Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2013 Jan 1.
Published in final edited form as: J Child Psychol Psychiatry. 2011 Jul 7;53(1):46–55. doi: 10.1111/j.1469-7610.2011.02437.x

Attachment security as a mechanism linking foster care placement to improved mental health outcomes in previously institutionalized children

Katie A McLaughlin 1, Charles H Zeanah 2, Nathan A Fox 3, Charles A Nelson 1
PMCID: PMC3193871  NIHMSID: NIHMS303072  PMID: 21733136

Abstract

Background

Children reared in institutions experience elevated rates of psychiatric disorders. Inability to form a secure attachment relationship to a primary caregiver is posited to be a central mechanism in this association. We determined whether the ameliorative effect of a foster care (FC) intervention on internalizing disorders in previously institutionalized children was explained by the development of secure attachment among children placed in FC and evaluated the role of lack of attachment in an institutionalized sample on the etiology of internalizing disorders within the context of a randomized trial.

Methods

A sample of 136 children (aged 6-30 months) residing in institutions was recruited in Bucharest, Romania. Children were randomized to FC (n=68) or to care as usual (CAU; n=68). Foster parents were recruited, trained, and overseen by the investigative team. Attachment security at 42 months was assessed using the Strange Situation Procedure, and internalizing disorders at 54 months were assessed using the Preschool Age Psychiatric Assessment.

Results

Girls in FC had fewer internalizing disorders than girls in CAU (OR=0.17, p=006). The intervention had no effect on internalizing disorders in boys (OR=0.47, p=.150). At 42 months, girls in FC were more likely to have secure attachment than girls in CAU (OR=12.5, p<.001), but no difference was observed in boys (OR=2.0, p=.205). Greater attachment security predicted lower rates of internalizing disorders in both sexes. Development of attachment security fully mediated intervention effects on internalizing disorders in girls.

Conclusion

Placement into FC facilitated the development of secure attachment and prevented the onset of internalizing disorders in institutionalized girls. The differential effects of FC on attachment security in boys and girls explained gender differences in the intervention effects on psychopathology. Findings provide evidence for the critical role of disrupted attachment in the etiology of internalizing disorders in children exposed to institutionalization.

Keywords: institutionalization, childhood adversity, attachment, internalizing, depression, anxiety

In many parts of the world abandoned or orphaned children are raised in socially depriving institutions. UNICEF estimates that 8 million children worldwide reside in such institutional settings. The number of children who are being placed in institutions is increasing (UNICEF, 2002), while opportunities for international adoption are diminishing. The number of international adoptions in the U.S., for example, has declined by approximately 44% in the past 5 years (U.S. Department of State, 2009). These trends raise concerns about how societies will manage the substantial burden of health problems among previously institutionalized children.

Children raised in institutions exhibit myriad developmental abnormalities including stunted growth (Rutter & English Romanian Adoptees (ERA) Study Team, 1998) physical and mental health problems (Rutter & English Romanian Adoptees (ERA) Study Team, 1998; Rutter, Kreppner, O'Connor, & English Romanian Adoptees (ERA) Study Team, 2001), cognitive and language deficits (Nelson et al., 2007; Rutter & English Romanian Adoptees (ERA) Study Team, 1998), and atypical social and emotional development (O'Connor et al., 2003; Zeanah, Smyke, Koga, Carlson, & The Bucharest Early Intervention Project Core Group, 2005), which frequently persist after children are removed from institutional care (Kreppner et al., 2007). High rates of psychiatric disorders among previously institutionalized children are particularly striking (Gunnar, van Dulmen, & The International Adoption Project Team, 2007; Hoksbergen et al., 2003; Rutter & English Romanian Adoptees (ERA) Study Team, 1998; Rutter et al., 2001; Stevens et al., 2008). The absence of a primary attachment figure has been posited to play a central role in explaining the detrimental mental health effects of institutionalization, given the critical importance of attachment for adaptive child development (Bowlby, 1951, 1982; Erickson, Sroufe, & Egeland, 1985; Matas, Arend, & Sroufe, 1978; Sroufe, 1983) and the consistently documented attachment disturbances among children reared in institutions (Chisolm, 1998; O'Connor, Bredenkamp, Rutter, & The English and Romanian Adoptees Study Team, 1999; O'Connor et al., 2003; Smyke, Dumitrescu, & Zeanah, 2002; Zeanah et al., 2005). Empirical data to support this claim, however, have been lacking. As such, it is unknown whether psychiatric disorders among previously institutionalized children are attributable to the lack of available attachment figures or whether providing opportunities for developing attachment security in this population would result in concomitant improvements in mental health.

The Bucharest Early Intervention Project (BEIP), a randomized controlled trial of foster care for institutionalized children, presented an opportunity to examine these questions. The study identified a sample of children living in institutions in Bucharest, Romania, and randomized them to be placed in high quality foster care created and monitored by the investigators or to receive care as usual. The randomized design provided an opportunity to examine the ameliorating effect of placement into a family and, presumably, the formation of an attachment relationship with a caregiver on psychiatric outcomes. At 54 months, children in foster care had lower rates of internalizing disorders than children in care as usual (Zeanah et al., 2009). Notable gender differences in the effects of foster care on psychiatric morbidity were observed such that girls in foster care had lower levels of internalizing symptoms and disorders than girls in care as usual, but no intervention effects were evident among boys. There were no intervention effects on the rates of externalizing disorders amongst the sample.

In the current investigation, we explored the mechanisms that explained the intervention effects on internalizing psychopathology in the BEIP and the gender differences in those effects. A young child develops a secure attachment relationship with a primary caregiver when the care-giving relationship is predictable, sensitive, and responsive to the child's needs (De Wolff & van IJzendoorn, 1997; Koren-Karie, Oppenheim, & Doley, 2002; McElwain & Booth-LaForce, 2006; Sroufe, 1979; van IJzendoorn, 1995). Not surprisingly, attachment relationships fail to develop or are disrupted in children reared in institutional environments (Chisolm, 1998; O'Connor et al., 1999; O'Connor et al., 2003; Smyke et al., 2002; Zeanah et al., 2005) where child-to-caregiver to ratios are high and the quality of care-giving is poor (Zeanah et al., 2003). Moreover, in non-institutionalized samples, insecure or disorganized attachment has been associated with risk for internalizing psychopathology in children (Allen, Moore, Kupermine, & Bell, 1998; Lee & Hankin, 2009; Lyons-Ruth, Easterbrooks, & Cibelli, 1997; McCartney, Owen, Booth, Clarke-Stewart, & Vandell, 2004), including anxiety disorders and major depression (Abela et al., 2005; Warren, Huston, Egeland, & Sroufe, 1997), although findings have been somewhat inconsistent across studies (Brumariu & Kerns, 2010). We therefore examined whether development of attachment security explained the mental health benefits of the foster care intervention.

Because those benefits were observed only among girls, we evaluated whether girls were more likely to develop a secure attachment to their caregivers following foster care placement. Gender differences in attachment security have not been examined in studies of previously institutionalized children or of maltreated children whose caregivers participated in attachment-focused interventions (Cicchetti, Rogosch, & Toth, 2006; O'Connor et al., 1999; O'Connor et al., 2003; van den Boom, 1994). However, it is possible that girls are better able to elicit sensitive caregiving or to or benefit from such caregiving than boys after exposure to institutional rearing or other caregiver disruptions. We evaluated these two research questions in the current report.

Methods

Sample

The BEIP was the first randomized controlled trial of foster care among children reared in institutional settings (Zeanah et al., 2003). A sample of 136 children (aged 6-30 months) was recruited from the six institutions for young children in Bucharest, Romania. Children were randomized to a foster care intervention developed by the research team or to care as usual. Comprehensive assessments of each child's health, cognitive ability, and psychosocial functioning were completed at initial entry into the study and at 42 months. Psychiatric disorders were assessed at 54 months.

Procedures

Participants were selected from each of the six institutions for young children in Bucharest. Physical examinations were completed on 187 children. Of this group, 51 were excluded from participation for medical reasons, ranging from genetic syndromes (e.g., Down syndrome), to fetal alcohol syndrome, to microcephaly (Figure 1). The remaining 136 children, many abandoned at birth, had lived in an institution for an average of 89% of their lives. Following a baseline assessment, half of the children (n=68) were randomized to the foster care intervention and half (n=68) remained in institutional care. No differences were found between the intervention and control group in gender distribution, age, birth weight, or percentage of life spent in the institution, and no gender differences were observed in any of these baseline characteristics (Table 1). The mean age at foster care placement was 22.97 months. By the 54-month assessment, 15 children were lost to follow-up, primarily due to adoption or return to their biological parents. The study design and methods have been described in detail previously (Zeanah et al., 2003).

Figure 1.

Figure 1

Open in a new tab

Study flow of participants from screening to analysis.

Table 1.

Baseline Characteristics of Institutionalized Boys and Girls Randomized to the Foster Care Intervention or Usual Care1

Foster Care (n=68) Usual Care (n=68) p-value
Female, No. (%) 33 (49.3) 35 (51.5) .797
Age, mean (SD), months
    Age at Institutionalization 2.6 (3.9) 2.4 (3.0) .783
    Age at Randomization 20.9 (7.1) 20.8 (7.7) .947
Birth Weight, mean (SD), grams 2733 (576.2) 2847 (570.2) .307
Gestation Age, mean (SD), weeks 37.0 (2.4) 37.6 (1.5) .124
Head Circumference for age percentile, mean (SD), cm. 32.3 (28.1) 25.2 (22.7) .133
Height for age percentile, mean (SD), cm. 25.7 (22.5) 26.9 (23.1) .781
Weight for age percentile, mean (SD), kg. 18.2 (19.4) 22.7 (24.6) .264
Weight for height percentile, mean (SD), kg. 30.1 (27.2) 37.3 (29.6) .163
Duration of Institutionalization, mean (SD), weeks 85.2 (23.0) 87.9 (17.9) .470
Girls (n=68) Boys (n=68) p-value
Age, mean (SD), months
    Age at Institutionalization 2.1 (3.5) 1.7 (3.4) .321
    Age at Randomization 20.6 (8.1) 22.2 (7.2) .256
Birth Weight, mean (SD), grams 2808 (642.3) 2956 (492.4) .932
Gestation Age, mean (SD), weeks 37.2 (2.2) 37.7 (1.5) .409
Head Circumference for age percentile, mean (SD), cm. 36.9 (28.2) 36.3 (28.0) .895
Height for age percentile, mean (SD), cm. 38.7 (28.7) 32.8 (26.3) .138
Weight for age percentile, mean (SD), kg. 29.9 (28.4) 29.0 (27.2) .729
Weight for height percentile, mean (SD), kg. 37.5 (29.1) 42.6 (29.9) .283
Duration of Institutionalization, mean (SD), weeks 48.8 (46.2) 57.7 (47.9) .195
Open in a new tab
1

None of these baseline characteristics was associated with internalizing symptoms or externalizing symptoms at 54 months.

The BEIP was initiated with support from the Secretary of State for Child Protection in Romania. Study procedures were approved by the local commissions on child protection in Bucharest, the Romanian ministry of health, an ethics committee including appointees from government and Bucharest University academic departments, and the institutional review boards of the institutions of the three principal investigators. A complete description of procedures employed to ensure ethical integrity has been published previously (Miller, 2009; Millum & Emanuel, 2007; Zeanah, Koga, Simion, Stanescu, Tabacaru, Fox, & Nelson, 2006).

Intervention

Because foster care was virtually non-existent in Bucharest at the beginning of the study, the investigators created a network of foster homes with Romanian collaborators (Smyke, Zeanah, Fox, & Nelson, 2009; Zeanah et al., 2003). Foster parents were recruited through advertising, completed a background check, and were trained using a manual adapted for the study. Foster care was supported by social workers in Bucharest who received weekly consultation from U.S. clinicians. Social workers assisted foster parents in managing problem behaviors and facilitated the establishment of warm, supportive, and committed relationships with their foster children. The foster care intervention, described in more detail elsewhere (Smyke et al., 2009; Zeanah et al., 2003), was designed to be affordable, replicable, and grounded in developmental theory and research on care-giving quality and attachment.

Children not randomized to foster care typically remained in institutional care, though over time, some were adopted (n=7), some were returned to their biological families (n=2), and others were placed in government FC that was created after the study began (n=18). We implemented a policy of non-interference regarding the placement of children in the CAU group into families and did not interfere with the removal of these children from institutional care if they were selected for adoption or re-unification with their family of origin (for elaboration on the ethical issues involved in BEIP, see Zeanah et al., 2006).

Assessment

Attachment was assessed using the Strange Situation Procedure (Ainsworth, Blehar, Waters, & Wall, 1978) prior to randomization while all children were living in institutions (Zeanah et al., 2005) and the preschool version of the procedure at 42 months (Smyke et al., 2010). Coders blinded to children's group status were trained to reliability in the MacArthur Preschool System, which was used to determine classifications of attachment at 42 months of age: secure, and several types of insecure, including avoidant, ambivalent-dependent, disorganized-controlling, and insecure-other (Cassidy, Marvin, & The MacArthur Working Group, 1992). For analysis purposes, children were divided into secure versus insecure attachment groups. Three-quarters (n=120) of the assessments were double-coded to assess reliability. As reported previously, agreement for secure versus insecure coding was excellent (kappa=0.83, percent agreement=91.7%) (Smyke et al., 2010). Coders also assigned a continuous rating of security using a scale ranging from 1 (no security evident) to 9 (most secure) (Cassidy et al., 1992). Inter-rater reliability for this scale was excellent (r=0.87) (Smyke et al., 2010). We also created a dichotomous variable coding whether each child's attachment changed from absent or incompletely-formed at baseline to secure at 42 months.

Psychiatric disorders were assessed at 54 months using a structured diagnostic interview for young children, the Preschool Age Psychiatric Assessment (PAPA; (Egger, Ascher, & Angold, 1999; Egger et al., 2006). This instrument collects information from a caregiver about the presence, frequency, and duration of psychiatric symptoms and generates DSM-IV diagnoses (American Psychiatric Association, 1994). We examined both internalizing and externalizing disorders and symptom counts to increase power due to the small sample size (see Zeanah et al., 2009). Internalizing disorders considered here included major depression, dysthymia, separation anxiety disorder, social phobia, specific phobia, generalized anxiety disorder, and post-traumatic stress disorder. Externalizing disorders included attention-deficit/hyperactivity disorder (ADHD), oppositional defiant disorder (ODD), and conduct disorder. Symptom counts represent that total number of symptoms reported by the child's caregiver that were included in the diagnostic criteria for internalizing disorders and externalizing disorders. The PAPA has been demonstrated to have comparable reliability to structured diagnostic interviews used with adults and older children (Egger et al., 2006). The interview was translated into Romanian and back-translated into English and retention of original meaning was assessed by bilingual study staff. The PAPA was administered by trained study interviewers.

For both the PAPA and the Strange Situation Procedure, biological or foster care mothers completed the assessments for non-institutionalized children. An institutional caregiver completed the assessments for children living in institutions. Caregivers who worked with the child regularly and knew the child well were selected. If the child had a favorite caregiver, agreed upon by staff consensus, the favorite caregiver completed the assessments.

Analysis Methods

All analyses used an intent-to-treat approach. Group differences in attachment security were examined using cross-tabulations and univariate analysis of variance (ANOVA) with intervention group as the between-subjects factor. Associations of the intervention and of attachment with internalizing symptoms were examined using linear regression; association with internalizing disorders were examined using logistic regression. Logistic regression coefficients and their standard errors were exponentiated to create odds ratios (ORs) and 95% confidence intervals (CIs). Statistical significance was evaluated using 2-sided .05-level tests.

Results

Intervention Effects on Psychopathology

As we reported previously, the foster care intervention had ameliorative effects on internalizing symptoms and disorders in girls but not boys (Zeanah et al., 2009). Girls in the foster care group had lower levels of internalizing symptoms, β=-0.37, p=.004, and fewer symptoms of both anxiety, β=-0.34, p=.009, and depression, β=-0.34, p=.009, than girls in the care as usual group (Table 2). Females randomized to foster care also were less likely to have internalizing disorders than females in care as usual, OR=0.17, p=006. In contrast, boys in the foster care group had equivalent levels of internalizing symptoms, β=0.12, p=.372, anxiety, β=0.17, p=.190, depression, β=-0.02, p=.879, and internalizing disorders, OR=0.47, p=.150, as boys in care as usual. The foster care intervention had no effect on externalizing symptoms or on the prevalence of externalizing disorders in either girls or boys.

Table 2.

Psychopathology in Institutionalized Children Randomized to the Foster Care Intervention or Usual Care at 54 months, Stratified by Gender

Foster Care (n=63) Usual Care (n=58) p-value1
Females
    Internalizing Symptoms, mean (SD) 3.0 (1.6) 5.2 (3.8) .004
        Anxiety Symptoms 2.3 (1.0) 3.6 (2.5) .009
        Depression Symptoms 0.7 (0.8) 1.6 (1.7) .009
    Internalizing Disorders, No. (%)2 4 (13.3) 14 (48.3) .006
    Externalizing Symptoms, mean (SD) 4.9 (1.0) 6.6 (1.3) .278
        ADHD Symptoms 3.6 (0.8) 4.8 (1.1) .377
        ODD/Conduct Disorder Symptoms 1.3 (0.3) 1.9 (0.4) .296
    Externalizing Disorders, No. (%)3 5 (17.2) 4 (16.0) .903
Males
    Internalizing Symptoms, mean (SD) 5.6 (3.4) 4.9 (2.4) .372
        Anxiety Symptoms 4.2 (2.5) 3.5 (1.7) .190
        Depression Symptoms 1.3 (1.3) 1.4 (1.1) .879
    Internalizing Disorders, No. (%) 10 (30.3) 14 (48.3) .150
    Externalizing Symptoms, mean (SD) 9.5 (1.4) 9.7 (1.4) .931
        ADHD Symptoms 6.2 (0.9) 6.7 (1.0) .692
        ODD/Conduct Disorder Symptoms 3.3 (0.6) 3.0 (0.5) .657
    Externalizing Disorders, No. (%)3 10 (33.3) 11 (40.7) .563
Open in a new tab

Abbreviations: ADHD, attention-deficit/hyperactivity disorder; ODD, oppositional defiant disorder.

1

Based on linear regression models predicting symptom outcomes and logistic regression model predicting disorders.

2

Internalizing disorders included major depression, dysthymia, and the anxiety disorders (separation anxiety disorder, specific phobia, social phobia, generalized anxiety disorder, and post-traumatic stress disorder).

3

Externalizing disorders included ADHD, ODD, and conduct disorder.

Intervention Effects on Attachment

We examined the effect of the intervention on dichotomous indicators of attachment security (secure vs. absent or incompletely-formed) at baseline and 42 months, on the dichotomous variable indicating whether each child's attachment changed from absent or incompletely-formed at baseline to secure at 42 months, and on the continuous measure of security at 42 months. The proportion of females with secure attachment at baseline did not differ across groups, OR=1.1, p=.933, (Table 3). By 42 months, girls in foster care were more like to be securely attached than girls in care as usual, OR=12.5, p<.001. A greater proportion of girls in foster care became securely attached from baseline to 42 months than in care as usual, OR=6.6, p=.029. Finally, girls in foster care had higher scores on the continuous measure of attachment security than girls in care as usual, F(1,61)=31.2, p<.001.

Table 3.

Attachment Security of Institutionalized Children Randomized to the Foster Care Intervention or Usual Care at Baseline and 42 months, Stratified by Gender

Foster Care (n=63) Usual Care (n=58) p-value1
Females
    Baseline Secure Attachment, No. (%)2 5 (22.7) 5 (23.8) .933
    42 Month Secure Attachment, No. (%)3 19 (63.3) 4 (12.1) <.001
    Change from Insecure/Incompletely-Formed to Secure, No. (%)4 9 (40.9) 2 (9.5) .029
    42 Month Attachment Security, mean (SD) 5.0 (1.6) 2.9 (1.3) <.001
Males
    Baseline Attachment Security, No. (%) 4 (14.8) 4 (16.0) .906
    42 Month Secure Attachment, No. (%) 12 (35.3) 6 (20.7) .205
    Change from Insecure/Incompletely-Formed to Secure, No. (%) 8 (29.6) 4 (16.0) .250
    42 Month Attachment Security, mean (SD) 4.2 (1.7) 3.1 (1.4) .007
Open in a new tab
1

Based on logistic regression models predicting dichotomous attachment security and linear regression model predicting continuous attachment security.

2

At baseline, 49 children in the foster care group (22 females and 29 males) and 46 in usual care (21 females and 25 males) completed the Strange Situation Procedure.

3

At 42 months, 64 children in the foster care group (30 females and 34 males) and 62 in usual care (33 females and 29 males) completed the Strange Situation Procedure.

4

Percentages calculated from the total number of children for whom data on attachment security were available at both baseline and 42 months.

Intervention effects on attachment in boys were notably different. The proportion of boys with secure attachment did not differ in the foster care and care as usual groups at baseline, OR=1.1, p=.906, or 42 months, OR=2.1, p=.205. The proportion of boys who became securely attached from baseline to 42 months also did not differ in the foster care group compared to the care as usual group, OR=2.2, p=.250. In contrast, boys in foster care had higher scores on the continuous measure of attachment security at 42 months than boys in care as usual, F(1,61)=7.8, p=.007.1

Attachment and Psychopathology

Attachment security was associated with internalizing psychopathology in both girls and boys. Females with secure attachment at 42 months had lower levels of internalizing symptoms, β=-0.41, p<.001, including anxiety symptoms, β=-0.40, p<.001, and depression symptoms, β=-0.35, p=.002, at 54 months. The prevalence of internalizing disorders was lower among girls with secure attachment than among girls with insecure attachment, OR=0.29, p=.035. Higher scores on the continuous measure of attachment security were associated with lower levels of internalizing symptoms, β=-0.49, p<.001, and of anxiety, β=-0.42, p<.001, and depression, β=-0.49, p<.001.

Boys with secure attachment at 42 months had lower levels of anxiety symptoms, β=-0.30, p=.006, depressive symptoms, β=-0.33, p=.003, and internalizing symptoms, β=-0.35, p=.002, at 54 months. Internalizing disorders were less common among boys with secure attachment as compared to insecure attachment, OR=0.25, p=.021. Higher scores on the continuous attachment measure were associated with lower levels of internalizing symptoms, β=-0.41, p<.001, and of anxiety, β=-0.36, p<.001, and depression, β=-0.39, p<.001.

Mediating Role of Attachment in Intervention Effects for Females

To determine whether the intervention effects on internalizing psychopathology in girls were mediated through attachment security, we first examined our dichotomous attachment indicator at 42 months. Intervention effects on internalizing symptoms in females were fully mediated by attachment security (Table 4). Intervention effects on total internalizing symptoms, β=-1.42, p=.105, and symptoms of anxiety, β=-0.73, p=.194, and depression, β=-0.68, p=.090, were no longer significant when dichotomous attachment security was added to the model. Attachment security at 42 months explained 36.6% of the intervention effect on total internalizing symptoms, 44.7% of the intervention effect on anxiety symptoms, and 26.1% of the intervention effect on depressive symptoms. In contrast, the intervention effect on internalizing disorders remained significant when dichotomous attachment security at 42 months was added to the model, OR=0.19, p=.025. Attachment security at 42 months explained 10.5% of the intervention effect on internalizing disorders.

Table 4.

Mediating Role of Attachment Security in the Effect of the Foster Care Intervention on Internalizing Psychopathology of Institutionalized Females at 54 months

Intervention Model1 p-value Mediation Model2 p-value % Attenuation3
Dichotomous Attachment Security, 42 months
    Internalizing Symptoms, β (SE) -2.24 (.75) .004 -1.42 (.86) .105 36.6
        Anxiety Symptoms -1.32 (.49) .009 -0.73 (.56) .194 44.7
        Depression Symptoms -0.92 (.34) .009 -0.68 (.40) .090 26.1
    Internalizing Disorders, OR (95% CI) 0.17 (0.04, 0.6) .006 0.19 (0.05, 0.82) .025 10.5
Change in Attachment Security, Baseline to 42 months
    Internalizing Symptoms, β (SE) -2.24 (.75) .004 -1.03 (.81) .214 54.0
        Anxiety Symptoms -1.32 (.49) .009 -0.58 (.62) .359 56.1
        Depression Symptoms -0.92 (.34) .009 -0.45 (.31) .164 51.1
    Internalizing Disorders, OR (95% CI) 0.17 (0.04, 0.6) .006 0.33 (0.05, 2.16) .246 48.5
Continuous Attachment Security, 42 months
    Internalizing Symptoms, β (SE) -2.24 (.75) .004 -1.12 (.89) .217 50.0
        Anxiety Symptoms -1.32 (.49) .009 -0.74 (.59) .218 43.9
        Depression Symptoms -0.92 (.34) .009 -0.38 (.40) .353 58.7
    Internalizing Disorders, OR (95% CI) 0.17 (0.04, 0.6) .006 0.29 (0.07, 1.26) .098 41.4
Open in a new tab
1

Intervention model includes dummy variable for the intervention and shows the total intervention effect on each of the four internalizing outcomes.

2

Mediation model includes dummy variable for intervention effect and attachment security variable from Column 1 and shows the magnitude of the intervention effect after controlling for attachment security.

3

Proportion of intervention effect on each internalizing outcome that is explained by intervention effect on attachment security.

Intervention effects on both internalizing symptoms and disorders were fully mediated by our indicator of whether attachment security changed from absent or incompletely-formed at baseline to secure by 42 months. After accounting for changes from absent or incompletely-formed to secure attachment, intervention effects were no longer evident for internalizing symptoms, β=-1.03, p=.214, anxiety symptoms, β=-0.58, p=.359, depressive symptoms, β=-0.45, p=.164, or internalizing disorders, OR=0.33, p=.246. Change in attachment status from baseline to 42 months explained 54.0% of the intervention effect on total internalizing symptoms, 56.1% of the intervention effect on anxiety symptoms, 51.1% of the intervention effect on depressive symptoms, and 48.5% of the intervention effect on internalizing disorders.

Similarly, intervention effects on internalizing symptoms and disorders were fully mediated by the continuous measure of attachment security. Intervention effects for internalizing symptoms, β=-1.12, p=.217, anxiety symptoms, β=-0.74, p=.218, depressive symptoms, β=-0.38, p=.353, and internalizing disorders, OR=0.29, p=.098, were no longer significant after adding continuous attachment security to the model. The continuous measure of attachment security explained 50.0% of the intervention effect on total internalizing symptoms, 43.9% of the intervention effect on anxiety symptoms, 58.7% of the intervention effect on depressive symptoms, and 41.4% of the intervention effect on internalizing disorders.

Discussion

Girls reared in Romanian institutions who were placed in foster care evidenced lower levels of internalizing psychopathology than girls who remained in institutional care. Our findings suggest that the emergence or development of attachment security is the mechanism through which the intervention had ameliorative effects on anxiety and depression. Girls placed into families developed a secure attachment to a primary caregiver (Smyke, Zeanah, Fox, Nelson, & Guthrie, 2010). This improvement in attachment status, in turn, protected them against the development of internalizing symptoms and anxiety and mood disorders in early childhood.

These findings have implications for research, practice, and policy. First, we provide strong evidence for the critical role of the lack of attachment in the etiology of internalizing disorders in children exposed to institutionalization. Although numerous observational studies have documented an association between attachment security and child internalizing psychopathology (Abela et al., 2005; Allen et al., 1998; Lee & Hankin, 2009; Lyons-Ruth et al., 1997; McCartney et al., 2004; Warren et al., 1997), findings have been inconsistent across studies (Brumariu & Kerns, 2010), and firm conclusions regarding the explanatory role of attachment in the etiology of internalizing disorders cannot be made in the absence of an experimental design. Our use of a randomized intervention allows considerably stronger causal inferences to be drawn about the role of attachment in the onset of child internalizing pathology than these previous observational studies. Developmental processes that are altered as a result of a randomized intervention and that confer protection against negative outcomes can be inferred to be true risk factors for those outcomes (Cicchetti & Hinshaw, 2002; Coie et al., 1993; Kraemer et al., 1997). Here, we demonstrated experimentally that placement into a family allowed for the development of attachment security for girls and explained their lower rates of internalizing psychopathology. Foster care did not result in similar improvements in attachment security for boys and thus did not protect them against the development of internalizing disorders. Together, these findings provide clear evidence for the importance of lack of attachment figures—and hence the absence of attachment—in the development of anxiety and depression in children exposed to institutionalization.

Our findings suggest an explanation for the gender differences in the positive mental health effects of the intervention. At 42 months, a greater proportion of girls randomized to foster care had secure attachment than girls who remained in institutional care, whereas the proportion of boys with a secure attachment did not differ across conditions. These results suggest that boys had a more difficult time developing a secure attachment to a caregiver following removal from institutional care. Why might this have occurred? The most obvious explanations include gender differences in potential risk factors (e.g., birth weight, problem behaviors in infancy, age at placement) or greater caregiving quality for girls compared to boys either in the institutions prior to randomization or in foster family placements. Neither of these explanations is consistent with our data. As described previously, we found no gender differences in a wide range of pre-intervention risk markers, and further analysis suggested no gender differences in behavior problems or disorganized attachment at baseline. It is notable that prior research was unable to identify developmental precursors of attachment disturbances in previously-institutionalized Romanian children adopted into the UK (O'Connor et al., 1999), further suggesting that baseline developmental differences are not responsible for these effects. Subsequent analysis also revealed no gender differences in caregiving quality in institutional settings at baseline or in foster families at 42 months (assessed using the Observational Record of Caregiving Quality [NICHD Early Child Care Research Network, 1996; see Smyke et al., 2007 for details]). Thus, no differences were observed in the quantity of specific caregiving behaviors or the quality of caregiver responses to children's behavior in the foster families of boys and girls. However, it is possible that girls were better able to elicit more sensitive caregiving in nuanced ways that were not detectable by our observational measure. It is also possible that caregivers perceived boys to be more difficult to manage or found them more difficult to connect with emotionally, despite having comparable levels of psychopathology as girls at the time of foster care placement. Alternatively, it may be that institutional deprivation has a more profound deleterious effect on boys’ ability to develop a secure attachment post-institutionalization. Some studies have found that the absence of a father is associated with greater attachment disturbances in maltreated male infants compared to females (Carlson, Cicchetti, Barnett, & Braunwald, 1989). Because institutional caregivers were predominantly female, the absent of a male caretaker in the institution may have been more detrimental for boys. Although we were unable to determine conclusively the explanation for gender differences in the effects of our foster care intervention on attachment security, these findings raise a number of important questions that warrant examination in future research regarding gender differences in attachment development in children exposed to institutional rearing, psychosocial deprivation and maltreatment.

Practically, we identified a mechanism underlying the ameliorative effects of foster care on internalizing psychopathology. This mechanism has implications for interventions targeting children raised in deprived environments. Specifically, they point to the importance of promoting the development of a secure attachment relationship for children who have been orphaned or abandoned. Several existing interventions hold promise in this regard. Intervention techniques that teach caregivers to attend appropriately and respond contingently to infant needs have demonstrated efficacy in improving attachment security among children with difficult temperaments and in at-risk families (Heinicke et al., 1999; Klein Velderman, Bakermans-Kranenburgh, & Juffer, 2006; van den Boom, 1994). These techniques also have been applied successfully with foster parents and ameliorated attachment security in foster infants and preschoolers (Dozier, Higley, Albus, & Nutter, 2002; Fisher & Kim, 2007). This evidence indicates that attachment security can be ameliorated in children who have confronted a wide range of adverse early rearing environments. Our findings suggest that the implementation of such interventions may prevent the onset of anxiety and depression in children raised in such circumstances.

From a policy perspective, our results argue strongly for removing children from institutional care and placing them in families to promote the development of secure attachment to caregivers. Our findings indicate that secure attachment relationships are critical for adaptive mental health functioning in previously institutionalized children. In the absence of such relationships, the probability of adverse mental health outcomes in this population is quite high. Because institutionalization remains common in many parts of the world (UNICEF, 2002, 2007) policy efforts to promote the placement of orphaned children into foster families represent an important public health priority. These findings also have policy implications for the much larger population of children who have experienced serious disruptions in their relationships with caregivers. Our findings suggest that placement in families whenever possible remains the preferred option. In families, these children also likely have a considerably greater opportunity to form secure attachment relationships with caregivers.

Findings should be interpreted in light of the following limitations. First, the study included only children from Romanian institutions, which may be characterized by more extreme deprivation than in other settings (Zeanah et al., 2003). Findings may not generalize to populations of institutionalized children who experienced less severe deprivation, and therefore, warrant replication in other samples and cultural contexts. Second, because none of the children in our sample were younger than six months of age at baseline we could not evaluate the effects of early placement into foster care. Given that placement before six months is associated with advantageous attachment outcomes (Chisolm, 1998; Rutter et al., 2007), early foster care may result in even greater effects on attachment and psychopathology. Third, behaviors observed in the Strange Situation at the baseline assessment while all children lived in the institutions suggest that in the absence of consistent caregivers and in an environment of severe psychosocial deprivation infants and young children may be unable to effectively organize even basic attachment responses. Fourth, the nature and quality of the child-caregiver relationship was assuredly different for children living in institutional settings versus foster families. These differences were unlikely to create systematic reporting bias, however. Institutional caregivers engaged in virtually all activities of daily living with children in their care, and there is no reason to believe that they were unable to accurately report on individual differences in behavior and emotions among these children. There is even less reason to suspect reporting bias that differed by gender. Finally, despite randomization to groups a number of children in the care as usual group were removed from institutional care over the course of the study. As such, our estimates of the effects of institutionalization on attachment and psychopathology are likely conservative.

Girls removed from institutional care and placed into foster families were able to develop secure attachments to their new caregivers, and these attachment relationships protected them from the development of internalizing disorders. Foster care programs for orphaned and maltreated children represent important targets for research, practice, and policy attention.

KEY POINTS.

  • Inability to develop a secure attachment relationship to a primary caregiver is thought to explain the elevated rates of internalizing disorders in children exposed to institutionalization.

  • Study findings show that changes in attachment status is a primary mechanism explaining the ameliorative effects of a foster care intervention on internalizing disorders among previously institutionalized children.

  • Capitalizing on a randomized study design, we show that absence of or unformed attachment plays a critical role in the development of internalizing disorders among children exposed to institutional rearing.

ACKNOWLEDGEMENTS

Drs. McLaughlin, Zeanah, Fox, and Nelson all had access to the data and contributed to writing the manuscript. Drs. Zeanah, Fox, and Nelson conceived and implemented the BEIP study. Dr. McLaughlin conceived the current paper and conducted all data analysis.

This research was supported by a grant from the John D. and Catherine T. MacArthur Foundation Research Network on Early Experience and Brain Development (Charles A. Nelson, Network Chair), the National Institutes of Health (to Charles A. Nelson; 1R01MH091363-01), and research support from the Robert Wood Johnson Foundation to Kate McLaughlin (Grant Number 053572). These funders provided support for all data collection and analysis. We thank Sebastian Koga for overseeing the project in Romania; Hermi Woodward and members of the MacArthur Foundation Research Network on Early Experience and Brain Development for input regarding the conceptualization, design, and implementation of the project; the caregivers and children who participated in this project; and the Bucharest Early Intervention Project staff for their tireless work on our behalf.

Footnotes

1

These findings mirror those reported previously by Smyke and colleagues (2010) although the final sample reported here includes 10 children that were not included in this previous analysis. Those children were excluded from that study due to concerns about not meeting study inclusion criteria. Subsequent medical exam of these children revealed that they did, indeed, meet the criteria. Thus, they were included in the current analyses.

The authors have no financial disclosures or conflicts of interest to report.

REFERENCES

  1. Abela JR, Hankin BL, Haigh EP, Adams P, Vinokuroff T, Trayhern L. Interpersonal vulnerability to depression in high-risk children: The role of insecure attachment and reassurance seeking. Journal of Clinical Child and Adolescent Psychology. 2005;34:182–192. doi: 10.1207/s15374424jccp3401_17. [DOI] [PubMed] [Google Scholar]
  2. Ainsworth MDS, Blehar MC, Waters E, Wall S. Patterns of attachment: A psychological study of the Strange Situation. Erlbaum; Hillsdale, NJ: 1978. [Google Scholar]
  3. Allen JP, Moore C, Kupermine G, Bell K. Attachment and adolescent psychosocial functioning. Child Development. 1998;69:1406–1419. [PMC free article] [PubMed] [Google Scholar]
  4. American Psychiatric Association . Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV) American Psychiatric Press; Washington, DC: 1994. [Google Scholar]
  5. Bowlby J. Maternal care and mental health. World Health Organization; Geneva: 1951. [Google Scholar]
  6. Bowlby J. Attachment and loss: Vol 1. Attachment. Basic Books; New York: 1982. [Google Scholar]
  7. Brumariu LE, Kerns KA. Parent-child attachment and internalizing symptoms in childhood and adolescence: A review of empirical findings and future directions. Development and Psychopathology. 2010;22:177–203. doi: 10.1017/S0954579409990344. [DOI] [PubMed] [Google Scholar]
  8. Carlson V, Cicchetti D, Barnett D, Braunwald K. Disorganized/disoriented attachment relationships in maltreated infants. Developmental Psychology. 1989;25:525–531. [Google Scholar]
  9. Cassidy J, Marvin RS, The MacArthur Working Group . Attachment organization in preschool chilren: Procedures and coding manual. 4th ed. University of Virginia; 1992. [Google Scholar]
  10. Chisolm K. A three year follow-up of attachment and indiscriminate friendliness in children adopted from Romanian orphanages. Child Development. 1998;69:1092–1106. [PubMed] [Google Scholar]
  11. Cicchetti D, Hinshaw SP. Prevention and intervention science: Contributions to developmental theory. Development and Psychopathology. 2002;14:667–981. doi: 10.1017/s0954579402004017. [DOI] [PubMed] [Google Scholar]
  12. Cicchetti D, Rogosch FA, Toth SL. Fostering secure attachment in infants in maltreating families through preventive interventions. Development and Psychopathology. 2006;18:623–649. doi: 10.1017/s0954579406060329. [DOI] [PubMed] [Google Scholar]
  13. Coie JD, Watt NF, West SG, Hawkins D, Asarnow JR, Markman HJ, et al. The science of prevention: A conceptual framework and some directions for a National Research Program. American Psychologist. 1993;48:1013–1022. doi: 10.1037//0003-066x.48.10.1013. [DOI] [PubMed] [Google Scholar]
  14. De Wolff MS, van IJzendoorn MH. Sensitivity and attachment: A meta-analysis on parental antecedents of infant attachment. Child Development. 1997;68:571–591. [PubMed] [Google Scholar]
  15. Dozier M, Higley E, Albus KE, Nutter A. Intervening with foster infants’ caregivers: Targeting three critical needs. Infant Mental Health Journal. 2002;23:541–554. [Google Scholar]
  16. Egger HL, Ascher BH, Angold A. Preschool Age Psychiatric Assessment: Version 1.1. Duke University Medical Center; Durham, NC: 1999. [Google Scholar]
  17. Egger HL, Erkanli A, Keeler G, Potts E, Walter BK, Angold A. Test-retest reliability of the Preschool Age Psychiatric Assessment (PAPA). Journal of the American Academy of Child & Adolescent Psychiatry. 2006;45:538–549. doi: 10.1097/01.chi.0000205705.71194.b8. [DOI] [PubMed] [Google Scholar]
  18. Erickson MF, Sroufe LA, Egeland B. The relationship between quality of attachment and behavior problems in preschool in a high-risk sample. Monographs for the Society for Research in Child Development. 1985;50 1/2, Serial No. 209. [PubMed] [Google Scholar]
  19. Fisher PA, Kim HK. Intervention effects on foster preschoolers’ attachment-related behaviors from a randomized trial. Prevention Science. 2007;8:161–170. doi: 10.1007/s11121-007-0066-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Gunnar MR, van Dulmen MHM, The International Adoption Project Team Behavior problems in postinstitutionalized internationally adopted children. Development and Psychopathology. 2007;19:129–148. doi: 10.1017/S0954579407070071. [DOI] [PubMed] [Google Scholar]
  21. Heinicke CM, Fineman NR, Ruth G, Recchia SL, Guthrie D, Rodning C. Relationship-based intervention with at-risk mothers: Outcome in the first year of life. Infant Mental Health Journal. 1999;20:349–374. [Google Scholar]
  22. Hoksbergen RA, ter Laak J, van Dijkum C, Rijk S, Rijk K, Stoutjesdijk F. Posttraumatic stress disorder in adopted children from Romania. American Journal of Orthopsychiatry. 2003;73:255–265. doi: 10.1037/0002-9432.73.3.255. [DOI] [PubMed] [Google Scholar]
  23. Klein Velderman M, Bakermans-Kranenburgh MJ, Juffer F. Effects of attachment-based interventions on maternal sensitivity and infant attachment: Differential susceptibility of highly reactive infants. Journal of Family Psychology. 2006;20:266–274. doi: 10.1037/0893-3200.20.2.266. [DOI] [PubMed] [Google Scholar]
  24. Koren-Karie N, Oppenheim D, Doley S. Mother's insightfulness regarding their infants’ internal experience: Relations with maternal sensitivity and infant attachment. Developmental Psychology. 2002;38:534–542. doi: 10.1037//0012-1649.38.4.534. [DOI] [PubMed] [Google Scholar]
  25. Kraemer HC, Kazdin AE, Offord DR, Kessler RC, Jensen PS, Kupfer DJ. Coming to terms with the terms of risk. Archives of General Psychiatry. 1997;54:337–343. doi: 10.1001/archpsyc.1997.01830160065009. [DOI] [PubMed] [Google Scholar]
  26. Kreppner JM, Rutter M, Beckett C, Castle J, Colvert E, Groothues C, et al. Normality and impairment following profound early institutional deprivation: A longitudinal follow-up into early adolescence. Developmental Psychology. 2007;43:931–946. doi: 10.1037/0012-1649.43.4.93. [DOI] [PubMed] [Google Scholar]
  27. Lee A, Hankin BL. Insecure attachment, dysfunctional attitudes, and low self-esteem predicting prospective symptoms of depression and anxiety during adolescence. Journal of Clinical Child and Adolescent Psychology. 2009;38:219–231. doi: 10.1080/15374410802698396. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Lyons-Ruth K, Easterbrooks MA, Cibelli CD. Infant attachment strategies, infant mental lag, and maternal depressive symptoms: Predictors of internalizing and externalizing problems at age 7. Developmental Psychology. 1997;33:681–692. doi: 10.1037//0012-1649.33.4.681. [DOI] [PubMed] [Google Scholar]
  29. Matas L, Arend RA, Sroufe LA. Continuity of adaptation in the second year: The relationship between quality of attachment and later competence. Child Development. 1978;49:511–514. [Google Scholar]
  30. McCartney K, Owen MT, Booth CL, Clarke-Stewart A, Vandell DL. Testing a maternal attachment model of behavior problems in early childhood. Journal of Child Psychology and Psychiatry. 2004;45:765–778. doi: 10.1111/j.1469-7610.2004.00270.x. [DOI] [PubMed] [Google Scholar]
  31. McElwain NL, Booth-LaForce C. Maternal sensitivity to infant distress and nondistress as predictors of infant-mother attachment security. Journal of Family Psychology. 2006;20:247–255. doi: 10.1037/0893-3200.20.2.247. [DOI] [PubMed] [Google Scholar]
  32. Miller FG. The randomzied controlled trial as a demonstration project: An ethical perspective. American Journal of Psychiatry. 2009;166:743–745. doi: 10.1176/appi.ajp.2009.09040538. [DOI] [PubMed] [Google Scholar]
  33. Millum J, Emanuel EJ. The ethics of international research with abandoned children. Science. 2007;318:1874–1875. doi: 10.1126/science.1153822. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Nelson CA, Zeanah CH, Fox NA, Marshall PJ, Smyke AT, Guthrie D. Cognitive recovery in socially deprived young children: The Bucharest Early Intervention Project. Science. 2007;318:1937–1940. doi: 10.1126/science.1143921. [DOI] [PubMed] [Google Scholar]
  35. O'Connor T, Bredenkamp D, Rutter M, The English and Romanian Adoptees Study Team Attachment disturbances and disorders in children exposed to early severe deprivation. Infant Mental Health Journal. 1999;20:10–29. [Google Scholar]
  36. O'Connor T, Marvin RS, Rutter M, Olrick JT, Britner PA, The English and Romanian Adoptees Study Team Child-parent attachment following early institutional deprivation. Development and Psychopathology. 2003;15:19–38. doi: 10.1017/s0954579403000026. [DOI] [PubMed] [Google Scholar]
  37. Rutter M, Colvert E, Kreppner JM, Beckett C, Castle J, Groothues C, et al. Early adolescent outcomes for institutionally-deprived and non-deprived adoptees: I: Disinhibited attachment. Journal of Child Psychology and Psychiatry. 2007;48:17–30. doi: 10.1111/j.1469-7610.2006.01688.x. [DOI] [PubMed] [Google Scholar]
  38. Rutter M, English Romanian Adoptees (ERA) Study Team Developmental catch-up, and deficit, following adoption after severe global early privation. Journal of Child Psychology and Psychiatry. 1998;39:465–476. [PubMed] [Google Scholar]
  39. Rutter M, Kreppner JM, O'Connor TG, English Romanian Adoptees (ERA) Study Team Specificity and heterogeneity in children's responses to profound institutional privation. British Journal of Psychiatry. 2001;17:97–103. doi: 10.1192/bjp.179.2.97. [DOI] [PubMed] [Google Scholar]
  40. Smyke AT, Dumitrescu A, Zeanah CH. Attachment disturbances in young children. I: The continuum of caretaking casualty. Journal of the American Academy of Child & Adolescent Psychiatry. 2002;41:972–982. doi: 10.1097/00004583-200208000-00016. [DOI] [PubMed] [Google Scholar]
  41. Smyke AT, Zeanah CH, Fox NA, Nelson CA. A new model of foster care for young children: the Bucharest early intervention project. Child and Adolescent Psychiatric Clinics of North America. 2009;18:721–734. doi: 10.1016/j.chc.2009.03.003. [DOI] [PubMed] [Google Scholar]
  42. Smyke AT, Zeanah CH, Fox NA, Nelson CA, Guthrie D. Foster care enhances quality of attachment among young institutionalized children. Child Development. 2010;81:212–223. doi: 10.1111/j.1467-8624.2009.01390.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Sroufe LA. The coherence of individual development: Early care, attachment, and subsequent developmental issues. American Psychologist. 1979;34:834–841. [Google Scholar]
  44. Sroufe LA. Infant-caregiver attachment and patterns of adaptation in preschool: The roots of maladaption and competence. In: Perlmutter M, editor. The Minnesota Symposia on Child Psychology. Vol. 16. Erlbaum; Hillsdale, NJ: 1983. pp. 41–83. [Google Scholar]
  45. Stevens SE, Sonuga-Barke EJS, Kreppner JM, Beckett C, Castle J, Colvert E, et al. Inattention/overactivity following early severe institutional deprivation: Presentation and associations in early adolescence. Journal of Abnormal Child Psychology. 2008;36:385–398. doi: 10.1007/s10802-007-9185-5. [DOI] [PubMed] [Google Scholar]
  46. U.S. Department of State [November 23, 2009];Total Adoptions to the United States. 2009 from http://adoption.state.gov/news/total_chart.html.
  47. UNICEF . Children on the brink 2002: A joint report on orphan estimates and program strategies. New York: 2002. [Google Scholar]
  48. UNICEF [November 22, 2009];Children without parental care. 2007 from http://www.unicef.org/protection/index_orphans.html.
  49. van den Boom DC. The influence of temperament and mothering on attachment and exploration: An experimental manipulation of sensitive responsiveness among lower-class mothers with irritable infants. Child Development. 1994;(65) doi: 10.1111/j.1467-8624.1994.tb00829.x. [DOI] [PubMed] [Google Scholar]
  50. van IJzendoorn MH. Adult attachment representations, parental responsiveness, and infant attachment: A meta-analysis on the predictive validity of the Adult Attachment Interview. Psychological Bulletin. 1995;117:387–403. doi: 10.1037/0033-2909.117.3.387. [DOI] [PubMed] [Google Scholar]
  51. Warren SL, Huston L, Egeland B, Sroufe LA. Child and adolescent anxiety disorders and early attachment. Journal of the American Academy of Child & Adolescent Psychiatry. 1997;36:637–644. doi: 10.1097/00004583-199705000-00014. [DOI] [PubMed] [Google Scholar]
  52. Zeanah CH, Egger HL, Smyke AT, Nelson CA, Fox NA, Marshall PJ, et al. Institutional rearing and psychiatric disorders in Romanian preschool children. American Journal of Psychiatry. 2009;166:777–785. doi: 10.1176/appi.ajp.2009.08091438. [DOI] [PubMed] [Google Scholar]
  53. Zeanah CH, Koga S, Simion B, Stanescu A, Tabacaru CL, Fox NA, et al. Ethical considerations in international research collaboration: The Bucharest Early Intervention Project. Infant Mental Health Journal. 2006;27:559–576. doi: 10.1002/imhj.20107. [DOI] [PubMed] [Google Scholar]
  54. Zeanah CH, Koga SF, Simion B, Stanescu A, Tabacaru CL, Fox NA, et al. Ethical considerations in international research collaboration: The Bucharest Early Intervention Project. Infant Mental Health Journal. 2006;27:559–576. doi: 10.1002/imhj.20107. [DOI] [PubMed] [Google Scholar]
  55. Zeanah CH, Nelson CB, Fox NA, Smyke AT, Marshall PJ, Parker SW, et al. Designing research to study the effects of institutionalization on brain and behavioral development: The Bucharest Early Intervention Project. Development and Psychopathology. 2003;15:885–907. doi: 10.1017/s0954579403000452. [DOI] [PubMed] [Google Scholar]
  56. Zeanah CH, Smyke AT, Koga SF, Carlson E, The Bucharest Early Intervention Project Core Group Attachment in institutionalized and community children in Romania. Child Development. 2005;76:1015–1028. doi: 10.1111/j.1467-8624.2005.00894.x. [DOI] [PubMed] [Google Scholar]

RESOURCES