeatable Battery for Brief evaluation of Adult sample (Randolph, 1998)

94

the assessment of

Neuropsychological Status

(RBANS; Randolph, 1998)

cognitive function in

adults with neurological

disturbance

0.92-0.94 0.88 Construct validity demonstrated

in TBI population (McKay,

Casey, Wertheimer, &

Fichtenberg, 2007)

Neurobehavioral

Functioning

Inventory(Kreutzer et al.,

1999)

Self-report measure of

behaviours and symptoms

commonly associated

with TBI

Adult sample with TBI (Kreutzer et al., 1996)

0.97 Not known Construct validity: scores on

depression (p<0.002),

memory/attention (p<0.002),

communication (p<0.001),

aggression (p<0.002) and motor

(p<0.002) subscales

distinguished between

individuals with TBI who were

employed and unemployed

(Sander, Kreutzer, & Fernandez,

95

1997)

The Beck Anxiety

Inventory(Beck et al.,

1988)

21 item self-report

inventory of anxiety

severity

Adult psychiatric outpatients (Beck et al., 1988)

0.92 0.75 Construct validity: discriminates

anxious diagnostic groups from

non-anxious diagnostic groups

The Beck Depression

Inventory II (Beck et al.,

1996)

21 item self-report

inventory of anxiety

severity

Adult psychiatric outpatients (Beck et al., 1996)

0.92 0.93 Construct validity: discriminates

depressed individuals from non-

depressed individuals(Arnau,

Meagher, Norris, & Bramson,

2001)

The Impact of Events Scale

– Revised(Weiss &

Marmar, 1997)

Self-report measure

assessing intrusive and

avoidant reactions

associated with a

Adult population (Weiss & Marmar, 1997)

0.79-0.92 0.57-0.94 Construct validity: correlates

well with PTSD Checklist

(Creamer, Bell, & Failla, 2003)

96

particular event

97

Appendix IV

Semi structured Interview

98

Interview Schedule: Head Injury Research

Participant ID no.: ___________________

DOB: ________________________

Informed consent given for study: Yes No

Initial BISI completed: Yes No

Social History

2. How would you describe your ethnicity?

Asian or Asian British Black or Black British Mixed White

Other Ethnic Group I do not wish to disclose this

3. Are you:

Single Married Divorced Separated (but still married)

Widowed

3a. Do you have children?

Yes No

How many

3b

.

Accommodation Type: Before you came to be in prison were you living in:

Your own property Rented accommodation Shared House Hostel

Other

4. Do you have siblings?

Brothers Sisters

Have they been in trouble with police?

Yes No

5. Mother’s Occupation:

Father’s Occupation:

99

Education

6. How many years have you spent in education?

7. What is your highest level qualification?

GCSE/O Level A Level/NVQ/Vocational Award Degree

Masters Degree Doctorate/PHD No Qualifications Diploma

7.a Have you achieved any other qualifications since school/university?

Yes No

Specify (inc. if in prison):

8. What were you like at school?

Academically Good Disengaged/disinterested

Disruptive Truant Left Expelled

9. Did you require any help at school?

Yes No

9a. If yes what was this for?

Dyslexia Learning disability

Learning difficulty Behavioural problems Other

Occupation History

10. What has been your main occupation?

11. What was your job stability?

Stable Left of own accord Dismissed

100

12. Number of jobs had in working life:

Medical History

13. Have you ever had any operations/illnesses?

Yes No

If yes please specify

14. How old were you? – for major illnesses and significant hospitalisation

15. How often have you spent time in hospital?

16. Do you suffer from any long term illness?

Yes No

17. Do you have a physical disability?

Yes No

18. Have you been diagnosed with a mental health problem?

Yes No Age of onset:

19. Have you been diagnosed with a learning disability?

Yes No

20. Have you ever self-harmed

Yes No

If yes, frequency and type

Daily Weekly Monthly Occasionally

Cutting Burning Ingestion Banging head Other

21. Have you ever attempted suicide

101

Yes No

Details (inc. age):

21a

.

Have you ever seen:

Psychologist Psychiatrist Had contact with mental health services

When was this and what was it for (general details)?

Injury

22. How many times have you experienced a blow to the head?

Once Twice Three times Four times Five times >5 times

23. a. How old were you?

If you experienced a loss of consciousness, how long was this for?

Just went dizzy Less than 10 minutes 10 minutes to 6 hours Over 6 hours

What was the cause of your head injury?

RTA Sporting accident Fight Fight with partner Falls

(on substances) Joyriding Falls (sober) Other crimes

Other (non crime related)

Other significant details:___________________________________________________

_________________________________________________________________________

What happened after the injury?

A&E Spent time in hospital Spent time at home recovering

Imprisonment Nothing

102

Other significant details: ____________________________________________________

_________________________________________________________________________

Did you feel different after the injury?

Yes No

If yes, what parts of you were affected?

Physical aspects Emotions Anger Feeling irritable Feeling

anxious Memory Concentration Learning Speech Headaches

Other significant details: ____________________________________________________

_________________________________________________________________________

b. How old were you?

If you experienced a loss of consciousness, how long was this for?

Just went dizzy Less than 10 minutes 10 minutes to 6 hours Over 6 hours

What was the cause of your head injury?

RTA Sporting accident Fight Fight with partner Falls

(on substances) Joyriding Falls (sober) Other crimes

Other (non crime related)

Other significant details:___________________________________________________

_________________________________________________________________________

What happened after the injury?

A&E Spent time in hospital Spent time at home recovering

Imprisonment Nothing

Other significant details: ____________________________________________________

_________________________________________________________________________

103

Did you feel different after the injury?

Yes No

If yes, what parts of you were affected?

Physical aspects Emotions Anger Feeling irritable Feeling

anxious Memory Concentration Learning Speech Headaches

Other significant details: ____________________________________________________

_________________________________________________________________________

c. How old were you?

If you experienced a loss of consciousness, how long was this for?

Just went dizzy Less than 10 minutes 10 minutes to 6 hours Over 6 hours

What was the cause of your head injury?

RTA Sporting accident Fight Fight with partner

Falls (on drugs) Joyriding Falls (sober) Other crimes

Other (non crime related)

Other significant details: ____________________________________________________

_________________________________________________________________________

What happened after the injury?

A&E Spent time in hospital Spent time at home recovering

Imprisonment Nothing

Other significant details: ____________________________________________________

_________________________________________________________________________

Did you feel different after the injury?

Yes No

104

If yes, what parts of you were affected?

Physical aspects Emotions Anger Feeling irritable Feeling

anxious Memory Concentration Learning Speech Headaches

Other significant details: ____________________________________________________

_________________________________________________________________________

d. How old were you?

If you experienced a loss of consciousness, how long was this for?

Just went dizzy Less than 10 minutes 10 minutes to 6 hours Over 6 hours

What was the cause of your head injury?

RTA Sporting accident Fight Fight with partner Falls

(on substances) Joyriding Falls (sober) Other crimes

Other (non crime related)

Other significant details:___________________________________________________

_________________________________________________________________________

What happened after the injury?

A&E Spent time in hospital Spent time at home recovering

Imprisonment Nothing

Other significant details: ____________________________________________________

_________________________________________________________________________

Did you feel different after the injury?

Yes No

If yes, what parts of you were affected?

Physical aspects Emotions Anger Feeling irritable Feeling

105

anxious Memory Concentration Learning Speech Headaches

Other significant details: ____________________________________________________

_________________________________________________________________________

e. How old were you?

If you experienced a loss of consciousness, how long was this for?

Just went dizzy Less than 10 minutes 10 minutes to 6 hours Over 6 hours

What was the cause of your head injury?

RTA Sporting accident Fight Fight with partner Falls

(on substances) Joyriding Falls (sober) Other crimes

Other (non crime related)

Other significant details:___________________________________________________

_________________________________________________________________________

What happened after the injury?

A&E Spent time in hospital Spent time at home recovering

Imprisonment Nothing

Other significant details: ____________________________________________________

_________________________________________________________________________

Did you feel different after the injury?

Yes No

If yes, what parts of you were affected?

Physical aspects Emotions Anger Feeling irritable Feeling anxious

Memory Concentration Learning Speech Headaches

Other significant details: ____________________________________________________

106

_________________________________________________________________________

24. How old were you when you had your first injury?

25. How old were you at the age of your most serious injury?

Offences

26. How many times have you been in prison?

Once Twice Three times Four times Five times More than Five

27. How old were you when you first committed an offence?

28. How many offences (counts) have you had for the following? (0=None, 1= Once, 2 =

Twice, 3 = Three times, or 4= Four or more times)

Burglary Theft & handling Violent offences Driving offences

Fraud & forgery Drug offences Sexual offences Other

Criminal damage Robbery

29. How severe were your violent offences?

None (0) Violent threat (1) Assault without causing injury (2) Minor injury

(3) Serious injury (4) Severe injury (5) Murder (6) Multiple murders (7)

30. How many years in total would you say you have spent in custody in your adult life?

31. Do you attend any education classes?

Yes No Which?

If no, why? ________________________________________________________________

__________________________________________________________________________

107

32. Do you attend any work?

Yes No Which?

If no, why? ________________________________________________________________

__________________________________________________________________________

33. Do you attend any rehabilitation programmes?

Yes No Which?

If no, why? ________________________________________________________________

__________________________________________________________________________

35. How long is your current sentence? (inc. sentenced and to be served)

36. When you were last released from prison, what went wrong?

Return to alcohol/drugs Couldn’t cope Needed money Returned to old habits

other people influencing decisions N/A (1st offence) Nothing

Other, please specify ________________________________________________________

__________________________________________________________________________

What help would have made a difference?

Job Support (to help cope) Family contact/support Drug/alcohol support

Moving away from the area you were living in previously Mental Health

support/intervention Access to education

Other, please specify ________________________________________________________

__________________________________________________________________________

Drug/Alcohol Use

108

37. Have you ever used/taken:

Heroin

Everyday (5) Most days (4) Weekends (3) Once a month (2) Once (1) Never (0)

Drugs prescribed for someone else

Everyday (5) Most days (4) Weekends (3) Once a month (2) Once (1) Never (0)

Cocaine

Everyday (5) Most days (4) Weekends (3) Once a month (2) Once (1) Never (0)

Crack cocaine

Everyday (5) Most days (4) Weekends (3) Once a month (2) Once (1) Never (0)

Amphetamine

Everyday (5) Most days (4) Weekends (3) Once a month (2) Once (1) Never (0)

Ecstasy

Everyday (5) Most days (4) Weekends (3) Once a month (2) Once (1) Never (0)

Cannabis

Everyday (5) Most days (4) Weekends (3) Once a month (2) Once (1) Never (0)

Hallucinogens

Everyday (5) Most days (4) Weekends (3) Once a month (2) Once (1) Never (0)

Solvents

Everyday (5) Most days (4) Weekends (3) Once a month (2) Once (1) Never (0)

Sedatives/tranquilisers

Everyday (5) Most days (4) Weekends (3) Once a month (2) Once (1) Never (0)

Crystal meth

Everyday (5) Most days (4) Weekends (3) Once a month (2) Once (1) Never (0)

Other, please specify _____________________________________________________

_______________________________________________________________________

109

Everyday (5) Most days (4) Weekends (3) Once a month (2) Once (1) Never (0)

38. How often do you/did you drink alcohol?

Everyday (5) Most days (4) Weekends (3) Once a month (2) Once (1) Never (0)

38a How much alcohol would you drink in a day/period?

Other Information

Suitable for study: Yes No

If No, select reason(s) why:

Not interested Other problems more significant e.g. MH or LD? Soon to be

released?

If yes, select TBI or non-TBI:

Non-TBI TBI

Repeat BISI: Yes No

NFI completed: Yes No

BDI-II completed: Yes No

BAI completed: Yes No

IES-R completed: Yes No

110

TOPF completed: Yes No

WASI-II completed: Yes No

TOMM completed: Yes No

RBANS completed: Yes No

BADS completed: Yes No

Assessment feedback session requested: Yes No

Assessment results to be shared with prison/health services: Yes No

To be contacted with research results after study: Yes No

Researcher:_______________________________ Date:___/___/_____

111

Date of Stage 1 __/__/____

Date of Stage 2 __/__/____

Appendix V

The Brain Injury Screening Inventory

112

113

114

BISI Global Score Formula

TBI (Yes = 1; No = 0) + disorientation (Yes = 1; No = 0) + PTA (Yes = 1; No = 0) +

LOC (Yes = 4; No = 0) + other TBI (Yes = 1; No = 0) + ABI (Yes = 10; No = 0) +

memory difficulties (Yes = 1; No = 0) + concentration difficulties (Yes = 1; No = 0) +

speech difficulties (Yes = 1; No = 0) + other difficulties (Yes = 1; No = 0)

115

Appendix VI

Impact of Events Scale – Revised (IES-R)

116

117

Appendix VII

Recruitment Poster

118

119

Appendix VIII

Participant Information Sheet

120

PARTICIPANT INFORMATION SHEET

Traumatic Brain Injury in Adult Female Offenders in the UK

You are being invited to take part in a research study. Before you decide whether or not

to take part, it is important for you to understand why the research is being done and

what it will involve. Please take time to read the following information carefully and

discuss it with others if you wish. Ask the researcher if there is anything that is not clear

or if you would like more information. Take time to decide whether or not you wish to

take part.

This research has been approved by the National Offender Management Service

National Research Committee, reference no.: 2013-266.

Thank you for reading this.

What is the purpose of the study?

We have three aims for this study. Our first aim is to find out how many female

prisoners have a history of traumatic brain injury, and to check that our brief screening

tool is appropriate for use with female prisoners. We would also like to see if

participants who have had a traumatic brain injury have different physical and mental

health needs to those without a traumatic brain injury.

Why have I been invited?

We are asking all new receptions to HMP ____ to take part.

Do I have to take part?

No. Your participation in this research study is completely voluntary and you do not

have to take part if you do not wish to. If you decide to take part you are still free to

withdraw at any time, without giving a reason. This will have no effect on your current

121

or future treatment or medical care and will not be recorded in your record. If you do

decide to take part you will be given this information sheet to keep and be asked to sign

a consent form.

What will happen to me if I take part?

If you decide to take part, you will complete a brief questionnaire and have one

interview with a researcher which will last around an hour. This interview will include

questions about your social history, health, and offending history, but there will be no

compulsory questions. All you have to do is sit down and have a conversation with the

researcher. If there is anything you find uncomfortable you don’t have to talk about it.

This research is independent of the prison and will not affect your sentence in any way.

You can withdraw from the research at any time, without giving a reason.

You will then be invited to complete further questionnaires and for neuropsychological

testing. This can be completed on the same day or arranged for a later date if preferred.

The questionnaires will include measures relating to mental health. The

neuropsychological tests will examine your current and previous cognitive functioning.

After the assessment is completed you will have the option of obtaining feedback on the

findings. You will also have the option of sharing the results with the prison/health

service if you wish. If feedback is requested you will receive a letter saying whether or

not the assessment indicated there was a need for a formal clinical assessment. In the

event that clinically significant results are obtained, e.g. if you seem to be experiencing

significant low mood, you will receive a letter from the researcher suggesting you attend

an arranged feedback session, with an allocated time and date. This can be altered if it

does not suit you. Feedback sessions will occur in a private space designated by the

prison. The researcher will discuss the nature of the significant results with you and try

to answer any questions you may have. We will advise you to allow us to inform prison

122

staff if results suggest you are experiencing difficulties so you can access appropriate

support through the prison healthcare system. In the case of neuropsychological

assessments, with your consent, scores can be forwarded to the psychology staff within

the prison and clinicians can contact the primary researcher with questions if necessary

to support their clinical work.

If you want to, you can also give the researcher your details and they will contact you

after they have the complete results from the study.

Who else will the researchers talk to?

The researchers will access your prison file to collect some background information.

Just like what you tell the researcher yourself in the assessment, this information will be

kept confidential and anonymous.

Only essential background information such as any diagnoses, number and nature of

offences, and behavioural infractions whilst in prison, will be attained through your file.

This background information will be stored anonymously on a secure database that is

only available to the research team.

What are the possible disadvantages and risks of taking part?

There are no health risks to taking part, however if you find the topics of conversation a

bit difficult, you may access support from the prison’s medical unit.

What are the possible benefits of taking part?

The research team hope that you will find the research interesting, but they cannot

promise that the study will help you. As you have the option of sharing results with the

medical unit this may inform your rehabilitation. We hope the information the

researchers get from this study may help inform rehabilitation programs, and how to

support individuals both in prison and on release.

123

Will my taking part in this study be kept confidential?

All information which is collected in the course of this research will be kept strictly

confidential. During the study you will be allocated a participant number. This will

ensure that throughout the study, you will remain anonymous. If you tell the researchers

of behaviour that is against prison rules and can be adjudicated against, illegal acts, and

behaviour which could represent a risk to yourself or others, then this will be reported to

prison staff, who should record this in your notes. The data collected in this study will

be used only for the purpose described in this form. All records related to your

involvement in this research will be stored securely. Data gathered from this study will

be maintained as long as required by regulations.

This research is part of a doctoral project, and will be written up in the thesis, which is

kept at the University of Surrey. Any information which is disseminated will be

completely anonymised so that you cannot be recognised from it.

What if I am unhappy or there is a problem?

If you are unhappy, or if there is a problem, please feel free to let us know by contacting

a member of the research team through the prison, and we will try to help. If you remain

unhappy or have a complaint which you feel you cannot come to us with then you

should the prison staff can put you in contact with the Head of the School of

Psychology at the University.

Who is organising and funding the research?

This research is being organised by the Disabilities Trust.

OK, so what happens now?

If you’d like to take part, you need to sign the form saying that you agree to participate.

Further information and contact details.

124

For any more information, or to answer any questions you may have, please ask a

member of prison staff who can put you in contact with a member of the research team.

Michelle O’ Sullivan & Steven Fitzsimons

School of Psychology, Faculty of Arts & Human Sciences

University of Surrey

You will be given a copy of the information sheet and a signed consent form to keep, in case you wish to refer to it in the future.

125

Appendix IX

Participant Consent Form

126

Participant Consent Form

Traumatic Brain Injury in Adult Female Offenders in the UK

Name of lead researchers: Michelle O’ Sullivan & Steven Fitzsimons

Please initial the relevant boxes

1. I confirm that I have read or the form has been read to me and understand the information sheet for the above study. I have had the opportunity to consider the information, ask questions and have had those answered satisfactorily.

2. I understand that my participation is voluntary and that I am free to withdraw at any time without giving any reason.

3. I understand that data collected during the study may be looked at by individuals from the research team and from regulatory authorities where it is relevant to my taking part in this research. I give permission for these individuals to have access to my data.

4. I understand that my participation is voluntary and that I am free to withdraw at any time, without giving any reason, without my medical care or legal rights being affected.

5. I agree to take part in the above study.

6. I understand that if I tell the researchers of behaviour that is against prison rules and can be adjudicated against, illegal acts, and behaviour which could represent a risk to myself or others, then this will be reported to prison staff, who should record this in my notes.

Name of Participant Date Signature

________________________ __________________ _______________

I have explained the study to the participant and have answered their questions

honestly and fully

Researcher Date Signature

________________________ ________________

_______________

127

Appendix X

Participant Feedback Pro Forma

128

PARTICIPANT FEEDBACK

Date assessment completed:

Prisoner No. _________ participated in a research study which involved an assessment

of her mood and cognition. This is a standardised research protocol and has not been

designed to assess this individuals’ clinical needs. Assessment results suggest that this

participant may be experiencing difficulties in the domains highlighted below. The

assessment results have been explained to the participant and they have consented to

these results being recorded in their notes and highlighted to healthcare. Results that are

of particular concern are highlighted in yellow. These results indicate that the

participant would benefit from a more in-depth assessment of their clinical needs.

This research has been approved by the National Offender Management Service

National Research Committee, reference no.: 2013-266.

Beck Depression Inventory II

Total Score: (mild/moderate/severe depression)

Beck Anxiety Inventory

Total Score: (mild/moderate/severe anxiety)

The Impact of Events Scale

Total Score: (a score above 33 indicates that this participant may be at risk of

posttraumatic stress disorder)

129

Wechsler Abbreviated Scale of Intelligence II

Full Scale IQ: percentile (qualitative descriptor)

Perceptual Reasoning Index: percentile (qualitative descriptor)

Verbal Comprehension Index: percentile (qualitative descriptor)

Repeatable Battery for the Assessment of Neuropsychological Status

Immediate Memory Index: percentile (qualitative descriptor)

Visuospatial/constructional Index: percentile (qualitative descriptor)

Language Index: percentile (qualitative descriptor)

Attention Index: percentile (qualitative descriptor)

Delayed Memory Index: percentile (qualitative descriptor)

Total score: percentile (qualitative descriptor)

Behavioural Assessment of the Dysexecutive Syndrome

Classification:

Risk issues or other researcher observations:

130

Appendix XI

OASys Items Extracted

131

Item 1.1a category of index offence

Item 1.8 Age at first contact with police

Item 1.26 Number of violent previous offences (with current offence added if

also violent)

Item 1.7 Age at first conviction

Item 6.3 Experience of childhood abuse

Item 6.7 Victim of domestic violence

OGP probability of proven non-violent reoffending Year 1%

OVP probability of proven violent-type reoffending Year 1%

Severity of violent offences

Level of violence if robbery is index offence

132

Appendix XII

Correspondence Regarding Ethical Approval

133

134

135

Appendix XIII

Normality Data

136

Table 17. Normality Data for Reliability Variables

Variable Z score Interpretation

Skewness Kurtosis

BISI age at 1st TBI -.079 -.505 Normal

Total number of TBI .627 -.505 Normal

Longest LOC 3.645 2.700 Positive skew & positive kurtosis

TBI Time index 4.026 3.910 Positive skew & positive kurtosis

Repeat BISI Age at 1st TBI -.439 -.891 Normal

Repeat Total number of TBI -.065 -1.448 Normal

Repeat Longest LOC 4.166 3.945 Positive skew & positive kurtosis

Repeat TBI Time index 5.269 8.136 Positive skew & positive kurtosis

TBI Severity Index 1.671 .340 Normal

Repeat TBI Severity Index 2.900 3.840 Positive skew & positive kurtosis

BISI Global Score 1.020 -1.006 Normal

Repeat BISI Global Score 1.488 -.849 Normal

Total no. of episodes of

PTA

.952 -.589 Normal

BISI no. of episodes of

dizziness

.056 -.965 Normal

Repeat BISI total no. of

episodes of PTA

1.602 .024 Normal

Repeat BISI total no. of

dizziness episodes

.620 -1.154 Normal

137

Table 18. Normality Data for Standardised Measures

Variable TBI Group Non-TBI Group

Z score Interpretation Z score Interpretation

Skewness Kurtosis

Normal

Skewnes

s

Kurtosis

Age 0.427 -1.281 .493 -1.569 Normal

Education -.659 .713 Normal .045 -1.305 Normal

TOMM -6.007 11.262 Negative

skew &

positive

kurtosis

- - TOMM

value is

constant in

non-TBI

group

BAI .455 .523 Normal .775 -1.220 Normal

BDI .714 -.098 Normal -.246 -1.514 Normal

TOPF with

Demo

.570 -.436 Normal -.046 .024 Normal

TOPF raw .687 -.503 Normal -0.078 -0.434 Normal

RBANS

immediate

memory

-.083 0.109 Normal -1.682 0.287 Normal

RBANS

visuospatial

1.427 1.026 Normal .906 -.867 Normal

RBANS

language

.374 -.613 Normal 1.633 1.821 Normal

RBANS

attention

-.063 -1.148 Normal 2.392 2.709 Positive

skew &

138

positive

kurtosis

RBANS

delayed

memory

-1.556 .807 Normal -3.190 4.120 Negative

skew &

positive

kurtosis

RBANS total 0.345 .730 Normal -1.693 1.132 Normal

BADS age

corrected

-2.028 1.193 Negatively

skewed

-1.634 1.401 Normal

BADS raw -2.148 1.314 Negatively

skewed

-1.829 1.650 Normal

DEX .356 -.180 Normal .971 .007 Normal

IES -.393 -.230 Normal .675 .304 Normal

NFI Dep % -1.241 .403 Normal -.484 -1.376 Normal

NFI Dep T .097 0.350 Normal -.807 -.654 Normal

NFI Dep Raw .007 .263 Normal -.721 -.752 Normal

NFI Somatic

%

-.420 -1.345 Normal -.049 -.800 Normal

NFI Somatic

Raw

.659 -.870 Normal -.726 -.558 Normal

NFI Somatic T 1.214 -.154 Normal -.686 -.691 Normal

NFI Memory

%

.001 -1.203 Normal .581 -.865 Normal

NFI Memory T .001 -.554 Normal -.118 -.733 Normal

NFI Memory -.132 -.872 Normal -.804 .525 Normal

139

Raw

NFI

communication

%

-.636 -1.006 Normal 1.331 -.016 Normal

NFI

communication

T

.524 .125 Normal .935 -.320 Normal

NFI

communication

raw

.436 -.257 Normal .507 -.272 Normal

NFI aggression

%

.186 -1.114 Normal .437 -1.096 Normal

NFI aggression

T

.537 -.881 Normal .188 -1.086 Normal

NFI aggression

raw

1.232 -.560 Normal .575 -1.446 Normal

NFI motor % .413 -.804 Normal .508 .035 Normal

NFI motor T .395 -.549 Normal -.818 .858 Normal

NFI motor raw .404 -1.144 Normal -1.428 .954 Normal

WASI VC -.125 .490 Normal -.162 -1.437 Normal

WASI PR -1.008 -.299 Normal .104 .213 Normal

WASI FS4 .171 -.410 Normal .216 -.436 Normal

WASI-TOPF .021 -.224 Normal -.376 -.650 Normal

140

Figure 3. Normality Histogram for BISI Age of 1st TBI

141

Figure 4. Normality Histogram for BISI Total No. of TBI

142

Figure. Normality Histogram for BISI TBI Time Index Score

143

Figure 5. Normality Histogram for Repeat BISI Age of 1st TBI

144

Figure 6. Normality Histogram for Repeat BISI Longest LOC

145

Figure 7. Normality Histogram for Repeat BISI Total No. of TBI

146

Figure 8. Normality Histogram for Repeat BISI TBI Time Index Score

147

Figure 9. Normality Histogram for BISI TBI Severity Index Score

148

Figure 10. Normality Histogram for Repeat BISI TBI Severity Index Score

149

Figure 11. Normality Histogram for BISI Global Score

150

Figure 12. Normality Histogram for Repeat BISI Global Score

151

Figure 13. Normality Histogram for BISI Total No. of Episodes of PTA

152

Figure 14. Normality Histogram for BISI Total No. of Episodes of Disorientation

153

Figure 15. Normality Histogram for Repeat BISI Total No. of Episodes of PTA

154

Figure 16. Normality Histogram for Repeat BISI Total No. of Episodes of

Disorientation

155

Figure 17. Normality Histogram for WASI-TOPF with TBI group

156

Figure 18. Normality Histogram for WASI-TOPF with no TBI group

157

Figure 19. Normality Histogram for Age with TBI group

158

Figure 20. Normality Histogram for Age with no TBI group

159

Figure 21. Normality Histogram for Years in Education with TBI group

160

Figure 22. Normality Histogram for Years in Education with no TBI group

161

Figure 23. Normality Histogram for TOMM Trial 2 with TBI group

162

Figure 24. Normality Histogram for BAI with TBI group

163

Figure 25. Normality Histogram for BAI with no TBI group

164

Figure 26. Normality Histogram for BDI-II with TBI group

165

Figure 27. Normality Histogram for BDI-II with no TBI group

166

Figure 28. Normality Histogram for TOPF Raw with TBI group

167

Figure 29. Normality Histogram for TOPF Raw with no TBI group

168

Figure 30. Normality Histogram for RBANS Immediate Memory with TBI group

169

Figure 31. Normality Histogram for RBANS Immediate Memory with no TBI group

170

Figure 32. Normality Histogram for RBANS Visuospatial with TBI group

171

Figure 33. Normality Histogram for RBANS Visuospatial with no TBI group

172

Figure 34. Normality Histogram for RBANS Language with TBI group

173

Figure 35. Normality Histogram for RBANS Language with no TBI group

174

Figure 36. Normality Histogram for RBANS Attention with TBI group

175

Figure 37. Normality Histogram for RBANS Attention with no TBI group

176

Figure 38. Normality Histogram for RBANS Delayed Memory with TBI group

177

Figure 39. Normality Histogram for RBANS Delayed Memory with no TBI group

178

Figure 40. Normality Histogram for RBANS Total Score with TBI group

179

Figure 41. Normality Histogram for RBANS Total Score with no TBI group

180

Figure 42. Normality Histogram for IES-R with TBI group

181

Figure 43. Normality Histogram for IES-R with no TBI group

182

Figure 44. Normality Histogram for WASI Verbal Comprehension with TBI group

183

Figure 45. Normality Histogram for WASI Verbal Comprehension with no TBI group

184

Figure 46. Normality Histogram for WASI Perceptual Reasoning with TBI group

185

Figure 47. Normality Histogram for WASI Perceptual Reasoning with no TBI group

186

Figure 48. Normality Histogram for WASI Full Scale IQ with TBI group

187

Figure 48. Normality Histogram for WASI Full Scale IQ with no TBI group

188

Figure 49. Normality Histogram for NFI Depression Raw with TBI group

189

Figure 50. Normality Histogram for NFI Depression Raw with no TBI group

190

Figure 51. Normality Histogram for NFI Somatic Raw with TBI group

191

Figure 52. Normality Histogram for NFI Somatic Raw with no TBI group

192

Figure 53. Normality Histogram for NFI Memory Raw with TBI group

193

Figure 54. Normality Histogram for NFI Memory Raw with no TBI group

194

Figure 55. Normality Histogram for NFI Communication Raw with TBI group

195

Figure 56. Normality Histogram for NFI Communication Raw with no TBI group

196

Figure 57. Normality Histogram for NFI Aggression Raw with TBI group

197

Figure 58. Normality Histogram for NFI Aggression Raw with no TBI group

198

Figure 59. Normality Histogram for NFI Motor Raw with TBI group

199

Figure 60. Normality Histogram for NFI Motor Raw with no TBI group

200

Figure 61. Normality Histogram for BADS Total Score with TBI group

201

Figure 62. Normality Histogram for BADS Total Score with no TBI group

202

Major Research Project Proposal Form

This form should be completed by the trainee and signed by the University supervisor, and then submitted by the deadline.

Remember to give a draft to your supervisor for comments before submitting the final version.

When preparing this document it would be helpful to consider what you would include when writing the Introduction and Method

sections for your MRP.

* Please append your literature review to this proposal

URN: 6242697

Project Title: Traumatic Brain Injury in Adult Female Offenders in the UK

Introduction

Background and Theoretical Rationale

TBI: Definition and prevalence

Traumatic brain injury (TBI) is “an alteration in brain function, or other evidence of

brain pathology, caused by an external force” (Menon et al., 2010). TBI is the most

common form of acquired brain injury (ABI; Fleminger & Ponsford, 2005), with an

estimated prevalence of 8.5% in the general population (Silver et al., 2001) across all

levels of severity. Incidence ranges from 91-419 per 100,000 in England (Tennant,

2005).

Relationship between TBI and offending

Clinical opinion suggests that violence and impulsive behaviours are both antecedents

and consequences of TBI (Anderson et al., 1999). Violence following TBI has been

characterised as unpredictable, ill-directed, and can occur in the absence of clear

triggers or provocation (Eslinger et al., 1995; Wood & Liossi, 2006). Individuals with

TBI have a significantly increased risk of committing a violent crime (Fazel et al.,

2011). While TBI cannot be assumed to be the sole cause of offending, the cognitive

203

and behavioural sequelae of TBI may predispose some individuals (Brower & Price,

2001; deSouza, 2003; Greve et al., 2001; Kreutzer et al., 1995; Kreutzer et al., 1991;

Miller, 1999; Simpson et al., 1999).

Relationship between TBI in females and offending

A recent report commissioned by the Barrow Cadbury Trust emphasised the need for

research examining causes and consequences of TBI in female offenders specifically

(Williams, 2012). Shiroma, Ferguson and Picklesimer’s (2010a) study of TBI in US

prisoners revealed a male and female prevalence estimate of 64.41% (95% CI: 53.3 to

75.53%) and 69.98% (95% CI: 50.18-89.79%) respectively. The prevalence of TBI in

UK female offenders is currently unknown. A valid screen for TBI in female offenders

would facilitate research and support clinical practice. There are currently no validated

published screening tools for use with UK female offenders. The Brain Injury Screening

Index (BISI; Pitman et al., 2013) has been validated in male offenders in the UK, but

has yet to be extended to females.

Without adequate screening in female offenders, TBI is likely to go undetected, and

may impact on engagement in offender rehabilitation programs and the legal process

(Jackson & Hardy, 2011). Individuals with TBI may be more difficult to rehabilitate and

discharge (Hawley & Maden, 2003), with services ill-equipped to address their needs.

Research in this field is congruent with the Transforming Rehabilitation strategic

business priorities within the National Offender Management Service (NOMS; National

Offender Management Service, 2013) and may increase efficiency by informing

programs to reduce recidivism.

Officials within the criminal justice system may misinterpret behaviour of offenders

with TBI (Merbitz et al., 1995; Shiroma et al., 2010b). Research demonstrating

increased disciplinary incidents in prisoners with TBI (Merbitz et al., 1995; Morrell et

204

al., 1998; Shiroma et al., 2010b) suggests that they may have increased difficulty

adapting to prison life due to cognitive and behavioural sequelae. This has implications

for engagement in the legal process, prison management, and post-discharge and release

pathways (Jackson & Hardy, 2011). Under-identification is likely to perpetuate

inadequate resources, providing no incentive to fund appropriate interventions and

inefficient use of available resources. Screening is consistent with NOM’s Reducing

Prison Unit Costs strategic business priority by ensuring appropriate services are

commissioned and targeting the most appropriate offenders (NOMS, 2013), e.g.

screening can provide a cost-effective way of determining who is appropriate for

referral to more limited and expensive resources such as neuropsychologists, or

alternative care pathways.

Risk factors for offending in females with TBI

Females are reportedly less likely to offend than males, yet those who do are more

likely to be experiencing a mental illness (Butler et al., 2005; Fazel & Grann, 2006).

New-onset major depression post-TBI increases the risk of aggression for females

eightfold (Rao et al., 2009). PTSD may also be a significant risk factor (Johansson et

al., 2008). Comorbidities may moderate or mediate the relationship between TBI and

offending in females.

Physical and sexual abuse throughout the lifespan may be a risk factor (Brewer-Smyth

& Burgess, 2008; Brewer-Smyth, Burgess, & Shults, 2004; Shiroma et al., 2010b).

Abuse commences for female prisoners at a young age (Browne et al., 1999). Childhood

victimisation strongly predicts victimisation in adulthood (Browne et al., 1999), and

adult victimisation in turn increases the risk of TBI (Kwako et al., 2011), which may

lead to increased violent behaviour. This is consistent with Brewer-Smyth and Burgess’

205

(2008) findings that female prisoners with more TBIs and violent crime convictions

have increased childhood family sexual abuse.

Post-TBI aggression may be related to being a victim of intimate partner violence

specifically. Shiroma et al. (2010b) found a decreased likelihood of suffering a TBI

whilst incarcerated. Brewer-Smyth et al. (2004) found increased recency of abuse and

hospitalisations for abuse-related injuries in those with violent crime convictions, with

most TBIs occurring in the fronto-temporal region, which has been related to post-TBI

aggression in the literature (Daoust et al., 2006).This lends some tentative support for a

relationship between intimate partner violence and post-TBI aggression.

Brewer-Smyth et al. (2004) examined the cumulative effects of recurrent TBIs on

violence, finding evidence for a dose-response effect between number of TBIs and

violence. These findings are consistent with previous research which has found reported

prevalence rates of multiple TBIs in female offenders ranging from 35-48% (Ferguson

et al., 2012). Multiple mild TBIs can have similar cognitive and behavioural profiles to

individuals with more severe TBI (Diamond et al., 2007).

Research Questions

How prevalent is TBI and how does TBI present itself in the cognitive, psychiatric

and health needs of female prisoners in the UK?

Is self-reported TBI using the BISI associated with cognitive performance in

standardised questionnaires and neuropsychological tests?

Main Hypotheses

Female offenders in the UK demonstrate similar prevalence of TBI as males.

The BISI provides an estimate of prevalence within the confidence intervals of

studies using clinical interviewing.

206

Results of the BISI will be significantly associated with those obtained in the

standardised questionnaires and neuropsychological tests.

The BISI will have a test-retest reliability coefficient of at least 0.60.

Female offenders with self-reported TBI have significantly more cognitive,

psychiatric, and physical health difficulties than female offenders without TBI.

Female offenders with self-reported TBI have higher historical rates of recidivism

and behavioural infarctions that those without TBI.

Method

Participants

Participants will be recruited from prisons for women in southern England, with four

potential prison sites. These prisons have a combined operational capacity of

1,607women from which potential participants can be recruited.

To explore prevalence and severity of TBI in the UK (stage 1), using Daniel and Cross’

(2013) formula for sample size calculation for prevalence studies, on the basis of a level

of confidence of 95%, an expected prevalence of 66% in female offenders using the

gold standard of clinical interviews (Shiroma et al., 2010a), and a precision value of

0.125, a sample of 56 female offenders would be required for stage 1. A large precision

value was chosen due to feasibility related to resource limitations, and the preliminary

nature of this research. The precision value meets the assumption of normal

approximation. To assess the test re-test reliability of the BISI, based on a minimum

reliability of 0.6, an expected reliability of 0.8, α=0.05 and β=.20, an estimated sample

of 39 of the original 56 will be required (Walter, Eliasziw, & Donner, 1998).

To explore the differences in cognitive, psychiatric and health needs of female offenders

with TBI and those without TBI, as well as rates of recidivism and behavioural

infarctions, based on the number of variables being measured, an anticipated large effect

207

size, statistical power of .80 and type I error α of .003 (Bonferroni correction for 14

comparisons), a sub selection of 28 participants with TBI and 28 without TBI will be

required for stage 2.Based on similar research using the BISI with male prisoners

(Pitman et al., 2013) the estimated response rate is 73%, therefore a total of 76 prisoners

will be approached for consent.

Inclusion criteria include prisoners over 18, with an upper age limit of 80 years of age in

line with norms provided in the instruments to be used. Exclusion criteria include acute

symptoms of physical or mental illness or other indication that participants may not be

able to provide informed consent. This will be achieved by excluding participants in the

medical unit. Acute illness and ability to provide consent is also assessed in the clinical

interview, at which point the assessment will be terminated. Prisoners with a confirmed

diagnosis of dyslexia, are not fluent in English, or have reported acquiring a TBI in the

last 6 months, will be excluded from stage 2 due to validity limitations of measures.

Participants with a learning disability will be included unless queries regarding capacity

to consent are raised when are being briefed.

Design

Cross-sectional design using a semi-structured clinical interview, clinical

questionnaires, and neuropsychological measures.

Measures/Interviews/Stimuli/Apparatus

Please see Appendix I for the Participant Information Sheet and Appendix II for the

Participant Consent Form. For psychometric properties of measures please refer to

Table 1 Appendix III. The following assessment tools will be employed:

Semi-structured interview designed for use in male prisoners, which has been adapted to

extend to research to female prisoners (Appendix IV)

Pro forma for collecting data from participant files (Appendix V)

208

The Brain Injury Screening Index (BISI; Appendix VI)

The Wechsler Abbreviated Scale of Intelligence II (WASI-II; Wechsler & Zhou, 2011).

The Test of Premorbid Functioning – UK Version (TOPF; Wechsler, 2009)

The Behavioural Assessment of the Dysexecutive Syndrome (BADS; Wilson,

Alderman, Burgess, Emslie, & Evans, 1996)

The Repeatable Battery for the assessment of Neuropsychological Status (RBANS;

Randolph, 1998)

The Test of Memory Malingering (TOMM; Tombaugh, 1996)

Neurobehavioral Functioning Inventory (NFI; Kreutzer, Seel, & Marwitz, 1999)

The Beck Anxiety Inventory (BAI; Beck, Epstein, Brown, & Steer, 1988)

The Beck Depression Inventory II (BDI; Beck, Steer, & Brown, 1996)

The Impact of Events Scale – Revised (IES-R; Appendix VI; Weiss & Marmar, 1997)

Pro formas are not provided in the appendices for assessments restricted due to

copyright.

Procedure

1. Participants will be recruited through a convenience sample. A list of prisoners who

have been new receptions over a 1-2 month period (depending on average intake

rate- until 76 participants are obtained for sufficient power and response rate

considerations) will be provided by prison staff. Prison staff will act as gatekeepers

and be asked to identify participants under 18, over 80, admitted to the medical unit,

or who cannot provide informed consent.

209

2. Eligible participants will be provided with a copy of the participant information

sheet and consent form, i.e. a copy will be dropped into their cells by a member of

the research team.

3. Within a week of providing the information sheets prisoners will be contacted face-

to-face by a member of the research team to discuss the participation and consent. A

suitable time to commence assessment will be arranged with those who agree to

participate.

4. Participants will be allocated a participant number, which will be recorded with their

prison reference number, and recorded on an encrypted device. This will be separate

from the main data collection file.

5. At assessment, participants will complete the BISI and clinical interview (stage 1).

This will take approximately 1 hour. Participants will be provided with the clinical

questionnaires in the interview and asked to complete them in their own time. These

will then be collected at the end of each day of data collection. Participants will

complete all questionnaires themselves. Data on social history, abuse history,

clinical history, offence history, and behavioural infractions, will be obtained for

these participants from Offender Assessment System (OASys), probation and re-

offending records, and individual history files.

6. Participants will have the option of progressing on to stage 2 straight after stage 1,

or arranging a time to continue with the assessment.

7. Stage 2, consisting of the battery of neuropsychological measures, will occur at an

arranged time, with an expected sample of 28 participants with TBI and 28 without

a history of TBI. This will take 1½-2 hours.

210

8. Participants who progress from stage 1 to stage 2 within 1-2 weeks of completing

the BISI will be asked to complete the BISI again for the purpose of test-retest

validity, until the required sample of 39 is reached.

9. As data is gathered it will be encoded in an SPSS data file stored on an encrypted

device for analysis.

10. Once data collection is completed appointments will be arranged with participants

who request an assessment feedback session. Participants may also request to have

results shared with the prison/health service. A brief report with results will be

provided.

Ethical considerations

1. Great effort will be made to clarify the nature of the relationship between the

researchers conducting assessments and the participant; i.e. it is not an assessment

from a clinical referral, that the primary purpose is research. However, participants

will have the opportunity to have a feedback session after the data collection is

completed and/or to have results shared with the prison/health services. Feedback

will be primarily descriptive and can be used to indicate if support from health

services is required.

2. Feedback on neuropsychological tests may impact on self-esteem, as mean prison

IQ in the UK is 88 (±12.0) (Hayes, Shackell, Mottram, & Lancaster, 2007). It is

hoped that the provision of information and training to staff and the potential for

additional support following the research will enable participants to feel

empowered, with the resources to manage difficulties.

3. Participants may assume that test results can be used for secondary gain, however

when briefing the participants it will be emphasised that tests will not be for clinical

211

use, and only a recommendation for a comprehensive clinical examination can be

made. The TOMM (Tombaugh, 1996) will also be used to assess effort.

4. Due to the time required of participants motivation and fatigue may invalidate

results. To manage this, participants can complete stages on different days if they

wish, or have a brief break during testing if requested.

5. To ensure test validity the participating prisons will be requested to provide a quiet

environment, with adequate space, ventilation, lighting and furniture. We will also

liaise with staff to try to minimise potential interruptions during testing.

6. Information that is related to risk to self or others, including drug use and abuse,

may be disclosed. The research team will liaise with participating prisons to ensure

that the research protocol is congruent with prison policy to manage risk. Potential

limits to confidentiality will be made explicit to participants when obtaining

consent.

Name of Ethics Committee: University of Surrey Ethics Committee: Faculty of Arts and

Human Sciences; National Offender Management Service............................

R&D Considerations

Name of R&D department: R&D approval is not required however a letter of support

from governors of participating prisons will be obtained.

Proposed Data Analysis

All analyses will be done using IBM SPSS version 20 (IBM, 2011). Data will be

checked for normality of distribution and homogeneity of variance.

Descriptive statistics, including age, proportion of participants with a TBI, ethnicity,

education, learning disability, special support at school, diagnosis of mental health

problem, history of self-harm, suicide attempts, frequency of drug and alcohol use, age

212

at first offence, number of offences, number of violent convictions, number of

behavioural infractions, participation in rehabilitation programs, and years spent in

custody, will be provided for those with an identified TBI and those without. This will

include means and standard deviations, and t-tests comparing those with and without a

TBI where appropriate. Further descriptive statistics will be provided for those with a

TBI, namely: number of TBIs, age at earliest TBI, loss of consciousness, cause of

injury, and help-seeking behaviour after TBI.

Concurrent validity of the BISI will be studied by calculating Pearson’s r between the

BISI and clinical interview. Receiver Operating Curve analysis will also be conducted

to assess sensitivity and specificity rates against clinical interview as the gold standard.

Test-retest reliability will be assessed using Pearson’s r.

Between samples t-tests will explore differences participants with TBI and without TBI

on cognitive, psychiatric and physical health measures.

Service User and Carer Consultation / Involvement

Service user and carer consultation was undertaken on 6th August 2013 and informed

this proposal (see Appendix VI for feedback).

Feasibility Issues

As the study is highly dependent on the amenability of the prison involved, there

must be organisational incentive for participation. After the study is complete,

the research team will provide a training and information session with prison

staff to support them in their roles when working with female offenders with

TBI.

Prisons participating must be able to provide adequate support in the event that

participants are distressed after the study. To overcome this we are liaising with

potential prisons and will prioritise the research site accordingly. We are also

213

aiming to obtain sensitive data on abuse history from the OASys to reduce the

risk of distress.

Cost of pro formas for the copyrighted measures used is high (totalling

£1418.40), and exceeds the research budget provided by the University of

Surrey. The Disabilities Trust have agreed to cover the cost of pro formas

exceeding the research budget.

Dissemination strategy

Analysed data will be written up in a thesis as part of a doctorate, and disseminated

through research articles (e.g. The Journal of Head Trauma; Brain Injury) and

conferences (The Annual Division of Forensic Psychology conference; International

Association of Forensic Mental Health conference). A briefing can also be provided for

the prison’s newsletter. The Disabilities Trust Foundation will publish results through

reports, presentations, journal articles, and on their website.

Study Timeline

Please see Gantt chart in Appendix VII.

214

PSYCHD CLINICAL PSYCHOLOGY

MRP Literature Review

Title: ‘Traumatic brain injury and violent behaviour in females:

A systematic review’

Student URN: 6242697

April 2013

Word count: 7,997

215

Abstract

Research has demonstrated that individuals with traumatic brain injury (TBI) have a

significantly increased risk of committing a violent crime. Whether and how this increased

risk of violent behaviour may present itself in females is not clear. Given the heterogeneity

of study design and sample populations in the literature, a systematic review aiming at

synthesising the literature on the relationship between TBI and violence in females is

presented. Three main psychological and medical databases were searched (PsychINFO,

Scopus, and PubMed). Six studies met the inclusion criteria. Results across these studies were

inconsistent, and due to methodological limitations and scarcity of research, it was concluded

that there is insufficient evidence to determine if there is a relationship between TBI and

violence in females. However, the studies reviewed contributed knowledge of other factors

that may influence this relationship, including psychiatric comorbidities and childhood abuse.

Implications for future research and clinical practice are discussed.

1. Introduction

1.1. Traumatic Brain Injury

Traumatic brain injury (TBI) is defined as “an alteration in brain function, or other evidence

of brain pathology, caused by an external force” (Menon, Schwab, Wright, & Maas, 2010),

capturing the range of presentations which fit under the TBI diagnostic umbrella, including

loss of or decreased consciousness, any loss of memory, neurological deficits, and any

alteration in mental state e.g. confusion (Menon et al., 2010). TBI is the most common form

of acquired brain injury (Fleminger & Ponsford, 2005) (ABI), with an estimated prevalence

of 8.5% (Silver, Kramer, Greenwald, & Weissman, 2001) across all levels of severity.

Annual incidence of TBI ranges from 180-250 per 100,000 in the United States (US; Bruns &

Hauser, 2003), and 91-419 per 100,000 in England (Tennant, 2005); however, rates may

overlook milder TBI due to reliance on medical records (Tennant, 2005) and associated

diagnostic and selection biases (Feigin et al., 2013). TBI severity traditionally has been

classified by scores on the Glasgow Coma Scale (GCS; World Health Organization, 2006).

Other commonly used measures include post-traumatic amnesia (PTA) and length of loss of

consciousness (LOC; Sherer, Struchen, Yablon, Wang, & Nick, 2008). Table 1 Appendix I

summarises typical cut-offs used for differentiating mild, moderate and severe TBI. Many

research studies do not adhere to this definition and classification systems for TBI, making

comparison across studies difficult (Corrigan, Selassie, & Orman, 2010), highlighting the

need for systematic reviews to lend coherence to the literature.

Up to twice the rate of TBI has been found in males than females in the general population

(Hillbom & Holm, 1986; Hirtz et al., 2007). This apparent protective effect of female gender

appears attenuated in specific populations, including those with substance use disorder

(Felde, Westermeyer, & Thuras, 2006) and prisoners (Brain Injury Association of Wyoming,

2008; Ferguson, Pickelsimer, Corrigan, Bogner, & Wald, 2012). Risk of TBI in females

appears to be more pronounced with milder TBI, e.g. Diamond, Harzke, Magalett, Cummins

and Frankowski (Diamond, Harzke, Magaletta, Cummins, & Frankowski, 2007) found that

54.7% of females in a prison population self-reported TBI with no LOC in comparison to

40% of males. This fell to 35.6% of females in comparison to 47.8% of males with LOC of

less than one hour. Differences may be attributable to different gender-related behavioural

patterns, as well as decreased likelihood of reporting mild TBIs. TBI in females may be

underestimated due to unreported experiences of intimate partner violence (Valera &

Berenbaum, 2003). Due to the nature of intimate partner violence it is difficult to get an

accurate prevalence rate of TBI in this population, however reported rates vary from 30-74%

(Kwako et al., 2011). Meta-analysis has found that females tend to have worse outcomes after

TBI than males across all TBI types (Farace & Alves, 2000). Although reviews and meta-

analyses have examined some gender-related factors, e.g. intimate partner violence (Kwako

217

et al., 2011) and outcomes (Farace & Alves, 2000), there are an abundance of issues such as

violence in females with TBI which have yet to be explored systematically.

1.2. TBI and violence

For the purpose of this review, the definition of violence is confined to physical acts that

could cause harm to others. This maintains consistency with the forensic mental health

literature (Monahan et al., 2001). For context, this review draws upon the heterogeneous

literature available, examining related concepts such as verbal aggression and general

offending, in the introduction and discussion.

TBI is a complex condition that can result in an array of cognitive, emotional, physical and

behavioural sequelae. Clinical opinion suggests that violence and impulsive behaviours are

both antecedents and consequences of TBI (Anderson, Bechara, Damasio, Tranel, &

Damasio, 1999). TBI has been associated with increased risk of developing comorbid

conditions (Timonen et al., 2002), including substance abuse (Honer et al., 2005), epilepsy

(Ferguson et al., 2010), psychoses (McAllister, 1998), post-traumatic stress disorder (PTSD;

Rogers & Read, 2007) and aggression (Alderman, 2007; Cole et al., 2008; Rao et al., 2009;

Visscher, van Meijel, Stolker, Wiersma, & Nijman, 2011). Females may be particularly at

risk for developing depression (Whelan-Goodinson, Ponsford, Schönberger, & Johnston,

2010) and anxiety (Hibbard, Uysal, Kepler, Bogdany, & Silver, 1998) post-TBI. Aggression

and violence following TBI has been characterised as unpredictable, ill-directed, and can

occur in the absence of clear triggers or provocation (Eslinger, Grattan, & Geder, 1995;

Wood & Liossi, 2006). Research from the Swedish population registers, the only population-

based cohort study of its kind, found that individuals with TBI have a significantly increased

risk of committing a violent crime (Fazel, Lichtenstein, Grann, & Långström, 2011). While

TBI cannot be assumed to be the sole cause of violence, it does suggest that the cognitive and

behavioural sequelae of TBI may predispose some individuals to violence (Miller, 1999).

218

Over a decade ago the Aspen Neurobehavioural Conference (Filley et al., 2001) released a

consensus statement that epidemiological research of risk factors needed to be a research

priority. This review aims to further our understanding of the epidemiology and outcomes of

TBI and how it may relate to violence.

Research suggests that aggression is specifically related to the temporal and frontal lobes

(Daoust, Loper, Magaletta, & Diamond, 2006). While clinical profiles of TBI can be

variable, mild and moderate TBI, which victims of intimate partner violence may be more

susceptible to, are often localised in the orbito-frontal and temporal polar zones (Zappalà,

Thiebaut de Schotten, & Eslinger, 2012). Meta-analyses have demonstrated a medium effect

size for the relationship between antisocial behaviour and neuropsychological measures of

executive functioning (Morgan & Lilienfeld, 2000; Ogilvie, Stewart, Chan, & Shum, 2011),

however this area of neuropsychological research is marred by methodological problems,

rendering findings inconclusive (Morgan & Lilienfeld, 2000; Ogilvie et al., 2011). Similarly,

comparing aggressive with non-aggressive individuals with severe TBI, some research has

demonstrated no significant cognitive differences (Greve et al., 2001). However, another

study of severely injured individuals found significant deficits in verbal memory and visuo-

perceptual skills in the aggressive group (Wood & Liossi, 2006).

Without successful early intervention, individuals with TBI may pose a risk to others and

become involved in the criminal justice system (Brower & Price, 2001; deSouza, 2003;

Greve et al., 2001; Kreutzer, Marwitz, & Witol, 1995; Kreutzer, Wehman, Harris, Burns, &

Young, 1991; Miller, 1999; Simpson, Blaszczynski, & Hodgkinson, 1999). Once there,

individuals with TBI may be more difficult to rehabilitate and discharge, with services ill-

equipped to address their needs. Hawley and Maden’s (2003) study of TBI in medium secure

units (MSUs) indicated that 41.6% of service users had a history of TBI, and were

significantly more difficult to discharge into the community due to perceived greater risk of

219

violence to others and self-harm. Research demonstrating increased disciplinary incidents in

prisoners with TBI (Merbitz, Jain, Good, & Jain, 1995; Morrell, Merbitz, Jain, & Jain, 1998;

Shiroma et al., 2010b) suggests that they may also have increased difficulty adapting to

prison life due to cognitive and behavioural sequelae. This has implications for engagement

in the legal process, prison management, and post-discharge and release pathways (Jackson &

Hardy, 2011). Due to inadequate screening and identification of TBI, services are unable to

provide adapted rehabilitation for this population. Under-identification is likely to perpetuate

inadequate resources, providing no incentive to fund appropriate interventions. Evidence of

publication bias in studies on TBI and violence has also been found (Fazel, Philipson,

Gardiner, Merritt, & Grann, 2009), highlighting the difficulties of conducting research and

applying evidence-based practice.

Recent research in male populations in this area is leading to increased resources, e.g. the

placement of a prison brain injury linkworker for individuals with TBI (Pitman, Haddlesey,

Ramos, Oddy, & Fortescue, 2013). The next step for research and practice is to extend these

developments to a female population.

1.3. Females and violence

As of December 2012 there were 79,837 males and 3,920 females imprisoned in the UK

(Ministry of Justice, 2013a). In 2011, 34% of females and 31% of males were arrested for

violence against the person, as well as 26.7% and 28.5% of incarcerated females and males

respectively (Ministry of Justice, 2011). The similarity in prevalence across gender

challenges the stereotype that violence in prisoners is a male issue. A total of 9,018 females

were released between July 2011 and July 2012 (Ministry of Justice, 2013a), demonstrating

high turnover. 17.7% of females reoffend within twelve months of release (Ministry of

Justice, 2013b). Despite these figures, female offenders are a relatively understudied

population, with a research gender bias favouring males.

220

The literature has recently attempted to generate models and theories to understand female

offending and recidivism. This has ranged from the exploration of the utility of gender

neutral theories of criminal behaviour (Rettinger & Andrews, 2010), to feminist models

which subscribe to the idea that gender is key to female offending (Blanchette & Brown,

2006). While factors such as self-regulation, impulse control; and the role of intimate partner

violence and victimisation in the case of feminist models; are discussed in such theories

(Rettinger & Andrews, 2010), none address the role TBI may play.

1.4. TBI in females and violence

No large community-based epidemiological studies explore the relationship between violence

and TBI in females. Unfortunately, the 35-year Swedish population study only controlled for

gender through matching, rather than including it in the stratified analyses as they did with

age, severity and diagnostic sub-group (Fazel et al., 2011). Services are ill-informed to serve

female populations. This is particularly risky considering the reported gender differences in

the presentation of offenders, such as in psychiatric comorbidity (Zlotnick et al., 2008), as

well as the dominance of socio-economic and child-raising risk factors for females and

parental characteristics for males (Farrington, Painter, & Britain, 2004). A recent report

commissioned by the Barrow Cadbury Trust emphasised this need for research examining the

causes and consequences of TBI in female offenders specifically (Williams, 2012).

Although females have been reported to be less likely to offend than males, those who do, are

more likely to be experiencing a mental illness (Butler, Allnutt, Cain, Owens, & Muller,

2005). Females with a mental illness aged 25-39 demonstrate a population attributable risk-

fraction of 14% in comparison to 6.7% of males (Fazel & Grann, 2006). This increases to

19% over 40 years for females and 7.3% for males (Fazel & Grann, 2006). Rao et al.’s (2009)

study comparing individuals with and without verbal aggression post-TBI found that new-

onset major depression increased the risk of aggression for females with a TBI eightfold. The

221

issue of circular causality in the relationship between TBI and violent behaviours has been

raised before (Timonen et al., 2002), in that there are multiple associations between violence

and TBI and an array of other variables such as psychiatric comorbidities, making it difficult

to establish a clear causal pathway.

Prevalence of TBI in general offender populations varies, but a recent meta-analysis places it

at 60.25% (95% CI: 48-72%) (Shiroma, Ferguson, & Pickelsimer, 2010a). Shiroma, Ferguson

and Picklesimer (2010a) revealed a male and female prevalence estimate of 64.41% (95% CI:

53.3 to 75.53%) and 69.98% (95% CI: 50.18-89.79%) respectively. Once TBI definition was

limited to LOC, excluding milder TBIs, males demonstrated a higher prevalence than females

(59.31% vs. 55.28%). This decrease in prevalence once limited by LOC provides further

support for prevalence of mild TBIs in females, which as mentioned previously may result

from intimate partner violence. Prevalence of TBI in service users in secure forensic

psychiatric units is under-researched. Rates vary from 22% in the US (Martell, 1992) and

Canada (Colantonio, Stamenova, Abramowitz, Clarke, & Christensen, 2007), to 41.6% in a

UK MSU (Hawley & Maden, 2003). None of these studies examined females specifically,

most likely due to the small proportion of females. The Canadian study (Colantonio et al.,

2007) found evidence of TBI in 10% of females and 28% in males, while Hawley and Maden

(2003) found 35% in females and 43% in males. Difference in prevalence rates between

prisons and MSUs are surprising in light of the increased risk of comborbidity, particularly

depression in females post-TBI (Whelan-Goodinson et al., 2010). Unfortunately these studies

do not separate violent from non-violent offences, making it impossible to explore the impact

of TBI on violent offending specifically.

Research examining a more heterogeneous population of ABI, as opposed to TBI, suggests

female offenders with cognitive difficulties may require different management strategies

(Jackson & Hardy, 2011). Female prisoners with ABI demonstrate significantly different

222

cognitive impairments post-ABI than males, with females performing worse on perceptual

and spatial ability, complex visual memory and spatial working memory (Jackson & Hardy,

2011). In comparison to females without an ABI, they demonstrated cognitive difficulties,

including significantly worse performance on tests of perceptual intellectual and executive

functions, complex processing speed, and working memory (Jackson & Hardy, 2011). As

violent and non-violent offenders were not separated it is not possible to discuss the

neuropsychological relationship to violent behaviour specifically.

As demonstrated in the studies discussed, a plethora of possible mediators and moderators

appear to blur the causal pathway between TBI and violence in females. Research does not

appear to have begun systematically investigating this. This is confounded by differing

definitions of TBI and violence, populations sampled, insufficient investigation of females

specifically, and failure to present gender specific data. The literature needs to be synthesised

in a coherent manner to provide direction for future research. Research is needed to inform

prison services and clinical practice, thereby improving outcomes not only for individuals

with TBI but also reduce risk to others. This systematic review aims to answer the question:

what is the empirical evidence for a relationship between TBI in females and violent

behaviour?

2. Method

2.1. Selection of studies

A literature search of three psychological and medical electronic bibliographic databases was

conducted, namely PsycINFO, PubMed, and Scopus. Date range searched within the

databases was from first available (PsycINFO 1806; PubMed 1948; 1823 Scopus) to

February 2013. Title search terms included those pertaining to TBI (i) “brain injury” OR

“head injury”; violent behaviour (ii) viole* OR offend* OR forensic OR aggress* OR

prison*. Titles, abstracts and key words were searched for terms pertaining to female sex (iii)

223

sex OR gender OR female OR women. The database search yielded a total of 77 papers after

duplicates were removed. This was supplemented with searching of article reference lists to

ensure no relevant studies were missed, yielding a further 76 articles. Figure 1 outlines the

flow of studies through the search process.

Abstracts and titles of identified studies were read to determine if they met the following

inclusion criteria:

Original empirical published research presenting data relating to the

association between physical violence, physical aggression, or violent crime

and TBI in female populations.

Verbal threats and aggression were included for inpatient studies only, as these

behaviours may be unreliable and difficult to measure outside inpatient

settings (Monahan et al., 2001). Verbal incidents in inpatient settings may

represent an increased risk of physical violence and demonstrate the same

underlying mechanisms of physical violence due to the restrictive nature of the

environment which prevents verbal incidents from escalating.

The article was written in English

All systematic methodological approaches were included due to the limited

volume of research in this area.

Studies of populations under eighteen years were included but examined

separately from studies of adults.

Mixed-sex studies were included if female data was separated for relevant TBI

and violence variables.

Studies with anger, impulsivity, or verbal aggression specifically as outcomes outside

inpatient settings were excluded to maintain congruency with other research in forensic

mental health (Chambers et al., 2009; Fazel et al., 2009; Monahan et al., 2001; Walsh,

224

Buchanan, & Fahy, 2002), maintaining focus on physical violence. Studies investigating

women as victims only, rather than as perpetrators of violence were excluded. Only data

presented in the papers were extracted, as failure to report gender-specific data is evidence of

the gender bias and selective reporting in this area and a limitation of much of the published

studies.

This process yielded 6 empirical papers which are presented in the following sections: cross-

sectional studies investigating the relationship between violence and TBI in females; case-

control studies comparing groups of females with violent behaviours with controls; and

longitudinal studies of violence and TBI in females. To explore how these studies have

contributed to key research questions, sub-sections discussing the strength of the relationship

between violence and TBI in females; and where on the causal pathway variables may lie in

exploring the relationship between violence and TBI in women, are presented.

2.2. Quality assessment

Due to the non-experimental nature of the data, and the absence of a preferred tool for

evaluating such research (Jarde, Losilla, & Vives, 2012), an adaptation of the method

employed by King et al. (2008) for assessing the quality of non-experimental studies, based

on the Cochrane Handbook’s general guidance on non-experimental studies, was used to

determine quality indicators (2 indicating higher quality than 1) was employed (appendix I,

table 2). Factors examined were: sampling (non-random = 1, random = 2); representativeness

(response rates <60% = 1, ≥60% = 2); population definition (selected sample e.g. prison = 1,

general population = 2); and sample size (<100 females with TBI = 1, >100 females with TBI

= 2).

2.3. Data analysis

225

Due to the heterogeneity of the available research designs and quality, it was neither feasible

nor appropriate to conduct a formal meta-analysis (The Cochrane Collaboration, 2011).

Therefore, a narrative systematic review was conducted.

Figure 1. Selection of studies using PRISMA guidelines

3. Results

3.1. Sample

As demonstrated in figure 1, 109 titles were retrieved from database searches. In addition, 76

titles were obtained through reference checking. After de-duplication, 153 titles and abstracts

remained and were screened for relevance. This phase demonstrated the poor gender

226

reporting amongst studies, with many abstracts failing to report sample sex. Sixty-one studies

were excluded during the title and abstract screening. 84 potentially relevant studies were

selected for further examination. 6 articles could not be retrieved. 78 did not meet the

inclusion criteria. The most common reason for exclusion (n=54) was having male-only

samples, or failing to separate their data by gender for relevant variables, demonstrating the

substantial male gender bias. Other papers did not provide violence specific outcomes, e.g.

they presented variables that did not separate physical violence from other forms, or did not

separate violent offenders from non-violent (n = 17). Inpatient studies that provided gender

separated data did not measure participants consistently as inpatients, e.g. they were

inpatients at the first time point but not the second (n = 2). 5 studies were excluded because

they did not examine TBI specifically, e.g. examining ABI rather than TBI. This process

yielded 6 studies, of which 2 were cross-sectional designs, 1 was a case-control, and 3 were

longitudinal. Quality criteria for the selected studies are presented in table 2 (appendix I),

demonstrating the paucity of research conducted in females with head injury, with no study

obtaining a female TBI sample of over 100.

The following sections will discuss findings for each of the methodological designs regarding

the strength of the relationship between violence and TBI in females; as well as possible

confounders, moderators and mediators that may sit on the causal pathway.

3.2. Cross-sectional studies

3.2.1. Strength of the relationship

Of the two cross-sectional studies, one was conducted with a prison population (Brewer-

Smyth & Burgess, 2008) and the other within a clinical population (Johansson, Jamora, Ruff,

& Pack, 2008) (Table 3, Appendix I). Both studies were conducted within the USA.

Johansson et al. (2008) recruited consecutive male and female patients from an outpatient

neuropsychology service, to explore factors associated with aggression in TBI, including

227

gender in their analysis. Aggression included physical aggression, and acted as a surrogate

measure of violence. Generalisability and representativeness is limited. By recruiting from an

outpatient neuropsychology service milder TBIs with less pronounced sequelae are likely to

be excluded. Most participants were also in litigation, which can distort data e.g. by

increasing anxiety or malingering (King, 1997). Johansson et al. (2008) did not address pre-

morbid socioeconomic status of participants. Due to the nature of the healthcare system in the

USA access to outpatient neuropsychology services is likely dictated by expensive private

health insurance, thereby excluding vulnerable (Oddy, Moir, Fortescue, & Chadwick, 2012)

socio-economically deprived demographics. Participants with inadequate English

proficiency, and pre-morbid neurological and/or psychiatric history, were also excluded,

limiting ecological validity. With a sample of 27 females, and no reference to power

calculations for their research questions, it is unclear whether the sample was sufficient to

detect significant results, despite adequate available literature when looking beyond female

specific research (Wood & Liossi, 2006). The study lacked a non-TBI control group. TBI

severity was defined using the American Congress of Rehabilitation Medicine’s (ACRM)

guidelines (American Congress of Rehabilitation Medicine, 1993). They also supplemented

this with neuro-imaging. To measure aggression, Johansson et al. (2008) created ordinal

groups of participants by clinical ratings on a 4-point scale, with level 4 indicating overt

physical aggression. This clinical rating was validated against the anger subscale of the Ruff

Neurobehavioral Inventory (Ruff & Hibbard, 2003) (RNBI). Self-report scales were

corroborated with family reports, but this was only possible in 16% of cases. The authors’

reported clinical impressions were that many participants were under-reporting aggression

due to embarrassment and stigma. Cross-sectional designs which rely on self-report are

limited by recall bias. This has been reported in non-TBI studies (Houtveen & Oei, 2007),

and may be more pronounced in individuals with cognitive impairment. 40% of females and

228

32.5% of males reported overt physical aggression, however no statistically significant

relationship was found between gender and physical aggression (r = 0.141, p > 0.05). As

individuals with prior neurological histories were excluded, the impact of multiple TBIs on

aggression could not be examined. Johansson et al.’s (2008) research suggests that females

demonstrate similar levels of post-TBI physical aggression as males.

Brewer-Smyth and Burgess (2008) randomly recruited participants from minimum and

maximum security units of a women’s prison to examine if childhood familial sexual abuse

was related to increased neurological histories, including TBI. Brewer-Smyth and Burgess

(2008) described the study design as cross-sectional in their abstract and as a “modified case-

control design” (p.167) in the methods section. One of the defining characteristics of a case-

control study is that cases and controls are matched on the basis of outcome (Mann, 2003),

whereas Brewer-Smyth and Burgess (2008) matched on exposure to familial childhood

sexual abuse, making their design cross-sectional. Such inconsistent reporting of study

designs in the literature is problematic. Although this study is limited in generalisability due

to the prison sample, it provides a useful overview of a vulnerable population, with reports of

as many as 75% of incarcerated females experiencing severe physical violence by partners

(Browne, Miller, & Maguin, 1999). Although a power calculation was not completed for this

research question as it was a secondary analysis, power analysis for the same variables but

with different variables as exposures indicated that the sample (n= 149) was sufficient to

detect an effect. Brewer-Smyth and Burgess (2008) included non-criminal females in the

analysis, however this group was too small. TBI was dichotomously defined as physical head

trauma resulting in LOC, which risks excluding mild recurrent TBIs and did not examine TBI

severity. As demonstrated by Browne (1999), while 75% of incarcerated females experienced

physical violence, only 22% reported concussions, indicating the risk of missing more subtle

neurological impairment. Brewer-Smyth and Burgess (2008) corroborated reports with

229

criminal and medical records where possible, as well as physical examination. Definition of

violence was that of violent crime as defined by the criminal justice system. While this is a

common outcome measure in violence research, it is likely an underestimate of violent

behaviour as it relies on an individual receiving a conviction. It may also be problematic

when comparing across jurisdictions. Brewer-Smyth and Burgess’s (2008) results suggest

that females who experienced familial childhood abuse had more TBIs (OR = 1.49, p = 0.01)

and were convicted of more violent crime (OR = 1.67, p = 0.05) than those who were not

abused. However, it was not reported whether TBI occurred before or after the abuse,

blurring the causal pathway. This will be discussed further in the next section.

Clear conclusions from these two cross-sectional studies on the relationship between TBI in

females and violence cannot be drawn. While Johansson et al.’s (2008) study demonstrated

higher levels of post-TBI physical aggression in females, this was not statistically significant

(r = 0.14, p > 0.05). This may be attributable to methodological flaws and being

underpowered. Brewer-Smyth and Burgess (2008) found a significant relationship between

childhood familial abuse, TBI and violence (OR = 1.67, p = 0.05), however it suggests that

the relationship is more complex than TBI causing violence.

3.2.2. The causal pathway

While it is not possible to determine causality in cross-sectional designs, they can provide

insight into variables that may impact the causal pathway. Johansson et al. (2008) found that

groups with higher aggression demonstrated significantly elevated depression and PTSD on

subscales of the RNBI, which in a larger sample may have contributed to the relationship

between gender and post-TBI violence. In the Brewer-Smyth and Burgess (2008) study it was

unclear whether TBI was independently related to violence, but rather childhood familial

abuse may be a confounder, a moderator or a mediator between TBI and violence. Or that

230

TBI may be a confounder, moderator or mediator between childhood abuse and violence.

This was not explored within the study, but lends itself to consideration for future studies.

3.3. Case-control studies

3.3.1. Strength of the relationship

Brewer-Smyth, Burgess and Shults’(2004) study of females from minimum and maximum

security sections of a women’s prison in the USA, was the only case-control study which met

inclusion criteria (table 4, appendix I). Their objective was to examine the relationship

between basal resting salivary cortisol levels, abuse history, neurological history, and violent

crime. This data appears to be from the same project as the Brewer-Smyth and Burgess

(2008) cross-sectional study. Cases and controls were selected on the basis of the outcome of

committing a violent crime. The cross-sectional design (Brewer-Smyth & Burgess, 2008)

stated that participants were randomly recruited, this is unclear from the case-control design,

and it appears to be a convenience sample. The sample (n = 113) had sufficient power to

detect a significant effect. Definition of violence was that of violent crime as defined by

previous research, including murder, non-negligent manslaughter, robbery, assault, sexual

assault, and kidnapping. TBI was dichotomously defined as physical head trauma resulting in

LOC. “Other neurologic history”, consisting of TBI without LOC and sensory abnormalities,

was not significantly different between violent and non-violent offenders (48% and 53%

respectively, p > .05). There was no significant difference between violent (4%) and non-

violent (6%) offenders for severe brain injury, defined as a coma exceeding one day (p > .05).

TBI with LOC was significantly higher in participants convicted of a violent crime (56% vs.

38%), with a dose response effect in that for every additional TBI with LOC the odds of

being convicted of a violent crime in comparison to a non-violent crime increased

significantly (OR = 1.451, p = 0.012). Mean number of TBI with LOC for those with violent

crime conviction was 1.75 (± 2.9) in comparison to .74 (± 1.19) convicted of a non-violent

231

crime (p< .05). Most TBIs were acquired through violence perpetrated against the

participants or through high-risk behaviours such as substance use – the authors do not

provide figures for this. Morning cortisol levels, which the authors posit is related to chronic

stress associated with emotional and physical trauma, were significantly lower for those with

a violent conviction (mean .297 ± .23 vs .446 ± .47, p ≤ .055; OR = 0.036, p = 0.17).

Brewer-Smyth et al. (2004) also compared those with no known violent conviction, with a

current non-violent conviction but known past violent conviction, and those with current

violent convictions. Those with current violent convictions had significantly (1.75 ± 2.96, p <

.05) more TBIs per person, while those with past but not current violent convictions had less

(.61 ± 1.17, p < .05) than those with no known violent convictions (.86 ± 1.23, p < .05). This

was attributed to poor recording of criminal behaviour. It is impossible to ascertain from the

design whether TBIs occurred prior to violent behaviour, or whether those who engage in

violent behaviour are more at risk of TBI. Neurologic examination for evidence of TBI found

them to exist predominantly in the frontal-temporal region. This was limited to physical

examination. Those with violent crime convictions had significantly more hospital treatments

for abuse-related injuries (2.15 ± 3.84 vs .94 ± 2.02, p < .05). Only four participants reported

receiving any neuropsychological interventions post-TBI. In conclusion, while TBI was

prevalent amongst female offenders (42%), those with violent convictions had significantly

more (56% vs 38%), while odds of having a violent conviction increased with each TBI (OR

= 1.45, p = 0.12).

3.3.2. The causal pathway

As mentioned in the previously, the design of past studies has made it impossible to

determine the temporal sequence between violent behaviour and TBI. However, participants

with violent convictions had experienced abuse significantly more recently than those with

non-violent convictions (3.83 ± 4.15 years vs. 9.77 ± 9.96 years, p < .05), suggesting that

232

recency of abuse may be a moderator or mediator of the TBI and violence relationship.

Brewer-Smyth et al. (2004) found that although depression scores were not significantly

different between groups using the Beck Depression Inventory-II (BDI-II; Beck, Steer, &

Brown, 1996), those with violent convictions had significantly increased odds of suicide

attempts (OR = 1.249, p = 0.026) which tended to occur years before the study as opposed to

out of remorse for the violent crime. Brewer-Smyth et al. (2004) do not attempt to explain

this, however it appears that current depression as measured by the BDI-II may not be a

significant variable, but lifetime depression as indicated by suicide attempts may be on the

causal pathway.

3.4. Longitudinal studies

3.4.1. Strength of the relationship

Of the three longitudinal studies, one recruited from a prison, one from a clinical setting, and

one from a high-school population (table 5, appendix I). Two were conducted within the

USA, and one in Australia. Shiroma et al. (2010b) used a prospective cohort design,

retrospectively gathering data on TBI over 11 years (1996-2007), while prospectively

measuring in-prison behavioural infractions. Behavioural infractions were defined as a

violation of prisoner code of conduct, and were deemed violent or non-violent according to

criminal justice system definitions for violent crime. The aim was to compare the in-prison

behavioural infraction rate in prisoners with and without a history of TBI. The study

population included 20,098 males and 1,512 females. This was the largest sample of females

with TBI (n = 94) of the 6 studies. No information on power calculations was provided. TBI

was defined as patients discharged from hospital with a TBI-related International

Classification of Diseases, 9th revision, Clinical Modification (ICD-9-CM) code. Similarly to

Johansson et al. (2008) this is likely to under-report TBI. Those who suffered a TBI outside

the 11 year study period were excluded. Data was extracted from existing databases,

233

including the state-wide hospital discharge and emergency department data sets. In total,

6.22% of females had a history of TBI. Males and females without TBIs had a significantly

higher proportion of current violent convictions (for females 31% vs. 20%, p = .03). Analysis

of the number and/or severity of TBIs and relationship to violence was not conducted. 24% of

females with TBI (n = 94) in comparison to 34% without TBI (n = 1,418) had at least one

behavioural infraction. However, those with TBI had an increased violent infraction rate of

144% (mean with TBI 1.20 ± 1.27, mean without TBI 0.68 ± 0.76; RR = 2.44, 95% CI: 1.45-

4.12) after controlling for age, violent crime conviction, prior criminal history, security level,

sentence length, and race. This significantly increased rate was found between males with

and without a TBI (mean with TBI 1.04 ± 1.50, mean without TBI 0.81 ± 1.18; RR 1.86, 95%

CI: 1.54-2.24), but was lower than in female participants. Shiroma et al. (2010b) did not test

if this difference between males and females was statistically significant. Non-violent

infractions did not have a significant increased likelihood in females with TBI (mean with

TBI 1.09 ± 1.47, mean without TBI 0.91 ± 0.95; RR = 0.62; 95% CI: 0.36-1.08). The

difference between violent behavioural infractions and violent convictions suggests that

although fewer females with a violent conviction had a TBI, those with TBI may have an

increased rate of violent behaviour. Alternatively, the difference between this finding of

lower violent convictions in those with TBIs may be due to methodological limitations. It

suggests, however, that violent crime conviction may not be a reliable surrogate measure for

physical violence. In summary, while prisoners with TBI had significantly less violent

convictions (p = .03) and infractions, those with TBI had an increased rate of infractions, with

this finding more pronounced in females.

Baguley, Cooper and Felmingham’s (2006) retrospective cohort study recruited consecutive

inpatients admitted to an Australian tertiary brain injury rehabilitation hospital, with the aim

of assessing prevalence and predictors of aggressive behaviour. This sampling may over-

234

represent more severe TBIs, with 68% of participants falling into the severe range on the

GCS. Participants were followed up at 6, 24, and 60 months post-discharge. Attrition

occurred during the follow-up period, with the sample commencing at 228, and 67 partaking

in the 60 month follow-up. There was no reference to a power calculation. TBI was defined

using markers on the GCS, Glasgow outcome scale (GOS) and PTA. Aggression was

measured using the Overt Aggression Scale (OAS; Yudofsky, Silver, Jackson, Endicott, &

Williams, 1986), providing a global measure of aggression, which can be broken down into

subscales (verbal, physical toward objects, physical toward self, and physical toward others).

Rather than using the OAS as a continuous variable, Baguley et al. (2006) categorised

participants dichotomously as aggressive or not using a cut-off score of 7. To be classified as

aggressive the participant must have demonstrated aggression in at least one of the physical

subscales. They validated this cut-off against an age-matched control group of individuals

without TBI. The OAS was completed by participants in 66.6% of cases, next of kin in 8.5%

of cases, and by both in 25% of cases. There were no significant differences between scores

with the different methods of completing the scale, demonstrating the scales’ reliability. At

each time point approximately 25% of participants demonstrated substantial aggression

which would include physical violence (scoring >7). Baguley et al. (2006) did not examine

the impact of multiple TBIs on violence. They mention that injury pattern as seen on

computed tomography (CT) was not significant in predicting aggression, however they

provide no information as to how injury pattern was evaluated. Furthermore, mild TBI often

does not demonstrate any abnormalities in CT scans (Haydel et al., 2000). Gender was not

significantly associated with aggression at any of the time points. Data are not provided for

this. The authors attributed the failure to find a gender difference on the underrepresentation

of females in their study. Similarly to Johansson et al. (2008), these findings suggest that

females exhibit comparable prevalence of post-TBI violence as males.

235

Stoddard and Zimmerman’s (2011) retrospective cohort study used data collected to study

youth at risk of high school dropout from 4 public high schools, with the aim of examining

differences in interpersonal violence amongst those with and without a TBI. Participants were

recruited on the basis of a low academic grade point average. The exact age participants were

recruited was not reported, only that they examined mid-adolescence to young adulthood.

Unlike previous studies reported, males and females were equally represented (n=850). They

did not report any power calculations. Data was collected over 8 waves, with waves 1-4

corresponding to consecutive high school years. TBI was assessed at waves 5-7. TBI was

defined as having had a concussion, skull fracture or LOC, with no measure of severity, type

or number of TBIs. Interpersonal violence was assessed using a 4-item scale. Cronbach’s

alpha ranged from 0.62-0.76, suggesting questionable internal consistency (Carmines &

Zeller, 1979; Nunnally, 1978). Multivariate regression analyses demonstrated that TBI in

waves 5 and 6 (childhood TBI occurring during or before high school years) was not

predictive of violent behaviour by the final wave (wave 8) in young adulthood (β = .08, p

> .05) when previous violence was included (β = .36, p < .001), yet gender was a significant

variable in this model (β -.08, p <.05). The role gender plays in this model in not explained,

and as no information was provided on how gender was coded, conclusions cannot be made

from the data provided alone. When TBI is restricted to the last year (i.e. TBI acquired in

young adulthood, wave 7) it predicts violence (β = 1.07, p < .001), but once previous violence

and risk behaviours such as alcohol and marijuana use, delinquency, and violence observation

were added to the model, gender was not significant (β = -.03, p > .05). They did not examine

the role of multiple TBIs and its relationship with violence. Similarly to Johansson et al.

(2008) and Baguley et al. (2006), this research suggests that males and females demonstrate

similar levels of interpersonal violence post-TBI.

236

In summary, the longitudinal studies failed to provide sufficient evidence of a relationship

between violence and TBI in females specifically. Shiroma et al. (2010b) had the most

females with TBI, and found that although they had an increased rate of violent behavioural

infractions they did not have significantly more violent convictions. Baguley et al. (2006) did

not find a relationship between gender and aggression in a clinical sample. Stoddard and

Zimmerman’s (2011) high school cohort did not provide support for the role of gender in the

relationship between TBI and violence. Non-significant results may be due to methodological

limitations, or demonstrate that females exhibit similar levels of post-TBI violence as males.

If intimate partner violence is a key risk factor for TBI, which leads to subsequent violence,

participants’ in Stoddard and Zimmerman’s (2011) study may have been too young to

demonstrate a significant relationship.

3.4.2. The causal pathway

An interesting finding by Shiroma et al. (2010b) was that prison reduced the likelihood of

suffering a TBI, with an odds ratio of obtaining a TBI while not incarcerated of 5.88 (95%

CI: 5.26-6.67) in males and 16.67 (95% CI: 5.88-50) in females. This gender difference

suggests that female participants may have been living in a more harmful environment than

males prior to incarceration, and may be evidence of the risk of intimate partner violence.

This supports the idea that intimate partner violence or other environment risk factors

influence the causal pathway between TBI and violence in females. Similarly to Johansson et

al. (2008), using the BDI, Baguley et al.(2006) found depression to be a significant predictor

of aggression across their five year follow-up period, accounting for 24.9% (p < .001) of the

variance in the aggression score at 24 months and 15.9% (p = .002) at 60 months, irrespective

of gender.

4. Discussion

237

Overall, three of the six studies (Brewer-Smyth & Burgess, 2008; Brewer-Smyth et al., 2004;

Shiroma et al., 2010b) provided some evidence of a relationship between TBI and violence in

females. The remaining three (Baguley et al., 2006; Johansson et al., 2008; Stoddard &

Zimmerman, 2011) did not find significant differences between levels of post-TBI violence

in males and females, suggesting that females exhibit similar levels as males. Unfortunately,

all these studies have substantial methodological limitations, and none were designed

specifically to answer the research question at hand. Therefore, the main conclusion

regarding the relationship between TBI and violence in females is that there is insufficient

evidence to answer this question. Brewer-Smyth et al.’s (2004) study was the only one to

examine the cumulative effects of recurrent TBIs on violence, and found some evidence for a

dose-response effect between number of TBIs and violence. This did not include mild TBI.

Brewer-Smyth et al.’s (2004) findings are consistent with previous research which has found

reported prevalence rates of multiple TBIs in female offenders ranging from 35-48%

(Ferguson et al., 2012). Multiple mild TBIs can have similar cognitive and behavioural

profiles to individuals with more severe TBI (Diamond et al., 2007), and has been related to

increased likelihood of post-concussive syndrome (Miller, Ivins, & Schwab, 2013) and

chronic traumatic encephalopathy (Kelly, Amerson, & Barth, 2012). Brewer-Smyth et al.

(2004) found that most TBIs occurred in the fronto-temporal region, which as discussed

previously has been related to post-TBI aggression in the literature (Daoust et al., 2006).

Although none of the studies could provide conclusive results, they did provide some

provoking thoughts about what variables may confound, mediate or moderate the relationship

between TBI and violence in females. Physical and sexual abuse throughout the lifespan may

play a role (Brewer-Smyth & Burgess, 2008; Brewer-Smyth et al., 2004; Shiroma et al.,

2010b). Browne et al.’s (1999) study of victimisation experiences and criminal behaviour

demonstrated that abuse commences for female prisoners at a young age, with 66% of those

238

experiencing sexual abuse and 71% of those severely physically assaulted by caretakers

having experienced such abuse by 11 years of age. Childhood victimisation strongly predicts

victimisation in adulthood (Browne et al., 1999), and adult victimisation in turn increases the

risk of TBI (Kwako et al., 2011), which may lead to increased violent behaviour. Therefore,

childhood abuse may be an important confounder of TBI and violence. This is consistent with

Brewer-Smyth and Burgess’ (2008) findings of increased childhood family sexual abuse in

those with more TBIs and violence crime convictions.

Research has demonstrated that increased post-TBI aggression is related to perpetration of

intimate partner violence in males (Rosenbaum et al., 1994), it is unclear if aggression after

TBI is related to being a victim of intimate partner violence. None of the six studies reviewed

provided gender separated data for mechanism of injury. However, Shiroma et al.’s (2010b)

finding of decreased likelihood of suffering a TBI whilst incarcerated lends some tentative

support for this hypothesis, as does Brewer-Smyth et al.’s (2004) findings of increased

recency and hospitalisations for abuse-related injuries in those with violent crime convictions.

Depression appears to be an important factor in violence as an outcome post-TBI (Baguley et

al., 2006; Brewer-Smyth et al., 2004; Johansson et al., 2008), and PTSD may also be

significant (Johansson et al., 2008). It is unclear if there are gender differences in the impact

of psychiatric comorbidities, and a variety of psychiatric comorbidities have been reported in

those with TBI and a history of violence in other literature (Colantonio et al., 2007). Rao et

al. (2009) found verbal aggression was associated with post-TBI depression. This suggests

that comorbidities may moderate or mediate the relationship between TBI and violence in

females. Overall, the studies reviewed suggest that TBI may be one of many factors that

contribute to violence in females.

4.1. Methodological considerations

239

The studies reviewed had a number of limitations. With regards to sampling, none of the

studies was generalisable as they were conducted within selected populations, and 5 were

conducted in the USA. No study had a sample with over 100 females with TBI, making it

difficult to have sufficient power to detect significant results. However, the studies that

included mixed-sex samples should be commended for providing gender separated data. Of

the 6 studies, only two, which came from the same original data-set, examined females

specifically (Brewer-Smyth & Burgess, 2008; Brewer-Smyth et al., 2004), demonstrating the

paucity of research in females.

There were limitations regarding the definitions and instruments used to measure TBI and

violence. These limitations are not specific to female studies, and have been reported in

research reviewing the relationship between TBI and violence irrespective of gender (Fazel et

al., 2009). All the studies relied on LOC as an indicator of TBI and only Brewer-Smyth et al.

(2004) attempted to exam the role of injury severity in the relationship. As discussed

previously, females may be at a higher risk of recurrent mild TBIs and have a higher

prevalence of TBI when not limited to LOC (Shiroma et al., 2010a). Milder TBIs should not

be dismissed considering the suggested impact of mild recurrent TBIs (Diamond et al., 2007).

However, Brewer-Smyth et al. (2004) did not find evidence supporting a relationship

between violence and TBI without LOC in females. Two of the three studies that provided

support for a relationship between violence and TBI in females used a combination of

methods for identifying TBI, corroborating self-report with medical and criminal records, as

well as physical examination. Only Stoddard and Zimmerman (2011) relied solely on self-

report, not supporting the hypothesis that there is an increased association between violence

and TBI in females. Aside from Shiroma et al. (2010b), the remaining studies which relied on

medical records only did not support this hypothesis. Reliance on medical records risks

under-identification of TBI, with reports of up to 43% of individuals with a TBI not seeking

240

medical attention (Setnik & Bazarian, 2007). This is further compounded by risk of errors

and insufficient recording (Horwitz & Yu, 1984). Reliance on self-report is problematic, with

brief scales and surveys such as that used by Stoddard and Zimmerman (2011) being at risk

for not detecting all but the most recent or severe TBIs (Corrigan et al., 2010). None of the

reported studies used neuropsychological assessment which is considered the gold-standard

of examining the sequelae of TBI (Shiroma et al., 2010a). Furthermore, identifying TBI by

simply using LOC does not assist in determining the prevalence of ongoing sequelae. This is

particularly important in forensic contexts considering research has demonstrated that

females have more ongoing symptoms post-TBI, including difficulties controlling substance

use, temper and emotions (Ferguson et al., 2012).

Three of the 6 studies relied on violent convictions as a measure of violent behaviour

(Brewer-Smyth & Burgess, 2008; Brewer-Smyth et al., 2004; Shiroma et al., 2010b). These

studies that provided the most support for a relationship between violence and TBI in

females. Convictions are limited by how states define violent crime, and may limit

comparison with other jurisdictions. It also does not include crimes that some may perceive

to be violent, and the crime committed may differ from the conviction crime (Shiroma et al.,

2010b). Furthermore, as identified in the mental health literature (Hodgins, 1998), those with

TBI may be at risk of higher convictions due to ease of detection related to cognitive

difficulties. Only Brewer-Smyth et al. (2004) included an analysis of past violent convictions,

alongside current violent convictions, which suggested that current violent crime may not be

a reliable surrogate for violence. The remaining three studies used self-report measures of

violence and physical aggression (Baguley et al., 2006; Johansson et al., 2008; Stoddard &

Zimmerman, 2011). Johansson et al. (2008) supplemented clinical ratings with a valid self-

report scale, however Stoddard and Zimmerman (2011) used a one-item rating scale with

questionable validity. Baguley et al. (2006) attempted to control for bias with the OAS by

241

incorporating informant (Yudofsky et al., 1986). Violent behaviour outside of current

convictions needs to be accounted for, as well as managing the bias of self-report measures.

Ideally this could be achieved using resource intensive methodologies such as those used in

the MacArthur study, combining self-report, informant-report, arrest and hospitalisation

records (Monahan et al., 2001).

Other measures of potentially relevant variables used in the studies reviewed also differed.

Overall, across studies there was insufficient investigation and methodological rigour in

examining the association between TBI and violence in females, and the potential

confounders, mediators and moderators of that relationship.

4.2. Future research

Future research can build upon the studies reviewed. A previous meta-analysis has already

identified the scarcity of studies on female populations (Farace & Alves, 2000), yet this

research need appears to have been relatively ignored, with an ongoing gender bias. These

studies suggest that there is a need for quality epidemiological research examining the

relationship between TBI and violence in females using appropriate valid measures. Ideally

this would be achieved by a longitudinal birth cohort study such as the Swedish population

study (Fazel et al., 2011), but examining gender-specific data. Where studies do include

females in their samples, gender separated data and analyses should be provided where

sample size allows. Observational research needs to ensure that reporting standards reach a

high quality, using guidelines such as STROBE (von Elm et al., 2007). This will facilitate a

meta-analysis of this research question when there are sufficient studies. Finally, researchers

and clinicians in both the fields of TBI and forensic mental health would benefit from agreed

measures of both variables to facilitate comparison between studies.

To facilitate future research and support clinical practice, a valid screen for TBI in females

needs to be available. Currently many studies rely on instruments developed for specific

242

individual studies, with little consideration of reliability and validity (Diamond et al., 2007).

There are currently 2 published valid screening tools developed for use with prisoners, the

Traumatic Brain Injury Questionnaire (TBIQ; Diamond et al., 2007) and the Ohio State

University TBI Identification Method (OSU TBI-ID; Bogner & Corrigan, 2009). These

measures have not been validated within a female UK prison population. Indices on the OSU

TBI-ID which required an estimate of mild TBIs, relating to episodes such as intimate

partner violence, were unreliable (Bogner & Corrigan, 2009). Therefore, this may not be

inappropriate for female prison populations. Also, the TBIQ has only been validated against

short rating scales. Validation of a TBI screen will enable researchers to determine the

prevalence of TBI in UK female offenders, which is currently unknown. Within the UK, the

Disabilities Trust Foundation are examining the validity of a TBI screen with prison

populations and have already examined its use in a male population (Pitman et al., 2013).

Although data is not yet available, preliminary reports suggest that this will be a useful tool

for clinical practice within the UK, and would benefit from validation in a female prison

population.

As indicated from Brewer-Smyth et al.’s (2004) study, future research needs to explore the

impact of recurrent TBIs and how this compares to the neuropsychological profiles of more

severe injuries. Research on neuropsychological profiles should also examine differences

between those with reported violent behaviour and those without violent behaviour,

particularly in females considering the cognitive differences between male and female

offenders with ABI discussed previously (Jackson & Hardy, 2011).

4.3. Clinical implications

Understanding the relationship between TBI and violence in females has many clinical

implications. Without adequate screening in female offenders, TBI is likely to go undetected,

and may impact on engagement in offender rehabilitation programs and the legal process

243

(Jackson & Hardy, 2011). Officials within the criminal justice system may misinterpret

behaviour (Merbitz et al., 1995; Shiroma et al., 2010b). Furthermore, screening can provide a

cost-effective way of determining who is appropriate for referral to more limited and

expensive resources such as neuropsychologists, or alternative care pathways such as MSUs.

As demonstrated by Brewer-Smyth et al. (2004), female offenders with TBI often have had

little if any access to neuropsychological interventions. Such screening can demonstrate

treatment needs and inform policy.

Without an understanding of the impact of TBI on violence in females, offender rehabilitation

programmes are likely to have a limited effect on this population (Jackson & Hardy, 2011).

Further research can inform rehabilitation programmes, facilitate engagement, inform

community placement (Hawley & Maden, 2003) and thereby reduce recidivism (León-

Carrión & Ramos, 2003). This will inform staff training, ensuring staff working with these

individuals understand presentations, are skilled in appropriate behaviour management

techniques, and can make appropriate adaptations to service delivery (Jackson & Hardy,

2011; Merbitz et al., 1995; Morrell et al., 1998).

Given the complex relationship that TBI appears to have with psychiatric comorbidities

(Rogers & Read, 2007; Silver et al., 2001; Timonen et al., 2002), this review highlights the

need for services to address psychiatric comorbidities in individuals with TBI rather than

focusing on the TBI or violent behaviour exclusively, and adopt a multidisciplinary approach

(Jackson & Hardy, 2011). Time spent in prison, away from risky environments (Shiroma et

al., 2010b), may provide offenders and services with a valuable opportunity for

individualised targeted interventions that may decrease victimisation and recidivism (Browne

et al., 1999).

5. Conclusions

244

Although there is some evidence to suggest a relationship between TBI and violence in

females, the inconsistency across studies, methodological limitations, and scarcity of research

in this area does not permit any firm conclusions regarding the nature of this relationship. The

studies reviewed suggest this is a complex relationship, with many variables possibly

impacting on the causal pathway, including psychiatric comorbidities and history of abuse.

These studies can inform future research, and thereby inform clinical practice.

245

Appendix I

Tables

246

Table 1. TBI classifications systems

GCS score (World Health Organization, 2006)

PTA (Lezak, 2004) LOC (Greenwald, Burnett, & Miller, 2003)

Mild TBI 13-15 < 1 hour <30 minutesModerate TBI 9-12 1-24 hours ≥ 30 minutes ≤ 6

hoursSevere TBI 3-8 > 24 hours > 6 hours

Table 2. Classification of quality indicators of studies included in the review

Sampling 1= Non-random 2= Random

Participation rate1= <60%2= ≥60%

Population1= Selected2= General

Female sample size1= <1002= ≥100

Brewer-Smyth & Burgess (2008)

2 2 1 1

Shiroma et al. (2010b) 1 2 1 1Johansson et al.(2008) 1 2 1 1Baguley et al .(2006) 1 2 1 1Brewer-Smyth et al.(2004)

1 2 1 1

Stoddard & Zimmerman (2011)

1 2 1 1

247

Table 3. Cross-sectional studies of violent behaviour and TBIs in females

Authors & Country

Sampling Gender (M=male; F=female); age (mean years)

Participation rate

Violent behaviour Article quality score

TBI Results

Prison populationBrewer-Smyth & Burgess (2008) USA

Minimum and maximum security units of women’s prisons

89F with no childhood family sexual abuse (34.59 years); 60F with childhood family sexual abuse (34.16 years)

81% Criminal offences defined by criminal justice system

6/8 LOC Females who experienced childhood family sexual abuse experienced more TBIs (OR = 1.49, p = .01) and were convicted of more violent crimes (OR = 1.67, p = .05)

Clinical population

Johansson et al. (2008) USA

Sample of consecutive patients seen at an outpatient neuropsychology office in San Francisco

40M; 27F; Mean age for total sample was 40 years

65% Anger severity rated by clinical interview on a four point ordinal scale. Points 3 and 4 indicate physical aggression; Anger and aggression scores from RNBI (Ruff & Hibbard, 2003) also obtained

5/8 GCS, LOC, PTA, focal neurological deficits, and neuro-imaging results

No significant relationship between gender and anger groups (r = 0.14, p > 0.05) or post-morbid RNBI anger score (t = -1.52, p > 0.05)

Table 4. Case-control studies of violent behaviour and TBIs in females

Author & Country

Sampling Case gender (M=male; F=female); age (mean years)

Control gender (M=male; F=female); age (mean years)

Participation rate

Violent behaviour Article Quality Score

TBI Results

Prison Population

Brewer-Smyth et al. (2004)USA

Convenience sample from minimum and maximum security units of women’s prisons

27F (32.86 years)

86F (33.57 years)

81% Crimes groups as violent or non-violent based on criteria established in previous research (Volavka, 2002)

5/8 LOC TBI was significantly higher in females convicted of a violent in comparison to non-violent crime (OR = 1.45, p = .012).

Table 5. Longitudinal studies of violent behaviour and TBI in females

Country & author

Sampling Gender (M=male; F=female); age (mean years)

Participation rate

Violent behaviour

Article Quality Score

TBI Results

Prison Population

Shiroma et al (2010b) USA

Inmates census sample from SCDS

1,136M (median = 30 years) with TBI, 18,962M without TBI (median = 33 years); 94F (median = 34 years) with TBI; 1,418 without TBI (median = 36 years)

Complete histories available for 87% of sample

South Carolina state statute definition of violent vs. non-violent crime

5/8 Medically attended TBI using ICD-9-CM criteria

Higher proportion of current violent crime convictions (54% vs. 40% [p<.0001] and 31% vs. 20%[p=.03]) in males and females without medically attended TBI respectively; 25% and 18% of males and females with TBI respectively had violent infractions, in comparison to 30% and 18% of males and females without TBI respectively; compared to females without TBI, females with TBI had a significantly increased violent infraction rate (RR=2.44)

Clinical populationBaguley et al. (2006) Australia

Consecutive inpatients admitted to a tertiary hospital over 7 years

179M; 49F; Mean age for male and female combined was 34.3 at time of injury

71.5% OAS (Yudofsky et al., 1986)

5/8 GCS; PTA; GOS

Gender was not significantly associated with aggression at 6, 24 or 60 month follow-up (figures not provided)

High School PopulationStoddard & Cohort of 425M; 425F 68% Four-item 5/8 Concussion, Childhood TBI not predictive of violent

Zimmerman (2011) USA

youth selected by grade point average to study youth at-risk for high school dropout, followed over 8 years from mid-adolescence to young adulthood

scale of inter-personal violence

skull fracture or LOC

behaviour in adulthood (β = .08, p > .05). Gender (β -.08, p <.05) and previous violence (β = .36, p < .001) contribute significantly to model. TBI acquired in young adulthood predicted violence (β = 1.07, p < .001), but gender did not make a significant contribution (β = -.03, p > .05)

251

Major Research Project Proposal Form

This form should be completed by the trainee and signed by the University supervisor, and then submitted by the deadline.

Remember to give a draft to your supervisor for comments before submitting the final version.

When preparing this document it would be helpful to consider what you would include when writing the Introduction and Method

sections for your MRP.

* Please append your literature review to this proposal

URN: 6242697

Project Title: Traumatic Brain Injury in Adult Female Offenders in the UK

Introduction

Background and Theoretical Rationale

TBI: Definition and prevalence

Traumatic brain injury (TBI) is “an alteration in brain function, or other evidence of

brain pathology, caused by an external force” (Menon et al., 2010). TBI is the most

common form of acquired brain injury (ABI; Fleminger & Ponsford, 2005), with an

estimated prevalence of 8.5% in the general population (Silver et al., 2001) across all

levels of severity. Incidence ranges from 91-419 per 100,000 in England (Tennant,

2005).

Relationship between TBI and offending

Clinical opinion suggests that violence and impulsive behaviours are both antecedents

and consequences of TBI (Anderson et al., 1999). Violence following TBI has been

characterised as unpredictable, ill-directed, and can occur in the absence of clear

triggers or provocation (Eslinger et al., 1995; Wood & Liossi, 2006). Individuals with

TBI have a significantly increased risk of committing a violent crime (Fazel et al.,

2011). While TBI cannot be assumed to be the sole cause of offending, the cognitive

and behavioural sequelae of TBI may predispose some individuals (Brower & Price,

252

2001; deSouza, 2003; Greve et al., 2001; Kreutzer et al., 1995; Kreutzer et al., 1991;

Miller, 1999; Simpson et al., 1999).

Relationship between TBI in females and offending

A recent report commissioned by the Barrow Cadbury Trust emphasised the need for

research examining causes and consequences of TBI in female offenders specifically

(Williams, 2012). Shiroma, Ferguson and Picklesimer’s (2010a) study of TBI in US

prisoners revealed a male and female prevalence estimate of 64.41% (95% CI: 53.3 to

75.53%) and 69.98% (95% CI: 50.18-89.79%) respectively. The prevalence of TBI in

UK female offenders is currently unknown. A valid screen for TBI in female offenders

would facilitate research and support clinical practice. There are currently no validated

published screening tools for use with UK female offenders. The Brain Injury Screening

Index (BISI; Pitman et al., 2013) has been validated in male offenders in the UK, but

has yet to be extended to females.

Without adequate screening in female offenders, TBI is likely to go undetected, and

may impact on engagement in offender rehabilitation programs and the legal process

(Jackson & Hardy, 2011). Individuals with TBI may be more difficult to rehabilitate and

discharge (Hawley & Maden, 2003), with services ill-equipped to address their needs.

Research in this field is congruent with the Transforming Rehabilitation strategic

business priorities within the National Offender Management Service (NOMS; National

Offender Management Service, 2013) and may increase efficiency by informing

programs to reduce recidivism.

Officials within the criminal justice system may misinterpret behaviour of offenders

with TBI (Merbitz et al., 1995; Shiroma et al., 2010b). Research demonstrating

increased disciplinary incidents in prisoners with TBI (Merbitz et al., 1995; Morrell et

al., 1998; Shiroma et al., 2010b) suggests that they may have increased difficulty

253

adapting to prison life due to cognitive and behavioural sequelae. This has implications

for engagement in the legal process, prison management, and post-discharge and release

pathways (Jackson & Hardy, 2011). Under-identification is likely to perpetuate

inadequate resources, providing no incentive to fund appropriate interventions and

inefficient use of available resources. Screening is consistent with NOM’s Reducing

Prison Unit Costs strategic business priority by ensuring appropriate services are

commissioned and targeting the most appropriate offenders (NOMS, 2013), e.g.

screening can provide a cost-effective way of determining who is appropriate for

referral to more limited and expensive resources such as neuropsychologists, or

alternative care pathways.

Risk factors for offending in females with TBI

Females are reportedly less likely to offend than males, yet those who do are more

likely to be experiencing a mental illness (Butler et al., 2005; Fazel & Grann, 2006).

New-onset major depression post-TBI increases the risk of aggression for females

eightfold (Rao et al., 2009). PTSD may also be a significant risk factor (Johansson et

al., 2008). Comorbidities may moderate or mediate the relationship between TBI and

offending in females.

Physical and sexual abuse throughout the lifespan may be a risk factor (Brewer-Smyth

& Burgess, 2008; Brewer-Smyth, Burgess, & Shults, 2004; Shiroma et al., 2010b).

Abuse commences for female prisoners at a young age (Browne et al., 1999). Childhood

victimisation strongly predicts victimisation in adulthood (Browne et al., 1999), and

adult victimisation in turn increases the risk of TBI (Kwako et al., 2011), which may

lead to increased violent behaviour. This is consistent with Brewer-Smyth and Burgess’

(2008) findings that female prisoners with more TBIs and violent crime convictions

have increased childhood family sexual abuse.

254

Post-TBI aggression may be related to being a victim of intimate partner violence

specifically. Shiroma et al. (2010b) found a decreased likelihood of suffering a TBI

whilst incarcerated. Brewer-Smyth et al. (2004) found increased recency of abuse and

hospitalisations for abuse-related injuries in those with violent crime convictions, with

most TBIs occurring in the fronto-temporal region, which has been related to post-TBI

aggression in the literature (Daoust et al., 2006).This lends some tentative support for a

relationship between intimate partner violence and post-TBI aggression.

Brewer-Smyth et al. (2004) examined the cumulative effects of recurrent TBIs on

violence, finding evidence for a dose-response effect between number of TBIs and

violence. These findings are consistent with previous research which has found reported

prevalence rates of multiple TBIs in female offenders ranging from 35-48% (Ferguson

et al., 2012). Multiple mild TBIs can have similar cognitive and behavioural profiles to

individuals with more severe TBI (Diamond et al., 2007).

Research Questions

How prevalent is TBI and how does TBI present itself in the cognitive, psychiatric

and health needs of female prisoners in the UK?

Is self-reported TBI using the BISI associated with cognitive performance in

standardised questionnaires and neuropsychological tests?

Main Hypotheses

Female offenders in the UK demonstrate similar prevalence of TBI as males.

The BISI provides an estimate of prevalence within the confidence intervals of

studies using clinical interviewing.

Results of the BISI will be significantly associated with those obtained in the

standardised questionnaires and neuropsychological tests.

255

The BISI will have a test-retest reliability coefficient of at least 0.60.

Female offenders with self-reported TBI have significantly more cognitive,

psychiatric, and physical health difficulties than female offenders without TBI.

Female offenders with self-reported TBI have higher historical rates of recidivism

and behavioural infarctions that those without TBI.

Method

Participants

Participants will be recruited from prisons for women in southern England, with four

potential prison sites. These prisons have a combined operational capacity of

1,607women from which potential participants can be recruited.

To explore prevalence and severity of TBI in the UK (stage 1), using Daniel and Cross’

(2013) formula for sample size calculation for prevalence studies, on the basis of a level

of confidence of 95%, an expected prevalence of 66% in female offenders using the

gold standard of clinical interviews (Shiroma et al., 2010a), and a precision value of

0.125, a sample of 56 female offenders would be required for stage 1. A large precision

value was chosen due to feasibility related to resource limitations, and the preliminary

nature of this research. The precision value meets the assumption of normal

approximation. To assess the test re-test reliability of the BISI, based on a minimum

reliability of 0.6, an expected reliability of 0.8, α=0.05 and β=.20, an estimated sample

of 39 of the original 56 will be required (Walter, Eliasziw, & Donner, 1998).

To explore the differences in cognitive, psychiatric and health needs of female offenders

with TBI and those without TBI, as well as rates of recidivism and behavioural

infarctions, based on the number of variables being measured, an anticipated large effect

size, statistical power of .80 and type I error α of .003 (Bonferroni correction for 14

comparisons), a sub selection of 28 participants with TBI and 28 without TBI will be

256

required for stage 2.Based on similar research using the BISI with male prisoners

(Pitman et al., 2013) the estimated response rate is 73%, therefore a total of 76 prisoners

will be approached for consent.

Inclusion criteria include prisoners over 18, with an upper age limit of 80 years of age in

line with norms provided in the instruments to be used. Exclusion criteria include acute

symptoms of physical or mental illness or other indication that participants may not be

able to provide informed consent. This will be achieved by excluding participants in the

medical unit. Acute illness and ability to provide consent is also assessed in the clinical

interview, at which point the assessment will be terminated. Prisoners with a confirmed

diagnosis of dyslexia, are not fluent in English, or have reported acquiring a TBI in the

last 6 months, will be excluded from stage 2 due to validity limitations of measures.

Participants with a learning disability will be included unless queries regarding capacity

to consent are raised when are being briefed.

Design

Cross-sectional design using a semi-structured clinical interview, clinical

questionnaires, and neuropsychological measures.

Measures/Interviews/Stimuli/Apparatus

Please see Appendix I for the Participant Information Sheet and Appendix II for the

Participant Consent Form. For psychometric properties of measures please refer to

Table 1 Appendix III. The following assessment tools will be employed:

Semi-structured interview designed for use in male prisoners, which has been adapted to

extend to research to female prisoners (Appendix IV)

Pro forma for collecting data from participant files (Appendix V)

The Brain Injury Screening Index (BISI; Appendix VI)

257

The Wechsler Abbreviated Scale of Intelligence II (WASI-II; Wechsler & Zhou, 2011).

The Test of Premorbid Functioning – UK Version (TOPF; Wechsler, 2009)

The Behavioural Assessment of the Dysexecutive Syndrome (BADS; Wilson,

Alderman, Burgess, Emslie, & Evans, 1996)

The Repeatable Battery for the assessment of Neuropsychological Status (RBANS;

Randolph, 1998)

The Test of Memory Malingering (TOMM; Tombaugh, 1996)

Neurobehavioral Functioning Inventory (NFI; Kreutzer, Seel, & Marwitz, 1999)

The Beck Anxiety Inventory (BAI; Beck, Epstein, Brown, & Steer, 1988)

The Beck Depression Inventory II (BDI; Beck, Steer, & Brown, 1996)

The Impact of Events Scale – Revised (IES-R; Appendix VI; Weiss & Marmar, 1997)

Pro formas are not provided in the appendices for assessments restricted due to

copyright.

Procedure

11. Participants will be recruited through a convenience sample. A list of prisoners who

have been new receptions over a 1-2 month period (depending on average intake

rate- until 76 participants are obtained for sufficient power and response rate

considerations) will be provided by prison staff. Prison staff will act as gatekeepers

and be asked to identify participants under 18, over 80, admitted to the medical unit,

or who cannot provide informed consent.

258

12. Eligible participants will be provided with a copy of the participant information

sheet and consent form, i.e. a copy will be dropped into their cells by a member of

the research team.

13. Within a week of providing the information sheets prisoners will be contacted face-

to-face by a member of the research team to discuss the participation and consent. A

suitable time to commence assessment will be arranged with those who agree to

participate.

14. Participants will be allocated a participant number, which will be recorded with their

prison reference number, and recorded on an encrypted device. This will be separate

from the main data collection file.

15. At assessment, participants will complete the BISI and clinical interview (stage 1).

This will take approximately 1 hour. Participants will be provided with the clinical

questionnaires in the interview and asked to complete them in their own time. These

will then be collected at the end of each day of data collection. Participants will

complete all questionnaires themselves. Data on social history, abuse history,

clinical history, offence history, and behavioural infractions, will be obtained for

these participants from Offender Assessment System (OASys), probation and re-

offending records, and individual history files.

16. Participants will have the option of progressing on to stage 2 straight after stage 1,

or arranging a time to continue with the assessment.

17. Stage 2, consisting of the battery of neuropsychological measures, will occur at an

arranged time, with an expected sample of 28 participants with TBI and 28 without

a history of TBI. This will take 1½-2 hours.

259

18. Participants who progress from stage 1 to stage 2 within 1-2 weeks of completing

the BISI will be asked to complete the BISI again for the purpose of test-retest

validity, until the required sample of 39 is reached.

19. As data is gathered it will be encoded in an SPSS data file stored on an encrypted

device for analysis.

20. Once data collection is completed appointments will be arranged with participants

who request an assessment feedback session. Participants may also request to have

results shared with the prison/health service. A brief report with results will be

provided.

Ethical considerations

7. Great effort will be made to clarify the nature of the relationship between the

researchers conducting assessments and the participant; i.e. it is not an assessment

from a clinical referral, that the primary purpose is research. However, participants

will have the opportunity to have a feedback session after the data collection is

completed and/or to have results shared with the prison/health services. Feedback

will be primarily descriptive and can be used to indicate if support from health

services is required.

8. Feedback on neuropsychological tests may impact on self-esteem, as mean prison

IQ in the UK is 88 (±12.0) (Hayes, Shackell, Mottram, & Lancaster, 2007). It is

hoped that the provision of information and training to staff and the potential for

additional support following the research will enable participants to feel

empowered, with the resources to manage difficulties.

9. Participants may assume that test results can be used for secondary gain, however

when briefing the participants it will be emphasised that tests will not be for clinical

260

use, and only a recommendation for a comprehensive clinical examination can be

made. The TOMM (Tombaugh, 1996) will also be used to assess effort.

10. Due to the time required of participants motivation and fatigue may invalidate

results. To manage this, participants can complete stages on different days if they

wish, or have a brief break during testing if requested.

11. To ensure test validity the participating prisons will be requested to provide a quiet

environment, with adequate space, ventilation, lighting and furniture. We will also

liaise with staff to try to minimise potential interruptions during testing.

12. Information that is related to risk to self or others, including drug use and abuse,

may be disclosed. The research team will liaise with participating prisons to ensure

that the research protocol is congruent with prison policy to manage risk. Potential

limits to confidentiality will be made explicit to participants when obtaining

consent.

Name of Ethics Committee: University of Surrey Ethics Committee: Faculty of Arts and

Human Sciences; National Offender Management Service............................

R&D Considerations

Name of R&D department: R&D approval is not required however a letter of support

from governors of participating prisons will be obtained.

Proposed Data Analysis

All analyses will be done using IBM SPSS version 20 (IBM, 2011). Data will be

checked for normality of distribution and homogeneity of variance.

Descriptive statistics, including age, proportion of participants with a TBI, ethnicity,

education, learning disability, special support at school, diagnosis of mental health

problem, history of self-harm, suicide attempts, frequency of drug and alcohol use, age

261

at first offence, number of offences, number of violent convictions, number of

behavioural infractions, participation in rehabilitation programs, and years spent in

custody, will be provided for those with an identified TBI and those without. This will

include means and standard deviations, and t-tests comparing those with and without a

TBI where appropriate. Further descriptive statistics will be provided for those with a

TBI, namely: number of TBIs, age at earliest TBI, loss of consciousness, cause of

injury, and help-seeking behaviour after TBI.

Concurrent validity of the BISI will be studied by calculating Pearson’s r between the

BISI and clinical interview. Receiver Operating Curve analysis will also be conducted

to assess sensitivity and specificity rates against clinical interview as the gold standard.

Test-retest reliability will be assessed using Pearson’s r.

Between samples t-tests will explore differences participants with TBI and without TBI

on cognitive, psychiatric and physical health measures.

Service User and Carer Consultation / Involvement

Service user and carer consultation was undertaken on 6th August 2013 and informed

this proposal (see Appendix VI for feedback).

Feasibility Issues

As the study is highly dependent on the amenability of the prison involved, there

must be organisational incentive for participation. After the study is complete,

the research team will provide a training and information session with prison

staff to support them in their roles when working with female offenders with

TBI.

Prisons participating must be able to provide adequate support in the event that

participants are distressed after the study. To overcome this we are liaising with

potential prisons and will prioritise the research site accordingly. We are also

262

aiming to obtain sensitive data on abuse history from the OASys to reduce the

risk of distress.

Cost of pro formas for the copyrighted measures used is high (totalling

£1418.40), and exceeds the research budget provided by the University of

Surrey. The Disabilities Trust have agreed to cover the cost of pro formas

exceeding the research budget.

Dissemination strategy

Analysed data will be written up in a thesis as part of a doctorate, and disseminated

through research articles (e.g. The Journal of Head Trauma; Brain Injury) and

conferences (The Annual Division of Forensic Psychology conference; International

Association of Forensic Mental Health conference). A briefing can also be provided for

the prison’s newsletter. The Disabilities Trust Foundation will publish results through

reports, presentations, journal articles, and on their website.

Study Timeline

Please see Gantt chart in Appendix VII.

263

Clinical Experience Précis

The following is an overview of the experience gained from clinical placements.

Year One:

Community Mental Health Team - 12 months.

This placement involved offering psychological assessments and interventions

to adults presenting with a range of mental health difficulties in the community at a

secondary care level. The main model utilised was CBT, however I also got to draw on

third wave models such as Acceptance and Commitment Therapy and systemic model.

The multi-disciplinary team consisted of psychiatrists, psychologists, nurses, social

workers, and an occupational therapist. I also got to run relaxation groups in the

affiliated acute ward, which was valuable in obtaining an understanding of the clinical

pathway and progression of the client group. Presenting problems included anxiety,

depression, obsessive-compulsive disorder, psychosis, bipolar affective disorder and

interpersonal issues. During this placement I learned complex formulation skills, and

the importance of perceiving an individual’s life narrative rather a reductionist

diagnostic category.

Year Two:

Community Mental Health and Learning Disabilities Team – 6 months

This placement involved offering assessment, treatment and consultation for

those with a Learning Disability and additional mental health needs, and their families,

carers and staff teams. This placement has involved using CBT, systemic work, and

narrative therapy, as well as neuropsychological assessment and formulation.

264

On this placement I have learned the true meaning of a strengths-based approach

and the power of working with families and wider systems. I learnt how to support

individuals in finding their voice, and developed the confidence in being a voice for

those who could not. I learnt how to adapt how I communicate complex ideas, and

synthesise vast amounts of clinical information. I learnt the power of mentalising

everyone in the system, and uniting people in visions they did not realise they shared.

Child and Adolescent Mental Health Service (CAMHS) and Youth Offending Team

(YOT) – 6 months

My CAMHS placement offered assessment and treatment to children and their

families using a range of approaches. During this placement I used CBT, Solution-

focused approaches, and a great deal of systemic thinking. During this placement I

worked with children who were presenting with a range of problems with anxiety,

mood, cognitive, and behavioural difficulties.

As part of the YOT, I worked with two young people whose complex

presentations contributed to involvement in the criminal justice system. This really

taught me about working with complexity, acknowledging the limits of the therapy

room, and the role of the psychologist in providing containment for the system. I had

the opportunity to work with a range of professionals in this placement including those

in mental health services, youth offending officers, and schools.

Year Three:

Older Adults Community Mental Health Team, Memory Assessment Service and

Challenging Behaviour Service – 6 months

265

This placement was split between three roles. The Community Mental Health

Team involved offering individual therapeutic intervention to older adults presenting

with a range of low mood, anxiety, and interpersonal problems. A further role was

offering assessment and consultation to care homes which were reporting experiencing

challenging behaviour from residents, specifically with a diagnosis of dementia. The

third role on this placement was assessment and intervention for those experiencing

cognitive problems. Assessment involved a full and comprehensive battery of

neuropsychological tests, designed to profile cognitive strengths and weaknesses and to

compare this to premorbid functioning to aim to assess for dementia or other cognitive

changes in conjunction with colleagues in psychiatry and neurology.

During this placement I particularly learnt how to create safe non-blaming

spaces for care staff to reflect on their clinical practice, and reconnect with their

compassion which can be compromised in such demanding work environments.

Specialist Trauma Service for Veterans – 6 months

This is a service offering psychological support for veterans experiencing mental

health difficulties, particularly post-traumatic stress disorder (PTSD). The service

provides an intensive residential trauma-focussed CBT programme, as well as brief two

week psychoeducational residential programmes. I also had the opportunity to work

with clients on an outpatient basis. During this placement I offered trauma focussed

CBT, and compassion focussed therapy particularly for those experiencing high levels

of guilt and shame. I also ran a range of psychoeducational groups and provided group

nursing supervision.

This placement taught me to understand and work with complex group

dynamics, and to be aware of issues around transference and counter-transference, and

266

how to use these therapeutically. Understanding how an individuals’ history before

exposure to a trauma contribute to their responses to trauma has been important

learning. One of the most valuable elements was learning to utilise micro-expressions in

the therapy room, and working within the therapeutic window of distress particularly

when with clients at high risk of disassociation.

267

Summary of Assessments

Year I Assessments

PROGRAMME

COMPONENT

TITLE OF ASSIGNMENT

Fundamentals of Theory

and Practice in Clinical

Psychology (FTPCP)

Short report of WAIS-IV data and practice

administration

Practice case report ‘Assessment with a 35 year old man presenting with

depression and multiple sclerosis, employing an

Acceptance and Commitment Therapy based

formulation’

Problem Based Learning

– Reflective Account

Problem Based Learning – Reflective Account. The

Relationship to Change

Research – Literature

Review

‘Traumatic brain injury and violent behaviour in

females: A systematic review’

Adult – case report ‘Assessment and intervention with a 35 year old man

presenting with depression and multiple sclerosis using

an Acceptance and Commitment Therapy approach:

Reflections on treatment challenges

Adult – case report ‘Assessment and intervention with a 36 year old man

presenting with psychosis using a Cognitive

Behavioural Therapy (CBT) approach: Incorporating an

interventionist-causal model with first generation CBT’

Research – Qualitative Experiences of Romantic Relationship Formation Using

268

PROGRAMME

COMPONENT

TITLE OF ASSIGNMENT

Research Project Computer-Mediated Communication: A Thematic

Analysis

Research – Major

Research Project

Proposal

Traumatic Brain Injury in Adult Female Offenders in

the UK

Year II Assessments

PROGRAMME

COMPONENT

TITLE OF ASSESSMENT

Research - SRRP An evaluation of the prevalence of poor sleep quality and

provision of treatment amongst service users under a

Community Mental Health Team

Research Research Methods and Statistics test

Professional Issues

Essay

It has been six years since Mencap’s report, Death by

Indifference, highlighted unequal healthcare and

institutional discrimination that people with learning

disabilities can experience within the NHS. What remains

to be done to achieve a more responsive and

compassionate culture of care? What is the role of clinical

psychology in initiating, sustaining and supporting this

culture?

Problem Based Problem Based Learning (PBL) Reflective Account. The

269

Learning – Reflective

Account

Stride Family

People with Learning

Disabilities – Case

Report

Extended assessment and formulation with a 22 year old

man presenting with a moderate learning disability and

behaviour that challenged his college

Personal and

Professional Learning

Discussion Groups –

Process Account

Personal and Professional Learning Discussion Group

Process Account

Child and Family –

Oral Presentation of

Clinical Activity

Working with diversity: from individual cognitive

behavioural (CBT) therapy to systemically informed CBT

Year III Assessments

PROGRAMME

COMPONENT

ASSESSMENT TITLE

Research – MRP

Portfolio

Utility of the Brain Injury Screening Index in Identifying

Female Prisoners with a Traumatic Brain Injury and

Associated Cognitive Impairment

Personal and

Professional Learning –

Final Reflective

On becoming a clinical psychologist: A retrospective,

developmental, reflective account of the experience of

training

270

Account

Older People – Case

Report

Neuropsychological assessment with a woman in her

early seventies presenting with memory problems and a

diagnosis of mild cognitive impairment

271