Recent Advances in the Theories of and Interventions with Attachment Disorders

Louise Newman; Sarah Mares

Curr Opin Psychiatry. 2007;20(4):343-348. 

Abstract and Introduction

Purpose of review: The present review examines the available literature to consider recent advances in the theories of, and interventions with, disorders of attachment. We discuss the existing evidence to argue that the conceptualization of attachment disorders remains problematic despite their clinical significance.
Recent findings: Research into clinical disorders of attachment is limited, but there is some recent evidence that a standardized approach to diagnosis may improve identification. The current classificatory systems do not accommodate all clinical presentations and proposed broader approaches have greater clinical utility. A recent review of interventions for attachment disordered children confirms that 'coercive' interventions have no empirical support and are not biased on coherent principles of attachment theory.
Summary: Clinical interest in identification and treatment of attachment disorders has remained high. This is despite continuing confusion regarding terminology, phenomenology and diagnosis. The lack of agreed definition and conceptualization of these conditions is reflected in disparate approaches to intervention and limited empirical data. Interventions have little demonstrated benefit and some potential risk.

Disturbances of attachment behavior and social functioning in early childhood have been described in the clinical literature for over 50 years,[1,2] predominantly in groups of neglected and institutionalized children. In the last decade, these descriptions have included children in foster-care, orphanages and high-risk groups such as Romanian orphans.[3-5] Despite this, there remains considerable confusion regarding attachment disorders including terminology, classification and boundaries for these conditions. Current debates focus on the relationship between attachment classifications, particularly attachment disorganization, and clinical disorders of attachment behavior, and the accuracy and utility of current diagnostic criteria for attachment disorders. While there is an increasing focus on the association between attachment disorganization and developmental psychopathology, it is not clear to what extent attachment disorganization overlaps with, or is related to, current classifications of attachment disorders.[6,7]

Research into clinical disorders of attachment is limited by the lack of a universally accepted diagnostic protocol. There is, however, preliminary support for one proposed set of practice parameters.[8,9] The clinical significance of attachment disturbance is highlighted by the fact there has been a recent proliferation of interventions that are claimed to modify attachment disordered behaviors. Some of these approaches lack empirical support and may pose significant physical and psychological risks for children.[10]

Diagnosis and Classification

There is no accepted definition of the term 'attachment disorder'. Reactive attachment disorder (RAD) was first described in formal nosological systems in the Diagnostic and Statistical Manual of Mental Disorders (DSM)-III in 1980, but there remain few epidemiological or follow-up data about the disorder after early childhood. The disorder is currently defined[11] as 'markedly disturbed and developmentally inappropriate social relatedness in most contexts that begins before age 5 years and is associated with grossly pathological care'. The disturbance is not accounted for solely by developmental delay or pervasive developmental disorder. Very similar criteria are included in The International Statistical Classification of Disease and Related Health Problems-10,[12] which shares the focus on individual psychopathology, and Diagnostic Classification 0-3,[13] which includes description of the relational context of this disorder. There is consensus that RAD, as a narrowly defined condition, is rare and occurs in very high-risk populations, and general consensus that RAD describes distinctive symptom clusters of aberrant behaviors in young children that are not adequately captured by other diagnostic categories.[7,8] There is evidence that attachment disorders in young children can be reliably diagnosed by experienced clinicians in high risk populations.[9] Despite this, a number of significant practical and conceptual difficulties or deficiencies are identified with the current diagnostic criteria. Boris, Zeanah and the American Academy of Child and Adolescent Psychiatry taskforce[8] reviewed the literature on this issue from 1980 to 2003, and provide comprehensive recommendations on approaches to assessment and treatment of these disorders.

In summary, adequate assessment involves direct observation of the children interacting with their caregivers, using a structured observational paradigm and a detailed history, particularly of each child's early caregiving environment. Any current or past maltreatment must be reported. Comorbid conditions should be diagnosed and appropriate adjunctive treatments implemented. Ensuring the child is in a stable placement with an emotionally available caregiver is a prerequisite to intervention. They conclude that 'International involving noncontingent restraint or coercion have no empirical support and have been associated with serious harm'.

The DSM diagnosis of RAD distinguishes two distinct subtypes: inhibited, with persistent failure to initiate and respond to most social interactions in a developmentally appropriate way, and dis-inhibited, with indiscriminate sociability or lack of selectivity in the choice of attachment figures. Both subtypes are identified in children exposed to maltreatment and to institutionalization. A mixed picture is seen in a substantial minority of maltreated children and this cannot be accommodated within the current classificatory system.[8,14]

A number of authors have argued that symptoms of the inhibited behavior subtype overlap with behaviors associated with disorganized patterns of attachment in children with an identified attachment figure.[6,15] This behavior is noted to resolve in most previously institutionalized children once they are placed in adequate family circumstances. Indiscriminate behavior persists in a minority of children, including those who have a preferred attachment figure.[14,16] The different course of the two subtypes has led to suggestions that they may be better understood as two separate disorders. A detailed discussion of the subtypes is found in O'Connor and Zeanah,[17] and the potential contribution of temperament to the development of subtypes of RAD in children exposed to similar pathogenic environments is explored in Zeanah and Fox[18] and Heller et al.[19]

The current diagnostic focus on children with a lack of a discriminated attachment[20] figure, as opposed to a specific, but disturbed attachment relationship, appears too narrow to have broad clinical utility, particularly with children with disturbed attachment histories who are presented to clinical services. Zeanah et al.[21,22] have developed a broad typology of attachment disorders and disturbances, and propose a continuum that includes secure base distortions and disrupted attachments. This has not been taken up within current classificatory systems, although it is argued to have clinical utility,[23] accords with developmental understanding of attachment relationships,[7] and, in one study, had better interrater reliability than current DSM-IV.[9] Current diagnostic systems have only limited capacity to address comorbidity and differential diagnosis, especially in older children.[10]

Crittenden[20] has developed the Dynamic Maturational Model of Attachment as an alternate approach to attachment across the lifespan. The model integrates an attachment based approach to formulation of behavioral and psychiatric disorders with an emphasis on developmental maturation of information processing strategies. The model proposes that an increasing range of attachment-related behaviors, adaptive and psychopathological, becomes possible as individuals mature. Crittenden[20] differentiates a number of atypical attachment strategies as opposed to 'disorganization', but does not specifically address limitations or confusion in the current classification of attachment disorders.

The strange situation procedure, which is the generally accepted standardized assessment of attachment quality in developmental and population studies of infants and young children, is of limited use clinically and there is no agreed method as yet for assessing attachment status of school age children, although some have been proposed.[20] Boris et al.[8,9] suggest and provide support for the usefulness of a structured interview and observations process, at least in younger children. Stafford and Zeanah[24] recommend that assessment include 'detailed observations in naturalistic and clinical settings as well as obtaining multiple focused reports from caregivers'.

Neglect and maltreatment have developmental and behavioral consequences,.[16,25] Therefore, children who fit the diagnosis of RAD often have comorbid conditions. It is argued that the symptoms and signs of attachment disorder symptoms are distinguishable from symptoms and signs that are attributable to other disorders.[9] Accurate diagnosis of comorbid conditions is necessary to ensure that children with multiple difficulties receive appropriate, safe and evidence-based treatments.[7,8,10] Autism, for example, leads to abnormal social relatedness and must be excluded before RAD is diagnosed. The symptoms of the inhibited subtype of RAD overlap with the hyperarousal, numbing and emotional withdrawal seen in traumatized preschool children, and this means that posttraumatic stress disorder and other anxiety disorders should also be adequately diagnosed and treated, if present.[8]

In older children, most diagnostic confusion occurs in relation to the controlling, aggressive and delinquent behaviors, and peer-related social difficulties associated with oppositional defiant disorder and conduct disorder. These symptoms and behaviors also commonly occur in children exposed to abuse, particularly violence, and in children who have demonstrated disorganized attachment patterns to caregivers in early childhood. Aggression is a common comorbid symptom in children with histories of maltreatment or institutionalization,[25,26] but should not be seen as an element or symptom of RAD per se.[7]

Green and Goldwyn[15] and Green[27] postulate that attachment disorders are better understood as 'disorders of current social impairment' with 'patterns of comorbidity' identified above, and Minnis et al.[28] propose a reconceptualization of RAD as a primary disorder of intersubjectivity. Intersubjectivity refers to the communicative processes which shape infants' brain development and development of self-consciousness and regulation. Disrupted communications between infants and their caretakers impact on the emerging capacity for attachment and complex social functioning.[15,27] This hypothesis has become prominent in the neurodevelopmental literature.[29] Minnis et al.[28] argue cogently that attachment disordered behavior, particularly indiscriminate behavior, may be best understood as a consequence of disrupted intersubjectivity in children who have been exposed to significant emotional neglect.

Recent Studies on Reactive Attachment Disorder

The course of RAD is not well documented beyond early childhood. Romanian orphans are the most studied group of attachment disordered children.[5,16,25,30,31] It is unclear to what extent this research can be generalized to maltreated children who have had a small number of consistent, albeit inadequate and/or abusive caregivers. We found only one published longitudinal case report of maltreated noninstitutionalized children.[19] A history of maltreatment does not imply disorder per se because resilient children may show mild and clinically nonsignificant problems.[10,16] The duration and severity of early deprivation have a prognostic impact on later functioning,[5,16,25] and there is evidence that children removed from institutional care before 6 months of age do better than those in care for longer.[16,25]

The absence of a consistently available attachment figure in early childhood appeared as the greatest etiological link with attachment disorders as currently defined, and indiscriminate behavior was the most common and persistent symptom. Rutter et al.[16] suggest the possibility of a sensitive period in relation to the development of disinhibited attachment behaviors and, more recently, this is supported by Zeanah[32] who reports a possible critical period for the development of disinhibited behaviors which were more likely to persist in children who remained in institutional care past late in the second year of life. It is likely there is a critical period for the development of neurological pathways underlying the organization of the attachment system and that early care experiences are central to this development. If empirical research confirms the existence of such a sensitive period, it is imperative to facilitate early (by 2 years) placement of children with alternate attachment figures.

Attachment Disorganization and Attachment Disorder

The disorganized category of attachment emerged in the attachment theory literature to describe a group of infants who demonstrated a lack of an organized strategy to deal with stress and emotional regulation.[33] Many of the indices of disorganized behavior, including contradictory and incomplete behaviors, are thought to reflect a state of 'fright without solution'.[34,35] Pathogenic parenting, particularly abuse and trauma, is associated with infant disorganization,[36] with evidence of ongoing difficulties in emotional regulation, behavioral disorders, and aggression and dissociative symptomatology.[37] Although these experiences may clearly be associated with difficulties of social and interpersonal functioning, the concepts of attachment disorganization and attachment disorder have developed largely independently, and are not necessarily equivalent.[6,21,22]

There are difficulties in defining the core characteristics of 'attachment disorders' as opposed to 'disorders of nonattachment' and of defining the severity of symptoms that constitute disorder. Evidence for the association between attachment disorganization and psychopathology has grown, and a series of papers has documented the association between aberrant parenting behavior and disorganized attachment (attachment status, socialization and externalizing behaviors, aimed at pediatricians)[38] with more recent attempts to describe subtypes of interactive disturbance.[39]

Children with both RAD and disorganization can exhibit simultaneous approach avoidance and contradictory behaviors, and disorganized/controlling behaviors may cooccur with RAD. There is, as yet, little overlap between the frames of reference of attachment theory and the clinical approach to disorders of attachment that might help to clarify this matter. The significance of conflicted behavior may relate to its association with stress and neurobiological development with preliminary findings of differential effects of maternal interactional style on infant brain function.[40]

Interventions for Children with Attachment Disorders

Considerable controversy exists in the literature about assessment of and intervention with children with RADs as defined in DSM-IV, and with children with what are known clinically as 'attachment disordered behaviors'. This latter term is often used very broadly to include older children with a wide range of social and behavioral difficulties, including controlling and conduct disordered symptoms.[7,10]

The diagnostic and conceptual difficulties in relation to attachment disorders are due in part to a convergence between the developmental and clinical attachment literatures.[7,10,22] Most of the literature on attachment theory concerns developmental rather than pathological or clinical populations, and yet ideas and concepts from attachment theory are regularly used to inform work with clinical populations from infancy to adulthood.[41-43]

A range of attachment-based interventions for parents with infants and young children has been found to be effective in altering parental sensitivity and behavioral symptoms in infants and young children,[44] but the relevance of these approaches to intervention with fostered and adopted children with RAD or older children with significant histories of maltreatment is unclear.

Multiple interventions, some justifiable theoretically and clinically, others with little empirical basis, have been proposed for children with reactive attachment disorder. Given the extent of comorbidity in this population, multifaceted interventions are inevitable and necessary.[7,8,10]

A number of so called 'Attachment Therapies' have gained credibility particularly with parents and clinicians involved with fostering and adoption of maltreated or previously institutionalized children. The Report of the American Professional Society on the Abuse of Children Task Force on Attachment Therapy, RAD and Attachment Problems[10] provides a comprehensive summary of and response to the debate in relation to 'controversial theories and corresponding controversial treatments' or what Zilberstein[7] calls 'egregious misapplication of attachment theory in relation to so called Attachment Therapies'.

Proponents offer alternate conceptualizations of attachment problems found in fostered and adoptive children, the most controversial of which include an etiological assumption that children exposed to significant early neglect or maltreatment have failed to develop the capacity to attach to others and seek control rather than closeness in their relationships. These authors describe the children as manipulative and predatory, with dire prognostications for psychopathy. There is a central focus on aggressive and conduct disordered symptomatology, and a range of coercive and sometimes punitive interventions have been advocated, including so-called 'holding therapies' for what might more appropriately be understood as comorbid externalizing behaviors. The Attachment Task Force Report[10] suggests that overdiagnosis of RAD is likely and that 'it should not be assumed that RAD underlies all or even most of the behavioral and emotional problems seen in foster children, adoptive children, or children who are maltreated'. Most aggressive children retain the ability to form attachments, and the aggressive and conduct disorders in these children and adolescents can be best understood as resulting from effects of abuse rather than as constituting evidence of primary attachment disorders.[7,45]

The American Professional Society on the Abuse of Children Task Force Recommendations conclude that the 'controversial therapies', including holding, restraint and other coercive strategies, are seen to represent a significant physical and psychological risk to children with little evidence of therapeutic benefit. It is argued that coercive interventions are the antithesis of sensitive responsiveness necessary to promote attachment.[7,10]

Several authors have outlined the principles necessary to inform interventions for children with RAD.[8,24,46] These include a structured assessment/observation paradigm; identification and report of any current maltreatment; advocating for a safe and stable placement with an appropriate attachment figure, and assessment of the caregivers in relation to treatment selection; creating opportunities for positive interactions with caregivers; assessment, referral and adjunctive treatment for any comorbid problems; and cautious use of pharmacotherapy especially in preschool age and younger children. They (along with Chaffin et al.[10]) conclude that noncontingent responses and coercive therapies have no empirical support and are associated with harm.

A number of comprehensive approaches to intervention with these children are in the process of evaluation.[30,47-49] These programmes address 'ecological issues',[23] including the stability of the current care-giving environment, attachment disordered symptoms in the children, the nature and quality of the caregiver and child interactions, and assessment and treatment of comorbid disorders. Zeanah and Smyke[49] outline a foster care programme that is currently being implemented and evaluated with maltreated young children in New Orleans and the Bucharest Early Intervention Project with Romanian Orphans.


While RAD has validity, there is a requirement to further refine classification of the broad spectrum of attachment disturbances seen in clinical practice. Clinical experience suggests that disorders of attachment exist beyond the confines of RAD as currently defined and modifications in diagnostic criteria are necessary to reflect this finding. Further exploration of possible underlying mechanisms, including disturbances of intersubjectivity, social cognition, attachment disorganization, and the neurodevelopment effects of abuse and trauma, should guide ongoing research. Further studies that compare symptom patterns in children exposed to deprivation with children exposed to trauma and abuse are needed to facilitate diagnostic and etiological clarity, and to integrate developmental and clinical approaches to disturbed attachment. This will enable better identification of children in need, and evaluation of the appropriateness and effectiveness of a range of interventions.


  1. Freud A, Burlingham D. The writings of Anna Freud: volume III: infants without families 1939-1945. Madison: International Universities Press; 1973.

  2. Spitz R. Hospitalism: an inquiry into the genesis of psychiatric conditions in early childhood. Psychoanal Study Child 1945; 1:53-74.

  3. Tizard B, Rees J. The effect of early institutional rearing on the behaviour problems of 4 year old children. J Child Psychiatry Psychol 1975; 16:61-73.

  4. Tizard B, Hodges J. The effect of early institutional rearing on the development of eight year old children. J Child Psychiatry Psychol 1978; 19:99-118.

  5. Chilshom K, Carter M, Amec E, Morison X. Attachment security and indiscriminately friendly behavior in children adopted from Romanian orphanages. Dev Psychopathol 1995; 7:283-294.

  6. Van Ijzendoorn M, Bakermans-Kranenburg M. Attachment disorders and disorganized attachment; similar and different. Attach Hum Dev 2003; 5:313-320.

  7. Zilberstein K. Clarifying core characteristics of attachment disorders: a review of current research and theory. Am J Orthopsychiatry 2006; 76:55-64.

  8. Boris N, Zeanah C. Practice parameters for the assessment and treatment of children and adolescents with reactive attachment disorder of infancy and early childhood. J Am Acad Child Adolesc Psychiatry 2005; 44:1206-1219.

  9. Boris N, Hinshaw-Fuselier S, Smyke A, et al. Comparing criteria for attachment disorders: establishing reliability and validity in high-risk samples. J Am Acad Child Adolesc Psychiatry 2004; 43:568-577.

  10. Chaffin M, Hanson R, Saunders B, et al. Report of the APSAC task force on attachment therapy, reactive attachment disorder, and attachment problems. Child Maltreat 2006; 11:76-89.

  11. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th edition, text revision 2000. Washington: American Psychiatric Association; 2004.

  12. World Health Organization. The international statistical classification of disease and related health problems. 10th review. Geneva: WHO; 1992.

  13. Zero to Three Centre for Infants and Toddlers and Families. Diagnostic classification of mental health and developmental disorders of infancy and early childhood (revised). Arlington: Zero to Three Centre for Infants and Toddlers and Families; 2005.

  14. Zeanah C, Scheeringa M, Boris N, et al. Reactive attachment disorder in maltreated toddlers. Child Abuse Neglect 2004; 28:877-888.

  15. Green J, Goldwyn R. Annotation: attachment disorganization and psychopathology: new findings in attachment research and their potential implications for developmental psychopathology in childhood. J Child Psychol Psychiatry 2002; 43:83-847.

  16. Rutter M, O'Connor T, and the ERA Study Team. Are there biological programming effects for psychological development? Findings from a study of Romanian adoptees. Dev Psychol 2004; 40:81-94.

  17. O'Connor T, Zeanah C. Attachment disorder: assessment strategies and treatment approaches. Attach Hum Dev 2003; 5:223-244.

  18. Zeanah C, Fox N. Temperament and attachment disorders. J Clin Child Adolesc Psychol 2004; 33:34-41.

  19. Heller S, Boris N, Fuselier S, et al. Reactive attachment disorder in maltreated twins follow-up: from 18 months to 8 years. Attach Hum Dev 2006; 8:63-86.

  20. Crittenden PM. The dynamic maturational model of attachment. Aust N Z J Family Ther 2006; 27:105-115. Provides an overview of an attachment-based approach to formulation of behavioral and psychiatric disorder.

  21. Lieberman A, Zeanah C. Disorders of attachment in infancy. Child Adolesc Psychiatr Clin North Am 1995; 4:571-587.

  22. Boris N, Zeanah C. Disturbances and disorders of attachment in infancy: an overview. J Infant Ment Health 1999; 20:1-9.

  23. Minde K. Attachment problems as a spectrum disorder: implications for diagnosis and treatment. Attach Hum Dev 2003; 5:289-296.

  24. Stafford B, Zeanah C. Attachment disorders. In: Luby JL, editor. Handbook of preschool mental health: development, disorders and treatment. New York: Guildford; 2006. pp. 231-251. This useful chapter provides a developmental perspective on attachment, reviews the phenomenology of RAD, and provides guidelines for assessment and treatment.

  25. Johnson R, Browne K, Hamilton-Giachritsis C. Young children in institutional care at risk of harm. Trauma Violence Abuse 2006; 7:34-60.

  26. Minnis H, Keck G. A clinical/research and dialogue on reactive attachment disorder. Attach Hum Dev 2003; 5:297-302.

  27. Green J. Are attachment disorders best seen as social impairment syndromes? Attach Hum Dev 2003; 5:259-264.

  28. Minnis H, Marwick H, Arthur J, McLaughlin A. Reactive attachment disorder - a theoretical model beyond attachment. Eur Child Adolesc Psychiatry 2006; 15:336-342.

  29. Schore AN. Effects of a secure attachment relationship on right brain development affect regulation and infant mental health. Infant Ment Health J 2001; 22:7-66.

  30. Zeanah C, Smyke A, Dumitrescu A. Attachment disturbances in young children; indiscriminate behavior and institutional care. J Am Acad Child Adolesc Psychiatry 2002; 41:983-989.

  31. Weitzman C, Albers L. Long-term developmental, behavioral and attachment outcomes after international adoption. Pediatr Clin North Am 2005; 52:1395-1419.

  32. Zeanah C. Preventing adverse outcomes in abandoned children: the Bucharest early intervention project. Irving Phillips Prevention Award Address. Annual Meeting of the American Academy of Child and Adolescent Psychiatry. San Diego; 2006.

  33. Main M, Solomon J. Discovery of a new insecure-disorganized/disoriented attachment pattern. In: Yogmon M, Brazelton TB, editors. Affective development in infancy. Norwood: Ablex; 1986. pp. 95-124.

  34. Main M, Solomon J. Procedures for identifying infants as disorganized/disoriented during the Ainsworth Strange Situation. In: Greenberg MT, Cicchettio D, Cummings EM, editors. Attachment in the preschool years: theory, research and intervention. Chicago: University of Chicago Press; 1990. pp. 121-160.

  35. Hess E, Main M. Disorganized infant, child and adult attachment: collapse in the behavioral and attentional strategies. J Am Psychoanal Assoc 2000; 48:1109-1127.

  36. Lyons-Ruth K, Yellin C, Helnick S, Atwood G. Expanding the concept of unresolved mental states: Hostile/Helpless states of mind on the Adult Attachment Interview are associated with disrupted mother-infant communication and infant disorganization. Dev Psychopathol 2005; 17:1-23.

  37. Lyons-Ruth K, Dutra L, Schuder M, Bianchi I. From infant attachment disorganization to adult dissociation: relational adaptations or traumatic experiences? Psychiatr Clin North Am 2006; 29:63-86.

  38. Rees C. Thinking about children's attachments. Arch Dis Childhood 2005; 90:1058-1065.

  39. Lyons-Ruth K, Madigan S, Muran G, Pederson D. Unresolved states of mind, attachment relationships and disrupted interactions of adolescent mothers and their infants. Dev Psychol 2006; 42:293-304. Description of the impact of maternal state of mind on quality of interaction.

  40. Diego M, Filed T, Jones N, Hernandez-Reif M. Withdrawn and intrusive maternal interaction style and infant frontal EEG asymmetry shifts in infants of depressed and nondepressed mothers. Infant Behav Dev 2006; 29:220-229.

  41. Berlin L, Ziv Y, Amaya-Jackson L, Greenberg L, editors. Enhancing early attachments: theory, research, intervention and policy. New York: Guildford Press; 2005.

  42. Holmes J. The search for the secure base: attachment theory and psychotherapy. New York: Brunner Routledge; 2001.

  43. Fonagy P. Attachment theory and psychoanalysis. New York: Other Press; 2001.

  44. Velderman M, Bakermans-Kranenburg M, Jutter F, Van Ijzendoorn M. Effects of attachment based interventions on maternal sensitivity and infant attachment. Differential susceptibility of highly reactive infants. J Family Psychol 2006; 20:266-274.

  45. Gutman-Steinmetz S, Crowell JA. Attachment and externalizing disorders; a developmental psychopathology perspective. J Am Acad Child Adolesc Psychiatry 2006; 45:440-451.

  46. Hughes DA. Psychological interventions for the spectrum of attachment disorders and intrafamilial trauma. Attach Hum Dev 2003; 5:271-278.

  47. Marvin R, Cooper G, Hoffman K, Powell B. The Circle of Security project: attachment based intervention with caregiver preschool dyads. Attach Hum Dev 2002; 4:107-124.

  48. Marvin R, Whelan W. Disordered attachments: towards evidence based clinical practice. Attach Hum Dev 2003; 5:284-299.

  49. Zeanah CH, Smyke AT. Building attachment relationships following maltreatment and severe deprivation. In: Berlin L, Ziv Y, Amaya-Jackson L, Greenberg L, editors. Enhancing early attachments: theory, research, intervention and policy. New York: Guildford Press; 2005. pp. 195-216.