A brief literature review on mental illness and violent behavior

Stereotypes about mental illness and violence are common among the general public. Link, Phelan, Bresnahan, Stueve and Pescosolido [1] presented a large sample (N = 1444) with vignettes of people with mental illness, in which no violent behavior or thoughts were described, and inquired how likely it was that the “patient” would be violent. Many participants believed it likely that the hypothetical mentally ill individual would perpetrate violence: 17% of respondents endorsed violence as likely among those with minor interpersonal problems, and 33% and 61% thought violence was likely among people with major depression or schizophrenia, respectively. Individuals with mental illness are frequently aware of others’ negative perceptions of them, which can worsen isolation, negative affect and treatment adherence [2, 3].

In research on the prevalence of mental illness among violent offenders, multiple studies have found the highest rates of violence among individuals with substance use disorders, rather than schizophrenia, BPD and other psychotic disorders [7, 8, 9, 10, 11] (Tables 2 and 3). Rates of substance use disorders (including alcohol use disorders and illicit substance use disorders) among self-reported violent offenders range from 20 to 42% [7, 11, 12] (Table 2). Rates of substance use disorders among convicted homicide offenders are lower but still noteworthy, ranging from 1 to 20% [8, 9, 13, 14] (Table 3).

Other studies have approached the question of how mental illness intersects with violence through examining rates of violent behavior among clinical populations. These studies tend to focus on severe/serious mental illness (SMI), that is, disorders which cause or are associated with serious functional impairment or limitations on major life activities [15]. The majority of studies on violent behavior among SMI patients focus on schizophrenia, although some also include other SMIs such as bipolar disorder and antisocial personality disorder (Table 4). Studies on violent behavior and homicide among individuals with schizophrenia indicate these individuals are at increased risk for both violence perpetration and victimization, but that violence is often predicted by comorbid substance use, medication noncompliance and a recent history of being assaulted [16, 17, 18]. Studies on violent behavior among individuals with BPD indicate that emotion dysregulation is a longitudinal mediator of violent behavior and may be a primary mechanism that increases risk for violence in this population [19, 20]. The complex DDs, including DID, have been conceptualized as disorders of emotional dysregulation and are often highly comorbid with BPD [21]. The association of emotion dysregulation with violence in DDs should be further examined.

Dissociative disorders and violent behavior

Notably missing from almost all studies on the intersection of mental illness and violent crime are individuals with dissociative disorders (DDs), including DID and DD not otherwise specified (DDNOS in DSM-IV)/other specified DD (OSDD in DSM-5). This is true of mixed clinical population studies [22, 23, 24, 25], studies on violence and mental illness in the general population [7, 11, 12, 26], as well as forensic studies of convicted violent offenders [8, 9, 13, 14, 27]. Although DID is missing from almost all the research on mental illness and violence, it gets an inordinate amount of focus in films about mental illness, particularly those in the horror and thriller genres such as Split, Psycho, Fight Club or Secret Window which portray people with dissociative self-states as prone to violence including homicide, or within comedies that poke fun at the “outlandishness” of dissociative self-states, such as Me, Myself and Irene. Given the paucity of research on violent behavior among individuals with DDs, coupled with the saturation of stereotypic portrayals of DDs in the media, misunderstanding abounds regarding what role dissociation plays in violent behavior, if any.

A few studies have examined dissociative symptoms, rather than DDs, as a predictor of violent interpersonal behavior within mixed clinical populations (Table 4). They typically focus on trait dissociation, that is, chronic and enduring dissociative experiences across multiple contexts [28], compared to state dissociation, e.g., transient, not enduring and time-limited dissociative experiences [29], the latter of which are often anecdotally reported by violent offenders, such as amnesia for a violent episode and violence-related dissociative episodes [30]. Quimby and Putnam [31] found that among adult psychiatric inpatients, trait dissociation was positively correlated with patient sexual aggression via staff reports. Kaplan and colleagues [32] found a positive correlation between trait dissociation and patient-reported general aggression among psychiatric outpatients. Dissociation has also been posited to play a role in the intergenerational transmission of domestic violence: grouping young mothers who were survivors of childhood maltreatment based on whether or not they abused their own children, Egeland and Susman-Stillman [33] found significantly greater trait dissociation among mothers who were abusive as compared to those who were not.

A number of case study reviews, conducted nearly three decades ago, reported high rates of violent behavior among patients with DID, according to reports by their treating clinicians [34, 35, 36, 37, 38] (Table 5). These studies were typically conducted with small samples derived from the authoring clinician’s case load, relied on clinician reports rather than patient self-report, utilized adult lifetime reporting timeframes rather than specified time frames (the latter is more typical of current studies on violence and mental illness), and did not attempt to objectively verify violent behavior through criminal records or other official documentation. Many studies inquired about DID patients’ violent and/or homicidal dissociative self-states.¹ Therapists reported that between 33 and 70% of DID patients had violent self-states [34, 35, 36, 37]. At times, aggressive self-states within individuals with DID threaten other self-states, which some patients perceive to be internalized homicidal ideation and/or threats, but if carried out, would result in suicide and not homicide. Some of the studies reviewed above did not distinguish violent self-states who were violent towards the individual themselves versus those who were externally violent toward others [34, 35, 36]. Putnam and colleagues [37] make the distinction that while 70% of those with DID had violent or homicidal self-states, 53% of the aggressive self-states were “internally homicidal,” that is, with homicidal ideation toward another self-state. Some DID patients may misperceive these internally aggressive self-states as external violent people, rather than the patient being self-destructive or suicidal [39]. Putnam and colleagues [37] describe internalized homicidal behavior as occurring among 53% of their 100 DID patient sample. Some DID patients may also experience flashbacks of past violence perpetrated by another person against them and mistakenly believe they are perpetrating violence against someone else when in fact they are experiencing an intrusive recollection of the past [39].

Within these aforementioned case studies, clinicians reported that 38–55% of their DID patients had a history of any violent behavior [34, 36, 37, 38]. Ross and Norton [38] reported that of 236 DID patients, 29% of the males and 10% of females reported being convicted of a crime, and the same percentage reported a history of incarceration. While the type of conviction and reason for incarceration were not specified, Ross and Norton [38] describe more antisocial behavior among men than women. Loewenstein and Putnam [36] and Putnam and colleagues [37] report high rates of sexual assault perpetration among their DID patient samples. Among an all-male sample, Loewenstein and Putnam [36] reported 13% of patients reported having perpetrated a sexual assault, while in a predominately female sample, Putnam and colleagues [37] reported 20% of patients reported having perpetrated sexual assault. Lewis, Yeager, Swica, Pincus and Lewis [40] reported severe childhood maltreatment and adult psychopathology among 12 DID inmates who were incarcerated for homicide. Two studies found 19% of DID patients had completed homicide [36, 37]. Loewenstein and Putnam [36] attribute this extremely high rate of violent behavior to the childhood maltreatment these patients experienced which increases their risk for aggression and violence, as well as their reliance on an all-male sample, who have higher rates of violence. Alternatively, Putnam and colleagues [37] describe confusion about “personified intraphysic conflicts” among the patients leading to misperceptions about the degree of actual violence among DID patients, as described above.

These numbers are concerning, but they are not consistent with more recent studies of DD patients and clinicians that utilize different sampling techniques and designs. Within the international prospective Treatment of Patients with DD (TOP DD) Network Study, only 2% of clinicians and 4–7% of patients report that DD (including both DID and DDNOS/OSDD) patients perpetrated sexual coercion or sexual assault toward a partner in their adult lifetime [41]. Additionally, rates of perpetration of intimate partner violence were low among DD patients, according to therapists: only 3.5% of DD patients were reported by their TOP DD therapists as having perpetrated physical or sexual abuse toward a partner in their adult lifetime [6].

To date, no studies have examined the variables that might contribute to violence and/or criminal behavior among individuals with DDs. Given the important role that emotion dysregulation has had in predicting violence among individuals with BPD, emotion dysregulation should be examined as a possible contributing factor among individuals with DDs. Dissociative and PTSD symptoms may also be associated with violence or criminal behavior due to the possibility that when highly symptomatic, individuals with DDs may be overwhelmed and unable to manage their symptoms such that they become vulnerable to dyscontrol. Lastly, potential psychological comordities to DDs related to violent behavior within the literature, such as mood and substance use disorders, should be examined as potential explanatorty variables for recent criminal justice involvement.

The present study

Many questions remain regarding what role mental illness plays in violence. Are mentally ill individuals more likely to perpetrate violence compared to people who do not have mental illness? What psychiatric diagnoses are most highly associated with violent behavior and crime? Are individuals with DDs particularly likely to engage in violent and/or criminal behavior? The present study attempts to provide evidence on violent behavior and crime among individuals with DDs engaged in outpatient treatment.

The purpose of our study was threefold; first, to provide a review of the extant literature on DDs and violent behavior; second, to describe the prevalence of recent criminal justice involvement among a sample of treatment-engaged individuals with DDs; and third, to assess symptomatic predictors of violent behavior and crime among individuals with DDs, including dissociative, emotion dysregulation, posttraumatic stress disorder (PTSD) and depressive symptoms, as well as problematic substance use. We hypothesized that crime rates would be low in our sample of individuals with DDs, with the majority of patients reporting no recent criminal history or involvement with the criminal justice system, unless their involvement was as victims of crime. Additionally, we hypothesized that the aforementioned symptoms (dissociation, emotion dysregulation, PTSD, depression and substance use) would not be significantly associated with recent criminal behavior and justice system involvement.

Procedure

Overview and recruitment

Clinician and patient participants were recruited through the Treatment of Patients with Dissociative Disorders (TOP DD) Network study. The TOP DD Network study is a longitudinal educational intervention study of patients with DDs who are diagnosed with either DID or DDNOS/OSDD. Over the course of 1 year, patients and clinicians watched weekly 7–15 min psychoeducational and skills training videos and completed written reflection and behavioral exercises. Additionally, therapist and patient participants completed surveys every 6 months (at baseline, 6, 12, 18 and 24 months) that provided additional clinical and behavioral data.

Clinicians were recruited through listservs for mental health professionals, professional trauma conferences and emails to participated in the first TOP DD study [42, 43]. Clinicians were asked to enroll as a dyad with one DD patient from their caseload. All clinician and patient participants completed a voluntary consent process, and the study was approved by the Towson University Institutional Review Board. Eligibility requirements for patients in the TOP DD Network study included a DD diagnosis (DID, DDNOS or OSDD); being in treatment with their current clinician for at least 3 months prior to starting the study; reading English at an 8^th grade level; being willing to continue in individual therapy and complete approximately 2 ½ hours weekly of study activities; and being able to tolerate references to trauma, dissociation and safety struggles.

Participants

The total TOP DD Network study included 242 patients who completed baseline measures, presented after the screening measures which verified study eligibility. TOP DD Network study patient participants were majority female (88.6%), Caucasian (82.1%), middle-aged (Median = 41), highly-educated (50.9% had at least a college diploma), and primarily resided in the United States (42.3%), although the study recruited internationally with a sizeable portion derived from Norway (27.5%) as well as other countries (30.2%). About half of participants (55.2%) were either in a dating or married relationship. Patients were primarily diagnosed by their therapists as having DID (63.4%). Clinician participants were primarily female (80%) and Caucasian (91.3%). Most reported years of experience as therapists (Median = 15), as well as in treating trauma (Median = 13), and dissociation (Median = 8). Clinicians primarily worked in private practice (81.1%) or in an outpatient clinic or hospital (41.6%).