The University of Newcastle, Australia

*Corresponding author: Dr Julie Casey, The University of Newcastle, Australia.Tel: 0061 02 0411271561, E-mail: [email protected]

Received: August 08, 2022 Published: November 16, 2022

Citation: Casey J, et al. (2022). Risk Factors in Abusive Parent and Child Relationships. Neonatal. 3(1):6

Copyright: Casey J. © (2022).

ABSTRACT

Background: In New South Wales Australia, relevant child protection services employ a Structured Decision-Making assessment system in order to triage ‘at risk’ children. Children and families can be designated as high or low-risk. Objective: The present study aims to explore maternal mental health, reflective functioning and personality symptoms that are suggested as precursor variables in explaining the perpetration of parental child abuse. Participants and setting: The study examined a sample of groups from the community of assessed high-risk mothers (n=21) and low-risk mothers (n=19), along with a comparison group of volunteer mothers from the community who had no recorded risk reports (n=10). Methods: The participants that had been designated high and low-risk groups as determined Structured Decision-Making assessment system in order to triage ‘at risk’ children were evaluated for their mental health status (DASS-21); personality (ZAN-BPD) and Reflective Functioning ability (Parent Development Interview). Results: The results revealed significant differences between the Community group and the two ‘at risk’ groups only. The results also suggest that the Structured Decision-Making system does not adequately evaluate mental health status; personality and parent-child attachment capacity during triage. Conclusion: Based on this study, child protection agencies should re-think how “at-risk” parents are defined. The risk (re)assessment tools used include measures of a parent’s depression, stress and borderline symptoms, as well as reflective capacity. Further research on defining low and high-risk pants on a larger sample size is also warranted.

Keywords: attachment, parenting, child abuse, reflective functioning

INTRODUCTION

Child abuse and neglect represent a significant and persistent problem for social services both locally (Stoltenborgh, Bakermans-Kranenburg, Alink, & van Ijzendoorn, 2015) and internationally (World Health Organisation, 1999, 2006). According to Stoltenborgh et al., prevalence rates for abuse – divided into sexual abuse, physical abuse, emotional abuse and neglect – in Australia are reported to be between 13% – 14%. Meaning that more than one in ten children suffer some form of maltreatment during their development (Stoltenborgh et al., 2015).

In New South Wales the State government report from Family and Community Services (New South Wales Government, Family and Community Services) ‘Child Deaths (NSW FACS) Annual Report 2015’, recorded the deaths of 79 children who were known to FACS and most of whom a “risk of significant harm’ report had been recorded (New South Wales Government Family and Community Services, 2016). A report from Family and Community Services Child Deaths 2018-2019 showed that twenty children died from injuries that were the result of abuse (11/20) or neglect (2/20), or other suspicious circumstances (7/20). Families known to child protective services were overrepresented in injury related deaths (New South Wales Government Child Death Review Team, 2019).

Measuring child risk is a complicated problem with risk factors involving many biological, psychological, and social variables. Social variables that have been associated with abuse include financial hardship, low socioeconomic status; violence in the home and housing stress (Nitahara, Tachibana, & Okuyama, 2018). Psychological variables relate to problems such as parents’ low level of education; parents’ substance abuse; poor mental health, parental stress levels and personality characteristics (van IJzendoorn, Bakermans?Kranenburg, Coughlan, & Reijman, 2020). For example, Sidebotham and Heron (2006) in their study of child abuse of “children in the nineties” suggest the “at-risk” factors for child maltreatment include parental low educational achievement, parental past psychiatric history and parents with their history of beined as children (Sidebotham and Heron, 2006). Hamilton and Brown (1999) suggest the following as predictive risk warning signs: - re-referral to child protective services; parents’ mental health problems; parents’ substance abuse; child’s temperament and learning difficulties (Hamilton and Browne, 1999).

The identification of risk factors is of practical importance since greater accuracy means that welfare services can better focus their intervention, treatment and protection regimens (Pecora, Chahine, & Graham, 2013). Moreover, the typical high demands and under-resourcing of social services means priorities need to be established to effectively and efficiently target service provision. In New South Wales the Child and Family Services partial out children “at-risk” into high-risk and low-risk categories, based on a structured questionnaire – the Structured Decision-Making System (NSW SDM) (Barlow, Fisher, & Jones, 2012; New South Wales Government Family and Community Services, 2012). Such systems have been found to show predictive validity and good inter-rater reliability (Barlow et al., 2012) however, they focus on certain predictive and contextual variables (Shlonsky and Wagner, 2005) perhaps at the expense of other problems.

One issue that has long-term implications is that the intergenerational nature of abuse (Widom, Czaja, & DuMont, 2015). Does a history of abuse as a child presage abuse as a parent? Such a concept can all too easily descend into “blaming the parent” – a trajectory that needs to be avoided. Nevertheless, the longer-term effects of abuse on personality development and mental health are parameters that might be expected to influence parenting abilities. Such seemingly speculative relationships are given a convincing theoretical framework by Bowlby’s attachment theory (Bowlby, 1978).

A recent review by Levy et al., (2015) captures the relationship between attachment in childhood and the impact of inadequate attachment on the development of Personality Disorders and longer-term mental health problems. Such relationships provide the background to the current study in which mothers identified as being “at-risk” of abusing their infants were assessed for Borderline Personality Disorder; mental health difficulties and Reflective Functioning. Reflective Functioning, measured by the Parent Development Interview, is taken here as an index of the mother’s ability to show insight into both her own, and her child’s, psychological makeup and mental processing (Levy, Johnson, Scala, Temes, & Clouthier, 2015).

The study recruited mothers of young children (including “at risk” families known to child protection services in New South Wales, Australia) with a view to explore relationships between intergenerational abuse, reflective functioning, attachment, mental health and trauma to evaluate the impact of parental personality on risk for child abuse. Of particular interest were the possible contributions of mental health and personality symptoms that have been shown to have an impact on bonding and mother-child relationships -variables such as Reflective Functioning and Borderline Personality Disorder.

METHOD

Participants

The participants in this study comprised 50 mothers (age 18-40 years) each of whom had at least one dependent child under the age of one year. Target mothers were grouped either as at highrisk of abusing (n=21); or at low-risk of abusing (n=19) a child. A third “community” group of volunteers formed the comparison group (n=10). The target mothers had all been reported to, and assessed by, New South Wales (NSW) Family and Community Services, with those designated as high-risk receiving a statutory intervention order or an Assessment Order and Care Plan. Those designated as low-risk being referred to non-government agencies (such as Family Support Services) for assistance.

The study adopted a cross sectional design, with three groups of mothers. The first group was a convenience sample of community mothers taken as representative of a normal baseline control group. The other two groups were “at risk” mothers, one designated as high-risk (current involvement with Family and Community Services; FACS) and the other designated low-risk (referred by FACS to non-government agencies), according to the New South Wales Structured Decision Making assessment protocol “SARA” (guided by the New South Wales Department of Family and Community Services, The Structured Decision Making System 2012).

The designations were made on the basis of the NSW SDM. This system uses a question and answer format to determine the risks of child abuse (of physical, sexual and psychological abuse or neglect) on the basis of responses to questions relating to details of the household; allegations in the household; details of the child; cultural background and vulnerability factors (New South Wales Government Family and Community Services, 2012). Caseworkers use the system in order to substantiate harm and evaluate risk of harm to the child. The comparison “community” group of mothers were recruited from local nongovernment and health agencies and had no abuse history and no referrals to Family and Community Services for risks of abuse.

Instruments

Maternal mental health was evaluated by the 21 item Depression, Anxiety and Stress Scale (DASS-21) (Lovibond and Lovibond, 1995). This provides a screen for general mental wellbeing and effectively evaluates affect, anxiety and stress. Maternal personality was assessed using the Zanarini Rating Scale for Borderline Personality Disorder (ZAN-BPD(Zanarini, 2003). This is a nine-item scale designed to measure Borderline Personality traits. The scale rates anger, mood, feelings of emptiness; sense of self, dissociation, suicide ideation, feelings of abandonment, impulsivity, interpersonal relationships and unstable relationships as experienced during the previous two weeks (Zanarini, 2003; Zanarini, Frankenburg, & Vujanovic, 2002). Maternal reflective functioning was assessed using the Parent Development Interview revised short version (PDIR2S(Slade et al., 2003). This is a semi-structured interview which is designed to elicit responses from the mother that illustrate representations of herself as a parent; her reflections of what her child is like and of her relationship with the child. The Parent Development Interview is intended to examine the parent’s internal working models of parenting and to capture how these relate to the parent’s own attachment mental representations and to their own parenting behaviour (Slade et al., 2003).

Procedure

Following ethical approval from the University of Newcastle Human Research Ethics Committee and the Hunter New England Health Research Ethics Committee (H20112), recruitment of “at-risk” mothers was made via non-government agency case co-ordinators and case managers. These case managers and co-ordinators provided a description of the study and provided the contact details of the researcher if any possible participant wished to volunteer. Subsequent contact identified appropriate volunteers and scheduled interviews. The “community” mothers were recruited through Day Care Centres, pre-schools and nongovernment agencies.

All interviews were undertaken in private, comfortable and child-friendly clinical rooms. The measurement instruments were completed in the same order for all participants – DASS21 (Lovibond and Lovibond, 1995); ZAN-BPD (Zanarini, 2003); PDI (Slade et al., 2003). The PDI responses were recorded on a Dictaphone voice recorder and subsequently transcribed. The transcriptions were then forwarded to the Coding Consortium at the Anna Freud Centre, London, for independent coding on level of Reflective Functioning.

RESULTS

Statistical analysis was performed using the Welch t-test for unequal variances in order to make initial comparisons of group means between the three study groups. Where the Welch test indicated a difference, post hoc testing was conducted to find out which groups were different to each other and the size of that difference. ANOVA was then used to compare differences between the three groups if the Welch test allowed.

Demographic profiles of the three groups are shown in Table 1 which shows differences between the “at-risk” groups and the “community” group. The demographic profiles show that both the low and high-risk mothers report experiencing trauma, domestic violence, were in a de-facto relationships or were single parents. The community mothers by contrast reported no trauma experiences (one exception), and/or domestic violence relationships.

[‘Demographic data for the Community, Low-risk and High-risk mothers’]

Table 1. Demographic data for the Community, Low and High-Risk Mothers

		Community	Low Risk	High Risk	Total
		n (%)	n (%)	n (%)	n (%)
Sample		10 (20%)	19 (38%)	21 (42%)	50 (100%)
Relationship	Married	7 (70%)	0 (0%)	0 (0%)	7 (14%)
	Defacto	2 (20%)	3 (16%)	6 (28%)	11 (22%)
	Single	1 (10%)	15 (79%)	13 (62%)	29 (58%)
	Unknown	0 (0%)	1 (5%)	1 (5%)	2 (4%)
	Widowed	0 (0%)	0 (0%)	1 (5%)	1 (2%)
Residential
Status	Home/rental	10 (100%)	17 (90%)	2 (9%)	29 (58%)
	Unknown	0 (0%)	2 (10%)	1 (5%)	3 (6%)
	Refuge/Residential(rehab)	0 (0%)	0 (0%)	18 (86%)	18 (36%)
Domestic Violence	Yes	1 (10%)	15 (79%)	16 (76%)	32 (64%)
Other	No	9 (90%)	4 (21%)	5 (24%)	18 (36%)
Trauma	Yes	1 (10%)	11 (58%)	18 (86%)	30 (60%)
	No	9 (90%)	7 (37%)	3(14%)	19 (38%)
	Unknown	0 (0%)	1 (5%)	0 (0%)	1 (2%)

In particular, the “at-risk” mothers were more likely to report lower socioeconomic status; to sustain a transitory lifestyle; and to have suffered domestic violence and other traumatic experiences. The demographic data show that there were no clear differences between the high-risk and low-risk groups except for their residential status. Many of the high-risk group were accommodated in refuge and drug/alcohol rehabilitation facilities.

The psychometric results of mental health symptoms (DASS-21); borderline personality symptoms (ZAN-BPD) and attachment (PDI) are shown in Tables 2, 3 and 4.

[Group Means, SD’s and statistical comparisons for the DASS-21 Scale]

Table 2. Depression, Anxiety and Stress Scale mean, standard deviation comparisons across community, low and high-risk study groups

	Community	Low Risk	High Risk			Comm vs Low Risk		Comm vs High Risk		Low vs High Risk
	Mean (SD)	Mean (SD)	Mean (SD)	Welch F	p	Welch F	p	Welch F	p	Welch F	p
Depression	2.50 (2.55)	7.06 (6.84)	6.35 (5.49)	5.21	0.012*	-4.56	0.14	-3.85	0.253	0.70	0.999
Anxiety	1.90 (2.60)	5.50 (5.35)	6.48 (5.20)	6.20	0.006*	-3.60	0.20	-4.58	0.055*	0.978	0.999
Stress	3.80 (2.82)	9.47 (5.00)	8.35 (4.61)	8.95	0.001*	-5.67	0.008*	-4.55	0.035*	1.12	0.999

Note. Comm = Community Group. vs = versus

Table shows that the DASS-21 subscale for Depression shows overall significance (Welch statistic F=5.21; p=0.012) but no statistically significant differences on a group-by-group comparison. The source of the overall significance would appear to be the lower levels of depression shown by the community group compared with the two “at-risk” groups – community group (M=2.5; SD= 2.55); high-risk group (M=6.35; SD=5.49); low-risk group (M=7.06; SD=6.84). Clearly, the greater variance of the two “at-risk” groups compared to the community group responses needs to be borne in mind here.

The DASS-21 subscale for Anxiety produced a similar result to that for Depression. The overall Welch statistic (F=6.2; p=0.006) is highly significant but the group-by-group comparisons do not reach significance. Even so, the community group by highrisk group comparison approaches significance (p=0.055). The inference again is that the significance of the overall Welch statistic derives from the differences in anxiety shown between the community group and the two “at-risk” groups. There was no difference between the high-risk and low-risk groups.

The DASS-21 subscale for Stress is significant overall (Welch F= 8.95; p=0.001) and the group-by-group comparisons show differences between both the Community and high-risk group (p=0.035) and the community and low-risk group (p=0.008). Again, there are no differences between the high-risk and lowrisk groups.

The ZAN-BDP Total scores show a highly significant overall difference (Welch F=28.95; p<0.001) and significant differences between both the community group and the high-risk group (F=-6.83; p<0.001) and between the community and the low-risk group (F=-8.36; p<0.001). There were no statistical differences between the two “at-risk” groups. The results for the ZAN-BPD are shown in Table 3.

[Group Means, SD’s and statistical comparisons for the ZAN-BPD scale]

Table 3. Zanarini-Borderline Personality Disorder symptoms total and subscale mean, standard deviation and Welch test comparisons across community, low and high-risk study groups

	Community	Low risk	High risk	Welsh Test		Comm vs low risk		Comm vs high risk		Low vs high risk
	M (SD)	M (SD)	M (SD)	M	p	M	p	M	p	M	p
ZANTotal	1.22 (1.716)	9.58 (6.149)	8.05 (4.140)	28.95	<.001*	-8.36	<.001*	-6.83	<.001*	1.54	.948
Aff/Dis	0.6 (0.699)	3.47 (2.412)	3.48 (1.601)	29.73	<.001*	-2.87	<.001*	-2.88	<.001*	-0.003	.999
Cog/Dis	0.4 (0.699)	2.84 (2.63)	1.71 (1.978)	9.31	.001*	-2.44	.014*	-1.31	.330	1.13	.289
Impul	0 (0)	1.26 (1.628)	0.52 (0.981)	4.07	0.024*	-1.26	0.028*	-0.52	.777	0.74	.169
Disturb	0.22 (0.667)	2 (1.491)	2.33 (1.317)	19.74	<.001*	-1.78	.005*	-2.11	.001*	-0.33	.999

Note. Comm = Community. Vs = versus. Affect Dis = Affective Disturbance, Cog Disturb = Cognitive Disturbance

The ZAN-BPD subscale for Affective Disturbance shows a significant overall difference (F=29.73; p<0.001); a significant difference between the community group and the high-risk group (F=-2.88; p<0.001) and between the community group and the low-risk group (F=-2.87; p<0.001). There were no differences between the high-risk and low-risk groups (p=0.99).

The ZAN-BPD subscale for Cognitive Disturbance yielded overall significance (F=9.31; p=0.001) and a significant difference between the community group and the low-risk group (F=- 2.44; p=0.014). There was no significant difference between the community and high-risk group (p=0.33), nor between the highrisk and low-risk groups (p=0.29).

The ZAN-BPD subscale for Impulsivity was significant overall (F=4.07; p=0.02) with a significant difference between the community group and the low-risk group (F=-1.26; p=0.03) but no significant difference between the community group and the high-risk group (p=0.78) nor between the high-risk group and the low-risk group.

The ZAN-BPD subscale for Disturbed Relationships showed overall significance (F=19.74; p<0.001) together with significant differences between the community and both the low-risk (p=0.005) and the high-risk (p=0.001) groups. Again, there was no difference between the low-risk and high-risk groups (p=0.99).

The coding procedure of the PDI interview questions is aimed at eliciting the level of Reflective Functioning (RF) revealed in the responses. In all cases a higher score represents a higher rating for the level of RF shown by the response. Table 4 represents the separate coded themes as subscales with attached scores. The combined Overall RF score shows a statistical difference using the Welch statistic (F=3.88; p=0.036). There are also significant differences between the community group and the high-risk group (F=1.67; p=0.003) and between the community group and the low-risk group (F=1.47; p=0.01), but no difference between the high-risk and the low-risk groups (p=0.99). The coded data from responses to the PDI are shown in Table 4.

[Group Means, SD’s and statistical comparison for the PDI]

Table 4. The Parent Development Interview mean, standard deviation comparisons across community, low and high-risk study groups

	Community	Low Risk	High Risk	Welch F		Comm vs Low Risk	Welch   F	Comm vs High Risk	Welch   F	Low to High Risk
	M (SD)	M (SD)	M (SD)		p		p		p		p
PDI Overall	5 (1.70)	3.53 (0.96)	3.33 (1.155)	3.878	0.036*	1.474	0.010*	1.667	0.003*	0.193	0.999
PDI Clicked	4.9 (2.13)	3.53 (1.26)	3.29 (0.845)	2.649	0.096	1.37368	0.0344	1.61429	0.009	0.2406	0.999
PDI Not Clicked	5 (1.83)	3.95 (1.43)	3.67 (1.528)	1.951	0.165	1.05263	0.269	1.33333	0.091	0.2807	1
PDI Rel Aff Pers	3.33 (1.50)	2.72 (1.23)	2.86 (1.153)	0.543	0.589	0.611	0.711	0.476	1	-0.135	.1
PDI Joy	3.2 (1.69)	2.47 (1.31)	1.9 (0.831)	3.326	.056	0.726	0.404	1.295	0.025	0.569	0.314
PDI Pain/ Diff	3.5 (1.27)	2.42 (0.84)	2.52 (1.762)	2.896	0.076	1.07	0.062	0.97619	0.097	-0.10276	.1
Changed You	3.1 (1.52)	2.42 (0.69)	2.05 (0.740)	2.787	0.085	0.679	0.202	1.052	0.015*	0.373	0.63
PDI Angry	5.7 (1.89)	4.21 (1.48)	4.05 (1.359)	3.052	0.067	1.489	0.046	1.652	0.02	0.163	0.999
PDI Guilty	5 (2.055)	3.95 (1.47)	3.57 (1.363)	1.982	0.162	1.053	0.271	1.429	0.063	0.376	0.999
PDI Needy	4.56 (1.81)	3.68 (1.60)	3.57 (1.599)	1.005	0.383	0.871	0.586	0.984	0.415	0.113	0.999

Note. PDI l = Parent Development Interview, Rel Aff pers= Relationship Affects Personality, PDIC Feelings Sep = Child’s feelings separation, PDIM Feelings Sep= Mothers feeling separation

The Overall statistical difference is a fair reflection of the raw data which show higher mean scores (i.e., ratings of a greater level of RF) for the community group on all variables except PDI-Losing, compared with the two “at-risk” groups. The PDILosing questions seek responses to the thought of the mother losing their child and the “at-risk” groups’ responses here may well reflect a heightened sensitivity owing to their genuine fears of having their child removed into protective care by social services; a fear not shared by the Community group. Inspection of the individual themes provides further insight into the source of the RF differences.

Of the separate themes, the ones either reaching or closely approaching significance include PDI-Joy; PDI-Pain/Diff; PDIChangedYou; PDI-Angry; PDI-Parents (p=0.049) and PDI-Losing (p=0.02). Group-by-group comparisons yielded significant differences between the community group and the high-risk group for PDI-ChangedYou (p=0.015); PDI-Parents (p=0.008) and PDI-Losing (p=0.052) and between the community group and the low-risk group for PDI-Parents (p=0.007). The subscale PDI-Joy concerns the positive feelings the mother has for the relationship with her child; PDI-Pain/Diff examines the mother’s pain and difficulty with being a parent, and PDI-ChangedYou looks at how the child has changed their relationship. The PDI-Angry captures instances of the child making the mother feel angry; PDI-Parents asks the mother to reflect on her own experience of being parented and PDI-Losing concerns thoughts of losing her child.

DISCUSSION

The data from this study need to be viewed with caution on account of the relatively small sample size. Nevertheless, the results show several important trends. In particular, the differences between the community group and the two “at-risk” groups on all variables; and secondly the absence of difference between the high-risk and low-risk groups, despite the highrisk mothers being mostly domiciled in refuges or addiction rehabilitation settings and also have experienced the direct involvement of child protection services.

On the DASS-21 scale, the at-risk groups both show greater variance than the community group but show consistently much lower mean scores. Likewise, on the ZAN-BPD the “atrisk” groups show greater variance and, on this variable, much higher mean scores compared with the community group. For the PDI, variances are more closely aligned, with the community group showing slightly greater variance than the two “at-risk” groups. Again, however, the mean scores show consistently greater Reflective Functioning capacity of the community group compared with the “at-risk” groups. In short, these data show that both the high-risk and low-risk groups demonstrate poorer mental health; higher levels of Personality Disorder and poorer Reflective Functioning compared with the community group.

In contrast to the significant differences between the community group and the two “at-risk” groups, these data show no significant difference between the high-risk and the low-risk groups in any of their responses. The only difference observed between the high and low-risk groups was in accommodation, with the high-risk group reporting living in refuge and drug and alcohol rehabilitation settings. The implication here is that the “at-risk” mothers demonstrate similar poor mental health; similar levels of personality disorder; and similar poor Reflective Functioning ability, regardless of their designated degree of risk.

In terms of the poor mental health of the “at-risk” mothers, the current data confirm the findings of other investigators (Jaffee, 2017; Li, D’Arcy, & Meng, 2016; Varese et al., 2012). The relative instability of the social circumstances of the “at-risk “mothers, shown in the demographic data, indicates a greater level of ongoing adversity compared with the mothers of the community group. These difficulties are then compounded by the relatively poor mental health of the “at-risk” mothers, demonstrated in the DASS-21 responses, inferring a lack of mental and emotional resources with which to deal with difficulties. These circumstances would place the “at-risk” mothers at a severe disadvantage in dealing with the day-to-day stresses and concerns of parenting an infant. However, these disadvantages are then amplified by the clear personality symptoms shown in the ZAN-BPD results. Borderline Personality is associated with poor emotional regulation, poor impulse control, poor self-image and difficulties with interpersonal relationships. The strong difference between the “at-risk” groups and the community group on the ZAN-BPD scale presages the “at-risk” mothers being more likely to have difficulties establishing positive, supportive and well-balanced relationships with their infants.

The results of the PDI reinforce the inadequacies of the mental and emotional processing of the “at-risk” mothers. Not only is the overall ability for Reflective Functioning shown to be poorer in the “at-risk” groups, but the differences in particular subscales are indicative of pinch-points of the mothers’ lack of relationship development capacity. The emotional toll of the mother-child relationship (PDI-Joy; PDI-Pain/Diff; PDI-Angry); the relationship’s requirement for personal adaptation (PDIChangedYou); and the recognition of threats to the relationship (PDI-Losing) were all illustrative of the “at-risk” mothers’ inability to reflect on their emotions insightfully. Also, the “at-risk” mothers appear to have little insight into their own experiences of being parented – (PDI-Parents). This result offers some evidence of the intergenerational nature of attachment disorders.

The finding here that the responses of the high-risk and lowrisk groups were not significantly different is of importance. The Structured Decision-Making System (SDM) currently used in New South Wales differentiated between the two groups and the intervention strategy adopted for each group depended on that segregation. The results of the current study do not negate the validity of the SDM rather they indicate a bias within the system. The SDM instructions guide caseworkers to take account of “a parent’s/carer’s current emotional, psychological status” (New South Wales Government Family and Community Services, 2012). However, the questions reflect concerns emphasising the social, environmental and circumstantial context of the child’s living conditions to assess the level of risk. This may be appropriate for evaluating immediate risk, but overlooking the mother’s mental health and personality characteristics is a major omission when devising intervention and treatment. Even appreciating the degree to which social services are oversubscribed, nevertheless, ignoring the mother’s personality and psychological vulnerabilities is not a way to devise longerterm holistic care strategies for the child.

REFERENCES

1. Barlow J, Fisher JD and Jones D. (2012). A systematic review of models of analysing significant harm. United King-dom.

2. Bowlby J. (1978). Attachment theory and its therapeutic implications. Adolescent Psychiatry. 6: 5-33.

3. Hamilton CE and Browne KD. (1999). Recurrent maltreatment during childhood: A survey of referrals to police child protection units in England. Child Maltreatment. 4(4): 275– 286. doi:10.1177/1077559599004004001

4. Jaffee SR. (2017). Child maltreatment and risk for psychopathology in childhood and adulthood. Annu Rev Clin Psy-chol. 13:525-551. doi:10.1146/annurevclinpsy-032816-045005

5. Levy KN, Johnson BN, Scala JW, Temes CM and Clouthier TL. (2015). The attachment theoretical framework for un-derstanding Personality Disorders: developmental, neuroscience and psychotherapeutic considerations. Psychologi-cal Topics. 24(1): 91–112.

6. Li M, D’Arcy CD and Meng X. (2016). Maltreatment in childhood substantially increases the risk of adult depression and anxiety in prospective cohort studies: systematic review, meta-analysis and proportional attributable fractions. Psychol Med. 46(4):717-30. doi:10.1017/S0033291715002743

7. Lovibond SH and Lovibond PF. (1995). Manual for Depression Anxiety and Stress scales (2nd ed. Sydney, Australia: Psychology Foundation of Australia.

8. New South Wales Government Child Death Review Team. (2019). Child death review team annual report (2018-20199 Sydney: Department of Family and Community Services. Sydney, Australia.

9. New South Wales Government Family and Community Services. (2012). The Structured Decision Making System. Policy and Procedures Manual Implementation. Sydney, Australia

10. New South Wales Government Family and Community Services. (2016). Child Deaths (2015) Annual Report, Learning to improve services. Sydney, Australia: D. o. F. a. C. Services.

11. Nitahara Y, Tachibana Y and Okuyama M. (2018). A retrospective study of the association between child abuse re-ports and risk factors. Int J Child Health Hum Dev. 11(3): 325-330. doi:10.1046/j.1525-1497.2003.20918.x

12. Pecora PJ, Chahine Z and Graham JC. (2013). Safety and risk assessment frameworks: Overview and implications for child maltreatment fatalities. Child Welfare. 92(2): 143-160.

13. Shlonsky A and Wagner D. (2005). The next step: integrating actuarial risk assessment and clinical judgement into an evidence -based practice framework in CPS case management. Children and Youth Services Review. 27: 409- 427. doi:10.1016/j.childyouth.2004.11.007

14. Sidebotham P and Heron J. (2006). Child maltreatment in the “children of the nineties”: A cohort study of risk fac-tors. Child Abuse Negl. 30(5):497-522. doi:10.1016/j. chiabu.2005.11.005

15. Slade A, Aber LJ, Berger B, Bresgi I and Kaplan M. (2003). Parent Development Interview Revised Short Version (PDIR2-S) New York, America: The Psychological Centre, The City College New York.

16. Stoltenborgh M, Bakermans-Kranenburg JJ, Alink LRA and van Ijzendoorn MH. (2015). The prevalence of child maltreatment across the globe: Review of a series of metaanalyses. Child Abuse Review. 24(1). 37-50. doi:10.1002/ car.2353

17. Varese F, Smeets F, Drukker M, Lieverse R, Lataster T et al. (2012). Childhood adversities increase the risk of psy-chosis: a meta-analysis of patient-control, prospective and crosssectional cohort studies. Schizophr Bull. 38(4):661-671. doi:10.1093/schbul/sbs050

18. Widom CS, Czaja SJ and DuMont K A. (2015). Intergenerational transmission of child abuse and neglect: real or de-tection bias? Science. 347(6229):1480-1485. doi:10.1126/science.1259917

19. World Health Organisation. (1999). Social change and mental health violence and injury Prevention: Report of the consultation on child abuse prevention. Geneva, Switzerland: World Health Organisation. 20. World Health Organisation. (2006). Preventing child maltreatment: A guide to taking action and generating evidence. Geneva, Switzerland: World Health Organisation.

21. Zanarini MC, Vujanovic AA, Parachini EA, Boulanger JL, Frankenburg FR. (2003). Zanarini rating scale for borderline personality disorder (ZAN-BPD): A continuous measure of DSM-IV borderline psychopathology. J Pers Disord. 17(3):233-242. doi:10.1521/pedi.17.3.233.22147

22. Zanarini MC, Frankenburg FR and Vujanovic AA. (2002). Inter-rater and test-retest reliability of the revised diagnostic interview for borderlines. J Pers Disord. 16(3): 270- 276. doi:10.1521/pedi.16.3.270.22538