Clinical Psychology Review 23 (2003) 35 – 56

The relevance of attachment theory to the philosophy,

organization, and practice of adult mental health care
Isabel Goodwin*
University of Birmingham, Edgbaston, Birmingham, UK

Received 11 July 2001; received in revised form 7 March 2002; accepted 2 May 2002

Abstract

This review is an inquiry into the relevance of attachment theory to the current philosophy,
organization, and practice of adult mental health care, via an examination of the literature relating to
attachment theory and, in particular, the literature relating to research into adult attachment. The review
does not seek to critique attachment theory itself, but considers carefully the relevance of the theory to
adulthood and to the field of adult mental health. In so doing, research into individual difficulties is
examined, as is the provision and delivery of therapeutic services. In addition, the literature regarding
the importance of mental health staff’s own attachments and the influence of attachment theory on
mental health service philosophy and organization are evaluated. Finally, potential areas for future
research and development in this field are suggested.
D 2003 Published by Elsevier Science Ltd.

Keywords: Attachment; Mental health

1. Introduction

Attachment theory was developed by psychiatrist and psychoanalyst John Bowlby during
the 1950s and 1960s. Bowlby began to formulate the main outlines of the theory between
1958 and 1963, and the hugely influential Attachment and Loss Volume 1: Attachment was
published in 1969. With the attachment theory, Bowlby sought to bring together psychoana-

* Field House, 1 Myddlewood, Myddle, Shrewsbury, SY4 3RY, UK. Tel.: +44-1-939-29-2109; fax: +44-1-
939-29-2025.
E-mail address: irg861@bham.ac.uk (I. Goodwin).

0272-7358/03/$ – see front matter D 2003 Published by Elsevier Science Ltd.

PII: S 0 2 7 2 - 7 3 5 8 ( 0 2 ) 0 0 1 4 5 - 9
36 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56

lytic concepts with those of ethology, to provide an explanation of the nature of the
fundamental affectional bond linking one human being to another (Holmes, 1993).
The original focus of attachment theorists was the mother–child relationship, but in recent
years, there has been a massive surge of both clinical and research interest in the field of adult
attachment (that is, adult–adult relationships). The field of adult mental health has proved of
immense interest to researchers exploring the causes and effects of insecure attachment
relationships, in terms of individual mental health problems, the role and purpose of
psychotherapeutic services, and—to a much lesser extent—the organization of mental health
care services. This review explores the relevance of attachment theory to adulthood and to the
field of adult mental health via a critical examination of the literature ranging from individual
difficulties and the provision and delivery of therapeutic services, to the importance of mental
health staff’s own attachments and the influence of attachment theory on mental health care
philosophy and organization.

2. Attachment theory—a resumé

Attachment behavior is conceived as any form of behavior that results in a person attaining
or retaining proximity to a differentiated and preferred individual (Bowlby, 1980). Put more
simply, attachment behavior involves individuals trying to get close to a person or persons
they feel most comfortable with. According to Bowlby (1980), attachment behavior has its
own dynamic and is distinct from both feeding and sexual behavior and of at least an equal
significance in human life. Thus, during the course of healthy development, attachment
behavior leads to the development of affectional bonds (or attachments) between child and
parent. In addition, Bowlby also proposed that attachment relationships were present and
active throughout the life cycle, so that attachment bonds would later develop between adult
and adult. A behavior complementary to attachment behavior, and serving a complementary
function, is caregiving. This is usually shown by a parent towards a child, but may also be
shown by one adult towards another (Bowlby, 1980). Attachment behavior is activated only
under certain conditions, in children particularly conditions involving strangeness, tiredness,
anything frightening, or unavailability or unresponsiveness of the attachment figure. In adults,
attachment behavior is more commonly activated in times of illness, stress, or old age.
Attachment behavior is assuaged by the proximity and responsiveness of attachment figures,
their comfort, and caregiving (Bowlby, 1979).
Weiss (1991) suggests that relationships can only be defined as attachments if they display a
number of key features, including elicitation of the behavior by threat, proximity-seeking, and
use of attachments as a secure base from which to explore the world. The theory also asserts
that it is through internal working models (developed via early social and parental experi-
ences) that childhood patterns of attachment are carried through into adult life, influencing the
quality of later personal relationships (Feeney & Noller, 1996; Holmes, 1993). Bowlby’s
(1980) theory proposes that attachment behavior functions as a kind of homeostatic
mechanism for modulating anxiety: increasing anxiety increases attachment behavior; thus,
the goal of attachment behavior can be seen as helping the individual to modulate their anxiety
 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56 37

and arousal (Adshead, 1998). In doing so, a preferred attachment figure is best but, failing this,
it has been argued, any attachment figure will do. That is, an unhelpful (or even damaging)
figure is likely to be seen as better than none at all (Adshead, 1998).
The further development of attachment theory by Bowlby and others has resulted in
identification of secure and insecure attachment patterns in children (Ainsworth, Blehar,
Waters, & Wall, 1978) and secure and insecure attachment styles in adults (Main & Goldwyn,
1989). Ainsworth’s ‘Strange Situation’ consists of a 20-min session involving a 12-month-old
child, the child’s mother, and an experimenter, which focuses on the response of the child to
separation and reunion, and aims to elicit individual differences in coping with the stress of
separation. From this, Ainsworth et al. (1978) identified initially three, and later four, major
patterns of response. Secure attachment is identified in children who are usually distressed by
separation, but on reunion greet their mother, allow themselves to be comforted if required,
and then return to continue their play. In Ainsworth et al.’s original sample, 66% of children
were classified as securely attached in this way. Three patterns of insecure attachment have
been identified, including avoidant, ambivalent, and disorganized attachment. Avoidant
attachment is identified when the child shows few signs of distress on separation, ignores
his/her mother on reunion, but remains watchful of her and inhibited in his or her play.
Ainsworth et al. found that 20% of their sample displayed this avoidant attachment pattern.
Ambivalent attachment is characterized by high distress on separation and an inability to be
pacified on reunion. Contact is sought with the mother, but then resisted and rejected, and the
child alternates between anger and clinging, and engages in little exploratory play. Twelve
percent of Ainsworth et al.’s sample was identified as ambivalent in their attachment pattern.
Disorganized attachment is characterized by a diverse range of confused behaviors including
‘‘freezing,’’ or stereotyped movements, on reunion with the parent. In a meta-analytic study
of the Strange Situation published in 1988, Ijzendoorn and Kroonenberg (1988) confirmed
the acceptance of the procedure and the attachment patterns identified as reliable and valid,
quoting extensive use and replication in over 30 different studies. A literature search from
1988 onwards, indicates that the Strange Situation continues to be widely utilized and cited in
attachment research.
Attachment theory’s insistence on the importance of the mother–child relationship over
and above all other conditions of childhood experience has been criticized on a number of
counts, particularly that it has excluded fathers, confined women with children to the
mothering role (referred to as the motherhood mandate), and has failed to acknowledge the
role of the child as an active agent in his or her social environment (Quiery, 1998). In
addition, Quiery (1998) criticizes Bowlby’s heavy reliance on animal research to draw
parallels with human mother–infant bonding, as it regards the bonding or attachment as an
imprinting event rather than an emotional process. A further criticism notes that although
much of the early evidence for the effect of maternal deprivation was drawn specifically from
children separated from their parents as a result of traumatic wartime experiences in Europe
during World War II, the conclusions drawn from this work have been transposed onto almost
any childhood experience of separation (Tizard, 1991).
Main, Kaplan, and Cassidy (1985) devised the Adult Attachment Interview (AAI) as a tool
for assessing the internal working models of adults with respect to attachment. The AAI is a
38 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56

semistructured interview conducted in a similar way to a psychotherapy assessment, focusing

on the individual’s childhood and relationships with each of their parents. The interviews are
subsequently rated to provide a classification of ‘state of mind’ with respect to attachment, or
attachment style. The four styles identified by Main et al. (1985) are autonomous–secure,
dismissing–detached, preoccupied–entangled, and unresolved–disorganized. These styles
may be considered to correspond to Ainsworth et al.’s (1978) childhood patterns of secure,
avoidant, ambivalent, and disorganized, respectively. A number of independent studies have
shown consistently high correlations of between 70% and 80% between the attachment status
of infants in the Strange Situation and that of their mothers in the AAI (including Fonagy,
Steele, Steele, Moran, & Higgins, 1991; Grossman & Grossman, 1991; Main & Goldwyn,
1984). Other models proposed, for instance, that of Bartholomew (1990), have sought to
expand the basic model of adult attachment patterns. Although Bartholomew’s model also
provides four categories (of secure, preoccupied, dismissing, and fearful), it situates them
within a framework relating to both the individual’s model of self and model of other
(Bartholomew, 1993). Bartholomew also points out that the patterns within the different
models do not correspond completely but do, in fact, conceptually overlap with each other.
The similarity in category names across models therefore has the potential to be misleading,
and the finer descriptive detail of any given classification requires careful attention.

3. The relevance of attachment theory to adulthood

Attachment theory has produced clear advances in our understanding of affective life and
relationships in childhood, but how successfully can it be utilized in furthering our
understanding of adult relationships? Can the theory be directly transferred from a parent–
child to an adult–adult context and, if so, should we reasonably expect to assess the latter in
similar ways to the former?
Despite the massive expansion of research into adult attachment in recent years, some
concerns have been expressed that attachment theory is essentially a child-centered theory
which is not readily adaptable to adult relationships. Birtchnell (1997) has argued that
Bowlby’s drawing of parallels between the attachment behavior of animals and that of
children (in his development of the theory) is incomplete, in that—for most animals—
survival into and through adulthood depends not only on a period of attachment but,
ultimately, upon ‘an early and successful separation from the parent’ (p.267). However,
Bowlby’s (1973) perspective on separation—as propounded in his book of the same name—
focused almost entirely on the different pathologies resulting from disruptions in the
attachment process. Birtchnell argues that the concept of attachment is essentially one of a
form of relating and, as such, it fits neatly into interpersonal theory (Freedman, Leary,
Ossorio, & Coffee, 1951), which preceded attachment theory by at least 10 years. The
essential feature of interpersonal theory, according to Birtchnell, is the proposition that
‘relating takes place along two orthogonal axes: a horizontal one, which represents some such
dichotomy as warm versus cold, loving versus hating, or friendly versus hostile; and a vertical
one, which represents some such dichotomy as active versus passive, powerful versus weak,
 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56 39

or dominant versus submitting’ (p. 265). He proposes that dependence is the adult equivalent
of attachment, and sees it not as a purely negative concept (as Bowlby suggested) but as
including both poles of a positive–negative continuum (Birtchnell, 1997): thus, negative
dependence is the (adult) equivalent of Bowlby’s insecure attachment, and positive depend-
ence the equivalent of secure attachment.
Birtchnell’s critique of attachment theory as applied to adults also includes Bowlby’s
failure to acknowledge that, just as attachment confers security, so detachment (separation)
may also confer security: he identifies ‘secure distance’ as feeling safely separate from
others (Birtchnell, 1997). In addition, he notes that desperation does not feature in
attachment theory, which he concludes is likely to be because it is more relevant to adults
than children, though he argues that it is highly relevant to understanding an adult
individual’s need to attain a particular state of relatedness to another and their subsequent
behavior (Birtchnell, 1997).
Birtchnell’s critique of attachment theory as not readily adaptable to adult relationships
appears logical, taking, as it does, a similar starting point to Bowlby, but taking the parallel
between animal and human behavior further. His emphasis on the interpersonal nature of
attachment theory is also an important one, and one which is being increasingly attended to
by contemporary researchers (e.g., Cook, 2000). However, it could also be argued that many
highly social animals do not seek a separation from the parental generation, but continue to
relate closely for the continued success of the social group. That man is also a social animal is
rarely disputed, and therefore it may be in his interest to retain such relationships—that is, to
remain attached—rather than to separate.
A fundamental tenet of attachment theory is that the attachment style developed in the
child–parent relationship influences future relationships (Bowlby, 1973). This influence is
believed to be mediated by the person’s ‘internal working model’ of the attachment
relationship (Main et al., 1985), which consists of internalized expectancies of an attachment
figure’s responses to oneself. According to Bowlby (1973), by late adolescence, early patterns
of interaction with attachment figures have become organized into generalized interactional
styles that are driven by the person’s internal working model. Thus, many self-report
measures concerning adult attachment styles are designed to measure people’s expectations
about close relationships in general, rather than in relation to specific others (Cook, 2000). In
contrast to this, studies of infant attachments tend to consider specific attachments (usually to
either the mother or father), and there is little in the literature to confirm that such a process of
generalization occurs. Indeed, it is becoming increasingly documented that the quality of an
infant’s relationships to its mother and father, as assessed in the Strange Situation, can be
markedly different from one another (Bretherton, 1985).
A criticism of contemporary research on adult attachment relationships linked to this is that
by focusing on internal working models of relationships, the cognitive rather than the
interpersonal sources of attachment security are emphasized (Cook, 2000). This debate about
the nature of adult attachment—the internal ‘individual differences’ or the interpersonal
aspects of relating—is one which preoccupies an increasing number of researchers (e.g.,
Banai, Weller, & Mikulincer, 1998; Birtchnell, 1997; Cook, 2000). The importance of the
consequences of such theoretical ambiguity in the conceptualization of adult attachment
40 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56

relationships is identified by Cook (2000), who notes that the nature of attachment relation-
ships determines the level of analysis appropriate for research in this area. Cook, in a
methodologically complex, but rigorous, study, tested three hypotheses about the interper-
sonal sources of adult attachment security, and concluded that greater emphasis on the
interpersonal sources of adult attachment security is warranted as internal working models of
relationships may not be so ‘internal’ after all, but rather, may be considerably dependent on
social processes.
Crowell and Treboux (1995) identify the reciprocal nature of adult–adult relationships as a
major difference between adult–adult attachment and the parent–child relationship: that is,
adult partners are not (usually) assigned to the different roles of ‘attachment figure’ or
‘caregiver’, and ‘attached individual’ or ‘care receiver’. Both attachment behavior and
serving as an attachment figure will commonly be observed in both adult partners, and often
shift between the two (Crowell & Treboux, 1995). Other differences that have been identified
are that attachment relationships between adults often serve many other functions, including
sexual bonds, companionship, sense of competence, and shared purpose or experience
(Ainsworth, 1985; Weiss, 1974).
Although, in much of both the theoretical and research literature, adult attachment appears
to be agreed as an extension of attachment in childhood, it is interesting that researchers have
not tended to utilize the same methods of assessment when investigating adult attachment
behaviors. The classic tradition in the field of childhood attachment research is of behavioral
observation in either naturalistic or laboratory situations (e.g., Ainsworth et al., 1978). In
adult attachment research, behavioral observations appear rare, with the emphasis instead on
assessments which utilize language and perceptions, usually via interviews and self-report
measures (e.g., Hazan & Shaver, 1987; Main et al., 1985). This may be a reflection of the
dominant forms of communication and expression in small children as opposed to adults (that
is, action as opposed to language), but although it may be inappropriate to attempt to assess
childhood attachments via language-based tools, it may not be inappropriate to assess the
attachment system in adult relationships via behavioral observation. However, it remains the
case that studies exploring the similarities between child and adult attachment tend not to
utilize comparable methodologies. For example, Johnston (1999), despite examining adults’
patterns of exploration that were operationalized in a manner similar to that used in studies of
earlier developmental periods (that is, in terms of childhood novelty-seeking, curiosity, and
impulsivity), used a specially designed self-report measure having little similarity to the
methodology used in the original childhood study.
One of the most prolific areas of research in adult attachment has focused on intimate,
romantic adult relationships, including marriage. Shaver and Hazan (1988), in an oft-cited
study, looked at romantic attachments and found very marked similarities between those and
the recognized infant attachment patterns identified by Ainsworth et al. (1978): 56% of
respondents demonstrated a secure attachment pattern, 25% showed an avoidant pattern, and
19% were anxious–ambivalent in their attachment relationship. Such findings have served as
great encouragement for those researchers in the area whose instincts coincide with Bowlby’s,
that attachment relationships continue through into adulthood with relative consistency.
However, there is research indicating that attachment style does not remain stable over time,
 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56 41

with reports of 30% of individuals changing their attachment style over a 4-year period
(Kirkpatrick & Hazan, 1994) and 40% changing over an 8-month period (Scharfe &
Bartholomew, 1994). More recently, Davila, Burge, and Hammen (1997) have found at-
tachment style change ranging from 28% (over 6 months) to 34% (over 2 years), findings
which are generally consistent with previous studies. Davila et al.’s (1997) findings largely
supported an individual difference conceptualization (rather than instability being a reaction
to current circumstances). That is, some individuals are more prone to attachment style
changes than others. Davila et al. conclude that this may be a manifestation of incoherent
working models which may be only tentatively held and therefore are more liable to change
over time. Although this study has a good sample size (155 female participants), Davila et al.
used only Hazan and Shaver’s (1987) single-item, self-report attachment measure relating to
romantic relationships. They comment that, at initial interview, 19% of the participants—all
of who were aged between 17 and 19 years—had never had a romantic relationship. This
casts significant doubt as to the validity of using such a measure with the particular
population under study.
Feeney (1996) tested the proposition that attachment and caregiving are central, inter-
related components of adult love relationships and found evidence of partner-matching in
terms of these dimensions in married couples. This supports previous findings from
longitudinal research by Kirkpatrick and Davis (1994), who found that—for both men and
women—attachment styles contributed significantly to the prediction of relationship stability
over 3 years. Feeney also found that secure attachment was associated with beneficial
caregiving to the spouse and that marital satisfaction was higher for securely attached spouses
and for those whose partners reported more beneficial caregiving. Such research supports the
relevance of attachment theory to close adult relationships.
In summary, despite some concerns that attachment theory is not readily adaptable to the
relating of adults, it remains widely agreed within the available literature that the conceptual
basis of attachment theory is one of relationships that develop in childhood and extend
through adulthood with relative (if not complete) consistency. Furthermore, it has been
suggested that attachment theory provides a useful framework for understanding adult
relationships independent of any claims of continuity between childhood and adult relation-
ship patterns, and that the qualitatively different patterns of attachment identified in childhood
may be parallel to patterns that characterize adults (Bartholomew, 1993). Longitudinal
research may offer further valuable insights into the developmental pathway of attachments
in the future. Meanwhile, attachment theory serves as a model for the meaningful study of
complex human behavior, through the life span, in a relatively rigorous manner (Eagle, 1997).

4. The relevance of attachment theory to adult mental health

4.1. The individual and mental health problems

A large body of research has focused on the significance of insecure attachment patterns in
adults. Bowlby (1942) himself first proposed a connection between adult mental health
42 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56

problems and early parental loss in his first book, Personality and Mental Illness. Attachment
theory suggests that attachment relationships and their difficulties might influence mental
health in three distinct, but interrelated, ways. Firstly, the breaking or disruption of bonds is
likely in itself to be a cause of disturbance. Secondly, the internalization of disturbed early
attachment patterns may influence subsequent relationships in a way that makes a person both
more exposed as well as more vulnerable to stress. And thirdly, a person’s current perception
of their relationships and the use they make of them may make them more or less vulnerable
to break down in the face of adversity (Holmes, 1993). Data from research into infant,
childhood, and adult attachments suggest that, in Western society, approximately one third of
adults will have relationships characterized by anxious, insecure attachment. It has been
argued that this may constitute a major vulnerability factor for mental health problems when
individuals encounter difficult or stressful life events (Holmes, 1993).

4.1.1. Depression
Researchers, from Bowlby onwards, have conducted investigations into a range of
different mental health problems. Depression is the most studied area, as the relationship
between loss and depression has long been acknowledged. Bowlby (1980) proposed that loss
of, or separation from, the mother in childhood, or poor parental relationships leading to
insecure attachment, leaves a person vulnerable to depression in adult life. This has been
repeatedly confirmed, for example, by Brown and Harris (1978) and Tennant (1988). Harris
and Bifulco (1991) found that women who had been bereaved of their mother as children had
significantly higher rates of depression (one in three) compared with nonbereaved women
(one in ten). Although it only involved a small subgroup of depressed individuals, their study
provides a good illustration of the interrelation of factors influencing adult mental health
described above: the experience of early loss of the mother (through either death or
separation) led commonly to lack of care in childhood accompanied by the development
of a sense of helplessness and hopelessness. Harris and Bifulco linked this with high
subsequent rates of premarital, teenage pregnancy. This, in turn, led to poor or limited choice
of life-partner, so that when the women—often living in disadvantaged situations, highly
vulnerable to stress, and retaining a continued sense of helplessness—experienced loss or
other major distress, they had little (if any) personal or social resources to call on, leaving
them much more likely to develop depression. However, Holmes (1993) provides a reminder
that most people who are identified as suffering from depression have come from ‘intact’
homes, and have not experienced significant parental loss (though may, of course, have
experienced other difficulties or disruptions in their early attachment relationships). Rose-
nfarb, Becker, and Khan (1994) found that severely depressed women and women with
bipolar depression reported little attachment to their mother at all ages, as measured by
Family Circle Drawings (Pipp, Shaver, Jennings, Lamborn, & Fischer, 1985), the Inventory
of Parent and Peer Attachment (Armsden & Greenberg, 1987), and the Parent–Child
Relations Questionnaire (Siegelman & Roe, 1979). All three of these measures are reported
to have good reliability and validity, and are an interesting selection in that they utilize
visual–perceptual assessment in addition to subjective and behavioral self-report methodo-
logies (respectively). In addition, Rosenfarb et al. (1994) found that severely depressed
 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56 43

women felt less attached than nonpsychiatric controls to their peers during development,
which may be expected if, as Bowlby suggests, they did not have the experience of a secure
base from which to explore the world. However, it has been argued that the quality of current
peer attachment in adulthood can mediate the impact of early parental bonding on levels of
depression (Strahan, 1995), confirming that present as well as past relationships may be
highly significant.

4.1.2. Sexual abuse

Although attachment theory posits links between early experiences with parents, adult
relationships, and adult mental health, research has only recently asked questions about
whether these experiences exert independent, mediating or moderating effects: that is,
whether adult relationships mediate the influences of early attachment experiences on later
mental health outcomes, and whether intimate (adult attachment) relationships may act as
moderators—either amplifying or attenuating the effects of early attachment experiences
(Gittleman, Klein, Smider, & Essex, 1998). One area, which has attracted recent interest in
this respect, is childhood sexual abuse. Schreiber and Lyddon (1998) found that female
survivors of childhood sexual abuse displayed significantly poorer psychological adjustment
than a control group, but that a high level of paternal care was significantly associated with
better psychological functioning among survivors of childhood sexual abuse. In a similar
vein, Whiffen, Judd, and Aube (1999) examined adult attachment, intimacy, and partner
physical abuse as potential mediators or moderators of the association between childhood
sexual abuse and depression. Their results support the view that attachment moderates the
relationship between abuse and depression rather than acting as a mediator. Specifically,
highly intimate attachment relationships appear to be protective for women with a sexual
abuse history, whereas low intimacy attachment relationships are a notable risk factor in
developing depression. Conversely, Roche, Runtz, and Hunter (1999), also investigating the
nature of the relationship between childhood sexual abuse, adult attachment style, and
psychological adjustment, found that a mediational model was supported by their results; that
is, attachment mediates between childhood sexual abuse and later psychological adjustment.
Roche et al.’s (1999) findings may be considered more robust, as their sample included
85 women with a history of childhood sexual abuse, compared to Whiffen et al. (1999) whose
study included only 22 abuse survivors.
The varying nature of the research findings available thus far in the area of sexual abuse
suggests that further research exploring the effects of both childhood and adult attachments
and their links with mental health is warranted.

4.1.3. Personality problems

Borderline personality and personality disorders have also received research attention in
terms of identifying attachment difficulties in a person’s early life, which may have
contributed to the diagnosis and associated mental health problems. Research has suggested
that people with a diagnosis of borderline personality or personality disorder often have a
childhood history of emotional neglect and trauma (Herman, Perry, & Kolk, 1989), leading to
fragile, unstable personality and highly insecure attachments. More recently, however, it has
44 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56

been suggested that clinical symptomatology thought be related to experiences of severe

relationship distress may, in fact, be best explained in terms of attachment disorganization
rather than as normative forms of attachment insecurity: that is, the absence of a coherent
attachment strategy as opposed to the presence of an identified insecure attachment style (a
fearful or avoidant style, for example) (George & West, 1999). A number of studies have
focused on the oscillations of attachment in borderline personality disorder, which have been
proposed as stemming from a central problem with regulation of interpersonal distance
derived from a conflict between fears of abandonment and domination (Melges & Swartz,
1989). It is suggested that oscillations between these two fears account for most of the
therapeutic problems encountered in the treatment of such people (Chabrol, 1997).

4.1.4. Agoraphobia and anxiety

Agoraphobia and other anxiety problems have also been the focus of theorizing and
research (Bowlby, 1973; Persons, Burns, Perloff, & Miranda, 1993). Bowlby saw ago-
raphobia as an example of separation anxiety arising from anxious attachment. He identified
three possible patterns of interrelating underlying agoraphobia: a role reversal whereby the
agoraphobic is recruited to alleviate the separation anxiety of the parent; the individual’s
own fears that something terrible may happen to the parent while they are apart; and the
individual’s fear of something awful befalling themselves while away from the protection of
their parent (Bowlby, 1973). In research terms, however, agoraphobia appears to have
inspired little interest, as the literature contains very few published empirical studies.
Anxiety problems have been more attended to in the literature, though they have also
received little exclusive attention from empirical researchers, tending to be incorporated into
studies with a broader focus, such as that of Fonagy et al. (1996), who found that anxiety
(as a diagnostic category) was associated with ‘unresolved’ status as rated in the AAI
(Main, 1985).

4.1.5. Psychoses
The diagnostic categories covering psychosis have also come to the attention of adult
attachment theorists and researchers. Holmes (1993) has suggested that it is at least possible
that there is a relationship between Expressed Emotion (EE, as described by Leff & Vaughn,
1983) and anxious attachment. Holmes implies some evidence for this link: first, that of
families of a person receiving a diagnosis of schizophrenia, approximately one third are
‘high’ in EE, and that around one third of the population is also considered as having anxious,
insecure attachments. Second, he notes, the two main patterns of high EE—hostility and
emotional overinvolvement—correspond with those found in anxious attachment, that is,
avoidant and ambivalent attachment (Holmes, 1993). Without any supporting empirical
evidence illustrating a relationship between EE and attachment insecurities, however, such
suggestions must necessarily be considered speculation.
Starkey and Flannery (1997) have presented a theoretical model for psychiatric rehab-
ilitation, which discusses the impairments of those diagnosed with schizophrenia in relation to
attachment (as well as mastery and meaning). However, as with much of the literature in this
 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56 45

field, the discussion stops at the theoretical level: they do not present any clinical evidence or
research to support their theses.

4.1.6. Attachment disorders

‘Reactive attachment disorder’ has recently become a new clinical category of diagnosis in
DSM-IV (American Psychiatric Association, 1994). One interpretation of this is that the
concept of adult attachment has thus received formal recognition by the psychiatric
establishment. In so doing, it remains to be seen if its psychological nature is retained and,
as has been anticipated, it becomes instrumental in developing effective management
strategies without the use of psychotropic medication (Minnis, Ramsay, & Capmbell,
1996). Alternatively, it may be viewed as yet another opportunity in the ever-increasing
tendency to medicalize peoples’ relationships difficulties, which will become susceptible to
standard psychiatric, medical treatment (Kutchins & Kirk, 1997).
In conclusion, the literature indicates that very many elements of mental health problems
have been subject to attachment-related research. Within some areas—such as depression—
research findings show general concurrence regarding the role of attachment in the
development of problems, while in other, ‘newer’ areas of study—such as sexual abuse—
findings appear conflicting. However, in many areas, research is still in its infancy, and future
study may provide further, more substantive evidence regarding the relevance of attachment
theory in this field.

4.2. Therapeutic service provision and delivery

In terms of research into attachment theory and the provision of psychological treatment
services for people encountering mental health problems as a result of insecure attachments,
studies have tended to focus on the psychotherapy process and relationship. Counseling and
psychotherapy can be conceived as being based on the movement from insecure to secure
attachment (Holmes, 1994), and the goals as being the achievement of intimacy and
autonomy (Holmes, 1997). An alternative aim of therapy involves helping the client mourn
the loss of that which they never fully experienced but yearned for deeply, that is, a secure
attachment relationship (West & Sheldon-Keller, 1994). The therapist’s ability to act as a
secure base, the importance of clients learning to tell a coherent narrative, the attunement of
the therapist in relation to their client, affective processing (especially of anger), and the
facilitation of self-exploration within the safety of the therapeutic relationship have all been
identified as essential ingredients of psychotherapy with vulnerable, insecure clients (Holmes,
1994). Many other researchers and clinicians have attested to the importance of these
elements of the therapy process and relationship (e.g., Ball & Legow, 1996; Byng-Hall,
1995; Sable, 1994, 1995), though the literature in this area appears to rely heavily on
theoretical discussion accompanied by illustrations from clinical practice, rather than em-
pirical research studies.
Holmes (1997) has suggested that attachment ideas are also valuable in their ability to
inform a therapist regarding assessment, in thinking about transference, and in devising
therapeutic strategies. He argues that attachment themes often stand out more clearly during
46 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56

the early, assessment phase of therapy than later in the process, and illustrates this using
narratives from clinical practice.

4.2.1. The client’s engagement in therapy

The client’s particular attachment style as an important variable in therapy has been
examined, resulting in the recommendation that therapists may need to tailor their interper-
sonal stance in psychotherapy according to the client’s attachment style (Dolan, Arnkoff, &
Glass, 1993). Dolan et al. (1993) report results from a psychotherapy research program,
including the finding that attachment style was related to therapists’ perceptions of the
therapeutic alliance. Evidence is also presented to illustrate how alteration in therapist
approach can enhance treatment outcome (Dolan et al., 1993), and although quite compelling,
the feasibility, for many therapists, of adjusting their therapeutic approach in such a way
(without having the resources to first formally assess the client’s attachment style) is likely to
be fairly limited.
The ways in which therapists function as attachment figures for their patients has also been
explored (Farber, Lippert, & Nevas, 1995). Here it has been argued that although there are
multiple commonalities between childhood attachments and the therapist–client relationship,
the latter is necessarily mediated by unique temporal, structural, ethical (and sometimes
financial) boundaries that render it significantly different from childhood attachment relation-
ships (Farber et al., 1995). Despite this caveat, it is generally accepted in the literature that the
therapist–client relationship can be defined as an attachment relationship.
Mallinckrodt, Gant, and Coble (1995) have developed an instrument to measure the
psychotherapy relationship from the perspective of attachment theory—the Client Attach-
ment to Therapist Scale—and report that the nature of the therapeutic attachment, client
transference, and the therapeutic strategies required may be significantly different across the
various types of client attachment. Although an interesting approach to the empirical
assessment of attachments in therapy, the stability of the measure may be open to question
owing to the limited ratio of participants to items reported in the factor analysis
(Mallinckrodt et al., 1995).
An adult attachment issue that applies across all therapy services is that of the individual
client’s ability to engage with services being offered. Korfmacher, Adam, Ogawa, and
Egeland (1997), in a well-designed, randomized trial involving 55 (female) participants in
the intervention group, found that women with more secure relationship representations
were more involved in the intervention and accepted more forms of treatment than those
with insecure (dismissing or unresolved) styles. More specifically, women with unresolved
attachment styles were more likely to have a ‘crisis’ orientation to the intervention, whereas
women identified as dismissing in attachment style had a more emotionally unengaged
involvement with the intervention (Korfmacher et al., 1997). Such findings may have
significant implications for clients’ engagement with mental health services in general, with
those already less vulnerable (more secure) individuals able to better utilize therapeutic
treatment services, and those already more vulnerable and less secure (and probably
therefore the majority of clients coming into contact with services), finding it more
difficult to utilize services in a planned and comprehensive manner. Replication of
 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56 47

Korfmacher et al.’s (1997) research with different client groups might give further valuable
insights here.
In a similar vein, Cortina (1999) has argued that ‘secure’ individuals—as defined by the
AAI—have greater ability to reflect on their experiences of past attachment relationships. In
contrast, insecure individuals are often too preoccupied with their experience or too detached
from it, and consequently, show a reduced capacity for reflection (Cortina, 1999). Such
findings are highly applicable clinically, as the capacity for reflection is, arguably, an essential
part of successful therapy. Similarly, Meyers (1998) found consistent relationships between
adult attachment styles and the ability to manage stress and anxiety. Although Meyers’
findings may be considered reasonably robust on the basis of a large sample size, the
participants were American university undergraduates, and thus, the findings are of
questionable generalizability. This does not preclude the potential implications for psycho-
logical therapies in general, particularly those of a more psychodynamic nature, which
necessarily involve the generation of anxiety in the client.

4.2.2. The therapeutic process

Holmes (1997) eloquently views the therapeutic process as ‘a microcosm of attachment
and separation, with its rhythm of regular sessions punctuated by endings and breaks’
(p.246). He also provides a pertinent analogy between therapy and the Strange Situation,
whereby the client is subjected to minor stresses, and the impact of these stresses are
examined in the ‘laboratory’ of therapy (Holmes, 1997). As with other work in this area,
Holmes’ opinions appear to be based on observations and interpretations of clinical
practice rather than research evidence. Although such observations and interpretations are
likely to be grounded in extensive clinical experience, they should perhaps be considered
with caution.
Where Holmes (1997) and others see attachment theory and AAI research as offering
scientific links between the biology of parent–child interaction and the narrative paradigms
of psychotherapy, other researchers question the validity of such links. Eagle (1997) argues
that there is no direct or compelling evidence that individuals who tell coherent, plausible
narratives (as recorded in the AAI) are, in fact, securely attached. This is an important
challenge, as a large body of adult attachment research utilizes Main et al.’s (1985) AAI.
Eagles’ grounds for this challenge are that it has not been shown that individuals who are
autobiographically competent on the AAI are securely attached as indicated by other
independent measures of attachment status, but that, conversely, attachment has become
operationally defined as the telling of coherent and plausible narratives in the AAI (Eagle,
1997). From this position, the validity of reporting the movement of clients from being
insecurely to securely attached—on the AAI—following a period of therapy (as, for instance,
Fonagy et al., 1995 do)—is called into question. Eagle suggests that such conclusions require
confirmation via independent and ecologically valid information regarding the individual
client’s relationships and how they subjectively experience them, though not making specific
suggestions as to what such sources of information might be. However, despite such concerns
regarding validity, research continues to utilize the AAI to provide an indicator of attachment
status (see, e.g., Diamond et al., 1999).
48 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56

4.2.3. Group therapy interventions

Although individual psychotherapy appears to be becoming increasingly available within
mental health services, a variety of therapeutic groups continue to be offered, regarding which
there is (so far) a small body of attachment literature. Kilmann et al. (1999) report the effects
of a recently developed attachment-focused group intervention designed to target the
attachment concerns of insecure individuals. In contrast to many of the studies of individual
therapy outcomes, this group intervention is manualized (involving four sequential segments)
and highly intensive-comprising 17 hours over a 3-day period. However, Kilmann et al.
report positive outcomes at 6 months follow-up after completion of the group intervention—
that is, improved interpersonal styles, enhanced satisfaction with family relationships,
decreased agreement with dysfunctional relationship beliefs, and less fearful and more secure
attachment patterns compared with controls. Further, Kilmann et al. remark that participants
with a fearful–avoidant attachment pattern reported the greatest gains. Limitations of the
intervention include the rather large group size (13 participants) and the reliance on self-
report measures with participants.
Another interesting study describes a group ‘course’ for lonely, isolated, older women
(aged 54–75) based on a self-help method and the principles of feminist therapy (Stevens,
1997). At 1-year follow-up, outcomes remained positive, though unlike Kilmann et al.’s
study, Stevens (1997) reports that interpersonal relationship changes appeared to be inde-
pendent of adult attachment style.
In summary, the wide variety of adult attachment literature examining psychological
therapies provided by mental health services suggests that attachment theory is highly
relevant in this area. The research evidence appears strongest in relation to client attachment
style and the individual’s ability to make use of therapies provided (be they individual or
group interventions). The literature exploring the therapeutic process in terms of attachment
issues currently relies more on observations drawn from clinical practice rather than empirical
research, and may usefully be a focus for future empirical exploration.

4.3. Taking account of staff attachments

Despite the suggestion that around one third of the adult population is likely to have
relationships which are characterized by anxious, insecure attachment and that this renders
them vulnerable to mental health problems (Holmes, 1993), very few studies have looked at
the impact the attachment style of the mental health professional (acting as caregiver) may
have on the attachment relationship formed with clients (who may be insecure and seeking a
secure and responsive attachment figure at a time of illness or distress). One exception to this
is Dozier, Cue, and Barnette’s (1994) thought-provoking study, where the relationships
between clinicians’ attachment style and interventions used with clients with serious mental
health problems were examined. AAIs were administered to both the clients and their ‘key’
clinicians, and clinicians were also interviewed regarding their interventions with clients.
The results indicated that clinicians with an insecure attachment style attended more to
dependency needs and intervened in greater depth with ‘preoccupied’ clients (those
remaining overinvolved with past conflicts and difficulties) than they did with ‘dismissing’
 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56 49

clients (those claiming few childhood memories and with a tendency to idealize the past).
Further, ‘preoccupied’ clinicians intervened with their clients in greater depth than did
‘dismissing’-style clinicians (Dozier et al., 1994). Dozier et al. (1994) do not provide details
of the percentage of clinicians identified as secure and insecure in this study, though this may
be a reflection of the nature and purpose of the study as well as the relatively small numbers
involved (18 clinicians and 27 clients). Nevertheless, this clearly suggests that the likelihood
of all mental health professionals being able to provide a secure base for their insecure
clients is questionable. It may be much more realistic to assume that at least a substantial
minority of clinicians may themselves have an insecure attachment style, and may be
seeking to provide care for others in their professional lives by way of compensation
(Vaillant, Sobowale, & McArthur 1972).
This pattern of insecure attachment is what Bowlby originally described as ‘compulsive
caregiving’, which has been validated conceptually and empirically against Ainsworth’s
patterns of attachment in infancy (Main, 1991). According to Bowlby (1977), compulsive
caregiving develops in response to a premature experience of a child being forced to care for a
parent (or other siblings), and only being able to attain sufficient proximity to the parent by
suppressing their own attachment behaviors: that is, the child’s need for care is renounced for
the sake of maintaining proximity to the attachment figure (the parent). West and Sheldon-
Keller (1994) note that it is important to differentiate this kind of caregiving from caregiving
initiatives that develop appropriately later in (adult) life in reciprocal relationships (such as
partnership or marriage) and true parental relationships. The latter form of caregiving
behaviors are considered complementary to the attachment system, whereas the former,
‘compulsive’ caregiving, which develops in the child and is directed towards an adult, arises
from the child’s attachment system and can lead to dysfunctional relationships later in life, as
the individual lacks the ability to express need or ask for care while retaining an unsatisfied
neediness and longing to receive care (West & Sheldon-Keller, 1994). It may be apposite that
Harris and Bifulco (1991) illustrate the category of compulsive caregiver in their research
with a vignette of a female research participant who had always wanted to be a nurse.
More recently, research by Tyrell and Dozier (1997) has explored the effects of clinician
and client attachment strategies on treatment outcome as well as process. Both clients and key
clinicians were administered the AAI, which were then coded using both Main and Gold-
wyn’s (1989) classification system as well as Kobak’s (1989) Q-set (a form of analysis using
the same AAI data, but adopting a quantitative approach). They found that, according to Main
and Goldwyn’s categorical system, 90% of clinicians were classified as autonomous (secure),
and 83% of clients were classified as insecure. Such proportions differ significantly from the
general population. One reason for this might be the very select nature of the groups: the
participants were clients with ‘serious psychopathological disorders’ and clinicians working
within an intensive clinical case management program with these clients. Within this service,
such polarized levels of security and insecurity might be anticipated. However, the additional
use of Kobak’s system provides a continuous measure of deactivation–hyperactivation
(categories corresponding to dismissing and preoccupied states of mind, respectively), which
directly assesses the extent to which particular (more specific) strategies are used by both
clients and clinicians. Thus, Tyrell and Dozier found a number of interesting interactions
50 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56

within their results: Clients who were more deactivating had stronger alliances with their
clinicians, better quality of life, and higher levels of functioning when working with clinicians
who were less deactivating, whereas clients who were less deactivating had stronger alliances,
better quality of life, and higher levels of functioning when working with clinicians who were
more deactivating. This suggests that client–clinician dissimilarity on the deactivation–
hyperactivation dimension of attachment was associated with the formation of stronger
alliances and more therapeutic gains, as reported by both client and clinician. Tyrell and
Dozier argue that this is due to a noncomplementary process, whereby clinicians can more
effectively challenge clients through their own interpersonal behavior when they are not
concordant with clients for attachment strategies. However, it is important to acknowledge, as
Dozier and Tyrell (1998) do, that when clinicians’ attachment strategies are identified within
an insecure attachment category, their findings indicate that such a pattern of results no longer
holds. They assert that the clinician must have the ego strength and flexibility necessary to
respond to the client in a noncomplementary manner, even if it is uncomfortable for the
clinician to do so. They further conclude that secure clinicians are likely to work better with
clients who differ from them in regard to deactivation–hyperactivation, and that—crucially—
insecure clinicians are likely to have difficulty providing effective interventions, regardless of
client attachment organization (Dozier & Tyrell, 1998).
As Dozier and Tyrell (1998) state, the effects of clinical staffs’ attachment styles on
treatment have only begun to be explored. However, findings so far are consistent with the
relatively untested assumption that staffs’ own attachment issues do influence the therapeutic
process and outcome. Tyrell and Dozier’s (1997) research is a valuable early step in this
exploration for a number of reasons: they are concerned with outcome as well as process; they
utilize the AAI but—importantly—use two complementary analytic processes to extract
qualitative and quantitative elements from the data; and, unlike many studies into the
therapeutic process, they consider the interrelation and impact of the styles of both client
and clinician in the attachment relationship.

4.4. Mental health service philosophy and organization

Mental health services in this country have been subject to massive change, particularly
during the latter part of the 20th century. It seems that no sooner has one directive been issued
from the Department of Health, another, different one is issued indicating a further shift in
direction that services are expected to follow. A question relevant here is, to what extent (if
any) has attachment theory influenced the philosophy and organization of mental health
services, either in terms of the explicit application of attachment theory to services or in terms
of principles of service provision that acknowledge the attachment needs of clients? In
attempting a response, it must be acknowledged that political and economic considerations
are, in reality, paramount in the development of services. Marrone (1998) opines that the
organization and administration of mental health services are often guided by a politics of
exclusion rather than by concern with the needs of a disadvantaged section of the population,
and that economic rather than humanitarian considerations usually retain the upper hand. In
addition, he notes that National Health Service provisions are often basic and are mainly
 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56 51

aimed at containing the most serious forms of behavioral disturbance (Marrone, 1998). This
view concurs with that of Pilgrim (2001) who argues that adult inpatient mental health units
need to seriously reflect on and review their current role and function which, he argues, is
more about social control than care.
Among adult attachment researchers there appears to be a belief that attachment theory
certainly can inform various aspects of mental health service philosophy and organization.
For example, Marrone (1998) identifies that attachment theory is making definite advances in
the study of etiological factors within adult psychiatry, adding an essential dimension that
could not be provided by the biological and neurochemical approaches. Inclusion of a
‘reactive attachment disorder’ in DSM-IV may be quoted as evidence of this influence.
Marrone (1998) also argues that attachment theory can usefully inform the study of
psychotherapeutic interventions aimed at helping very disturbed individuals, though acknow-
ledging that many senior clinicians (particularly psychiatrists) continue to argue that the
majority of seriously disturbed clients are unable to engage in formal psychotherapeutic work.
Nevertheless, he asserts that, for such people, the provision of a secure and caring
environment in which individuals can develop trusting relationships with staff so as to be
able to express their innermost feelings and anxieties, is essential (Marrone, 1998). This may
be true, though much of the service user literature (e.g., Jenkinson, 1999), along with the
author’s experience, suggests that such provision remains limited.
Marrone (1998) also believes that attachment theory is well placed to inform the study of
mental health care policies, and suggests that knowledge of attachment theory can be useful
for all those who hold decision-making power in this field, including governments, health
authorities, and hospital administrators. As noted however, the extent to which such
recommendations have been transformed into reality is unclear, and Marrone makes no
comments on this point. In contrast to this optimistic view, Holmes (1993) sees the major shift
from hospital to community care as a version of autonomy, which is akin to avoidance, and
one which has overlooked the continuing need for dependence. In addition, Holmes refers to
the development of a more subtle appreciation of the issues involved in attachment relation-
ships in creating an analogy with the problems within mental health services: he notes that
institutions are not necessarily intrinsically harmful any more than the ‘community’ is always
beneficial, rather, the problem is the way that care is often delivered in them. In this respect, it
is necessary to look carefully at the actual quality of the client’s experience, whether an
inpatient or an outpatient, before deciding whether it is bad or not (Holmes, 1993).
Some now commonplace elements of service philosophy and organization suggest they
have been influenced by attachment theory, though without explicit reference to the
theoretical underpinnings. For instance, the concept of continuity of care, where an individual
is commonly allocated a ‘keyworker’ whose responsibility is to be available regularly and
consistently, to take a special interest in getting to know the client and their problems, and to
coordinate the different elements of care provided. Research into service user views supports
the view that such consistency is valued by clients (Goodwin, Holmes, Newnes, & Waltho,
1999; Trinder, Mitchell, & Todd, 1994).
A current service issue illustrating Holmes’ (1993) point in relation to this is the
development of Assertive Outreach teams in the community. Many such teams specifically
52 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56

avoid having keyworkers, with the ‘attachment’ being considered as most usefully being to
the team rather than to an individual. As yet, there does not appear to be any published
literature available on the helpfulness (or otherwise) as perceived by clients of organizing and
providing services in this way, though it seems to contradict the fundamental tenet of
attachment theory, and to imply that the development of close, individual relationships is
either not necessary for people in distress or perhaps that the purpose of such teams is more
about monitoring and short-term practical assistance rather than long-term resolution of
distress and insecurity.
It has also been argued that policymakers have not sufficiently recognized the need for
support for carers and the fact that psychologically damaged individuals who have lost their
attachments may need many years of connection to a stable and secure person or place before
that experience can be hoped to be internalized enough for them to gain increased
independence (Holmes, 1993). Holmes (1993) sees this as the result of the search for quick
and easy solutions by policymakers to the problems of mental distress and personal growth. It
is also likely to be indicative of the fact that attachment theory has, as yet, had little real
influence on mental health service philosophy and organization.

5. Future research and development

Much research has been conducted into adult attachment and, more specifically, into
mental health problems arising out of insecure attachments. The majority of the research in
relation to adult mental health issues has focused on individuals and/or on the different
diagnostic categories of disorders that individuals fall within. In addition, in relation to
treatment provision, the focus has remained almost exclusively on issues of treatment with
individual psychotherapy.
The focus of research on psychotherapy omits acknowledgement that specialist psycho-
therapy may only be offered to a small minority of people referred to mental health services,
and that many other forms of care and treatment constitute mental health service provision.
Consequently, hospital-based inpatient mental health care, community-based services, and the
attachment relationships developed between mental health workers and the clients using these
services have not been a focus of study. It has been argued that there is no shared framework
of understanding between mental health disciplines to enable strategies to be designed aimed
at providing a secure base from which some reparation—however limited—may take place
(Marrone, 1998). Thus, mental health services leave a lot to be desired in terms of their
capacity to provide continuous, reliable, and sensitive attachments for people (Marrone,
1998). The conflict between the need for stable attachments and the complexity and pressures
affecting mental health services provision has been recognized (Holmes, 1994) and the
concern that current psychiatric practice offers little or no space for the opportunity to recount
painful experiences has been voiced (Adshead, 1998). Despite these concerns, no research
has yet examined the experiences that people have within the various models of mental health
care provision in terms of the services’ ability to meet clients’ attachment needs and serve as
secure attachment figures. Based on the assumption that secure attachment relationships are
 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56 53

crucial to psychological health, it becomes important to understand what role mental health
services play in providing and developing secure attachment relationships for people.
Research with this focus can also be justified in terms of the need to assess whether services
may be harming rather than helping clients, for example, by perpetuating or repeating their
experience of insecure, damaging attachments. Furthermore, if the elements or aspects of
different service models, which do facilitate the development of secure attachment relation-
ships, can be identified, then this could provide areas of focus for relevant improvement and
expansion of current adult mental health services.

Acknowledgements

I thank Dr. Guy Holmes, Professor Ray Cochrane, and Dr. Oliver Mason for their very
helpful comments on early drafts of this paper.

References

Adshead, G. (1998). Psychiatric staff as attachment figures. British Journal of Psychiatry, 172, 64 – 69.
Ainsworth, M. (1985). Attachments across the life span. Bulletin of the New York Academy of Medicine, 61(9),
792 – 811.
Ainsworth, M., Blehar, M., Waters, E., & Wall, S. (1978). Patterns of attachment: assessed in the strange
situation and at home. Hillsdale, NJ: Erlbaum.
American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Wash-
ington, DC: American Psychiatric Association.
Armsden, G. G., & Greenberg, M. T. (1987). The inventory of parent and peer attachment: individual differ-
ences and their relationship to psychological well-being in adolescence. Journal of Youth and Adolescence,
16, 427 – 454.
Ball, S. A., & Legow, N. E. (1996). Attachment theory as a working model for the therapist transitioning from
early to later recovery substance abuse treatment. American Journal of Drug and Alcohol Abuse, 22(4),
533 – 547.
Banai, E., Weller, A., & Mikulincer, M. (1998). Inter-judge agreement in evaluation of adult attachment style: the
impact of acquaintanceship. British Journal of Social Psychology, 37(1), 95 – 109.
Bartholomew, K. (1990). Avoidance of intimacy: an attachment perspective. Journal of Social and Personal
Relationships, 7, 147 – 178.
Bartholomew, K. (1993). From childhood to adult relationships: attachment theory and research. In Duck S.
(Ed.), Learning about relationships: understanding relationship processes series, vol. 2. Newbury Park,
CA: Sage.
Birtchnell, J. (1997). Attachment in an interpersonal context. British Journal of Medical Psychology, 70(3),
265 – 279.
Bowlby, J. (1942). Personality and mental illness. New York: Emerson.
Bowlby, J. (1969). Attachment and loss volume 1: attachment. London: Hogarth Press.
Bowlby, J. (1973). Attachment and loss volume 2: separation: anxiety and anger. London: Hogarth Press.
Bowlby, J. (1977). The making and breaking of affectional bonds: I. Aetiology and psychopathology in the light of
attachment theory. British Journal of Psychiatry, 130, 201 – 210.
Bowlby, J. (1979). The making and breaking of affectional bonds. London: Tavistock Publications.
Bowlby, J. (1980). Attachment and loss volume 3: loss: sadness and depression. London: Hogarth Press.
54 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56

Bretherton, I. (1985). Attachment theory: retrospect and prospect. Monographs of the Society for Research in
Child Development, 50(1 – 2), 3 – 35 (Serial No. 209).
Brown, G., & Harris, T. (1978). The social origins of depression. London: Tavistock.
Byng-Hall, J. (1995). Rewriting family scripts: improvisation and systems change. New York: Guildford Press.
Chabrol, H. (1997). Abandonment and intrusion fears in borderline personality disorder. American Journal of
Psychiatry, 154(9), 1329.
Cook, W. L. (2000). Understanding attachment security in family context. Journal of Personality and Social
Psychology, 78(2), 285 – 294.
Cortina, M. (1999). Causality, adaptation and meaning: a perspective from attachment theory and research.
Psychoanalytic Dialogues, 9(5), 557 – 596.
Crowell, J., & Treboux, D. (1995). A review of adult attachment measures: implications for theory and research.
Social Development, 4(3), 294 – 327.
Davila, J., Burge, D., & Hammen, C. (1997). Why does attachment style change? Journal of Personality and
Social Psychology, 73(4), 826 – 838.
Diamond, D., Clarkin, J., Levine, H., Levy, K., Foelsch, P., & Yeomans, F. (1999). Borderline conditions and
attachment: a preliminary report. Psychoanalytic Inquiry, 19(5), 831 – 884.
Dolan, R. T., Arnkoff, D. B., & Glass, C. R. (1993). Client attachment style and the psychotherapist’s inter-
personal stance. Psychotherapy, 30(3), 408 – 412.
Dozier, M., Cue, K. L., & Barnett, L. (1994). Clinicians as caregivers: role of attachment organisation in treatment.
Journal of Consulting and Clinical Psychology, 62(4), 793 – 800.
Dozier, M., & Tyrell, C. (1998). The role of attachment in therapeutic relationships. In J. A. Simpson, & W. S.
Rholes (Eds.), Attachment theory and close relationships. New York: Guilford Press.
Eagle, M. (1997). Attachment and psychoanalysis. British Journal of Medical Psychology, 70(3), 217 – 229.
Farber, B. A., Lippert, R. A., & Nevas, D. B. (1995). The therapist as attachment figure. Psychotherapy, 32(2),
204 – 212.
Feeney, J., & Noller, P. (1996). Adult attachment. Thousand Oaks, CA: Sage.
Feeney, J. A. (1996). Attachment, caregiving and marital satisfaction. Personal Relationships, 3(4), 401 – 416.
Fonagy, P., Leigh, T., Steele, M., Steele, H., Kennedy, R., Mattoon, G., Target, M., & Gerber, A. (1996). The
relation of attachment status, psychiatric classification, and response to psychotherapy. Journal of Consulting
and Clinical Psychology, 64(1), 22 – 31.
Fonagy, P., Steele, M., Steele, H., Leigh, T., Kennedy, R., Mattoon, G., & Target, M. (1995). Attachment, the
reflective self, and borderline states. In S. Goldberg, R. Muir, & J. Kerr (Eds.), Attachment theory: social,
developmental and clinical perspectives. Hillsdale, NJ: Analytic Press.
Fonagy, P., Steele, M., Steele, H., Moran, G., & Higgins, A. (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, 201 – 218.
Freedman, M. B., Leary, T., Ossorio, A. G., & Coffee, H. S. (1951). The interpersonal dimension of personality.
Journal of Personality, 20, 143 – 161.
George, C., & West, M. (1999). Developmental vs. social personality models of adult attachment and mental ill
health. British Journal of Medical Psychology, 72(3), 285 – 303.
Gittleman, M. G., Klein, M. H., Smider, N. A., & Essex, M. J. (1998). Recollections of parental behavior,
adult attachment and mental health: mediating and moderating effects. Psychological Medicine, 28(6),
1443 – 1455.
Goodwin, I., Holmes, G., Newnes, C., & Waltho, D. (1999). A qualitative analysis of the views of inpatient mental
health service users. Journal of Mental Health, 8(1), 43 – 54.
Grossman, K., & Grossman, K. (1991). Attachment quality as an organiser of emotional and behavioral responses
in a longitudinal perspective. In C. M. Parkes, J. Stevenson-Hinde, & P. Marris (Eds.), Attachment across the
life cycle. London: Routledge.
Harris, T., & Bifulco, A. (1991). Loss of parent in childhood, attachment style and depression in adulthood. In
C. M. Parkes, J. Stevenson-Hinde, & P. Marris (Eds.), Attachment across the life cycle. London: Routledge.
 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56 55

Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment process. Journal of Personality
and Social Psychology, 52, 511 – 524.
Herman, J., Perry, C., & Kolk, B. (1989). Childhood trauma in borderline personality disorder. American Journal
of Psychiatry, 146, 490 – 495.
Holmes, J. (1993). John Bowlby and attachment theory. London: Routledge.
Holmes, J. (1994). Attachment theory: a secure theoretical base for counselling? Psychodynamic Counselling,
1(1), 65 – 78.
Holmes, J. (1997). Attachment, autonomy and intimacy: some clinical implications of attachment theory. British
Journal of Medical Psychology, 70, 231 – 248.
Ijzendoorn, M., & Kroonenberg, P. (1988). Cross cultural patterns of attachment: a meta-analysis of the Strange
Situation. Child Development, 59, 147 – 156.
Jenkinson, P. (1999). The duty of community care: the Wokingham MIND crisis house. In C. Newnes, G.
Holmes, & C. Dunn (Eds.), This is madness: a critical look at psychiatry and the future of mental health
services. Ross-on-Wye: PCCS Books.
Johnston, M. A. (1999). Influences of adult attachment in exploration. Psychological Reports, 84(1), 31 – 34.
Kilmann, P. R., Laughlin, J. E., Carranza, L. V., Downer, J. T., Major, S., & Parnell, M. M. (1999). Effects of an
attachment-focused group preventive intervention on insecure women. Group Dynamics, 3(2), 138 – 147.
Kirkpatrick, L. A., & Davis, K. E. (1994). Attachment style, gender and relationship stability: a longitudinal
analysis. Journal of Personality and Social Psychology, 66(3), 502 – 512.
Kirkpatrick, L. A., & Hazan, C. (1994). Attachment styles and close relationships: a four-year prospective study.
Personal Relationships, 1, 123 – 142.
Kobak, R. R. (1989). The Attachment Interview Q-Set. University of Delaware, unpublished manuscript.
Korfmacher, J., Adam, E., Ogawa, J., & Egeland, B. (1997). Adult attachment: implications for the therapeutic
process in a home visitation intervention. Applied Developmental Science, 1(1), 43 – 52.
Kutchins, H., & Kirk, S. A. (1997). Making us crazy: DSM—the psychiatric bible and the creation of mental
disorders. London: Constable.
Leff, J., & Vaughn, C. (1983). Expressed emotion in families. New York: Guildford Press.
Main, M. (1985). An adult attachment classification system: its relation to infant – parent attachment. Paper
presented at the Biennial Meeting of the Society for Research in Child Development, Toronto, Canada.
Main, M. (1991). Metacognitive knowledge, metacognitive monitoring, and singular (coherent) vs. multiple
(incoherent) models of attachment: Findings and directions for future research. In C. M. Parkes, J. Steven-
son-Hinde, & P. Marris (Eds.), Attachment across the life cycle. London: Routledge.
Main, M., & Goldwyn, R. (1984). Predicting rejection of her infant from mother’s representation of her own
experience: implication for the abused – abuser intergenerational cycle. International Journal of Child Abuse
and Neglect, 8, 203 – 217.
Main, M., & Goldwyn, R. (1989). Adult attachment rating and classification system. Berkeley, CA: Department of
Psychology, University of California.
Main, M., Kaplan, K., & Cassidy, J. (1985). Security in infancy, childhood and adulthood. A move to the level of
representation. In I. Bretherton, & E. Waters (Eds.), Growing points of attachment theory and research.
Monographs of the Society for Research in Child Development, 50, 66 – 104.
Mallinckrodt, B., Gantt, D. L., & Coble, H. M. (1995). Attachment patterns in the psychotherapy relationship:
development of the client attachment to therapist scale. Journal of Counseling Psychology, 42(3), 307 – 317.
Marrone, M. (1998). Attachment and interaction. London: Jessica Kingsley.
Melges, F. T., & Swartz, M. S. (1989). Oscillations of attachment in borderline personality disorder. American
Journal of Psychiatry, 146(9), 1115 – 1120.
Meyers, S. A. (1998). Personality correlates of adult attachment style. Journal of Social Psychology, 138(3),
407 – 409.
Minnis, H., Ramsay, R., & Campbell, L. (1996). Reactive attachment disorder: usefulness of a new clinical
category. Journal of Nervous and Mental Disease, 184(7), 440.
Persons, J. B., Burns, D. D., Perloff, J. M., & Miranda, J. (1993). Relationships between symptoms of depression
56 I. Goodwin / Clinical Psychology Review 23 (2003) 35–56

and anxiety and dysfunctional beliefs about achievement and attachment. Journal of Abnormal Psychology,
102(4), 518 – 524.
Pilgrim, D. (2001). Too much to expect. Openmind, 107, 8.
Pipp, S., Shaver, P., Jennings, S., Lamborn, S., & Fischer, K. W. (1985). Adolescents’ theories about the develop-
ment of their relationships with parents. Journal of Personality and Social Psychology, 48, 991 – 1001.
Quiery, N. (1998). Parenting and the family. In K. Trew, & J. Kremer (Eds.), Gender and psychology. London:
Arnold.
Roche, D. N., Runtz, M. G., & Hunter, M. A. (1999). Adult attachment: a mediator between child sexual abuse
and later psychological adjustment. Journal of Interpersonal Violence, 14(2), 184 – 207.
Rosenfarb, I. S., Becker, J., & Khan, A. (1994). Perceptions of parental and peer attachments by women with
mood disorders. Journal of Abnormal Psychology, 103(4), 637 – 644.
Sable, P. (1994). Anxious attachment in adulthood: therapeutic implications. Journal of Analytic Social Work,
2(1), 5 – 24.
Sable, P. (1995). Attachment theory and post-traumatic stress disorder. Journal of Analytic Social Work, 2(4),
89 – 109.
Scharfe, E., & Bartholomew, K. (1994). Reliability and stability of adult attachment patterns. Personal Relation-
ships, 1, 23 – 43.
Schreiber, R., & Lyddon, W. J. (1998). Parental bonding and current psychological functioning among childhood
sexual abuse survivors. Journal of Counseling Psychology, 45(3), 358 – 362.
Shaver, P., & Hazan, C. (1988). A biased overview of the study of love. Journal of Social and Personal Relation-
ships, 5, 473 – 501.
Siegelman, M., & Roe, A. (1979). Manual: the Parent – Child Relations Questionnaire—II. New York: City
College of New York.
Starkey, D., & Flannery Jr., R. B. (1997). Schizophrenia, psychiatric rehabilitation and healthy development: a
theoretical framework. Psychiatric Quarterly, 68(2), 155 – 166.
Stevens, N. (1997). Friendship and well being: a course for older women. Tijdschrift voor Gerontologie en
Geriatrie, 28(1), 18 – 26.
Strahan, B. J. (1995). Predictors of depression: an attachment theoretical approach. Journal of Family Studies,
1(1), 33 – 47.
Tennant, C. (1988). Parental loss in childhood. Archives of General Psychiatry, 45, 1045 – 1055.
Tizard, B. (1991). Working mothers and the care of young children. In M. Woodhead, P. Light, & R. Carr (Eds.),
Growing up in a changing society: a reader. Child development in social context 3. London: Routledge.
Trinder, H., Mitchell, S. A., & Todd, L. M.-K. (1994). Valuing patient feedback: I. Shelton hospital’s acute wards.
Clinical Psychology Forum, 70, 6 – 11.
Tyrell, C., & Dozier, M. (1997). The role of attachment in therapeutic process and outcome for adults with serious
psychiatric disorders. Cited in Dozier, M., and Tyrell, C. (1998). The role of attachment in therapeutic
relationships. In J. A. Simpson & W. S. Rholes (Eds.), Attachment theory and close relationships. New York:
Guilford Press.
Vaillant, G., Sobowale, N., & McArthur, B. (1972). Some psychological vulnerabilities in physicians. New
England Journal of Medicine, 287, 372 – 375.
Weiss, R. (1974). The provisions of social relationships. In: Rubin Z. (Ed.), Doing unto others. Englewood Cliffs,
NJ: Prentice-Hall.
Weiss, R. (1991). The attachment bond in childhood and adulthood. In C. M. Parkes, J. Stevenson-Hinde, & P.
Marris (Eds.), Attachment across the life cycle. London: Routledge.
West, M., & Sheldon-Keller, A. (1994). Patterns of relating: an adult attachment perspective. New York:
Guilford Press.
Whiffen, V. E., Judd, M. E., & Aube, J. A. (1999). Intimate relationships moderate the association between
childhood sexual abuse and depression. Journal of Interpersonal Violence, 14(9), 940 – 954.