Abstract
It is not possible to predict homicide in an actuarial or statistical sense with any appreciable degree of accuracy. In an important and interesting study, Ferranti and colleagues highlight the centrality of religious delusions in women who kill children, consistent with the long-standing recognition that delusions are especially important in the context of violence by the mentally ill. They also note, among other findings, high rates of borderline personality disorder among female homicide offenders found not guilty by reason of insanity (60%) compared with their male counterparts (9%). As a result, the combination of religious delusions, unstable affect, access to children, and features of borderline personality disorder can usefully guide clinical decision-makers toward higher levels of treatment and follow-up, especially in women with aggressive tendencies. Despite the importance of this kind of risk stratification and treatment, however, homicide remains impossible to predict at an individual level.
In 1892, Dora, a 34-year-old servant in Dublin, Ireland, was convicted of the murder of her eight-month-old child.1 She was sent to prison and then onward to the Central Criminal Lunatic Asylum, Ireland's only inpatient forensic psychiatry facility. There, she was detained at the “Lord Lieutenant's pleasure” (Ref. 1, p 149) (i.e., indefinitely, at the discretion of the chief administrator of the British government in Ireland).
On admission, Dora was described as “pale and pasty, hair brown, eyes gray” (Ref. 1, p 149). The admitting officer recommended that she have a “hot bath and be put to bed” (Ref. 1, p 149). She was depressed: she cried “on the least provocation,” she was “subject to bad dreams,” and her diagnosis was “a case of melancholia.” (Ref. 1, p 150). She was also psychotic, with religious delusions: the admitting officer noted that she had “a delusion that she is lost—no heaven for her…. For the last seven years she has neglected her religious duties entirely; this upset and worried her greatly. She thought she was lost with no hope of being saved” (Ref. 1, p 150).
At the time she killed her child, Dora was especially disturbed:
Some days before the crime, she suffered from a violent pain in the head. She felt as if the top of her head was splitting open—she was very depressed and unable to attend her household duties. The loss of her family with very little hope of the last child living, together with her religious troubles, weighted on her mind so that she determined to drown herself and the child. On the morning of the crime, she took the child in her arms and left the house. She wandered off some distance from home, did not know where she was or what she was doing. She imagined that she was followed by a large crowd of soldiers and people. She had no idea of drowning the child when she left the house that morning [and] is now sorry that she did not drown herself—but wishes the child was alive [Ref. 1, p 150].
Over the course of the following years in the Central Criminal Lunatic Asylum, Dora developed further physical symptoms including epileptic seizures, laryngeal stridor, and various neurologic anomalies. Her mental state also worsened considerably despite, or possibly because of, treatment with two medications commonly used in 19th-century asylums: mercury and potassium iodide.2,–,4 Throughout this time, it became apparent that she had a brain tumor, for which there was little effective treatment. Nine years after admission, She was “demented and it [was] almost impossible to get her to speak” (Ref. 1, p 150).
Dora's story was not atypical in the 19th century, as various combinations of mental and physical illness, poverty, and social exclusion led many women into the courtrooms of various countries and onward to institutions such as the Central Criminal Lunatic Asylum in Ireland and equivalent establishments across Europe, the United States, and beyond.5 From a psychopathological perspective, the feature of Dora's story that resonates most strongly today, however, is the centrality of religious delusions in determining the fates of her unfortunate child and herself. Her religious troubles ensured that her child died tragically at the age of eight months and that she herself, also tragically, lived much of her life behind the walls of prisons and psychiatric institutions.
Delusions and Homicide in Women
Over a century later and a continent away, women today represent most of the United States' population (51.6%) but a minority of its prison population (17.7%).6 As a result, correctional policy, treatment, and programming in the United States, as elsewhere, are based, for the most part, on research performed predominantly on men.
Ferranti and colleagues7 go some way toward redressing this imbalance by focusing on the clinical characteristics of female homicide offenders found not guilty by reason of insanity. These women are the contemporary equivalent of Dora from 19th-century Ireland: women who engaged in offending behavior, appeared to be mentally ill, and now live lives shaped largely by governmental policies on imprisonment, forensic mental health care, and rehabilitation, as well as societal attitudes toward this troubled, troubling group.
This is a difficult population to study, not least because women commit fewer homicides than men. As a result, in the time it takes to accumulate a sufficient sample to draw reliable conclusions, laws and societal practices are likely to have changed, and psychiatric diagnostic systems will certainly have shifted direction in the unpredictable pattern that has become their trademark. Notwithstanding these challenges, this is an important area to research and Ferranti and colleagues have managed to assemble an impressive sample of 47 women who were found not guilty by reason of insanity. They were hospitalized at Napa State Hospital, California, between January 1991 and August 2005 for a homicide offense. For comparison, they also studied a random sample of 47 men committed during the same period.
The authors report that religious delusions were more common in women who killed infants (aged under one year) and children between the ages of 2 and 18. It is, of course, already recognized that delusions are important in the context of violence by the mentally ill. This was the case in the 19th century,1,5,8 and it remains the case today.9 In this study, the particular importance of religious delusions among women who kill children is, as the authors suggest “a striking finding” (Ref. 7, p 520). They relate this to “the psychoanalytic literature, which suggests that religious delusions may be derived from the psychological conflicts of mothering” (Ref. 7, p 520, citations omitted).
Notwithstanding this association, do religious delusions have any predictive power for the killing of children? In theory, such a question can be answered definitively only by a prospective study of risk factors and outcomes, and the logistic and ethics-related complexities of such a study are such as to render it essentially impossible. As a result, we need to work as best as possible with the imperfect data that are available to us, while remaining mindful of our extremely limited ability to predict any human behavior, let alone a rare outcome such as homicide. The impossibility of prediction is further underlined by Ferranti and colleagues' observation that “[f]or the majority of women in our sample, their homicide offense was their first and only arrest” (Ref. 7, p 521).
In this context, it is appropriate for Ferranti and colleagues to conclude:
That many female perpetrators of homicide against children have religious delusions at the time of their offenses is a red flag for investigating clinicians who are psychiatrically evaluating women at risk of engaging in violence against their children. Clinicians should be vigilant in asking about religious ideas and assessing for delusional thought content that could lead to violence [Ref. 7, p 520].
At present, it is simply not possible to be any more precise than this in our prediction of homicide.
Women, Crime, and Society
While the centrality of religious delusions is arguably the most arresting finding in their work, Ferranti and colleagues7 make many other interesting observations. They note, for example, that “[w]omen who commit any violent crime, especially homicide, are violating social and psychological norms to a much higher extent than are their male counterparts” (Ref. 7, p 516, citation omitted). Does this fact influence the ways in which our courts, prisons, and forensic mental health services treat women? Should it? There is certainly evidence that women present with a different range of social and psychological needs than do men, not least because women who commit violence demonstrate greater evidence of psychosocial stressors than men show.10 These facts can and should legitimately inform the nature of social and psychological services and supports offered to women in these circumstances.
There are many other interesting differences between male and female offenders identified by Ferranti and colleagues.7 For example, 66 percent of women were or had been married, compared with 38 percent of men (p < .01), and 60 percent of women had a diagnosis of borderline personality disorder, compared with just 9 percent of men (p < .05). As the authors noted:
The question of whether the diagnosis of borderline personality disorder and the psychopathy construct contain within them gender biases is an interesting one to consider, as is whether there are sociological biases that cause practitioners to view affective dyscontrol problems more readily as female disorders [Ref. 7, p 521].
While some research has been done on this question,11,12 further study is clearly needed, especially in light of the data of Ferranti and colleagues indicating “that 60 percent of our female homicide offenders carried a primary diagnosis of borderline personality disorder (compared with 9% of the matched male sample) [thus providing] evidence of its association in women with serious violence toward others” (Ref. 7, p 521).
Conclusions
Ferranti and colleagues7 concluded:
The presence of psychosis with religious delusions, unstable affect, access to children under the age of 18, and borderline traits on Axis II form a particularly ominous constellation of violence risk factors that warrants close clinical scrutiny. These factors, when they occur together in women with aggressive tendencies, should guide clinical decision-making in favor of higher levels of treatment to prevent acts of violence [Ref. 7, p 521].
This is a reasonable and proportionate reflection of their findings, as well as those in the relevant literature. These findings should help inform clinical practice and risk stratification in mental health settings in California and beyond.
Even in light of the findings from this and other studies, one important question remains stubbornly outstanding: can we predict homicide, even in relatively high-risk groups? Despite the admirable work of Ferranti and colleagues, the answer to this questions remains, as it always has been: no.
Identifying and treating risk factors for homicide is important. Identifying and treating mental illness is important. Providing social support and rehabilitation is important. These are all reasonable, sensible interventions, and it is possible, although unproven, that these interventions reduce risk of homicide and other unfavorable outcomes.
Despite the identified risk factors, reasonable interventions, and the myriad studies of homicide, it remains the case that we are unable to predict homicide in an actuarial, statistical sense at an individual level with any appreciable degree of accuracy. As a result, the stories of Dora in 19th-century Ireland and the women studied by Ferranti and colleagues in contemporary California will remain rare, unpredictable tragedies that we fail to predict and, for the most part, struggle to explain.
Footnotes
Disclosures of financial or other potential conflicts of interest: None.
- © 2013 American Academy of Psychiatry and the Law
