Introduction

Filicide—the murder of a child by their parent—is a rare but tragic event that occurs across geographic and social boundaries. Differences in definitions and categorisation, along with the hidden nature of many cases, make incidence estimation difficult (Putkonen et al. 2016). Different approaches to classification have stemmed from psychology, drawing on the characteristics and motivations of the perpetrators, and from sociology, child protection, and paediatrics, drawing more on the victims’ characteristics and the circumstances of the death (Bourget et al. 2007; Brown and Tyson 2014; Putkonen et al. 2016; Resnick 1969; Sidebotham 2013). Both mothers and fathers are known to kill their children, with individual studies finding different relative proportions and characteristics (Bourget et al. 2007).

Various perpetrator characteristics have been identified, including parental mental illness, psycho-social stressors, social isolation, domestic violence, and parental adverse childhood experiences (Bourget et al. 2007). Understanding these characteristics within different parent groups is crucial to any prevention of filicide. As Putkonen et al. (2016) state, different sub-groups of potential perpetrators require different preventive approaches: strategies that work for violent, impulsive parents are unlikely to be effective among socially isolated and depressed but sober, or psychotic parents, for example.

To better understand the characteristics and circumstances of maternally perpetrated filicides, we analysed data from English Serious Case Reviews (SCRs) from 2011 to 2014. In England, there is a statutory requirement for Local Safeguarding Children Boards (LSCBs) to undertake a SCR whenever a child (from birth to their 18th birthday) dies and abuse or neglect is known or suspected (irrespective of whether the abuse or neglect was the primary cause of death, or whether the abuse or neglect was known prior to the child’s death) (HM Government 2015). These SCRs are independent multi-agency case reviews focused on learning lessons about how organisations are working together to safeguard and promote the welfare of children. There is a requirement that these reviews should be published, unless there are good reasons not to, such as protecting the welfare of surviving children, or because of ongoing court processes. Published reviews are available on individual LSCB websites and on a national repository held by the National Society for the Prevention of Cruelty to Children (NSPCC). The primary aim of this study, which was embedded within a wider study of all SCRs, was to analyse the circumstances and background features of cases of maternal filicide to identify factors that may help prevent future child deaths. Secondary aims were to compare the circumstances and background features of cases of maternal filicide with those of paternal filicide, and to identify learning points for mental health professionals.

Materials and methods

Data were obtained for all SCRs notified to the Department for Education (DfE) relating to child deaths in England between 1 April 2011 and 31 March 2014. Basic demographic data and initial details of the circumstances of the death were provided by DfE. This included the ages and gender of the children, their ethnicity, and the Local Authority area in which they were resident. These cases were matched to published SCRs on the NSPCC national case review repository (NSPCC 2017), individual LSCB websites, or reports provided by LSCBs.

Each SCR was subject to a process of ‘layered reading’ (Brandon et al. 1999) to supplement the quantitative data and to develop a preliminary coding frame for qualitative analysis. Each case was categorised using our previously developed classification for violent and maltreatment-related deaths (Table 1) (Sidebotham et al. 2016). Cases where deaths were directly attributable to child maltreatment within the immediate family were included. Cases were excluded where the suspected perpetrator was outside the immediate family and where child maltreatment was not the primary cause of death. The relationship of the suspected perpetrator to the child was identified, where possible, from the information in the SCR report. Where one or more person(s) had been convicted of the child’s murder/manslaughter, this was taken as the suspected perpetrator; otherwise, this was based on the most likely perpetrator from the case information provided, or left as ‘unclear’ if there was insufficient information. For the purposes of this study, the immediate family was taken to be the birth or adoptive mother or father with whom the child was living, plus any partner of the birth or adoptive mother or father, and any siblings; parents were taken to be the biological parents of the child; and non-biological parent figures and partners of a biological parent are referred to as such.

Table 1 Classification system used for categorising violent and maltreatment-related deaths (Sidebotham et al. 2011; Sidebotham et al. 2016)
Full size table

Cases of maternal overt or covert filicide were subject to a second and third round of layered reading using principles of inductive thematic analysis (Braun and Clarke 2006). The circumstances of the death and the familial and environmental background factors were manually coded. The information obtained was dependent on what was included in the published Serious Case Reviews, and we did not have access to any original case files. Child, parent, and family characteristics were included if they were mentioned in a SCR. As such, there were no specific criteria for any variables such as drug and alcohol misuse, parental mental ill health, or social isolation. The data were then reviewed to identify consistencies and discontinuities in the emerging themes.

Quantitative data were analysed using IBM SPSS Statistics 24. Where appropriate, chi-square tests, Fisher exact tests, and Student t tests were used to compare the maternal with the paternal filicide cases.

The data obtained for this analysis were from published, anonymised reports available in the public domain. Ethical approval was obtained from the University of East Anglia Research Ethics Committee. This project adhered to the University of East Anglia’s and the University of Warwick’s Research Codes of Conduct and to the Department for Education and the Department of Health Research Governance Framework for Health and Social Care.

Results

The DfE was notified of 197 SCRs for children who had died between 1 April 2011 and 31 March 2014. A published report was available for 165 (83.8%). Six SCRs had not been published to protect the privacy of surviving children or other relatives; for 13, publication had been delayed, primarily because of ongoing criminal proceedings; for the remaining 13, no report had been published and no response was obtained from the LSCB (Fig. 1).

Fig. 1
figure 1

Serious case review data available for analysis

Full size image

Eighty-six cases (43.7%) were directly attributable to abuse or neglect within the context of the immediate family; 7 (3.6%) were attributed to extrafamilial homicide or fatal assaults; 91 (46.2%) were related to but not directly caused by child maltreatment; and 13 had insufficient information to categorise the death, or appeared unrelated to child maltreatment (Table 2).

Table 2 Category of death and suspected perpetrator as identified in the SCRs (whole cohort)
Full size table

The mother was the suspected perpetrator in 20 (23.3%) of the 86 deaths directly attributed to child maltreatment within the family; the father in 25 (29.1%); an unrelated male in 14 (16.3%); and both parents in eight (9.3%). In three cases, the suspected perpetrator was another close relative, and in 16 cases the suspected perpetrator was unclear.

Table 3 compares the characteristics of direct maltreatment deaths where the suspected perpetrator was the mother (n = 20) with those where it was the biological father (n = 25). Of the 12 overt filicide cases perpetrated by mothers, four involved the mother killing all children in the family and in a further six the victim was an only child. In all 12 cases, the mother committed or attempted suicide.

Table 3 Case characteristics identified from the SCRs for maternal and paternal filicide and direct fatal maltreatment cases
Full size table

Children killed by their mothers were more likely to be of Black and Minority Ethnic origin; their parents were more likely to be separated; and father perpetrators were more likely to have a history of violent crime. Over two thirds of cases occurred in a family context of known domestic violence. However, the cases varied in that 12 of the mother perpetrators were themselves victims of domestic violence, while 15 of the father perpetrators were known to be perpetrators of domestic violence. Other differences between cases perpetrated by mothers and fathers did not reach statistical significance.

Qualitative analysis

Full SCR reports were available for 10 of the 12 maternally perpetrated overt filicides and for all three covert filicides. These reports were analysed thematically and four core themes identified: domestic violence, maternal mental illness, separation and maternal isolation, and the invisibility of the child. Sub-themes were identified for each of these themes, and in all there were exceptions that did not fit the theme (Table 4).

Table 4 Themes and subthemes from qualitative analysis of the SCRs
Full size table

The role of domestic violence

In nine cases, including one case of covert filicide, the mother had made domestic violence allegations including direct physical assaults, and sustained emotional and sexual violence and coercive control. This culture of violence was noted to have a direct impact on mothers’ health and well-being, and in some cases was directly linked to deteriorating mental health or to self-harming behaviour. In one case, ongoing domestic violence-related court proceedings precluded the provision of mental health services for the mother. In other cases, it appeared that a focus on mothers’ immediate physical safety and her separation from the perpetrator meant her parenting capacity was overlooked. However, in six of the nine cases, partners denied the veracity of the allegations or made counter allegations of violence or controlling behaviour by the mother.

The role of maternal mental health and illness

While maternal mental illness was highly prevalent in these cases, severe mental illness was rarely identified before the fatal event, and in many cases, there were no indicators of significant mental illness. In some cases, viewed in retrospect, signs were missed by professionals, including indicators of delusional or paranoid thinking. In others, concerns had receded by the time the mother was assessed. With hindsight, relatives also identified areas of concern, though these had rarely been sufficient for them to seek help. The difficulties in identifying clear pointers of risk were raised by a psychiatric expert reviewing one case:

Her view is that it is more likely that [mother] had a more longstanding serious mental illness, possibly a chronic psychotic state which she was able to contain by her unusual lifestyle, which might have merely seemed eccentric to others, except at moments of great stress. [The doctor] has suggested that this would explain her fluctuating presentation and the fact that on occasion of contact with professionals she appeared to have none of the symptoms that a non-specialist would regard as indicative of a psychosis i.e. she did not appear to have hallucinations or thought disorder. Any evidence of delusional thinking or unreasonable beliefs would only have been evident if the reasons had been explored and then probably only by an experienced mental health practitioner. [Case A]

The role of separation and maternal isolation

Many mothers appeared isolated from their family and community. Among those who were isolated, this could be compounded by relationship breakdown, often precipitated by domestic violence. Separation from an abusive partner may result in safety for the mother and child but could also increase her sense of isolation and distress. One mother’s experience of emotional isolation due to her partner’s infidelity was considered a key factor in the mother’s stress and deteriorating mental health.

The invisibility of the child

With four important exceptions, the children killed by their mothers were previously unknown to social services as being at risk of harm. To professionals, they presented as healthy, thriving children with no indicators of concern. That their families were ‘well-educated’, ‘middle class’, and enjoyed ‘a good standard of living’ was specifically mentioned in some reviews. The relationship between the mother and the child was typically perceived—by professionals and other family members—as loving and warm, with the mother responding well to the physical and emotional needs of the child. In one case, it was postulated that the mother’s love for the child was a primary motivation for her taking the child’s life before committing suicide.

Of the three covert filicide cases, one involved a teenage mother with a concealed pregnancy who had suffocated her baby immediately after birth. There had been no previous concerns about the mother who was described as ‘an articulate and intelligent young woman’. She had not presented with any typical signs of pregnancy and denied that she was pregnant. The other two cases involved older women, one of whom had not presented with any prior concerns.

There were, however, four cases of maternal filicide where child welfare concerns had been raised before the fatal event. In one case, the child was subject to a child protection plan following a serious physical assault by the mother. In another, a child protection investigation had been initiated after the 5-year-old disclosed being slapped by his parents and witnessing arguments at home. In the third, long-standing concerns about the well-being of all the children in the family resulted in a series of child protection plans and interventions which were, however, limited by poor parental engagement. In the fourth, there were concerns of maternal ambivalence toward her child, and fears that she might abscond with the child to another country. A striking feature of all these cases was the mothers’ awareness of the involvement and impact of social services, and their fear of their child’s removal. This fear may have precipitated the fatal actions of one mother:

[Mother] went on to say that she ‘did what she did because it needed to be done’, and that ‘the system was corrupt; Social Workers were treating her badly and had taken her daughter’… [Mother] informed medical staff that she was dead at home because she had suffocated her. [Case B]

Another mother’s previous engagement with social services had resulted in her presenting a positive appearance thus avoiding arousing concern—a form of disguised compliance:

The contribution of family members and friends was extremely helpful, both in terms of understanding what the children and their mother were like, as well as understanding the mother’s experiences of what it was like to be involved with the services who were working to support her and the children. There were several key factors that came through in these conversations. The first and overriding message was that the children’s mother was a loving, caring and competent mother, who in her normal life would do anything to protect her children. The next was that she was very keen to display a positive aspect to anyone in a position of authority; she was very able to understand and display what was expected of her. Several of those we spoke with indicated that while she was able to display this positive aspect, because of early interactions with children’s services, she was in fear that if she did not do what was expected of her, her children might be taken back into the care of the local authority. [Case C]

Discussion and conclusions

In this national cohort of child maltreatment fatalities, males were the most common perpetrators of deaths directly caused by maltreatment. Nevertheless, mothers were implicated, alone or with their partner, in nearly one third of cases. This is similar to our previously published national analysis (Sidebotham et al. 2011). In comparing mothers with fathers or father figures, some important findings emerge. Those deaths resulting from impulsive violence or severe, persistent cruelty are almost exclusively perpetrated by males, while those with an apparent intent to kill the child (overt or covert filicide) are slightly more likely to be perpetrated by mothers. In keeping with this, paternal perpetrators were more likely to have a history of violent crime, suggesting that violence and impulsivity are prevalent characteristics. In contrast, none of the mothers had such a history.

It was striking that in nearly all maternally perpetrated filicides, the victim was either an only child or all children in the family were killed, and the mother committed or attempted suicide. In addition, the majority of mothers had underlying mental illness, and two thirds were victims of domestic violence.

These findings shed additional light on the nature and characteristics of maternal filicide, building on previous work (Bourget et al. 2007; Friedman et al. 2005; Mugavin 2005; Oberman 2003; Putkonen et al. 2016). In particular, our findings highlight the important role of domestic violence and its interaction with maternal mental health, echoing characteristics identified by Putkonen et al. (2016). Domestic violence was a common, though not universal, finding. While this often involved acts of physical violence, the pervasive nature of coercive control was clear (Stark 2007). However, many cases involved counter allegations of control and abuse perpetrated by the mothers. This highlights the damaging nature of these relationships where elements of violence, control, and domination can impact on the mothers’ emotional health and well-being and on their parenting. Separation of a mother from a violent partner may bring some immediate safety, but can also lead to increased stress and isolation. Some mothers may view filicide, together with suicide, as the only way of escape both for themselves and their children.

The role of maternal mental illness in these cases presents challenges to professionals. While a majority of the mother perpetrators had identifiable mental health problems, for many, these were not obvious prior to the fatal event. Our qualitative review demonstrated how, in many cases, the severity of the mothers’ mental health problems were unrecognised or did not lead to timely mental health assessments or intervention. This has implications for all those working in mental health, primary care services, and child and maternal welfare. Services must be accessible; concerns, including those expressed by relatives, should be taken seriously; and health and social care practitioners must be sensitive to signs indicating escalating stress or severity of mental illness, particularly where there are delusional thoughts relating to children, suicidal ideation, or self-harm.

It is important to note that most children in this cohort did not present as being at risk of harm. Practitioners should therefore focus on the wider family dynamics and parental risk indicators, rather than solely looking for evidence of harm in the children. This is even more important given that some mothers are particularly attuned to the possibility of their child’s removal and that this may be a factor in some filicides. This resonates with the conclusion of Oberman (2003) that ‘maternal filicide is committed by mothers who cannot parent their child under the circumstances dictated by their particular position in place and time’ (p494) and that ‘only when we come face to face with the desperation of these mothers can we begin to devise effective manners of protecting both them and their children’ (p514).

This study, drawing on a comprehensive national cohort of child maltreatment deaths provides a unique perspective on the nature and characteristics of maternally perpetrated filicides. Nevertheless, there are important limitations. While we are confident, because of the legislation surrounding the requirement for SCRs, that all cases were notified, this has not been independently verified, and there is currently no way of confirming whether this is indeed the case. We recognise that, even as a comprehensive national cohort, the number of cases included remains small, and some caution should therefore be applied in extrapolating from these findings. For our analysis, we were dependent on the information gathered for the SCR and published in the final report. We did not have access to any primary case files or coronial or legal outcomes in the cases. We were thus reliant on the information that the SCR authors chose to include in their reports, which could be subject to conscious or unconscious bias. The fact that a particular characteristic was not mentioned in a review could have been because it was not present, it had not been identified, or that the SCR author did not consider it significant. It is possible, therefore, that some characteristics had a higher prevalence than that recorded. In addition, there were no standard criteria for defining variables such as mental illness, and there could be considerable variation between SCRs as to what was or was not included within any such definition. The retrospective nature of this study meant that so we cannot provide comparator data for non-filicidal mothers. Even as a national cohort extending over 3 years, the number of maternal filicides is small, limiting the statistical significance of some of the observed trends. However, the addition of qualitative data strengthens the relevance and validity of these findings and contributes to our understanding of these cases.

Maternal filicides are, fortunately, rare. We identified less than five cases per year in England. They are extremely distressing, and those working in child and family welfare must do all we can to support families at risk, and to minimise those risks. In a systematic review of maternal filicide, Friedman et al. (2005) highlighted the limited state of knowledge of the maternal characteristics that distinguish mothers at risk of killing their children. The heterogeneous and often hidden nature of these risks emphasises that it will not be possible to fully prevent all maternal filicides. Nevertheless, a deeper understanding of the characteristics of these cases may facilitate strategies that help to minimise risk. Most importantly, practitioners must be aware of the impact of domestic violence on mothers and children, and the need to adopt a supportive but professionally curious stance; to be alert to signs of escalating stress or worsening mental ill-health; and to provide supportive and accessible structures for at-risk families.