Improving antenatal risk assessment in women exposed to high risks

Abstract

Antenatal substance use and related psychosocial risk factors are known to increase the likelihood of child protection involvement; less is known about the predictive nature of maternal reflective functioning (RF) in this population. This preliminary study assessed psychosocial and psychological risk factors for a group of substance dependent women exposed to high risks in pregnancy, and their impact on child protection involvement. Pregnant women on opiate substitution treatment (n = 11) and a comparison group (n = 15) were recruited during their third trimester to complete measures of RF (Pregnancy Interview), childhood trauma, mental health and psychosocial assessments. At postnatal follow-up, RF was reassessed (Parent Development Interview – Revised Short Version) and mother–infant dyads were videotaped to assess emotional availability (EA). Child protection services were contacted to determine if any concerns had been raised for infant safety. Significant between-group differences were observed for demographics, psychosocial factors, trauma and mental health symptoms. Unexpectedly, no significant differences were found for RF or EA between groups. Eight women in the ‘exposed to high risks’ group became involved with child protection services. Reflective functioning was not significantly associated with psychosocial risk factors, and therefore did not mediate the outcome of child protection involvement. Women ‘exposed to high risks’ were equally able to generate a model of their own and their infants’ mental states and should not be seen within a deficit perspective. Further research is required to better understand the range of risk factors that predict child protection involvement in high risk groups.

Keywords

Reflective functioning high-risk parenting child protection risk assessment substance abuse

High-risk parenting

Many parents experience a number of psychosocial risk factors, such as single parenting, social isolation and/or mental health issues that can limit parenting capacity. Early disturbances to ‘good enough’ parenting have been linked to poor developmental outcomes, attachment insecurity, and future psychopathology safety concerns for children (Cassidy & Mohr, 2001; Cyr, Euser, Bakermans-Kranenburg, & Van Ijzendoorn, 2010; Van Ijzendoorn, Schuengel, & Bakermans-Kranenburg, 1999). Parents with histories of attachment disruption are also likely to have difficulties acting as an attachment figure for their own child (Fonagy, Steele, Moran, Steele, & Higgit, 1993; Slade, 1999, 2007a; Slade, Grienenberger, Bernbach, Levy, & Lockyer, 2005; Zeanah, Benoit, Hirshberg, Barton, & Regan, 1994).

The World Health Organization (2010) reported that up to 50% of children worldwide experience some form of child maltreatment. Parental substance abuse has been identified as a major risk factor associated with child deaths and child protection concerns. Maternal substance use during pregnancy has been associated with poor psychosocial, developmental, emotional, behavioural and psychiatric outcomes for children (Appleyard, Egeland, van Dulmen, & Sroufe, 2005; Fraser, Harris-Britt, Thakkallapalli, Kurtz-Costes, & Martin, 2010; Mayes & Truman, 2002; Street, Harrington, Chiang, Cairns, & Ellis, 2004). Parents who use drugs and/or alcohol can also have a significantly impaired ability to understand and respond to their infant’s basic developmental needs (Cleaver, Unell, & Aldgate, 2011; Mayes & Truman, 2002). Despite the fact that substance use alone does not predict maltreatment (Doris, Meguid, Thomas, Blatt, & Echenrode, 2006; Jaudes, Ekwo, & Van Voorhis, 1995; Kroll, 2004; Smith, Johnson, Pears, Fisher, & De Garmo, 2007; Street et al., 2004; Velez et al., 2004) it is the single most common factor predicting removal of children from parental care (Banyard, Williams, & Siegal, 2003; Kelley, 1992; Smith & Testa, 2002; Suchman, McMahon, Zhang, Mayes, & Luthar, 2006), and is a distinguishing feature of unlikely family reunification (Reading et al., 2009) and child fatalities (Gibson & Vulliamy, 2010; NSW Government, 2011; NSW Ombudsman, 2009; Palmiere, Staub, La Harpe, & Mangin, 2010).

Parents who present with substance use disorders or mental health problems often experience psychosocial and demographic risks, trauma and victimisation (Appleyard, Berlin, Rosanbalm, & Dodge, 2011; Bromfield, Lamont, Parker, & Horsfall, 2010; Grant et al., 2011; Human Services, 2010; Mayes & Truman, 2002; Suchman et al., 2006; Taplin & Mattick, 2011; Velez et al., 2004). Such parents are frequently referred to as ‘high risk’ or ‘exposed to high risks’, as the complex array of multilevel problems and chronicity of parental issues increase the likelihood of children’s needs being unmet, impacting adversely on their development as well as heightening the risk of maltreatment (Bromfield et al., 2010; Cleaver, Nicholson, Tarr, & Cleaver, 2007). Parents draw largely on their own experiences of being parented and typically, within this population, few parents have experienced secure relationships with their own parents, resulting in emotional deprivation, adversity and a lack of a ‘good enough’ parental role model (Keen & Alison, 2001; Lier, Gammeltoft, & Knudsen, 1995). A focus on attachment during pregnancy is particularly pertinent as many caregivers with a history of child abuse will result in poor attachment, psychological disadvantage or maltreatment for their children (Begle, Dumas, & Hanson, 2010). The latest NSW child deaths report of children known to Child Services (CS) also revealed that 40% of parents who had a child that died were exposed to generational risk factors, with the majority known to CS when they themselves were children (NSW Government, 2011).

Parental reflective functioning

A number of studies (Fonagy, Steele, Steele, Moran, & Higgit, 1991; Grienenberger, Kelly, & Slade, 2005; Leigh, 2011; Meehan, Levy, Reynoso, Hill, & Clarkin, 2009; Reynolds, 2003; Rosenblum, McDonough, Sameroff, & Muzik, 2008; Schechter et al., 2005; Slade, 2006; Slade, Grienenberger, et al., 2005; Suchman, DeCoste, Leigh, & Borelli, 2010) have focused on the concept of mentalisation (defined as ‘the capacity to perceive and understand oneself and others in terms of mental states … feelings, beliefs, intentions, and desires’, Fonagy, 1995, p. 7), and parental reflective functioning (RF; defined as the parents ability to ‘hold the child in mind’ (Slade, 2005, p. 273), and understand the inner world of the child in terms of mental states). A parent’s capacity to develop a psychological understanding of their child positively influences behaviour and the quality of their emotional interaction, parental beliefs and internal representation of the child (Grienenberger et al., 2005; Slade, 2005; Slade, Grienenberger, et al., 2005), which provide the foundation for a secure attachment relationship (Fonagy, Steele, & Steele, 1991; Fonagy, Steele, Steele, Moran, et al., 1991; Levine & Tuber, 1991; Stronach et al., 2011).

Parents with low RF may fail to recognise their infants’ feelings (internal world/mental states) or comprehend the impact of their own behaviour on their child, leading to infant distress and welfare concerns (Fonagy, Steele, Steele, Moran, et al., 1991). Several studies have shown that low RF is associated with atypical maternal behaviour (Bronfman, Parsons, & Lyons-Ruth, 1999), lower levels of emotional availability (EA) (Suchman, Mayes, Conti, Slade, & Rounsaville, 2004) and maternal sensitivity, resulting in parental hostility and intrusive behaviours (Grienenberger et al., 2005). Parental emotional availability, that is, the capacity to tolerate their child’s emotional needs (Suchman et al., 2004), is considered to be the foundation for ‘good enough’ parenting and interaction.

Mothers with substance use disorders are more likely to have lower levels of reflective functioning (Cyr et al., 2010; Jenkins & Williams, 2008; Pajulo, Suchman, Kalland, & Mayes, 2006; Stronach et al., 2011; Suchman, McMahon, Slade, & Luthar, 2005; Suchman et al., 2006; Suchman et al., 2010; Van Ijzendoorn et al., 1999). Suchman and colleagues (2004) found that mothers with substance use problems also had a limited understanding of basic child development and a reduced capacity to think about their child’s needs, resulting in either harsh or permissive parenting styles, poor tolerance and limited active interaction (Suchman et al., 2004).

A parent’s reflection on their relationship with their own parents (as measured by the Adult Attachment Interview; George, Kaplan, & Main, 1984, 1988, 1996) has also been found to predict child attachment status at age one (Steele & Steele, 2008). The majority of RF studies focus on parent–infant relationships; less is known about RF during pregnancy and its effect on future parenting practices. Moreover, numerous studies have investigated mothers with substance use problems with multiple risk factors and established their associated parenting practices (Gilchrist & Taylor, 2009; Grella, Hser, & Huang, 2006; Nair et al., 1997), though few have included attachment related concepts as predictors of child protection involvement.

Current study

Parenting deficits associated with maternal substance use problems are often synchronous with comorbid physical and mental health conditions as well as psychosocial and demographic risks, which all significantly impact on a child’s emotional and behavioural development (Fraser et al., 2010). Maternal reflective capacity, in particular, has been highlighted as a key factor in the successful adaption to parenting in populations ‘exposed to high risks’ (Donald & Jureidini, 2004; Fonagy, Steele, Steele, Moran, et al., 1991; White, 2005). Low RF has been implicated as a significant mediating factor between maternal substance use and poor child outcomes (Mayes & Truman, 2002; Suchman et al., 2004). Yet, little attention has been paid to the extent to which parental RF is present as an overarching risk indicator in populations with substance use problems. The current study piloted the addition of psychological assessments to a pre-existing routine antenatal psychosocial interview to improve the detection of women in most need of intensive support and clinical parenting interventions in two groups of pregnant women – with and without current substance use problems – to better understand the multitude of factors associated with infants at risk of physical, sexual and psychological harm.

Hypotheses

Pregnant women with current substance use problems in drug treatment will exhibit higher rates of current psychosocial stress, mental health symptoms and past childhood trauma, as well as deficits in parenting capacity, when compared to a comparison group of pregnant women who do not have current substance use problems. It is also hypothesised that RF measured during the antenatal period will predict postnatal RF in both groups. Furthermore, it is predicted that low levels of parental RF will be significantly associated with current psychosocial stress, past and current trauma, and current mental health problems and mediate the relationship with child protection involvement.

Method

Ethics approval was obtained by the University of Newcastle Human Research Ethics Committee (approval number H-2008-0366).

Participants

Recruitment and assessment spanned from February 2009 to June 2011. Participants comprised 26 primiparous (n = 5) and multiparous (n = 21) women in their third trimester of pregnancy recruited through a tertiary hospital antenatal service in New South Wales, Australia. The group exposed to high risks were 11 pregnant women (M = 29.22 years old, SD = 4.35; M = 33.73 weeks pregnant, SD = 4.08) who were treated for heroin dependence with opioid substitution (methadone or buprenorphine maintenance). These women were monitored by a public drug and alcohol service or a general practitioner authorised to prescribe opioid substitution in conjunction with a specialist hospital-based multidisciplinary ‘drugs in pregnancy’ service. The comparison group comprised 15 pregnant women (M = 28.81 years old, SD = 4.89; M = 36.60 weeks pregnant, SD = 2.02), recruited from mainstream antenatal clinics contemporaneously at the same hospital. One woman in the comparison group was excluded, as she disclosed a significant history of substance use and child protection involvement during assessment. Women in the comparison group had no reported current substance use (including no tobacco use) and their children were not considered at risk of significant harm for their welfare. Women were excluded from participating in the study if they reported recent experience of acute or chronic psychosis; had a known current serious medical illness; had unmonitored drug and alcohol use; or were carrying a foetus with a known neurological condition. Employment and income were not included as demographics owing to the difficulty of maintaining normal work in the latter stages of pregnancy.

Data were collected by the study researcher, a clinical psychologist. The Pregnancy Interview – Revised, Parent Development Interview – Revised Short Version, and Emotional Availability Scale 4^th Edition were coded by independent trained and reliable raters. Of the women who volunteered for the study, eight did not participate in follow-up testing at Time 2 (T2), two women from the comparison group (unable to be contacted via mail or phone) and six women ‘exposed to high risks’ (one was unable to be contacted via mail or phone; one relocated interstate; two had their infants removed from their care; one who was working closely with CS was recommended by her caseworker not to participate and one was unable to participate due to tragic family circumstances). The attrition of eight mothers between T1 and T2 meant that a full data set was only available for 17 participants (12 ‘comparison’ and five ‘exposed to high risks’). Missing values were not replaced. Normality assumptions, as tested by the Kolmogorov–Smirnov test, were violated for the majority of outcome variables (with the exception of depression, newborn developmental knowledge, the social desirability subscale of the Childhood Trauma Questionnaire (CTQ) and the sum of psychosocial risk factors), so non-parametric tests were preferentially used for those outcomes.

Procedure

Potentially eligible women were given a brief outline of the study by their treating antenatal health professional (midwife, obstetrician, case worker or social worker). Interested participants were provided with further information; the study was explained (by a study researcher); eligible and interested women provided consent before participating in the study. Potentially eligible clinical participants may have declined involvement due to fear of further involvement with child protection services; an unwillingness to talk about their childhood and/or an inability to commit due to their current social circumstances (such as housing needs) and a lack of financial or material motivating factors. Following consent, routine screens (the Psychosocial Assessment interview (PA), Hunter New England Health, 2005; and Edinburgh Postnatal Depression scale (EPNDS), Cox, Holden, & Sagovsky, 1987) were accessed from the client’s antenatal file. Time 1 (T1) testing involved the completion of two self-report questionnaires (the CTQ, Bernstein & Fink, 1998; and the Newborn Developmental Knowledge Questionnaire (NDKQ), Newman, 2006); and participation in two audio-taped semi-structured clinical interviews (the Pregnancy Interview – Revised (PI), Slade, 2007b; and the Zanarini Rating Scale for Borderline Personality Disorder (ZAN-BPD), Zanarini et al., 2003) with one of three trained research assistants with an undergraduate degree in psychology.

When infants reached three to six months of age, contact was made with the mothers in order to complete T2 testing. The T2 follow-up testing included the Parent Development Interview (PDI; Slade, Aber, Berger, Bresgi, & Kaplan, 2003) and a videotaped unstructured free play/interaction session with their infant. Child protection outcomes were also recorded at T2. There was difficulty contacting some mothers for T2, resulting in a delay for follow-up. As a result, follow-up varied from the infant being three months old up to one year and eight months old (M = 20.35, SD = 14.30) at T2. Participants were sent a certificate of appreciation following completion.

Measures

Psychosocial Assessment Interview (Hunter New England Health, 2005)

The routine interview conducted at the first antenatal clinic consisted of 20 broad questions pertaining to medical, psychosocial history as well as substance use, involvement with child protection and present and previous pregnancy experiences. Domains such as practical and emotional support, life stressors, domestic violence, and mental illness were also explored. Each risk factor was coded as present or absent (Kratas, Matthey, & Barnett, 2009; Matthey et al., 2004).

Edinburgh Postnatal Depression Scale (Cox et al., 1987)

The EPNDS is a 10-item self-report measure of maternal depressive symptomatology reported over the past week, with total scores ranging from 0–30. The EPNDS is a valid and reliable measure in assessing antenatal mood (Bunevicius, Kusmunskas, Pop, Pedersen, & Bunevicius, 2009).

Childhood Trauma Questionnaire (Bernstein & Fink, 1998)

The CTQ is a 28-item self-report questionnaire designed to retrospectively assess maltreatment during childhood. Each subscale has five questions pertaining to emotional neglect (e.g. ‘I felt loved’), emotional abuse (e.g. ‘People in my family said hurtful or insulting things to me’), physical neglect (e.g. ‘I didn’t have enough to eat’), physical abuse (e.g. ‘I was punished with a belt, a board, a cord or some other hard object’) and sexual abuse (e.g. ‘Someone tried to make me do sexual things or watch sexual things’). The remaining three items assess minimisation and denial of abusive experiences. The questionnaire items are measured on a five-point Likert scale, taking about 15 minutes to complete. The CTQ has been validated in a number of clinical and community samples (Bernstein & Fink, 1998). It demonstrates good test–retest reliability (r = 0.79 to 0.86) and internal consistency (α = 0.66 to 0.92). The measure also possesses sound criterion validity and good convergent validity with clinical interviews and therapists ratings of child abuse (Scher, Stein, Asmundson, McCreary, & Forde, 2001).

Newborn Developmental Knowledge Questionnaire (Newman, 2006)

The NDKQ is a 35-item self-report instrument designed to assess parental knowledge of the developmental needs of infants aged to three months. The NDKQ is categorised into five subscales with questions pertaining to communication, visual and mutual gaze, tiredness, and regulation as well as verbal and non-verbal expression. Participants categorise each statement as ‘true’, ‘false’ or ‘unsure’. Correct answers attract a score of one with a total maximum score of 35. Incorrect and unsure answers are given a score of zero. The 10-minute questionnaire has demonstrated face validity and scores have been demonstrated to improve following education and training about infant development.

Pregnancy Interview – Revised (Slade, 2007a)

The 60 minute pregnancy interview consists of 22 semi-structured questions pertaining to the emotional experience of pregnancy as well as mothers’ expectations about their future as a parent. Third trimester mothers are also asked to describe their relationship with the foetus as well as ideas about their infant’s characteristics and personality. Overall, the interview is aimed at maternal self-representation, particularly on the mother’s capacity to reflect and act on the needs of both the unborn child and the newborn. The PI is scored for maternal reflective functioning using the same systematic coding process as the PDI, the Addendum to the Reflective Functioning Scoring Manual (Slade, Bernbach, Grienenberger, Levy, & Locker, 2005). Each question is given a score on an 11-point Likert scale ranging from −1 (negative reflective functioning) to 9 (exceptional reflective functioning). With Slade’s permission, Newman (2010) made linguistic changes to the Pregnancy Interview – Revised (2007b) to ensure questions were in Australian English and were understood by the Australian population.

Zanarini Rating Scale for Borderline Personality Disorder (Zanarini et al., 2003)

The ZAN-BPD is a nine item clinician-administered scale designed to measure changes on DSM-IV Borderline Personality psychopathology. The ZAN-BPD is modelled on the Borderline Personality Disorder (BPD) module of the Diagnostic Interview for DSM-IV Personality Disorders. The 30 minute clinical interview assesses anger, mood, feelings of emptiness, sense of self, dissociation, suicidality, abandonment, impulsivity and interpersonal relationships during the past week. The scale also yields four subscale scores for affect, cognition, impulsivity and relationships yielding a maximum score of 36. The ZAN-BPD has high discriminant validity and is sensitive to changes in symptom severity (r = 0.04–0.0001) (Zanarini et al., 2003).

Parent Development Interview – Revised Short Version (Slade et al., 2003)

The PDI is a semi-structured clinical interview containing 45 questions taking around 90 minutes to administer. The interview is designed to examine aspects of parental representations and reflective functioning. The 21 questions that require parents to reflect and/or consider the complexity and opacity of mental states are scored (Sharp & Fonagy, 2008). The PDI correlates with the Adult Attachment Interview and the Strange Situation (SS) classifications (Miller, 2008). In addition, RF predicts maternal behaviour and BPD in the presence of abuse (Grienenberger et al., 2005). Trained inter-rater reliability is good (r = 0.81 between 0.94; Bouchard et al., 2008) and the measure has been shown to be a valid and predictive tool in both clinical and non-clinical samples (Bouchard et al., 2008).

Emotional Availability Scales 4th Edition (Biringen, 2008)

This instrument is designed to evaluate the quality of communication, connection and emotional tone between an adult and a child (Biringen, 2008). The EA was used to assess the quality of mother–child connection during a 15-minute unstructured videotaped dyadic interaction between mother and infant (Biringen, 2008). Emotional availability was categorised into four adult dimensions – the parent’s sensitivity; structuring; non-intrusiveness and non-hostility toward the infant; and two child dimensions – responsiveness and involvement toward the adult. All scales were scored on a 7-point Likert scale where contextual cues and clinical judgements of the interaction were inferred (Biringen, 2008). All scales were scored by trained coders according to the EA Scales Manual, 4^th Edition (Biringen, 2008). Limited research data is available for the fourth edition (Biringen, 2011); nevertheless, earlier versions possess good inter-rater reliability (r = 0.80) (Biringen, Emde, Campos, & Appelbaum, 1995; Pressman, Pipp-Siegal, Yoshinaga-Itano, Kubick, & Emde, 1999).

Child protection involvement

The frequency of substantiated reports of significant harm to unborn children and/or infants was monitored through child protection services until the follow-up testing at T2. Involvement with community services was coded as ‘nil report’, ‘reports with no substantiation’, ‘substantiated reports requiring action’ and ‘removal of the child from parental care’.

For ease of comparison, a number of data manipulations were performed. The subscales in the CTQ, ZAN-BPD and NDKQ were summed to give a total score. In addition, the NDKQ subscales were calculated to give a percentage of correct, incorrect and unsure answers. To calculate the stability of RF across the two time points, a difference score was calculated; the T2 score minus the T1 score. An additional variable was created totalling the number of identified psychosocial risk factors identified at the first antenatal appointment for each participant. The child protection outcome variable was recoded into two categories: those with no child protection reports and those who had child protection involvement post-birth, as all reports were substantiated and resulted in allocation, referral or child removal by child protective services.

Results

The demographic comparisons between women ‘exposed to high risks’ and the ‘comparison’ group are shown in Tables 1 and 2. An independent samples t-test revealed no difference in maternal age between participants ‘exposed to high risks’ and the ‘comparison’ group, t (23) = 2.12, p = .83. There were significant differences on all demographic variables and some psychosocial risk factors. There were no between-group differences in emotional support, support at home, alcohol use, living with father, recent stress, feeling confident, anxious/depressed, work/relationship problems, having a history of physical abuse, being frightened by their partner or domestic violence as measured by the psychosocial interview.

Table 1.

Sample characteristics.

	n = 11 Exposed to risks	n = 14 Comparison

	M (SD)
Age in years	33.73 (4.08)	28.82 (4.90)
Number of pregnancies	2.56 (1.21)	1.86 (2.03)
Gestation at T1	33.73 (4.08)	36.60 (2.03)
	n
Primipara	2	9
Highest level of education
< Year 10	7	1
< Year 11/1	3	7
Tertiary	1	6
Marital status
Single	5	1
De facto/married	6	13
Housing
Homeless/social housing	7	0
Living with parents	3	0
Renting/mortgage	1	14
Born in Australia	10	11
Aboriginal	1	0

Table 2.

Demographic and psychosocial differences between high risk exposed and comparison groups.

	n = 11High-risk	n = 14Comparison	df (N)	χ²	p ^a
Demographics
Primi/multipara			1 (25)	5.31	.04*
Level of education			4 (20)	10.1	.01*
Marital status			3 (25)	16.9	<.001*
Housing			4 (23)	20.3	<.001*
Psychosocial factors
Hx forensic charges	9	0	1 (22)	18.3	<.001*
Hx mental illness	8	4	1 (25)	4.81	.03*
Illicit drug use	9	0	1 (25)	17.9	<.001*
OST	10	0	1 (25)	21.2	<.001*
Living with father	5	13	1 (25)	6.87	.02*
Recent stress	9	5	1 (24)	4.61	.05*
Pregnancy anxiety	5	0	1 (23)	6.30	.02*
Hx of child abuse	4	0	1 (24)	5.67	.03*
Child living away	2	0	1 (24)	5.67	.03*

Mann-Whitney-U p-value.

= p <.05, ** = p <.01.

Hx: History; OST: Opiate substitution treatment; Living with father: Living with father of the baby; Child living away: Child living away from home due to child protection concerns.

Statistical comparisons were made on T1 data (demographics, psychosocial stressors, childhood trauma, borderline symptoms and newborn developmental knowledge) in order to test whether or not women ‘exposed to high risks’ that completed T2 assessments were different from those who were lost to attrition. There were no statistically significant differences between those who participated in T2 and those who did not.

Differences were particularly evident for substance use and mental health disorders as well as both historical (trauma) and present stressors (e.g. living alone). In addition, women ‘exposed to high risks’ reported significantly more psychosocial stressors (M = 8.21, SD = 2.61) than the ‘comparison’ group (M = 1.43, SD = 1.95), as measured by an independent samples t-test, t (18) = 6.70, p = <.001.

Table 3 shows the measures of depression, childhood trauma, borderline personality disorder traits, reflective functioning and emotional availability for both groups. Women ‘exposed to high risks’ had significantly higher scores for depression (EPNDS), childhood trauma (CTQ) and borderline symptomatology (ZAN-BPD) compared with women not at risk. Women ‘exposed to high risks’ also achieved marginally lower scores for knowledge of newborn development (M = 24.00, SD = 4.36) than women in the comparison group (M = 26.79, SD = 3.45), t (23) = 1.79, p = .09.

Table 3.

Differences in outcome variables between high risk exposed and comparison groups.

	M (SD)				t	df		p ^a
	High-risk		Comparison
EPNDS	8.82	(5.91)	4.46	(2.60)	2.2	13	.04
CTQ	51.91	(15.76)	33.43	(3.52)	7		*	.001**
Physical abuse	8.36	(3.04)	5.36	(.75)				.01*
Sexual abuse	9.36	(7.19)	5.00	(0.00)				.03*
Emotional abuse	11.09	(4.76)	6.07	(1.73)				.001*
Emotional neglect	8.82	(3.52)	5.71	(.91)				.004**
Social desirability	7.73	(3.13)	6.14	(1.35)				.004**
ZAN-BPD	5.55	(4.72)	1.43	(2.14)		13		.02*
Affect	2.36	(2.16)	.71	(1.07)	1.5		.14	.02*
Impulsivity	.91	(1.05)	1.4	(.54)	7			.02*
Relationships	1.09	(1.05)	.43	(1.09)				.04*
Reflective functioning
Antenatal RF (T1)	4.64	(1.21)	5.07	(1.00)				.43
Postnatal RF (T2)	5.00	(1.23)	4.67	(1.07)				.58
Emotional availability
M sensitivity	5.60	(.65)	5.33	(3.89)				.47
M structuring	5.70	(.57)	5.21	(1.12)				.52
M non-intrusiveness	6.50	(.00)	6.33	(.25)				.25
M non-hostility	7.00	(.00)	6.92	(.20)				.55
C responsiveness	5.40	(.65)	5.04	(.50)				.28
C involvement	4.70	(.48)	4.38	(.53)				.23

Mann-Whitney-U p-value.

= p < .05; ** = p < .01.

EPNDS: Edinburgh Postnatal Depression Scale; CTQ: Childhood Trauma total score; ZAN-BPD: Zanarini Rating Scale for Borderline Personality Disorder total score; RF: Reflective functioning; T2: Time 2; M = Maternal; C = Child; RF = Reflective Functioning.

One comparison of interest is that the measures of reflective functioning and emotional availability showed no significant between-group differences. Risk-exposed women had a non-significant lower mean score than the comparison group for antenatal RF. There were no between-group differences detected in emotional availability across all subscales.

Pearson’s correlational analysis revealed a weak correlation between antenatal and postnatal reflective functioning (r = .33, p= .198). A linear regression analysis confirmed that T1 scores did not predict T2 scores (R² = .11, p = .198). The majority of the RF scores across groups remained stable from T1 to T2 (2/5 in the ‘exposed to high risk’ group and 7/13 in the ‘comparison group’). However, there was individual variation resulting in both an increase (1/5 ‘exposed to high risks’ and 1/13 in the ‘comparison group’) and decrease (1/5 ‘exposed to high risks’ and 5/13 in the ‘comparison group’) in scoring from T1 to T2 across both groups. Differences between antenatal RF and postnatal RF were no more than two points on the scale.

Antenatal RF was significantly negatively correlated with depression (r = –.48) and the ZAN-BPD subscale of impulsivity (r = .54). Antenatal or postnatal RF was not significantly associated with any of the other expected variables (NDKQ, total CTQ and ZAN-BPD, EA scales or psychosocial stress). The frequency of psychosocial risks significantly correlated with depression (r = .60), childhood trauma (r = .84) and borderline symptomatology (r = .67).

The difference in child protection involvement (post-birth) between groups was highly significant. No woman in the ‘comparison’ group had any reported child protection concerns compared to 8/11 of the women in the ‘high risk exposed’ group, (M–W, p = <.001). Due to the small sample size, there was a lack of power to obtain comparisons between women ‘exposed to high risks’ who had child protection involvement and those who did not have any concerns reported. Therefore, the child protection outcome variable comprised women who had involvement with Community Services (n = 8) compared to those who did not (n = 17; ‘exposed to high risks’, n = 3 and the entire ‘comparison’ group, n = 14). As most women ‘exposed to high risks’ had substantiated reports of harm, the between-group differences for the child protection outcome variable was essentially determined by the demographic differences (such as relationship status, housing) previously reported in Tables 1, 2 and 3. The difference in psychological measures between groups was also driven by the existing differences between groups. All significant differences previously reported between women ‘exposed to high risks’ and the ‘comparison’ group remained for the child protection outcome variable. Significant between-group differences were detected for antenatal depression, childhood trauma and borderline personality disorder symptoms. No differences were detected for newborn developmental knowledge, RF or EA. There were no significant differences in RF or EA between those who had child protection involvement and those who did not on the CTQ or ZAN-BPD.

Case vignette

Antenatal presentation

Kylie (not participant’s real name) (32 years) was single, unemployed, homeless, and pregnant with her first child on presentation, an unplanned pregnancy. She had an extensive history of opiate and cannabis dependence and amphetamine abuse. Kylie completed year 10 at school. She had a past history of depression and anxiety. She commenced methadone maintenance during pregnancy, stabilising on 120mg daily. Kylie described her mother as being ‘mostly supportive’ and cautiously expressed confidence in becoming a parent, though she held great concern for her child remaining in her care due to pressure from Community Services. Kylie had maintained minimal contact with the baby’s father.

Community Services’ safety and risk assessment revealed that the unborn child met the criteria for ‘risk of significant harm’. On the NDKQ, Kylie scored low on developmental knowledge; on the ‘Pregnancy Interview’ she recorded a moderate score, suggesting she had a sound understanding of the impact of her mental state and behaviour on her unborn infant. She displayed some affective and interpersonal disturbance on the ZAN-BPD and mild depression on the EPNDS. She disclosed a history of childhood trauma including physical abuse and neglect, emotional abuse and neglect and sexual abuse. During the PDI, Kylie stated that ‘the baby’s number one to me and I’m trying to make up for what I’ve done… trying to be even more better at… doing the right thing’ and ‘I’m afraid I’ll just fail as a parent’. Throughout the interview Kylie was concerned about her homelessness and lack of support from the baby’s father. She remained reflective and committed to adjusting to her role as a parent. She was able to find stable private rental housing after assistance with a local housing service.

Postnatal presentation and follow-up

Kylie gave birth to her son at term with no major complications. She remained stable on methadone; she had two lapses of episodic amphetamine use. Her housing remained stable. Kylie reported feeling extremely guilty and disappointed about the baby’s father having very little involvement with her son. Community services closed the case as the infant no longer met the criteria for significant risk of harm. Postnatally, Kylie scored moderately for RF and within acceptable ranges for EA. Eighteen months after the birth she remained in sustained remission from opioid use and continued in methadone treatment. A public health outreach parenting support service continued to provide support and her son’s development age one was normal. Kylie maintained engagement with drug and alcohol clinical services throughout the perinatal period.

The case example provided illustrates that while a parent with a current substance use problem can have several risk factors for poor parenting, with support they are able to provide a ‘good enough’ standard of parental care. As demonstrated by the vignette above, Kylie demonstrated adequate RF and EA and was able to maintain parental responsibility for her son despite his being initially assessed by protective services as being ‘at risk of significant harm’.

Discussion

The current study elucidated numerous risk factors among expectant mothers with current substance use problems. Significant between-group differences were detected between ‘high risk exposed’ women and the comparison group for demographic factors, psychosocial problems, trauma and mental health symptoms. Contrary to the study hypothesis, parenting ability, as measured by RF assessments and EA observations, did not differ between groups. However, child protection involvement was significantly different between groups.

Differences between women ‘exposed to high risks’ and ‘comparisons’

As expected, pregnant women with substance use problems were more likely to be single, less educated and residing in poorer accommodation or being homeless compared to the ‘comparison’ group. Women ‘exposed to high risks’ experienced significantly more antenatal psychosocial problems, at almost six times more than the ‘comparison’ group. In a study that focused on stress and parenting, Suchman and Luthar (2001) found that stress mediated 69% of the variance between sociodemographic risk and maladaptive parenting. In the current study, as hypothesised, women in the high risk exposed group had a higher prevalence of substance use, mental health disorders and forensic history compared to the comparison group. Consistent with previous research, women ‘exposed to high risks’ also experienced more childhood trauma than the comparison group (Grella et al., 2006; Lyons-Ruth & Block, 1998; Mayes & Truman, 2002).

The results of the current study indicate that women ‘exposed to high risks’, in this sample, possess RF abilities akin to that of the comparison group. Contrary to expectation, there were no statistically significant differences in RF or EA scores between groups. The reflective capacity scores only ranged between a score of three (low RF) to six (moderate RF) across both groups. The lack of between-group RF differences suggests that the ‘high risk’ women were equally able to generate a ‘mental model’ of their baby’s needs. The question then arises as to what factors interfere with RF stability; it is likely that the psychosocial risks and stress experienced by mothers ‘at risk’ may interfere with RF stability (Appleyard et al., 2005; Epstein, 2001; Nair et al., 1997; NSW Department of Community Services, 2007; Price-Robertson & Bromfield, 2011; White & Walsh, 2006a, 2006b). However, there is a paucity of research that explains why RF is fluctuant or how it may change or be superseded by psychosocial risk factors such as single parenting, unstable housing or mental state (Theran, Levendosky, Bogat, & Huth-Bocks, 2005). Whilst the results were an accurate depiction of this sample, it is likely that they may not be an accurate representation of a broader population of pregnant women with substance use problems who are not engaged in treatment. Substance dependent women not engaged in treatment may often have more psychosocial risk factors and be less motivated to improve their own and their child’s quality of life (Zilberman & Blume, 2005).

Leigh (2011) suggests that the capacity for mentalisation naturally fluctuates and is greatly impacted upon by emotional arousal, for example stress, mental health disorders or substance related symptoms. The high prevalence of these issues within the ‘high risk’ group may have affected parental RF and their behaviour toward their infants at some stage between Time 1 (T1) and Time 2 (T2). Specifically, at times women may have become mentally unwell or relapsed to substance using which could have resulted in a lack of insight and parental reflectivity. This would have likely limited their capacity to attend to their own (and others’) states of mind and impacted upon ‘good enough’ parenting. Despite demonstrating definite evidence of reflective function antenatally, the majority of women in the ‘high risk’ group became enmixed in the child protection system, resulting in involuntary loss of child custody and out of home care or significant monitoring and/or referrals for intervention.

The lack of between-group differences in EA was also notable, and while unexpected, has been previously observed. Tronick (2005) did not find any differences in EA in cocaine-exposed infants compared to controls, while Moehler found women who had experienced trauma scored significantly lower only on the intrusiveness EA subscale (Moehler, Biringen, & Poustka, 2007). Fraser and colleagues (2010) found that only the maternal sensitivity scale differed between women with substance use problems in treatment and a comparison group.

Pajulo et al. (2006) acknowledged that there is minimal research providing evidence that reflective abilities among problematic substance users are generally low (Truman, Levy, & Mayes, 2004, as cited in Pajulo et al., 2006; Suchman et al., 2004; Levy, Truman, & Mayes, 2001, as cited in Suchman et al., 2005 ). Nevertheless, despite demonstrating definite evidence of RF and EA, most women ‘exposed to high risks’ became embroiled in the child protection system, implying that other demographic and psychosocial factors overrule these state-dependent psychological constructs.

The RF results in the current study raise a number of issues, since many women at risk became involved in the child protection system. Based on these findings, reflective functioning cannot be supported as a reliable predictive measure of long-term parenting ability. Despite the similarity of the RF results between the groups there is likely to be dissimilar themes between the groups’ transcripts of the PI and the PDI.

The psychosocial stress under which high risk exposed women live can have a significant impact on parental mental state and parenting capacity (Appleyard et al., 2005; Epstein, 2001; Nair et al., 1997; NSW Department of Community Services, 2007; Price-Robertson & Bromfield, 2011; White & Walsh, 2006a, 2006b). The lack of between-group differences may be attributable to the lack of sensitivity of the measures. Also, Slade (2005) notes that coding and scoring of RF measures is impacted by the developmental stage of young infants, highlighting the challenges in understanding the mental state of toddlers compared to infants.

Reflective functioning

The current study also demonstrated the feasibility of implementing the PI (Slade, 2007b) among women exposed to multiple risks. Antenatal RF was significantly negatively correlated with depression and impulsivity. Condon and Corkindale (1997) also found that depression had a significant impact on antenatal reflective functioning and attachment. It was further hypothesised in the current study that antenatal RF would predict postnatal functioning; however, a correlational analysis revealed an insignificant weak relationship between baseline and follow-up. Subsequently, antenatal RF scores did not predict postnatal RF scores and had no predictive value for child protection involvement. Upon investigation of the scores an interesting trend between the groups was detected. The majority (50%) of women’s antenatal and postnatal RF scores remained unchanged. Despite the increases (17%) and decreases (33%) within the sample from T1 to T2, the changes in scores were no more than two points. Of the five women who followed up in the ‘exposed to high risks’ group, two women’s scores increased from T1 to T2 showing that mothers ‘exposed to high risks’ had the capacity for improvement in RF, with a demonstrable and more sophisticated understanding about the relational impact of their own and their infant’s mental states. In the current study, women ‘exposed to high risks’ that completed follow-up assessments were engaged in intensive parenting and clinical services that may have reduced the impact of psychosocial and demographic risk factors and increased postnatal RF abilities. A similar study by Theran and colleagues (2005), who also assessed the stability and changes of maternal representation in pregnant women, found that 62% of mothers’ RF categories remained stable from the third trimester to follow-up when their child was aged one. They also found that changes were attributable to psychosocial and demographic factors; however, such variables were not assessed at follow-up during this study (Theran et al., 2005). Nevertheless, the stability of the measure is important, as it means either the measure is insensitive to change or that the capacity it is measuring is a stable phenomenon.

The concept of reflective functioning is complex and multifaceted. Only recently have we learned of the preliminary evidence of the two-dimensional nature of RF (self and child mentalisation) within the PDI (Suchman et al., 2010). Further studies examining these two types of RF antenatally may shed more light into the aetiology of child maltreatment prevention and could have significant implications on the future of risk assessments.

Child protection involvement

A primary aim of this study was to assess a number of known predictors of risk as well as determining the impact of psychological concepts as potential indicators of child protection involvement. The majority of women ‘exposed to high risks’ were investigated by statutory child protection, unlike women in the ‘comparison’ group. We would have liked to use the same methods as Taplin and Mattick (2011) and compare the differences between women ‘exposed to high risks’, namely those with and without child protection involvement, but our small sample size prevented this statistical analysis. Therefore, we were compelled to group all participants by involvement or no involvement with child protection services.

It was hypothesised that lower levels of RF would mediate the relationship between psychosocial risk factors and child protection involvement. This hypothesis was not supported as pre and postnatal RF was not significantly associated with the numerous historical and antenatal demographic and psychosocial risks as anticipated. Reflective functioning was not significantly associated with child protection involvement; therefore, the proposed mediation could not be evaluated. The differences between women with child protection involvement within the ‘exposed to high risks’ group (8/11) and without (entire ‘comparison’ group and three women ‘exposed to high risks’) child protection involvement highlighted many differences in demographic, psychosocial and outcome variables. Mothers with child protection involvement presented with more risk factors such as single parenting, low education attainment, housing instability, forensic problems, mental health disorders and substance use problems compared to women without child protection involvement. Moreover, women involved in the child protection system had higher levels of childhood trauma, psychological symptoms of depression and BPD symptoms than those who were not, consistent with previous research in families frequently reported to child protection services (Hopkins & Smoothy, 2007; Nivison-Smith & Chilvers, 2007; NSW Government, 2011; Price-Robertson & Bromfield, 2011; Taplin & Mattick, 2011). However, the differences between those with and without child protection involvement were largely driven by the aforementioned between-group differences.

The demographic differences between women enmeshed in child protection in the current study were comparable to those described by Taplin and Mattick (2011). Their results highlight the synergy of both individual and environmental factors that contribute to child protection involvement, rather than substance using problems alone. For instance, symptoms of borderline personality disorder, such as impulsivity, are known to have a significant impact on parenting consistency and sensitivity (Newman, Stevenson, Bergman, & Boyce, 2007), which could be a significant predictor of potential risk of harm to an infant or child. While we were unable to build a model of predictors due to the small sample size (Hosmer & Lemeshow, 2000), further research into impulsive behaviour in populations with concurrent substance use is warranted. Previous research has supported the link between impulsivity, depression and disturbances in affect and cognitions (Fonagy et al., 1995; Newman et al., 2007; Whiteside & Lynam, 2001; Zanarini et al., 2003) and the current study showed a significant negative correlation between antenatal RF scores and impulsivity. Impulsivity has also been linked to poor adaptive functioning which may adversely impact on the transition to parenthood (Brain Injury Association of Qld Inc., 2011). Low antenatal RF concurrently with impulsive behaviours could be an indicator of risk, better signifying a parent’s current functioning. Previous studies have also provided evidence of the link between impulsivity and substance use disorders (Brain Injury Association of Qld Inc., 2011).

Predicting child protection involvement

Child protection involvement is stimulated mostly by professionals within government or non-government bodies (health, education and/or police) or community members making reports to statutory agencies based on observed or suspected concerns. Therefore, it may be that psychological concepts such as RF may never accurately predict child protection involvement, as existing parenting risk assessments often do not assess the potential impact of such constructs; rather, they typically focus on understandable and measurable demographic and psychosocial risk factors. Moreover, we did not include any psychometric measures of protective factors, perceived support or resilience. A recent study found that women with substances use problems who had children remaining in their care had higher resilience and more support than those who had children removed (Grant et al., 2011).

Study limitations

Substance-using pregnant women can have unstable lifestyles and are unreliable to recruit. The small sample size was due to low incentives, fear of child protection involvement and being monitored, distrust in services and an unwillingness to discuss their past. The sample size in this pilot, together with the high attrition rate, limits the generalisability of this study. In addition, employment status or income was not assessed, which may have impacted between-group comparisons, as income may be a predictor for sensitive parenting (Fraser et al., 2010). The reason for the reported concern/s to the statutory agency could not be obtained, so we cannot be sure why CS became involved. Unlike Taplin and Mattick (2011), we were not able to thoroughly assess the frequency or type of substances used by participants.

Future research

There is a paucity of prenatal studies that assess the effectiveness of measures predicting child outcomes. Further research in risk exposed populations is required, especially due to the numerous reports of young infants reported to child protection services for risk of significant harm (Fraser et al., 2010). Researchers must continue to work collaboratively with child protection services to develop more accurate risk assessments to ensure the safety of children, in utero and after birth.

We have learnt from this study that a number of factors are more prevalent amongst women with child protection outcome involvement; however, we do yet understand which factors are moderators or mediators and which factors combined are the most predictive of risk. Future studies should delve further into the pernicious combination of factors that are most predictive of risk during the antenatal period to better predict infants’ postnatal safety. Assessing psychosocial risk factors and their association with RF, EA and behavioural attachment (Goldberg, Benoit, Blokland, & Madigan, 2003) should continue to reveal more information about maternal antenatal attachment. The identified psychosocial risk factors for substance dependent women in treatment are likely to be somewhat different for the other subgroups of parents ‘exposed to high risks’; therefore, studying diverse populations during the antenatal period is imperative. The ongoing development of antenatal measures that accurately identify at-risk parents (prior to postnatal harm) could not only have an immense impact on the safety and wellbeing of infants, but could continue to benefit children throughout their lives, particularly if a positive early attachment relationship could be harnessed.

Clinical implications

Improving the identification of women who may benefit from targeted clinical and parenting intervention may enhance RF during pregnancy and reduce infants exposed to risk of harm (Reynolds, 2003; Slade et al., 2006). While the current antenatal psychosocial assessment (Hunter New England Health, 2005) is comprehensive and has success in determining the occurrence of psychosocial factors, drugs in pregnancy services should consider including psychological screens, particularly for impulsivity, to detect risk and identify women in need of intervention.

Perinatal and family services should aim to attend to both infants/children and parents to address multiple risks to improve dyadic and relational functioning. Effective care pathways, as termed by Callagahan and colleagues (2011), strive to prevent child maltreatment by improving parental capacity and reduce the impact of psychosocial stressors.

Commencing parenting interventions during pregnancy is beneficial as pregnant women may be more motivated and willing to change previously resistant negative behaviours to become better mothers and improve relationships with their children (Miller, 2008). The emphasis on the parent–child relationship is acknowledged throughout the literature but it remains difficult to accurately assess (Budd, 2001). Despite this, therapies targeting RF have shown significant capacity to improve the quality of relationships in mother–infant dyads (Sadler, Slade, & Mayes, 2006; Slade et al., 2006). Such therapies encourage the development of reflection about children and enhance internal working models and sensitive parenting practices, reducing the impact of the intergenerational transmission of parenting styles and poor attachment (Fonagy & Target, 1998; Miller, 2008).

Many stressors within disadvantaged populations are not mutually exclusive; indeed, many are strongly inter-related. It is known that the more risk factors a mother reveals prenatally (Putnam-Hornstein & Needell, 2011), the more likely she is to require targeted clinical parenting intervention. Prevention oriented intervention for at-risk families could improve the wellbeing and safety of children and foster healthy attachment relationships, and consequently improve child development and protect against future psychopathology (Budd, 2001). Moreover, prevention-focused efforts would also result in saving the substantial financial costs that are associated with child protection investigations, costs of assessments and monitoring, out of home placements as well as children placed under the guardianship of the minister (Putnam-Hornstein & Needell, 2011).

Women ‘exposed to high risks’ were equally able to generate a model of their own and their infant’s mental states and should not be seen within a deficit perspective. Further research is required to assess the combination of risk factors to predict child protection involvement. This preliminary Australian study adds to important clinical information for practising clinicians and researchers working in the realm of high risk parenting and child protection.

Footnotes

Acknowledgements

This research would not be possible without the support of NSW Health, John Hunter Hospital Antenatal Clinic, NSW Family and Community Services. Thank you to the women who participated in the study, without them we would not be able to continue to better understand the risk factors associated with child protection involvement. Thanks must also be extended to Josie Byrne for greatly assisting me with recruitment and Joanne Allen and Melissa Harris for their extensive work in launching the project in its early phase. I must also thank the Anna Freud Centre and the International Center for Excellence in Emotional Availability for blindly scoring my materials.

Funding

This research received a small amount of funding from NSW Health.

Author biographies

Natasha Perry is a Clinical and Health Psychologist working in Keep Them Safe – Whole Family Team (Hunter New England Health) and a Conjoint Lecturer within the School of Psychology at the University of Newcastle. Natasha is working and conducting research for high-risk families, addressing the needs of families where parents have mental health and/or drug and alcohol concerns have led to significant child protection concerns. Natasha has completed a Doctor of Clinical and Health Psychology at the University of Newcastle and has published in the areas of high-risk parenting. Natasha is interested in increasing the safety, wellbeing and improving developmental outcomes for children by developing parental reflective capacity and strengthening parent-child relationships.

Louise K Newman is the Professor of Developmental Psychiatry and Director of the Monash University Centre for Developmental Psychiatry & Psychology. Prior to this appointment she was the Chair of Perinatal and Infant Psychiatry at the University of Newcastle and the previous Director of the New South Wales Institute of Psychiatry. In January 2011 she was appointed as a Member in the General Division of the Order of Australia. She is a practising infant psychiatrist with expertise in the area of disorders of early parenting and attachment difficulties in infants. She has undertaken research into the issues confronting parents with histories of early trauma and neglect. Her current research is focussing on the evaluation of infant-parent interventions in high-risk populations, the concept of parental reflective functioning in mothers with borderline personality disorder and the neurobiology of parenting disturbance. She has published in the areas of infant mental health, attachment disorders trauma, and prevention of child abuse. She is co-author of the textbooks Clinical Skills in Infant Mental Health and Contemporary Approaches in Child and Adolescent Mental Health.

Mick Hunter’s research interests concern the function of brain in conditions of health and illness. Prof. Hunter maintains an interest in the areas of clinical psychology and neuroscience. His interests in clinical and therapeutic work include those related to the use of evidenced-based therapeutic techniques; the testing of theories concerning the development of mental illness; and the pathways of child development. Prof. Hunter is a registered psychologist and a member of the Clinical College of the Australian Psychological Society. He has over 30 years of experience in teaching and research; over 50 articles published in academic journals; and successfully supervised 15 Doctoral and over 40 Masters research students.

Adrian Dunlop is Area Director and Senior Staff Specialist, Drug and Alcohol Clinical Services for the Hunter New England Local Health District, NSW Australia, Clinical Advisor for the Mental Health Drug & Alcohol Office, NSW Ministry of Health, Conjoint Associate Professor in the School of Medicine and Public Health, Faculty of Health, University of Newcastle, Australia and a member of the Priority Research Centre for Translational Neuroscience and Mental Health with the Hunter Medical Research Institute and the University of Newcastle. He is a Foundation Fellow of the Australasian Chapter of Addiction Medicine, Royal Australasian College of Physicians and Past-President of the Australasian Professional Society on Alcohol and other Drugs. In 2005 he was awarded a Churchill Fellowship to investigate the treatment of opiate dependence in pregnancy.

References

Appleyard

Egeland

van Dulmen

M. H. M.

Sroufe

L. A.

(2005). When more is not better: The role of cumulative risk in child behaviour outcomes. Journal of Child Psychology and Psychiatry, 46(3), 235-245.

Appleyard

Berlin

L. J.

Rosanbalm

K. D.

Dodge

K. A.

(2011). Preventing early child maltreatment: Implications from a longitudinal study of maternal abuse history, substance use problems, and offspring victimization. Preventative Science, 12(2), 139-149.

Banyard

V. L.

Williams

L. M.

Siegal

J. A.

(2003). The impact of complex trauma and depression on parenting: An exploration of mediating risk and protective factors. Child Maltreatment, 8(4), 334-349.

Begle

A. M.

Dumas

J. E.

Hanson

R. F.

(2010). Predicting child abuse potential: An empirical investigation of two theoretical frameworks. Journal of Clinical Child and Adolescent Psychology, 39(2), 208-219.

Bernstein

D. P.

Fink

(1998). Childhood trauma questionnaire: A retrospective self-report manual. San Antonio, Texas: The Psychological Corporation.

Biringen

(2008). The emotional availability (EA) scales (4th edition). Boulder, CO.

Biringen

(2011). The emotional availability (full, 4th ed.) as well as EA intervention training. Brisbane, Australia.

Biringen

Emde

R. N.

Campos

J. J.

Appelbaum

M. I.

(1995). Affective reorganization in the infant, the mother, and the dyad: The role of upright locomotion and its timing. Child Development, 66, 499-514.

Bouchard

M. -A.

Target

Lecours

Fonagy

Tremblay

L. -M.

Schacter

Stein

(2008). Mentalization in adult attachment narratives: Reflective functioning, mental states, and affect elaboration compared. Psychoanalytic Psychology, 25(1), 47-66.

10.

Brain Injury Association of Qld Inc. (2011). Impulsivity fact sheet. Synapse (pp. 1–2). South Brisbane, QLD: Synapse.

11.

Bromfield

Lamont

Parker

Horsfall

(2010). Issues for the safety and wellbeing of children in families with multiple and complex problems: The co-occurrence of domestic violence, parental substance misuse, and mental health problems. Melbourne: Australian Institute of Family Studies.

12.

Bronfman

Parsons

Lyons-Ruth

(1999). Atypical Maternal Behavior Instrument for Assessment and Classification (AMBIANCE): Manual for coding disrupted affective communication, Version 2. Unpublished manuscript: Harvard Medical School, Cambridge, MA.

13.

Budd

K. S.

(2001). Assessing parenting comptence in child protection cases: A clinical practice model. Clinical Child and Family Psychology Review, 4, 1-18.

14.

Bunevicius

Kusmunskas

Pop

V. J.

Pedersen

C. A.

Bunevicius

(2009). Screening for antenatal depression with the Edinburgh depression scale. Journal of Psychosomatic Obstetrics and Gynecology, 30(4), 238-243.

15.

Callaghan

Crimmins

Schweitzer

R. D.

(2011). Children of substance-using mothers: Child health engagement and child protection outcomes. Journal of Paediatrics and Child Health, 47, 223-227.

16.

Cassidy

Mohr

J. J.

(2001). Unresolvable fear, trauma and psychopathology: Theory, research and clinical considerations related to disorganized attachment across the lifespan. Clinical Psychology: Science and Practice, 8, 275-298.

17.

Cleaver

Nicholson

Tarr

Cleaver

(2007). Child protection, domestic violence and parental substance misuse: Family experiences and effective practice. London, UK: Jessica Kingsley Publishers.

18.

Cleaver

Unell

Aldgate

(2011). Children’s needs – Parenting capacity. Child abuse: Parental mental illness, learning disability, substance misuse and domestic violence. Second Edition. London: The Stationery Office. Retrieved from: https://www.education.gov.uk/publications/eOrderingDownload/Childrens%20Needs%20Parenting%20Capacity.pdf.

19.

Condon

J. T.

Corkindale

(1997). The correlates of antenatal attachment in pregnant women. British Journal of Medical Psychology, 70, 359-372.

20.

Cox

J. L.

Holden

J. M.

Sagovsky

(1987). Detection of postnatal depression. Development of the 10-item Edinburgh postnatal depression scale. The British Journal of Psychiatry, 150, 782-786.

21.

Cyr

Euser

E. M.

Bakermans-Kranenburg

M. J.

Van Ijzendoorn

M. H.

(2010). Attachment security and disorganization in maltreating and high-risk families: A series of meta analyses. Development and Psychopathology, 22, 87-108.

22.

Donald

Jureidini

(2004). Parenting capacity. Child Abuse Review, 13, 5-17.

23.

Doris

J. L.

Meguid

Thomas

Blatt

Echenrode

(2006). Prenatal cocaine exposure and child welfare outcomes. Child Maltreatment, 11, 326-337.

24.

Epstein

M. R.

(2001). Predicting abuse and neglect in the first two years of life from risk assessments during the prenatal and perinatal period (Thesis). California: University of California.

25.

Fonagy

Steele

Moran

Steele

Higgit

(1993). Measuring the ghost in the nursery: An empirical study of the relation between parents’ mental representations of childhood experiences and their infants’ security of attachment. Journal of the American Psychoanalytic Association, 41, 957-989.

26.

Fonagy

Steele

(1991). Maternal representations of attachment during pregnancy predict the organization of infant-mother attachment at one year of age. Child Development, 62, 891-905.

27.

Fonagy

Steele

Moran

G. S.

Higgit

A. C.

(1991). The capacity for understanding mental states: The reflective self in parent and child and its significance for security of attachment. Infant Mental Health Journal, 12(3), 201-218.

28.

Fonagy

Steele

Leigh

Kennedy

Mattoon

Target

(1995). Attachment, the reflective self, and borderline states: The predictive specificity of the Adult Attachment Interview and pathological emotional development. Hillsdale, NJ: Analytic Press.

29.

Fonagy

Target

(1998). Mentalization and the changing aims of child psychoanalysis. Psychoanalytic Dialogues, 8, 87-114.

30.

Fraser

J. G.

Harris-Britt

Thakkallapalli

E. L.

Kurtz-Costes

Martin

(2010). Emotional availability and psychosocial correlates among mothers in substance-abuse treatment and their young infants. Infant Mental Health Journal, 31(1), 1-15.

31.

George

Kaplan

Main

(1984, 1988, 1996). The adult attachment interview. Unpublished protocol (3rd ed.): Department of Psychology, University of California, Berkeley.

32.

Gibson

J. C.

Vulliamy

(2010). Accidental methadone poisoning in children: A call for Canadian research action. Child Abuse & Neglect, 34 (8), 553-554. Retrieved from http://www.researchgate.net/publication/45096625_Accidental_methadone_poisoning_in_children_a_call_for_Canadian_research_action

33.

Gilchrist

Taylor

(2009). Drug-using mothers: Factors associated with retaining care of their children. Drug and Alcohol Review, VI(2), 53-82.

34.

Goldberg

Benoit

Blokland

Madigan

(2003). Atypical maternal behavior, maternal representations, and infant disorganized attachment. Development Psychopathology, 15(2), 239-257.

35.

Grant

Huggins

Graham

J. C.

Ernst

Whitney

Wilson

(2011). Maternal substance abuse and disrupted parenting: Distinguishing mothers who keep their children from those who do not. Children and Youth Services Review, 33, 2176-2185.

36.

Grella

C. E.

Hser

Y. I.

Huang

Y. C.

(2006). Mothers in substance abuse treatment: Differences in characteristics based on involvement with child welfare services. Child Abuse and Neglect, 30, 55-73.

37.

Grienenberger

Kelly

Slade

(2005). Maternal reflective functioning, mother–infant affective communication, and infant attachment: Exploring the link between mental states and observed caregiving behavior in the intergenerational transmission of attachment. Attachment and Human Development, 7(3), 299-311.

38.

Hopkins

Smoothy

(2007). A close look: Recent trends in child protection reports to DOCS. Ashfield, Australia: NSW Department of Community Services.

39.

Hosmer

D. W.

Lemeshow

(2000). Applied logistic regression, Second edition. New York, NY: John Wiley & Sons.

40.

Human Services. (2010). Working with parental substance misuse, Research to practice notes. Ashfield, Australia: Community Services, NSW Government.

41.

Hunter New England Health. (2005). Antenatal psychosocial risk assessment. In John Hunter Hospital (ed.), Obstetrics. (pp. 1–2) Newcastle: Hunter New England Health.

42.

Jaudes

P. K.

Ekwo

Van Voorhis

(1995). Association of drug abuse and child abuse. Child Abuse and Neglect, 19, 1065-1075.

43.

Jenkins

Williams

(2008). The mother-baby prenatal group: Nurturing reflective functioning in a methadone-maintenance clinic. Journal of Prenatal and Perinatal Psychology and Health, 22(3), 163-179.

44.

Keen

Alison

L. H.

(2001). Drug misusing parents: Key points for health professionals. Archives of Diseases in Childhood, 85, 296-299.

45.

Kelley

S. J.

(1992). Parenting stress and child maltreatment in drug-exposed children. Child Abuse and Neglect, 16(3), 317-328.

46.

Kratas

J. C.

Matthey

Barnett

(2009). Antenatal psychosocial assessment: How accurate are we in determining ‘low risk’ status? A pilot study. Archives of Women’s Mental Health, 12, 97-103.

47.

Kroll

(2004). Living with an elephant: Growing up with parental substance misuse. Child and Family Social Work, 9, 129-140.

48.

Leigh

(2011). Mentalization and reflective functioning: An introduction. Australian Psychological Society: Perinatal and Infant Psychology Interest Group Presentation (pp. 269–281). Melbourne: Australian Psychological Society.

49.

Levine

L. V.

Tuber

S. B.

(1991). Mothers’ mental representations and their relationship to mother-infant attachment. Bulletin of the Menninger Clinic, 55(4), 454-469.

50.

Lier

Gammeltoft

Knudsen

I. J.

(1995). Early mother-child relationship. The Copenhagen model of early preventive intervention towards mother-infant relationship disturbances. Arctic Medical Research, 15(1), 15-23.

51.

Lyons-Ruth

Block

(1998). The disturbed caregiving system: Relations among childhood trauma, maternal caregiving, and infant affect and attachment. Infant Mental Health Journal, 17(3), 257-275.

52.

Matthey

Phillips

White

Glossop

Hopper

Panasetis

… Barnett

(2004). Routine psychosocial assessment of women in the antenatal period: Frequency of risk factors and implications for clinical services. Archives of Women’s Mental Health, 7, 223-229.

53.

Mayes

Truman

(2002). Substance abuse and parenting. In Bornstein

(Ed.), Handbook of parenting, Vol 4: Social conditions and applied parenting (pp. 329–359). Mahwah, NJ: Lawrence Erlbaum Associates.

54.

Meehan

K. B.

Levy

K. N.

Reynoso

J. S.

Hill

L. L.

Clarkin

J. K.

(2009). Measuring reflective function with a multidimensional scale: Comparison with scoring reflective function on the AAI. Journal of the American Psychoanalytic Association, 57, 208-213.

55.

Miller

M. R.

(2008). The impact of a home-based intervention program on maternal reflective functioning in first-time mothers (Thesis). New York: Psychology, The City University of New York.

56.

Moehler

Biringen

Poustka

(2007). Emotional availability in a sample of mothers with a history of abuse. American Journal of Orthopsychiatry, 77(4), 624-628.

57.

Nair

Black

M. M.

Schuler

Keane

Snow

Rigney

B. A.

(1997). Risk factors for disruption in primary caregiving among infants of substance abusing women. Child Abuse and Neglect, 21(11), 1039-1051.

58.

Newman

(2006). Newborn developmental knowledge questionnaire. Sydney, NSW: NSW Institute of Psychiatry.

59.

Newman

(2010). Pregnancy Interview – Revised – Australian Version: Unpublished work.

60.

Newman

Stevenson

C. S.

Bergman

L. R.

Boyce

(2007). Borderline personality disorder, mother–infant interaction and parenting perceptions: Preliminary findings. Australian and New Zealand Journal of Psychiatry, 41, 598-605.

61.

Nivison-Smith

Chilvers

(2007). Child protection reports in context. Ashfield, Australia: NSW Department of Community Services.

62.

NSW Department of Community Services. (2007). Risk, protection and resilience in children and families. Ashfield, Australia: Centre for Parenting and Research.

63.

NSW Government. (2011). Child deaths 2010 annual report. Learning to improve services. Ashfield, Australia: Department of Family and Community Services.

64.

NSW Ombudsman. (2009). Report of reviewable deaths in 2007. Volume 2: child deaths, NSW, Australia: Crown.

65.

Pajulo

Suchman

Kalland

Mayes

(2006). Enhancing the effectiveness of residential treatment for substance abusing pregnant and parenting women: Focus on maternal reflective functioning and mother–child relationship. Infant Mental Health Journal, 27(5), 448-465.

66.

Palmiere

Staub

La Harpe

Mangin

(2010). Parental substance abuse and accidental death in children. Journal of Forensic Sciences, 55(3), 819-821. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20345788

67.

Pressman

Pipp-Siegel

Yoshinaga-Itano

Kubick

Emde

(1998). A comparison of the link between emotional availability and language gain in young children with and without hearing loss. Volta Review, 100, 251-277.

68.

Price-Robertson

Bromfield

(2011). Risk assessment in child protection. Melbourne, Australia: Australian Institute of Family Studies.

69.

Putnam-Hornstein

Needell

(2011). Predictors of child protective service contact between birth and age five: An examination of California’s 2002 birth cohort. Children and Youth Services Review, 33, 1337-1344.

70.

Reading

Bissell

Goldhagen

Harwin

Mason

Moynihan

… Webb

(2009). Promotion of children’s rights and prevention of child maltreatment. Lancet, 373, 332-343.

71.

Reynolds

(2003). Mindful parenting: A group approach to enhancing reflective capacity in parents. Journal of Child Psychotherapy, 29(3), 357-374.

72.

Rosenblum

K. L.

McDonough

S. C.

Sameroff

A. J.

Muzik

(2008). Reflection in thought and action: Maternal parenting reflectivity predicts mind-minded comments and interactive behaviour. Infant Mental Health Journal, 29(4), 362-376.

73.

Sadler

L. S.

Slade

Mayes

L. C.

(2006). Minding the baby: A mentalization-based parenting program. In Allen

J. G.

Fonagy

(Eds.), Handbook of mentalization-based treatments (pp. 47–65). Chichester, UK: Wiley & Sons.

74.

Schechter

D. S.

Coots

Zeanah

C. H.

Davies

Coates

S. W.

Trabka

K. A.

… Myers

M. M.

(2005). Maternal mental representations of the child in an inter-city clinical sample: Violence-related posttraumatic stress and reflective functioning. Attachment and Human Development, 7(3), 313-331.

75.

Scher

C. D.

Stein

M. B.

Asmundson

G. J.

McCreary

D. R.

Forde

D. R.

(2001). The childhood trauma questionnaire in a community sample: Psychometric properties and normative data. Journal of Traumatic Stress, 14(4), 843-857.

76.

Sharp

Fonagy

(2008). The parent’s capacity to treat the child as a psychological agent: Constructs, measures and implications for developmental psychopathology. Social Development, 17(3), 737-754.

77.

Slade

(1999). Representation, symbolization and affect regulation in the concomitant treatment of a mother and child: Attachment theory and child psychotherapy. Psychoanalytic Inquiry, 19, 797-830.

78.

Slade

(2005). Parental reflective functioning: An introduction. Attachment and Human Development, 7(3), 269-281.

79.

Slade

(2006). Reflective parenting programs: Theory and development. Psychoanalytic Inquiry, 26, 640-657.

80.

Slade

(2007a). Disorganized mother, disorganized child: The mentalization of affective dysregulation and therapeutic change. In Oppenheim

Goldsmith

(Eds.), Attachment theory in clinical work with children: Bridging the gap between research and practice (Vol. 226–250). New York, NY: Guilford Publications.

81.

Slade

(2007b). The pregnancy interview – Revised. Manuscript.

82.

Slade

Aber

J. L.

Berger

Bresgi

Kaplan

(2003). Parent development interview – Revised: Short version. New York, NY: City College of New York.

83.

Slade

Bernbach

Grienenberger

Levy

D. W.

Locker

(2005). Addendum to reflective functioning scoring manual. New York, NY: City University of New York.

84.

Slade

Grienenberger

Bernbach

Levy

Lockyer

(2005). Maternal reflective functioning, attachment, and the transmission gap: A preliminary study. Attachment and Human Development, 7(5), 283-298.

85.

Slade

Sadler

L. S.

de Dios-Kenn

Webb

Currier-Ezepchick

Mayes

(2006). Minding the baby: A reflective parenting program. Psychoanalytic Study of the Child, 60, 74-100.

86.

Smith

B. D.

Testa

M. F.

(2002). The risk of subsequent maltreatment in families with substance-exposed infants. Child Abuse and Neglect, 26, 97-114.

87.

Smith

D. K.

Johnson

A. B.

Pears

K. C.

Fisher

P. A.

De Garmo

D. S.

(2007). Child maltreatment and foster care: Unpacking the effects of prenatal and postnatal parental substance use. Child Maltreatment, 12(2), 150-160.

88.

Steele

(2008). On the origins of reflective functioning. In Busch

(Ed.), Mentalization: Theoretical considerations, research findings, and clinical implications (pp. 133–258). New York, NY: Taylor and Francis.

89.

Street

Harrington

Chiang

Cairns

Ellis

(2004). How great is the risk of abuse in infants to drug-using mothers. Child: Care, Health and Development, 30(4), 325-330.

90.

Stronach

E. P.

Toth

S. L.

Rogosch

Oshri

Manly

J. T.

Cicchetti

(2011). Child maltreatment, attachment security, and internal representations of mother and mother–child relationships. Child Maltreatment, 16(2), 137-145.

91.

Suchman

DeCoste

Leigh

Borelli

(2010). Reflective functioning in mothers with drug use disorders: Implications for dyadic interactions with infants and toddlers. Attachment and Human Development, 12, 567-585.

92.

Suchman

Luthar

(2001). The mediating role of parenting stress in methadone-maintained mothers’ parenting. Parenting: Science and Practice, 1(4), 285-315.

93.

Suchman

Mayes

Conti

Slade

Rounsaville

(2004). Rethinking parenting interventions for drug-dependent mothers: From behavior management to fostering emotional bonds. Journal of Substance Abuse Treatment, 27, 179-185.

94.

Suchman

McMahon

T. J.

Slade

Luthar

(2005). How early bonding, depression, illicit drug use, and perceived support work together to influence drug-dependent mothers’ caregiving. American Journal of Orthopsychiatry, 75(3), 431-445.

95.

Suchman

McMahon

T. J.

Zhang

Mayes

L. C.

Luthar

(2006). Substance-abusing mothers and disruptions in child custody: An attachment perspective. Journal of Substance Abuse Treatment, 30, 197-204.

96.

Taplin

Mattick

R. P.

(2011). Child protection and mothers in substance abuse treatment, Technical Report 30. Sydney, NSW: National Drug and Alcohol Research Centre.

97.

Theran

S. A.

Levendosky

A. A.

Bogat

G. A.

Huth-Bocks

A. C.

(2005). Stability and change in mothers’ internal representations of their infants over time. Attachment and Human Development, 7(3), 253-268.

98.

Tronick

E. Z.

Messinger

D. S.

Weinberg

M. K.

Lester

B. M.

LaGasse

Seifer

(2005). Cocaine exposure is associated with subtle compromises of infants’ and mothers’ social–emotional behavior and dyadic features of their interaction in the face-to-face still-face paradigm. Developmental Psychology, 41(5), 711-722.

99.

Van Ijzendoorn

M. H.

Schuengel

Bakermans-Kranenburg

M. J

. (1999). Disorganized attachment in early childhood: Meta-analysis of precursors, concomitants and sequelae. Development and Psychopathology, 11, 225-249.

100.

Velez

M. L.

Jansson

L. M.

Montoya

I. D.

Schweitzer

Golden

Svikis

(2004). Parenting knowledge among substance abusing women in treatment. Journal of Substance Abuse Treatment, 27, 215-222.

101.

White

(2005). Literature review: Assessment of parenting capacity. Ashfield, Australia: Centre for Parenting Research.

102.

White

Walsh

(2006a). An issues paper: risk assessment in child welfare. Ashfield, Australia: Centre for Parenting and Research & NSW Department of Community Services.

103.

White

Walsh

(2006b). Risk assessment in child welfare. Ashfield, NSW Australia: NSW Department of Community Services.

104.

Whiteside

S. P.

Lynam

D. R.

(2001). The five factor model and impulsivity: Using a structural model of personality to understand impulsivity. Personality and Individual Differences, 30, 669-689.

105.

World Health Organization. (2010). Child maltreatment. Geneva, Switzerland: World Health Organization.

106.

Zanarini

M. C.

Vujanovic

A. A.

Parachini

E. A.

Boulanger

J. L.

Frankenburg

F. R.

Hennen

(2003). Zanarini rating scale for borderline personality disorder (ZAN-BPD): A continuous measure of DSM-IV borderline psychopathology. Journal of Personality Disorders, 17(3), 233-242.

107.

Zeanah

C. H.

Benoit

Hirshberg

Barton

M. L.

Regan

(1994). Mothers’ representations of their infants are concordant with attachment classifications. Developmental Issues in Psychiatry and Psychology, 1, 9-18.

108.

Zilberman

M. L.

Blume

S. B.

(2005). Drugs and women. In Lowinson

J. H.

Ruiz

Millman

Langord

J. G.

(Eds.), Substance abuse: A comprehensive textbook (pp. 1075–1092). Philadelphia, PA: Lippincott Williams and Wilkins.