Case vignette 1
Mr G. is a 39-year-old man who has had episodes of schizophrenia since the age of 17. He lives on his own. Since his diagnosis, he has been hospitalised six times because of uncontrolled auditory hallucinations and delusions of persecution; the most recent hospitalisation occurred 2 years ago. For the past 7 years, his key-worker has been Mr S., a community psychiatric nurse with whom Mr G. has formed a trusting relationship. Owing to staff shortage, Mr G. is reviewed medically by the senior house officer (SHO) on the team rather than the consultant; this means that he has to meet a new doctor every 6 months. At Mr G.’s latest out-patient psychiatric appointment, the SHO (who was meeting the patient for the first time) suggested that Mr G.’s antipsychotic medication be increased, as it was reported that his auditory hallucinations were becoming more prominent. Mr G. at first denied the increase in positive symptoms and objected strenuously to the proposed medication change. However, in consultation with Mr S., who attended the appointment with him, he admitted to a deterioration in his mental state and accepted the increase in medication.
Staff as attachment figures
The idea that patients can come to regard mental health professionals as attachment figures (even if for a limited time period) has been argued cogently by a number of authors (Reference BowlbyBowlby, 1988; Reference Dozier, Cue and BarnettDozier et al, 1994; Reference AdsheadAdshead, 1998; Reference GoodwinGoodwin, 2003; Reference WilkinsonWilkinson, 2003), although this has not always been backed up by empirical research. Psychiatric staff who function as caregivers may play an important role in providing both a secure base for patients whose attachment needs are activated during periods of distress and corrective experiences that disconfirm patients’ insecure ‘internal working models’ of attachment relationships (see Reference MaMa, 2006), thus enabling more secure ways of interacting with others. The case vignette of Mr G. with which I opened this article illustrates the secure base function that can be fulfilled by individual keyworkers. During his illness, Mr G. has met a series of different doctors at different stages of their training. His attachment system is activated during an appointment with another new doctor, which makes it difficult for him to accept medical advice from this relative stranger. However, the presence of Mr S., with whom he has developed a secure attachment over the years, enabled Mr G. to accept the change in medication. This might not have occurred had the keyworker not been present in the consulting room.
Case vignette 2
Mrs A. is a 40-year-old woman who was assessed by psychiatric services following a paracetamol overdose. The overdose was precipitated when her husband of 20 years left her for another woman. She was diagnosed with a moderate depressive disorder, with prominent anxiety features. The only psychiatric history of note was one previous overdose 15 years earlier, during a brief marital separation. However, Mrs A. described herself as having always been anxious. There was also a significant family history of anxiety disorders; her mother suffered agoraphobia for as long as Mrs A. could remember. At the point of engagement with services, Mrs A. had been isolating herself socially. She also admitted that she had for some time lived in constant fear of her husband leaving, describing him as ‘cold’ and ‘distant’.
Mrs A. was put on antidepressant medication, and was followed up jointly by a trainee psychiatrist and a community psychiatric nurse, Ms T. The first antidepressant tried did not help; the second was of limited benefit. The patient developed a number of physical symptoms about which she would telephone Ms T. on a regular basis. The mental health team thought that these symptoms were psychosomatic. For a while, despite a colleague's concerns that she would foster overdependence in the patient, Ms T. decided to increase the frequency and regularity of her appointments. Mrs A. was appreciative of this, and the calls in between appointments reduced. She was also more able to work through her grief over the marital separation during their appointments, as she developed a sense of trust in the availability and responsiveness of her keyworker. Some work was also done in re-establishing Mrs A.’s neglected social relationships.
Mrs A.’s case vignette illustrates the provision of a corrective experience. Judging from her upbringing in an anxiety-ridden family, and her premorbid personality, it is quite possible that Mrs A. might be ‘preoccupied’ in her state of mind with respect to attachment (Reference Main and GoldwynMain & Goldwyn, 1998) (see Reference MaMa, 2006: Table 2). Clinicians might regard patients such as Mrs A. as ‘clingy’ and ‘overdependent’, as demonstrated by the feelings of Ms T.’s colleague. Under the stress of marital separation, Mrs A.’s overdoses could be construed as attempts to re-establish proximity to her secure base. When the secure base of her husband was lost (however inadequate he might have been in this role), Ms T., the keyworker, took over this function. The multiple psychosomatic symptoms, likewise, could be seen as attachment behaviours (Reference Schuengel and van IJzendoornSchuengel & van IJzendoorn, 2001) with the same purpose of maintaining proximity to the attachment figure.
Ms T. seemed to have intuitively understood Mrs A.’s needs. By providing extra regular appointments, the keyworker did not ‘reject’ Mrs A. as her husband might perhaps have done. Thus, Ms T. was able to provide a corrective attachment experience for her patient (Reference Lieberman, Zeanah, Cassidy and ShaverLieberman & Zeanah, 1999). This resulted in fewer unscheduled telephone calls, and perhaps even led to a lasting revision of Mrs A.’s internal working models of attachment. With Ms T. acting as a secure base, Mrs A. was able to explore her grief over her marital breakdown. This ability to explore in the presence of one's secure base has been emphasised since the early days of attachment theory (Reference Cassidy, Cassidy and ShaverCassidy, 1999).
When is a therapeutic relationship an attachment relationship?
An insightful critique of the assumption that psychiatric staff can function as attachment figures has been provided by Reference Schuengel and van IJzendoornSchuengel & van IJzendoorn (2001), who start by questioning whether relationships that patients form with staff can indeed be seen as attachment relationships. The answer appears to be that some therapeutic relationships are attachment relationships, whereas some are not:
‘[a] relationship with a staff member would only be called an attachment relationship … if using the staff member as a secure base would be characteristic of their “history of interactions” that is, would be displayed during an extended period of time’ (Reference Schuengel and van IJzendoornSchuengel & van IJzendoorn (2001: p. 308).
What constitutes an ‘extended period of time’ is not defined. They also discuss the importance of attachment bonds (that is, one person's experience of feeling attached to another). The characteristics of attachment bonds are summarised in Box 1. It is easy to see that bonds forged in some therapeutic relationships can possess most (but perhaps not all) of these characteristics. However, this begs the question of whether all the listed characteristics are necessary for a therapeutic bond to be classed as an attachment bond and the therapeutic relationship consisting of such a bond to be classed as an attachment relationship.
Box 1 Characteristics of attachment bonds
-
• Attachment bonds are persistent
-
• Attachment bonds involve a specific figure who is not interchangeable
-
• The relationship within the dyad is emotionally significant
-
• The individual wishes to maintain proximity to or contact with the attachment figure
-
• The individual feels distress at involuntary separation from the attachment figure
-
• The individual seeks security and comfort in the relationship with the attachment figure
Reference Schuengel and van IJzendoornSchuengel & van IJzendoorn (2001) pose three questions about the circumstances under which attachment relationships are more or less likely to form in the institutional setting:
-
• Is the patient cut off from their pre-existing network of attachment relationships, such that new attachments are needed?
-
• Is their attachment system activated, or can they cope adequately with their own internal resources?
-
• Is a stable secure base feasible within the particular healthcare setting?
The last question is especially pertinent to the organisation of mental healthcare, given that a significant number of psychiatric patients will need to form more than one new attachment as their journey begins in the mental healthcare system. In a study of missed appointments within a community mental health team over a 21-month period, Reference McIvor, Ek and CarsonMcIvor et al(2004) found that mean monthly patient non-attendance rates for appointments with psychiatric trainees were roughly double those for the two consultant psychiatrists, which were in turn double that for the consultant clinical psychologist. The reasons for non-attendance were not examined, and there might have been several explanations for the findings. However, it is possible that, through their greater experience of and longer periods of contact with senior clinicians, patients found it easier to form attachment relationships with these professionals than with more junior clinicians. Indeed, as Mr G. in Case vignette 1 demonstrated, the stress of seeing a new clinician may well activate the patient's attachment system, with consequences that might partially be predicted by the patient's state of mind with respect to attachment (see below).
It is assumed that corrective attachment experiences are provided within one-to-one relationships. In this respect, Reference GoodwinGoodwin (2003) has argued that the concept of individual keyworkers in community teams is positive in attachment terms, whereas the practice of many assertive outreach teams of not assigning keyworkers may be less so. The question here is whether patients can form an attachment to a team (or institution) rather than to an individual; thus far this has been little explored. To facilitate empirical research in this area, Goodwin and colleagues developed the 25-item Service Attachment Questionnaire (SAQ; Reference Goodwin, Holmes and CochraneGoodwin et al, 2003). Focus groups of existing service users participated in the initial development process. The sub-scales identified on the SAQ (e.g. consistency, comfort) correspond well to dimensions of attachment found in the literature. In the validation part of the study, the researchers found that psychiatric in-patients scored lower on the SAQ (i.e. felt that the in-patient service was less able to meet their attachment needs) than three other groups of clients (users of community health team, psychological and day centre services), a finding possibly explained by the fact that in-patients come into contact with a wide number of workers on different shifts and that one of the focus of weekly ward rounds is the discussion of discharge. The SAQ requires further validation, and the hypothesised correlation between clients’ responses on the SAQ and their attachment style needs to be tested empirically. Nevertheless, it is clear that how mental health services are organised may have significant effects on whether the attachment needs of patients can be met.
Case vignette 3
Ms J. is a 22-year-old with a diagnosis of borderline personality disorder, an extensive history of being sexually and physically abused, and a history of repeated presentation following threatened or actual self-harm. Until recently, she was well-known to the trainee psychiatrists on the on-call rota, who were called regularly to assess her mental state out-of-hours at the local accident and emergency department. Sometimes, apparently depending on the experience (and personality) of the assessing doctor, Ms J. would be admitted overnight to a psychiatric ward, only to be discharged the next day by her local consultant. A plan eventually was drawn up by the team not to admit Ms J. psychiatrically on an emergency basis. However, this apparently caused her self-harming behaviour to increase, culminating in a significant paracetamol overdose that resulted in a psychiatric admission.
Given her past history and current presentation, it is likely that Ms J. would be classified as having an unresolved/preoccupied attachment if Reference Main and GoldwynMain & Goldwyn's (1998) Adult Attachment Interview were administered (Reference Fonagy, Leigh and SteeleFonagy et al, 1996; Reference HolmesHolmes, 2004). On the one hand, rejection would serve to reinforce her attachment behaviour (i.e. self-harming), while on the other this same behaviour is likely to lead to strong countertransferential reactions in mental health professionals – some of which may be negative and rejecting. A vicious cycle is thus established. Some members of the ward staff believe that her ‘attention-seeking’ behaviour should be counteracted by ignoring it. However, in light of her significantly rejecting and abusive early attachment experiences, her current way of interacting with significant others (including mental health staff) may be the only way she knows; her internal working models are highly maladaptive.
Ms A. is referred to the local dialectical behavioural therapy programme, with the recognition that the crucial establishment of a working alliance and any therapeutic change will occur only after some time.
Patient attachment, the working alliance and how patients make use of treatment
If the assumption that patients may form attachments to mental health staff is correct, it follows that patients’ unconscious attachment strategies might influence the working alliance that they form with staff. This has received some empirical support in the psychotherapeutic literature (Reference Dolan, Arnkoff and GlassDolan et al, 1993; Reference Eames and RothEames & Roth, 2000). In a study of 30 clinical psychology out-patients, Eames & Roth found that patients classified as fearful using Bartholomew's classification (Reference Bartholomew and HorowitzBartholomew & Horowitz, 1991), i.e. those with negative internal working models of both self and others, had the greatest difficulty forming a therapeutic alliance. However, psychotherapeutic relationships may well be more intense than the therapeutic relationships formed within the general adult psychiatric setting, and thus may be more likely to become attachment relationships for psychotherapy clients.
Dozier and colleagues in the USA have undertaken a series of different research projects looking at the influence of attachment on the relationship between patients with serious psychopathology and their case managers (Reference DozierDozier, 1990; Reference Dozier, Cue and BarnettDozier et al, 1994; Reference Tyrrell, Dozier and TeagueTyrrell et al, 1999). Case managers share some of the characteristics of community keyworkers in the UK, so this research may be relevant to the British setting. The studies, which included significant proportions of people with schizophrenia or bipolar disorder, made use of the Adult Attachment Interview (Reference Hesse, Cassidy and ShaverHesse, 1999; discussed in Reference MaMa, 2006). However, their use of the Q-set rating method (further details available on request) meant that unresolved attachment was not considered; the contribution of this attachment classification, which might be especially associated with psychopathology, could therefore not be determined. The studies also involved small numbers of participants, in common with other work using the Adult Attachment Interview.
Dozier and her team found that insecure individuals were less likely than secure patients to take their psychotropic medication as prescribed (Reference DozierDozier, 1990). Furthermore, those with deactivating (i.e. avoidant or dismissing) strategies were more likely to show greater rejection of treatment providers, less self-disclosure and poorer use of treatment. Medication adherence was not identified as a greater problem in avoidant than in preoccupied patients. Given the over-representation of insecurely attached individuals in clinical populations (Reference van IJzendoorn and Bakersman-Kranenburgvan IJzendoorn & Bakersman-Kranenburg, 1999), these findings have important implications for the treatment of psychiatric patients. According to Reference DozierDozier (1990), clients who use avoidant strategies might be helped by treatment that allows more ‘interpersonal distance’, whereas preoccupied clients (e.g. Mrs A. in Case vignette 2) might benefit from more interaction and supervision. Dozier spoke especially of the challenges of working with clients with avoidant attachment strategies, which she regards as self-perpetuating:
‘[avoidant] strategies are designed to suggest that the individual does not need anything from the attachment figure … If the clinician responds to the client's self-presentation by withdrawing help, the client loses needed support … In addition, the experience of having support withdrawn will confirm the client's expectancies that others are unavailable … and perpetuate the help-rejecting strategies employed’ (Reference DozierDozier, 1990: p. 57).
Unfortunately, medication adherence was clinician-rated in Dozier's study and did not appear to be assessed using objective criteria. Whether adherence is affected by, among other factors, the patient's attachment insecurity is an important clinical question. Evidence from the diabetes literature has shown that a dismissing attachment style (using Bartholomew‘s classification) is associated with poorer medication adherence and significantly higher mean glycated haemoglobin HbA1c levels, especially in patients who rate their clinician's communication as poor (Reference Ciechanowski, Katon and RussoCiechanowski et al, 2001). Dismissing individuals have a defensively positive internal working model of the self and a negative internal working model of others as attachment figures; they minimise the need for attachments (Reference Bartholomew and HorowitzBartholomew & Horowitz, 1991), and may have difficulty forming a trusting patient–clinician relationship. Research now needs to address whether a similar association is seen in patients on psychotropic medication.
Attachment strategies of staff and effects on treatment
The relationships that staff form with clients cannot be construed as attachment relationships, despite the possibility of the converse. Generally, clinicians do not (or should not) look to their patients to fulfil their own attachment needs. Nevertheless, the state of mind of mental health professionals with respect to their own attachments may be relevant, if staff are expected to fulfil the dual function of providing a secure base for and trying to modify the internal working models of their clients. Part of the clinician's task is to help clients to modify their strategies for approaching interpersonal relationships and regulating emotions, thus promoting more adaptive functioning (Reference Tyrrell, Dozier and TeagueTyrrell et al, 1999: p. 726). The attachment strategies of staff may partially underlie what is understood as countertransference (Reference GoodwinGoodwin, 2003), at least insofar as the term can denote feelings aroused by a patient in the therapist owing to the latter‘s own unresolved conflicts and problems (Reference Brown and PedderBrown & Pedder, 1991).
Reference Dozier, Cue and BarnettDozier et al(1994) studied 27 patients (8 with paranoid schizophrenia and 9 with bipolar disorder) and 18 case managers. They found that, regardless of client attachment, case managers with preoccupied attachment strategies intervened more intensively with clients than case managers who had dismissing strategies. The authors commented that ’variability along the preoccupied dimension appears to be associated with countertransference in which the clinicians’ own issues predominate’ (p. 799). Attachment theory may in time offer a valuable framework for psychiatrists in understanding their own reactions within the doctor–patient relationship and in preventing potential therapeutic pitfalls (Reference WilkinsonWilkinson, 2003). Reference GoodwinGoodwin (2003) noted that a substantial minority of clinicians may have insecure attachment and may be seeking to provide professional care for others by way of compensation. However, this contention is not backed by directly applicable empirical evidence. Clearly, not all clinicians can or would want to take the Adult Attachment Interview, and self-report measures have not as yet gained wide currency. There would also be ethical considerations in routinely assessing the attachment of clinicians, depending on how this information is used. Reference WilkinsonWilkinson (2003), however, emphasises the need during training for all doctors (not just psychiatrists) to learn about ‘their own attachment strategies, and perhaps how their own traumatic experiences can come to affect their capabilities’ (pp. 239–240).
Case vignette 4
Dr C. has just started as a senior house officer on the ward where Ms J. (see Case vignette 3) is still an in-patient. This is her second job in psychiatry. In Dr C.’s previous posting, her consultant noted that she had an excessive tendency to identify with clients emotionally. On one occasion she was discovered to have made a disclosure to a patient about her own history of mental illness (although there was little to suggest that Dr C. was experiencing a psychiatric disorder at the time). The College tutor was alerted. Beyond being counselled unequivocally on the inappropriateness of her disclosure, no further action was taken.
Soon after the start of her new post, Dr C. is noted to ‘take her patients’ problems home’ with her. She forges a particularly strong alliance with Ms J., for whom she becomes a steadfast advocate during ward reviews, to the bewilderment of other staff. After consulting with some team members who have raised concerns, her new consultant decides to speak to Dr C. about her overinvolvement with patients. Dr C. becomes highly anxious at this ‘criticism’, but undertakes to maintain professional boundaries. Soon afterwards, Dr C. begins to disclose her history of neglect and emotional abuse to a couple of recently qualified nurses with whom she is forming a good relationship; these nurses are unsure how to handle the information, being unwilling to disclose it to senior colleagues because of a felt need to maintain Dr C.’s confidentiality. Of even more concern is the extent to which Dr C. has begun to rely on these nursing colleagues for emotional support. Questions again arise as to how much personal information is being divulged to Ms J.
The above is a case of a mental health professional inappropriately using other staff, and potentially a patient, as a secure base for her unfulfilled attachment needs. Although the case vignette represents an extreme scenario, clinicians reading this article could perhaps think of colleagues whose insecure attachment strategies might to a greater or lesser extent impinge on their ability to carry out their professional duties. In this context, Reference WilkinsonWilkinson (2003) discusses the compulsive caregiver. In Patricia Crittenden's ‘dynamic-maturational’ approach to attachment classification (Reference Crittenden, Goldberg, Muir and KerrCrittenden, 1995), ‘compulsive caregiving’ is one of the dismissing attachment strategies. It is thought to evolve from rejection by a ‘sick’ (e.g. depressed) parent. The child protects and cares for the parent at the expense of his or her own needs, thus maximising closeness to the parent (Reference WilkinsonWilkinson, 2003). There is in effect role reversal. The compulsive caregiving strategy may survive into adulthood, and may dictate the choice of a career in the caring professions. In the healthcare context, Wilkinson believes that staff using this strategy make enormous efforts for their patients, but are vulnerable to burnout, sometimes without warning (p. 133). Furthermore, they are more likely to blame themselves than system errors for their professional failures. Clinical supervision is one forum where the influences of a clinician's attachment pattern on their work may be addressed.
Interaction between the attachment strategies of patients and staff
In a study of 54 clients with serious psycho-pathology and their 21 case managers, Reference Tyrrell, Dozier and TeagueTyrrell et al(1999) hypothesised that clients would have better outcomes when managed by clinicians with dissimilar attachment strategies along the hyper-activation–deactivation spectrum, as assessed on the Adult Attachment Interview using Kobak‘s Q-set. For instance, the pairing of a more deactivating client with a less deactivating case manager would be more productive than if the clinician also scored highly on deactivation. (High deactivation corresponds in this context to dismissing attachment, low deactivation to attachment preoccupation.) The rationale was that a case manager who possessed a non-complementary attachment pattern would best be able to disconfirm the client's usual interpersonal and emotional strategies. If, on the other hand, the attachment patterns of the clinician and client are complementary, the client's internal working models may be perpetuated (Reference Dozier, Cue and BarnettDozier et al, 1994).
When the data were analysed, statistically and clinically significant interactions occurred for the working alliance and the client‘s quality of life (Reference Tyrrell, Dozier and TeagueTyrrell et al, 1999). Less deactivating case managers formed stronger alliances with more deactivating clients than with less deactivating clients. Compared with less deactivating clients, more deactivating clients reported higher general life satisfaction when working with less deactivating case managers. There were no significant interactions between client and clinician attachment deactivation when hospitalisation and depressive symptoms were considered, although the study may not have possessed enough statistical power to detect these differences.
The authors’ positive conclusions regarding their original hypothesis were perhaps stronger than their results allowed. Indeed, Reference WilkinsonWilkinson (2003) appears to hold a different viewpoint to that of Tyrrell and colleagues: secure clinicians may be expected to respond most flexibly to their patients’ varying attachment needs. (Tyrrell et al did not consider the effect of secure v. insecure attachment in their data analysis.) On the basis of clinical experience, Wilkinson believes that ‘insecure’ clinicians may actually do better with clients who have similar (rather than dissimilar) attachment strategies along the deactivation–hyperactivation dimension. Further empirical work is needed in this area to clarify which hypothesis is correct. What is clear is that the interaction of clinician and patient attachment strategies may well have important clinical consequences.
Conclusions
The main emphasis of my two articles has been on the potential usefulness of attachment theory in the general adult psychiatric setting. Attachment theory may afford valuable insight into the developmental trajectories of at least some common psychiatric disorders. Although most research in this area has focused on depressive disorder, the aetiology of anxiety, eating and personality disorders may also benefit from being examined through the lens of attachment theory. At the same time, the theory has been used to further our understanding of the therapeutic relationship, with empirical work beginning to demonstrate the importance of both client and clinician attachment within this relationship. More rigorous research is needed in all areas covered by these two articles. None the less, it is hoped that readers, if they were unfamiliar with the attachment literature before, will begin to appreciate the value that an attachment perspective may bring to clinical practice.
Declaration of interest
None.
MCQs
-
1 Attachment theory:
-
a promotes a developmental understanding of psycho-pathology
-
b cannot be applied to romantic relationships
-
c is the only theory for understanding the doctor–patient relationship
-
d has little relevance to the organisation of mental health services
-
e is irrelevant to the understanding of help-seeking for mental health problems.
-
-
2 The following are correctly matched:
-
a Bowlby: Service Attachment Questionnaire
-
b Bartholomew: childhood attachment research
-
c Brown & Harris: Adult Attachment Interview
-
d Ainsworth: Strange Situation Procedure
-
e Main: Relationship Styles Questionnaire.
-
-
3 A clinician's attachment pattern:
-
a has no relevance to the countertransference
-
b may affect how he or she acts towards clients
-
c cannot change with personal experiences
-
d is determined solely by clinical training
-
e should be routinely measured.
-
MCQ answers
| 1 | 2 | 3 | |||
|---|---|---|---|---|---|
| a | T | a | F | a | F |
| b | F | b | F | b | T |
| c | F | c | F | c | F |
| d | F | d | T | d | F |
| e | F | e | F | e | F |
eLetters
No eLetters have been published for this article.