Sophie Davison; Aleksandar Janca
Curr Opin Psychiatry. 2012;25(1):39-45.
Purpose of review: There is a well established association between personality disorder and offending but the nature of the relationship is less well understood. We reviewed the recent literature on personality disorder and offending, picking out studies that examined the relationship between the two.
Recent findings: Cluster A, B and C personality disorders are each associated with different types of offences. Although rates of personality disorder are high in all serious offenders, the role played by personality disorder may be greater in some offences than others, for example, in rapists compared with child molesters, men who kill their fathers rather than their mothers, men who kill their children compared with mothers who kill their children; and in less severe stalking behaviour compared with those who get convictions. Three articles suggested frameworks for understanding how personality disorder may interact with other factors to contribute to offending.
Summary: Frameworks that integrate personality traits; comorbid problems such as substance misuse, mood disorders and attention deficit hyperactivity disorder (ADHD) symptoms; motivation for offending; maladaptive cognitions; beliefs and attitudes; anger and arousal; and situational factors are helpful when considering risk assessment, risk management and treatment. More empirical research is needed to test these theories.
It is well established that people in correctional and forensic mental health settings have higher rates of personality disorder, especially antisocial personality disorder, than people in the general community.The relationship between personality disorder and violent re- offending is so well established that the presence of personality disorder has been incorporated as a risk factor in structured risk assessment tools such as the Historical Clinical Risk Management-20 and Violence Risk Assessment Guide (VRAG). The Hare Psychopathy Checklist revised (PCL-R), which measures psychopathy, a particular subtype of antisocial personality, has become well established as an actuarial tool for predicting the risk of violent re-offending.
The relationship between antisocial personality disorder and offending is not surprising given the rather tautological definition. Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) includes repeated acts that are grounds for arrest; and central to the ICD 10 diagnosis is the gross disparity between behaviour and the prevailing social norms.
A clear understanding of the nature of the link between personality disorder and offending has important implications for treatment and risk management. Until recently, however, less attention has been paid to types and aspects of personality disorder other than antisocial and their relationship with offending behaviour; and between personality disorders with particular aspects and types of offending behaviour, violent and nonviolent.
We reviewed the recent peer reviewed literature (2009 to present) on personality disorder and offending by searching for the term personality disorder combined with the terms offending, criminality, criminal behaviour, violence, homicide, prison and forensic, respectively, in the MEDLINE and psycINFO online databases.
We picked out studies that examined the relationship between personality disorder and offending.
Relationship Between Personality Disorder Subtypes and Different Offences
Roberts and Coid[5••] used data from the National Survey of Psychiatric Morbidity in Prisoners in England and Wales to examine the relationships between different personality disorder subtype scores and lifetime offences. A representative sample of prisoners were interviewed using self-report questionnaires about previous conviction history; and a subsample of 391 prisoners were interviewed by clinicians using the Structured Clinical Interview for Axis II Disorders (SCID-II), using continuous scores for the different personality disorder subscales. They used logistic regression to examine the relationships.
Unsurprisingly, they found that conduct disorder scores were significantly related to all offence categories and adult antisocial personality disorder scores were associated with most offences, especially obstruction of justice, firearms, robbery and blackmail, escape and breach, fraud, burglary and theft as well as violence. With regards to other cluster B disorders, narcissistic personality disorder scores were associated with fraud and forgery.
A surprising finding was that there was no association between borderline personality disorder scores and types of offending, despite the high rates of borderline personality disorder found in prisoners, especially in women. The authors attribute this to the high rates of comorbidity of borderline with antisocial personality disorder, such that any relationship was lost after controlling for other confounding factors.
In the cluster C disorders, they found that avoidant personality disorder scores were associated with criminal damage but negatively associated with firearm offences.
Obsessive–compulsive personality disorder scores were associated with firearm offences and dependent personality disorder scores were significantly associated with firearm offences and violence but negatively associated with criminal damage.
In the cluster A disorders, paranoid personality disorder scores were associated with robbery and blackmail but negatively associated with driving offences. Schizotypal personality disorder scores were significantly associated with arson but negatively associated with robbery and blackmail. Schizoid personality disorder scores were associated with kidnap, burglary and theft.
Interestingly, homicide offences and sex offending were not associated with any personality disorder scores. In the case of homicide, this may be due to the small numbers.
This is a novel look at the relationship between personality disorders other than antisocial and different offending behaviours using a large sample. It confirms that antisocial personality disorder accounts for most of the relationship between offending and personality disorder but throws up some interesting associations with personality disorders, including from clusters A and C. The authors speculate about why some of these relationships might arise, but ultimately the method used can only point to associations and not to causality or to the mechanism.
Warren and South looked at the relationship between cluster B personality disorders only and patterns of criminality and violence in a sample of women incarcerated in a maximum secure prison. They found no relationship between antisocial, borderline, histrionic or narcissistic personality disorder scores and history of convictions apart from a negative relationship between antisocial personality disorder scores and homicide and a positive relationship between borderline personality disorder scores and prostitution.
However, antisocial, borderline and narcissistic scores all showed a significant relationship with perpetrating threats and physical assaults in the prison environment. In addition, borderline and histrionic scores were associated with perpetrating sexual assaults in prisons. Narcissistic and antisocial scores were associated with institutional infractions.
Male Sex Offending: Rapists versus Child Molesters
Francia et al. [8•] examined personality disorders and their features in two groups of incarcerated male sexual offenders in Colorado, USA. They looked at 251 rapists and 311 child molesters and compared them with a group of nonsexual offenders, using the Coolidge Correctional Inventory, a 250-item self-report DSM-IV aligned personality and neuropsychological inventory. They found that rapists had significantly higher levels of antisocial personality traits than child molesters but no difference in narcissistic traits.
Nonsexual offenders also scored higher on the antisocial scale than child molesters but not higher than rapists. Child molesters had higher levels of avoidant traits than rapists and higher levels of avoidant, dependent, and schizoid traits than nonsexual offenders. When they looked at the number of individuals who reached the cut-off for different personality disorder diagnoses they found that the greatest prevalence for nonsexual offenders was for obsessive compulsive, antisocial, avoidant and narcissistic personality disorders. For sexual offenders, if considered together, avoidant personality disorder had the highest prevalence followed by obsessive compulsive, schizoid, paranoid and borderline personality disorders.
Eher et al.[9•] did a similar study of 807 incarcerated sexual offenders admitted to Austria’s prisons between 2002 and 2009. They found that sexual offenders displayed high rates of mental illness, sexual disorders, personality disorders and substance abuse. However, there were differences between rapists and child molesters consistent with the finding of the study by Francia et al. [8•] Rapists had higher rates of personality disorder overall (76 versus 60%) and alcohol and substance abuse than child molesters, who had higher rates of sexual disorders (78 versus 24%). Rapists had higher rates of cluster B personality disorders, especially antisocial and borderline, than child molesters (66 versus 40%), whereas child molesters had higher rates of cluster C disorder (14 versus 6%).
These two studies have treatment implications, as they suggest that child molesters may benefit from treatment aimed at sexual disorders and cluster C traits, whereas rapists may need an approach aimed at the features of cluster B personality disorders more generally and not just at deviant sexual behaviour.
Stalking and Borderline Personality Disorder
Sansone and Sansone[10•] examined the relationship between stalking and borderline personality disorder. Stalking consists of chronic, repeated, unwanted nuisance behaviours by an offender, which have adverse psychological and/or physical effects on victims.
Because borderline personality disorder is characterized by a pattern of unstable and intense interpersonal relationships coupled with frantic efforts to avoid real or imagined abandonment, one might expect this personality disorder to be prominent amongst stalkers. The authors report that previous studies have examined the prevalence of axis II disorders amongst stalkers but grouped them all together or just reported on cluster B as a cluster. The prevalence of borderline personality disorder has not often been reported on specifically. They located five studies between 1995 and 2008 that reported rates of borderline personality disorders amongst samples of stalkers. Four studies found prevalence between 4 and 15%. They consisted of three studies of forensic populations and one of patients who stalked their psychiatrist. A fifth study, of women in the USA looking at retrospective cases of mental health and law enforcement professionals, where having a conviction was not necessary to be included in the study, found a rate of 45%, with borderline personality disorder being the most common diagnosis. The authors pointed out that the rates are likely to vary because of the different methodologies and sample types. They concluded that the differing rates may suggest that stalkers who end up in mental health treatment and forensic settings are more likely to suffer from axis I disorders, in contrast to community samples of less severe stalking, where a significant minority of individuals appear to suffer from borderline personality disorder.
Representative studies of offenders convicted of homicide have consistently found personality disorder to be the most common mental health diagnosis among these offenders, with 6% of all homicide offenders over a three-year period in the UK having a primary diagnosis of personality disorder and 20% of homicide offenders in Angers, France over a ten-year period having a diagnosis of personality disorder. In the last couple of years, there have been a number of studies looking at particular aspects of killing and particular types of homicide and reporting on rates of personality disorder.
Homicide followed by suicide is a rare but tragic phenomenon. Flynn et al. reported that studies of homicide followed by suicide have not examined mental health characteristics in details. They conducted a national cross-section study of perpetrators in England and Wales 1996–2005 and found 203 incidents recorded. The median age was 41 and most were men. Men more often killed a spouse or partner and women more often killed their children. The most common primary diagnoses were personality disorder (32%) and affective disorder (26%). There was no difference in the proportion of offenders with personality disorder in the homicide–suicide sample compared with their large sample of homicide-only offenders. The characteristics of those with personality disorder were not reported on separately. Overall few homicide–suicide perpetrators had been in contact with services prior to the event. In fact, fewer perpetrators of homicide followed by suicide had been under the care of mental health services before the incident than in cases of suicide or homicide only. Interestingly, multivariate analysis showed the risk of suicide was six times greater when firearms were used in the homicide.
Filicide and Personality Disorder
Kauppi et al. in Finland[14•] examined filicides (killing one’s child) in Finland over a 25-year period to look at the difference between those perpetrated by the mother and by the father. In Finland, a psychiatric examination is conducted on all individuals charged with homicide and other serious offences so the sample can be expected to be representative. They found 200 filicides, of which 59% were committed by mothers, 39% by fathers and 2% by stepfathers. Fifty-six percent of cases involved the killing of neonates. There were 75 filicide–suicides. There were 65 filicides left over, which were classified as ‘other filicides’. It was only for these cases that the authors examined the diagnosis. Fifty-one percent of maternal perpetrators had psychosis or psychotic depression and 76% were deemed not responsible for their actions by reason of insanity. On the contrary, 67% of paternal perpetrators had personality disorder, 45% abused alcohol and there were high rates of marital jealousy and domestic violence. Eighteen percent of the male perpetrators were deemed not responsible on the grounds of insanity. The presence of significant life stressors amongst the male perpetrators included marital break-up, jealousy, fear of separation, long-term substance misuse and low level of education and socioeconomic status. The mean age of the paternal victims (4.2 years) was higher than for those killed by their mothers (2.2). This is evidence suggesting different mechanisms linking personality disorder and killing one’s child from those at play in psychosis. It also shows a different view taken by the courts regarding the responsibility for one’s actions in personality disorder as opposed to psychosis
Matricide, Patricide and Personality Disorder
Liettu et al. looked at all male matricidal (n = 86) and patricidal (n = 106) offenders referred for forensic psychiatric examinations between 1973 and 2004 in Finland. This included all offenders who had killed a parent as well as all those convicted of aggravated assault or attempted murder against a parent. They found that matricidal offenders suffered more commonly from psychotic disorders than did patricidal offenders (46 versus 26%, respectively). A greater proportion of patricidal offenders had a personality disorder (64 versus 45%, respectively), with borderline personality disorder being more common amongst the patricidal than matricidal offenders (29 versus 10%, respectively). For matricidal offenders the most common motive was a mental disorder, whereas for patricidal offenders the offences were most commonly motivated by long-term conflict.
Once again this hints at a different set of internal and environmental factors interacting to lead to killing in personality disorder compared with an axis I psychotic disorder.
Familicide is defined in this article as the killing of an intimate partner and at least one child. Léveillée et al. studied sixteen cases of familicide in Quebec between 1986 and 2000; all were perpetrated by men. Sixty-two percent of the perpetrators used firearms and many (68%) of them killed themselves after the act. The authors reported that 19% had borderline personality disorder and 38% borderline traits. Twelve percent had antisocial personality disorder or antisocial traits and 68% had a lifetime history of depressive symptoms. The four main motivations were: intimate partner loss, social loss, mental state perturbations and economical motivation. Once again this starts to hint at the mechanism for the link between personality disorder and this sort of offending. The authors underline the apparent link between borderline personality disorder, impulsivity and intense fear of loss and abandonment and familicide. One can start to hypothesise how intimate partner loss, social loss and economic motivation might interact with mood, maladaptive thinking, coping and behaviour styles to lead to violence. The sample was, however, very small as this is a rare crime, making conclusive statements about the role of different mental disorders difficult.
Several of the articles we have reviewed show an association between personality disorder and particular types of offending or maladaptive behaviour. Several of the articles lump all personality disorders together without distinguishing between the clusters or subtypes. Some start to provide a hint as to how an individual’s personality disorder might relate to the motivation for particular behaviours. However, all they show is an association between a type of crime and a personality disorder; they do not explain the nature of that link or indeed whether it is causal. We found three articles that tried to elucidate in more depth the link between personality disorder and criminality, in particular violence.
Motivation for Offending in Personality Disorder
Gudjonsson et al.[17•] examined the link between personality disorder and offending in a slightly different way by examining the relationship between motivation for offending, axis I clinical syndromes, personality disorder (axis II) and attention deficit hyperactivity disorder (ADHD) in 196 prisoners. The Millon Clinical Multiaxial Inventory and Offending Motivation Questionnaire were completed by 196 prisoners. Each type of motivation (financial, provocation, excitement and compliance) had a specific relationship with axis I and axis II disorders. Antisocial and sadistic personality disorders were significantly correlated with financial and provocation motivations. Sadistic and dependent personality disorders correlated significantly with excitement motivation; and avoidant, dependent and paranoid personality disorders correlated significantly with compliance motivation. However, hierarchical multiple regressions showed that personality disorders did not predict motivation for offending beyond that of axis I clinical syndromes and ADHD symptoms.
Drug dependence and ADHD symptoms were the most powerful predictors of financial motivation, alcohol of provocation and anxiety of compliance motivation. The authors conclude that this underscores the importance of drug and alcohol dependence, anxiety and ADHD symptoms in predicting the motivation for offending.
This is further support for suggesting that simply having a diagnosis of personality disorder per se does not explain offending behaviour. To understand the link between personality disorder and different offending behaviours one has to look at the associated co-morbid symptoms, including substance misuse, as well as the actual traits displayed and how these interact in particular circumstances to make it more likely the individual will choose to engage in a particular behaviour.
Explaining the Link Between Personality Disorder and Violence
The mechanism of the link between personality disorder and offending is addressed in an article by Logan and Johnstone.[18••] They provide a conceptual framework for understanding the link between personality disorder and risk of violence in individuals when doing structured professional judgement risk assessment. They argue that when assessing risk of violence in personality disorder there is a need to go beyond just listing risk factors such as having a personality disorder and instead produce detailed, comprehensive, individual explanations of risk potential and risk management needs. They outline the principles and stages of risk formulation. They then consider the four higher order dimensions into which DSM-V will organize personality pathologies (dissocial, emotional deregulation, social avoidance and compulsive dimensions) and how these dimensions might relate to violence if doing a risk formulation, drawing on the literature to back up their hypotheses. They suggest that people who are on the emotional dysregulation dimension experience extreme and fluctuating moods, dejection and disillusionment, impulsive anger and irritability and fear of abandonment and rejection.
Where these features co-exist with dissocial characteristics violence can occur when triggered by perceived injury or abandonment by another person. The motivation for injuring others may be more motivated by a desire to prevent loss and express intolerable emotions rather than the desire for retribution and to restore self esteem that is seen in dissocial individuals. They suggest that dissocial individuals may be violent when they perceive themselves to have been injured by another person. This triggers feelings of shame, guilt, humiliation or envy, which are highly threatening to their fragile self esteem. They hypothesize that violence may arise in the socially inhibited dimension when an individual is confronted with unexpected, unwanted or overwhelming evidence of their inadequacy. Violence may arise in the compulsive dimensions when control over their environment or those within it is threatened by the actions of others.
Gilbert and Daffern[19••] wrote a conceptual article integrating contemporary psychological aggression theory and the extant personality disorder–violence literature. They proposed that the general aggression model (GAM) should be drawn on to elaborate on the link between personality disorder and violence. The article points out that there is increasing acknowledgement that violent behaviours result from a complex interaction of the individual and contextual factors. They explain the GAM, which contends that aggressive acts rarely occur without the convergences of multiple precipitating situational factors and predisposing personal characteristics. The GAM postulates that people who hold entrenched and accessible aggression-related cognitions and who are susceptible to experiencing the internal states that activate these cognitions are more prone to act aggressively. The GAM specifies several key constructs that are important in influencing aggression: maladaptive cognitions (these influence how the environment is perceived and interpreted), aggressive behavioural scripts (acquired through observation of others), aggression-supportive beliefs (beliefs about the appropriateness of aggressive behaviour in particular circumstances) and anger and affective arousal (which serves as a cue for retrieval of cognitive schema and aggressive scripts). The authors review the evidence for the importance of these constructs in the personality disorder–violence literature. They conclude that people with personality disorders associated with aggression have more of the constructs described in GAM. They conclude that the empirical evidence is strongest for antisocial personality disorder but a number of constructs, including violent scripts, have been rarely studied. They suggest that their conceptualization provides a focus for researchers to further elucidate the relationship between personality disorder and violence and for clinicians to more systematically assess relevant constructs to determine violence potential in people with personality disorder and to focus their violence reduction efforts.
There is now much evidence that personality disorder is related to offending. The studies above show that some personality disorders other than antisocial are related to particular types of offending behaviour. The studies also demonstrate that, although rates of personality disorder are high in all serious offenders, the role played by personality disorder may be greater in some offences than others: for example, in rapists compared with child molesters, men who kill their fathers rather than their mothers, men who kill their children compared with mothers who kill their children; and in less severe stalking behaviour compared with those who get convictions. These types of studies are only able to show an association between personality disorder and offending but tell us nothing of the causal link. Elucidating the nature of the link is complicated. Any act of criminal behaviour usually arises from a complex interaction between individual predisposing characteristics and a particular set of circumstances, for example, a physical and social context. This is illustrated by the fact that a couple of the studies suggested that, when comorbid axis I and axis II problems such as substance misuse are controlled for, the association between personality disorder and offending or motivation for offending is no longer significant. Thus, it could be that personality disorder is associated with increased substance misuse, and it is substance misuse that leads to offending. The last three articles reviewed are interesting in that they start to suggest a framework for understanding how personality disorder may contribute to criminal behaviour that integrates the personality traits; comorbid problems that are commonly associated with personality disorder such as substance misuse, mood disorders and ADHD symptoms; motivation for offending; maladaptive cognitions; beliefs and attitudes facilitating offending and anger and arousal. These frameworks are helpful when considering risk assessment, risk management and treatment. However, more empirical research is needed to test these hypotheses.
- Fazel S, Danesh J. Serious mental disorder in 23000 prisoners: a systematic review of 62 surveys. Lancet 2002; 359:545–550.
- Webster CD, Douglas KS, Eaves D, Hart HCR-20: Assessing the risk for violence. Version 2. Vancouver, British Columbia, Canada: Mental Health, Law, and Policy Institute, Simon Fraser University; 1997.
- Quinsey VL, Harris GT, Rice ME, Cormier CA. Violent offenders: appraising and managing risk. 2nd ed. Washington DC: American Psychological Association;
- Hare RD, The Hare Psychopathy Checklist-Revised. Toronto, Ontario: Multi-Health Systems; 1991.
- Roberts ADL, Coid Personality disorder and offending behaviour: findings from the national survey of male prisoners in England and Wales. J Forensic Psychiatry Psychol 2010; 21:221–237.
This article examines the relationship between different personality subtypes and different crimes in a nationally representative sample of prisoners in the UK.
- First M, Gibbon M, Spitzer RL, et Structured clinical interview for DSM-IV Axis II personality disorders (SCID-II). Washington, DC: American Psychiatric Press Inc.; 1997.
- Warren JI, South A symptom level examination of the relationship between Cluster B personality disorders and patterns of criminality and violence in women. Int J Law Psychiatry 2009; 32:10–17.
- Francia CA, Coolidge FL, White LA, et al. Personality disorder profiles in incarcerated male rapists and child molesters. Am J Forensic Psychol 2010; 28:55–68.
This study found that rapists have higher antisocial personality disorder scores than child molesters and more similarities to nonsexual offenders in terms of antisocial personality disorder.
- Eher R, Rettenberger M, Schilling Psychiatric diagnoses of sexual offenders. An empirical investigation of 807 delinquents imprisoned for child abuse and rape [in German]. Z Sexualforsch 2010; 23:23–35.
They found that rapists had more personality disorder than child molesters and child molesters had a higher prevalence of sexual disorders.
- Sansone RA, Sansone Fatal attraction syndrome: stalking behavior and borderline personality. Psychiatry 2010; 7:42–46.
This is a review of five studies of borderline personality disorder and stalking which found varying prevalence depending on the sample. The results suggest that borderline personality disorder is more prevalent in community samples of less severe stalking.
- Shaw J, Hunt IM, Flynn S, et al. Rates of mental disorder in people convicted of homicide: national clinical Br J Psychiatry 2006; 188:143–147.
- Richard-Devantoy S, Gohier B, Chocard AS, et Sociodemographic, clinical, criminological and forensic characteristics of 210 murderers [in French]. Ann Méd-Psychol 2009; 167:568–575.
- Flynn S, Swinson N, While D, et al. Homicide followed by suicide: a cross-sectional J Forensic Psychiatry Psychol 2009; 20:306–321.
- Kauppi A, Kumpulainen K, Karkola K, et Maternal and paternal filicides: a retrospective review of filicides in Finland. J Am Acad Psychiatry Law 2010; 38:229–238.
This study looked at filicides in Finland and found male perpetrators had high rates of personality disorder and alcohol abuse whereas the female perpetrators had high rates of psychosis and psychotic depression.
- Liettu A, Säävälä H, Hakko H, et al. Mental disorders of male parricidal offenders: a study of offenders in forensic psychiatric examination in Finland during 1973–2004. Soc Psychiatry Psychiatr Epidemiol 2009; 44:96– 103.
- Léveillée S, Lefebvre J, Marleau Familicide in Quebec – 1986 to 2000 [in French]. Ann Méd-Psychol 2009; 167:591–596.
- Gudjonsson GH, Wells J, Young Motivation for offending among prisoners and the relationship with Axis I and Axis II disorders and ADHD symptoms. Personality and Individual Differences 2011; 50:64–68.
They looked at the motivation for offending among prisoners and found using logistic regression that anxiety, ADHD and substance misuse symptoms were more strongly related to motivation than personality disorders.
- Logan C, Johnstone L. Personality disorder and violence: making the link through risk formulation. J Personality Disord 2010; 24:610–633.
This study provides a conceptual framework for understanding the link between personality disorder and violence, particularly in relation to structured professional judgement assessment of risk.
- Gilbert F, Daffern M. Illuminating the relationship between personality disorder and violence: contributions of the General Aggression Model. Psychol Violence 2011; 1:230–244.
This is a study which draws together the literature of personality disorder and violence and the literature on a contemporary aggression theory. They propose the General Aggression Model should be drawn upon to understand the link between personality disorder and violence.
Papers of particular interest, published within the annual period of review, have been highlighted as:
- of special interest
- • of outstanding interest
Additional references related to this topic can also be found in the Current World Literature section in this issue (p. 68)
�Personality disorder is associated with criminal behaviour.
�Understanding the link will help risk assessment and management and treatment.
�Considering the interaction of maladaptive traits and comorbid problems with situational factors may be more helpful than making categorical diagnoses.
�Contemporary theories about aggression may provide a useful framework for understanding the link between personality disorder and offending.