Attachment & Psychosis: Implications for Therapeutic Alliance


In what they say is the first study to investigate relationships between emotion regulation, attachment and the therapeutic alliance, researchers in Northwest England recruited 49 persons with diagnoses of schizophrenia and assessed the quality of alliance with therapists, psychotic symptoms, positive and negative affect,  emotional regulation, and anxiety and avoidance within attachment relationships. They conclude that “attachment is a useful theoretical construct for understanding psychosis, with evidence for a link between a positive staff–patient relationship and enhanced emotion regulation.” Results will appear in Clinical Psychology and Psychotherapy.

Abstract → 

Related Blogs:
Responding to Madness With Loving Receptivity: a Practical Guide

Related research:
From Infant Attachment Disorganization to Adult Dissociation: Relational Adaptations or Traumatic Experiences?
The Interface Between Attachment and Intersubjectivity:Perspective from the Longitudinal Study of Disorganized Attachment
Quality of Early Care and Childhood Trauma: A Prospective Study of Developmental Pathways to Dissociation

Related “In the News” Items:
APA Proposes Alternative to Juvenile Bipolar
Loneliness and Mental Illness

This study’s introduction:

Cognitive models recognise the role of emotional dysfunction and dysregulation in the development, onset and course of psychosis (Garety, Kuipers, Fowler, Freeman & Bebbington, 2001). However, although cognitive models of psychosis have been of immense value in increasing knowledge of the development and maintenance of symptoms, the precise role of emotion regulation in these processes remains poorly understood.

Emotion regulation is developing as a field of psychological theory and research. It broadly refers to the various strategies individuals use to manipulate their emotional experiences and expressions. Regulatory efforts may be directed at various parts of the emotion process, altering appraisals, arousal levels, thoughts, expressions and actions (Mikulincer & Shaver, 2007).

Attachment theory is a developmental theory that provides a useful framework for understanding the development of emotion regulation across the lifespan (Mikulincer, Shaver & Pereg, 2003). This model views emotion regulation as developing largely in the context of the relationship between child and primary caregiver, with attachment figure availability as one of the major sources of variation in the development of attachment patterns and emotion regulation strategies.

Attachment figure availability facilitates the development of a secure attachment pattern. The caregiver responds promptly, sensitively and consistently to the infant’s needs, and the infant experiences a positive emotional state. Conversely, attachment figure unavailability facilitates the development of an insecure attachment pattern, and the infant experiences an increase in the frequency and intensity of negative emotional states.

There are two principal insecure attachment patterns. The avoidant strategy occurs when caregivers are dismissive of the infant’s distress. The infant learns to inhibit or suppress emotion in order to elicit the caregiver’s response. The anxious strategy occurs when caregivers are inconsistent to the infant’s distress, and the infant learns to maintain or exaggerate emotional distress in order to elicit the caregiver’s response.

Such experiences contribute to the working model of emotion-related expectations that will transfer from the immediate caregiving environment to the larger social world. Studies fairly consistently yield a moderate degree of stability in attachment patterns from infancy to adulthood and throughout the adult years (Fraley, 2002; Waters, Hamilton & Weinfield, 2000); however, discontinuity in attachment patterns can result to changing life experiences (Hamilton, 2000).

Over the last two decades, attachment theory, with its focus on emotional regulation, has begun to make major contributions to our understanding of how earlier caregiving relationships may influence the development of psychopathology. However, there is limited research investigating the relevance of attachment theory to psychosis despite emotional disturbance predating and occurring alongside this illness (Freeman & Garety, 2003).

Contemporary research evidence suggests that there are significantly higher levels of insecure attachment in a psychosis sample when compared with non-clinical controls (Ponizovsky, Nechamkin & Rosca, 2007). Further, individuals with schizophrenia have higher levels of insecure attachment, particularly avoidant attachment, than those with bipolar disorder and major depression (Dozier, 1990; Dozier, Stevenson, Lee & Velligan, 1991).

A number of researchers are beginning to investigate emotion regulation strategies in psychosis samples. This small body of research has largely focused on two specific emotion regulation strategies, cognitive reappraisal and suppression, with mixed results. Individuals with psychosis have been found to use suppressing strategies more frequently and appraisal strategies less frequently than non-clinical controls (Livingstone, Harper & Gillanders, 2009; van der Meer, van’t Wout & Aleman, 2009); however, another study reported that individuals with schizophrenia did not differ from non-clinical controls with regards to the use of suppression or reappraisal strategies (Henry, Rendell, Green, McDonald & O’Donnell, 2008). Such inconsistencies may be due to differences in sample selection criteria, sample size variability and differences in conceptualising emotion regulation. Future research may build upon current findings by assessing other dimensions of emotion regulation.

The ability to identify attachment and emotion regulation patterns in psychosis has important implications in clinical practice. Although it is recognised that not all staff–patient relationships are attachment relationships, because of limited social networks, mental health professionals play a central role in the lives of people with psychosis and have been conceptualised by some authors as key attachment figures for this population (Adshead, 1998). There is emerging evidence of associations between attachment and the concept of the therapeutic alliance in psychosis (Berry, Barrowclough & Wearden, 2008). The therapeutic alliance refers to the quality of the working relationship between client and psychiatric staff. A stronger therapeutic alliance is associated with higher client functioning and lower symptom severity (Neale & Rosenheck, 1995).

One mechanism that may account for the effects of the therapeutic alliance on treatment outcome in psychosis is the ability of the care provider to help the client regulate emotions. In infancy, emotional regulation is acquired through the interactions represented in attentive and sensitive caregiving. The sensitive caregiver recognises changes in emotional states of the infant based on subtle shifts in expression and body language. Signs of distress are attended to and managed before they become overwhelming. Similarly, a sophisticated mental health worker may process details of the client’s posture, gaze and tone of voice, and use this rich information to inform the therapy (Meares, 2005). Through this, the client begins to recognise changes in his/her own internal state and becomes more efficient at managing dysregulating symptomatology.

To date, emotion regulation has not been examined alongside the therapeutic alliance in people diagnosed with psychosis. It is anticipated that the current study may provide insights into the types of therapeutic relationships and interventions that would regulate emotion and modify attachment styles in individuals with insecure attachment.

This study’s conclusions:

The findings of links between attachment and emotion regulation in this sample of individuals with psychosis suggest that attachment theory provides a useful framework in which to conceptualise emotion regulation difficulties in this client group. Associations between attachment avoidance and emotion regulation are consistent with previous research in non-clinical populations, which demonstrate that insecure-avoidant individuals employ avoidant defences, such as denial and repression, to inhibit emotional states that may activate attachment behaviours such as vigilance to the presence of the attachment figure, seeking out the attachment figure and distress following separation from the attachment figure (Shapiro & Levendosky, 1999). Similarly, associations between anxious attachment and emotion regulation reflect previous research demonstrating that people who score high on attachment anxiety tend to react to stressful events with intense distress and ruminate on their emotions instead of focusing on what to do about the problematic situation (Mikulincer & Florian, 1998).

As predicted, the more positive patients perceived their relationship with staff, the less emotion regulation difficulties they reported. From an attachment theory perspective, the child comes to understand his or her emotions by the mother’s ability to accurately ‘reflect back’ her child’s internal experience. This ‘emotional attunement’ or ‘mirroring’ is primarily a non-verbal mode of communication between infant and caregiver, which is essential for optimal emotional development (Schore, 2007). A similar process may therefore be occurring in the patient–staff relationship.

The lack of association between staff and patient ratings of the therapeutic alliance suggests that a mental health worker’s perception of how their client values the relationship may not always be an accurate reflection of the patient’s perception. Previous studies have found only modest correlations between patient and therapist ratings of alliance using psychosis samples, with a tendency for the patient to view the relationship more positively than the therapist (Barrowclough et al., 2010; Couture et al., 2006). In the current study, patient ratings were lower than staff ratings, which may reflect social desirability bias or differences in sample selection criteria given that the staff sample in the current study were key workers.

Attachment anxiety predicted emotion regulation over and above the variance contributed by psychotic symptoms and negative emotion; however, attachment avoidance did not. Attachment was assessed using the two dimensions of anxiety and avoidance; therefore, any associations between an individual’s level of attachment avoidance and specific emotion regulation strategies may be accounted for by their levels of attachment anxiety.

Criticisms of the study design include that it was cross-sectional. It is therefore not possible to determine the direction of associations between variables. From a theoretical perspective, early attachment experiences are more likely to precede and contribute to the development of emotion regulation. Associations between emotion regulation and the therapeutic alliance are more likely to be dynamic and bi-directional, with emotion regulation exerting an influence on the quality of the therapeutic alliance and vice versa.

The authors also acknowledge that the study conducts a number of analyses without correcting for multiple comparisons. As this is the first study to investigate relationships between emotion regulation, attachment and the therapeutic alliance, the hypotheses were considered exploratory in nature and it was felt better not to be overly conservative when conducting the analyses. However, this can be taken into account when interpreting the results without necessarily needing to perform post hoc adjustments such as Bonferroni’s method (Perneger, 1998).

Despite these caveats, this study furthers our understanding of the role of attachment and the therapeutic alliance in individuals with psychosis. It makes an important contribution to the literature by providing preliminary evidence to suggest that attachment and a positive patient–staff relationship facilitates the development of emotion regulation in individuals with psychosis. Following on from this, it would seem logical to consider the implications for therapeutic interventions that are both informed by attachment theory and practical to implement into the wider multidisciplinary team (MDT).

The links between attachment and emotion regulation in psychosis highlight a need to vary therapeutic approaches in accordance with patients’ attachment styles. This may be achieved by managing the ‘therapeutic distance’ (Daly & Mallinckrodt, 2009). For example, a gradual increase in the therapeutic distance is required for patients with anxious attachment who must then manage the resulting frustration. Conversely, therapeutic distance is decreased for avoidantly attached patients who must then manage their fears of intimacy. Increased distance is facilitated by setting firmer boundaries and gratifying patients’ needs less frequently, whereas decreased distance is created by insisting on higher levels of intimacy and focusing on ‘here and now’ aspects of therapy. Attachment theory would predict that changes in attachment style in therapeutic relationships would change an individual’s internal working model of relationships, which would transfer to relationships outside of therapy. There is some evidence that attachment styles can change as a result of therapy (Daniel, 2006); however, this is an area that needs further research.

The links between the therapeutic alliance and emotion regulation in psychosis highlight a need to train MDT staff in ways of facilitating emotion regulation and managing difficulties in attachment relationships. Gumley and Schwannauer (2006) provide a comprehensive treatment model, which incorporates an interpersonal focus into existing evidence-based psychological therapies for psychosis – Cognitive Interpersonal Therapy (CIT). A key feature of this treatment protocol is that it is designed to be incorporated into the wider MDT, and training aimed at encouraging staff to reflect on their own beliefs about psychosis, relapse and recovery is provided.

Developing staff self-reflection would likely promote self-reflection in their patients. For example, key workers who are attuned to their own feelings and internal states may be more likely to facilitate the recognition and labelling of internal states in their patients, ultimately enhancing emotion regulation. This ability to understand the internal states of oneself and others is known as ‘mentalisation.’ The discrepancy between staff and patient ratings of the alliance in the current study might suggest that staff and service users do not talk openly about their feelings in relationships. This ability may be another way of helping clients to recognise their internal states and thereby enhance emotion regulation.

There is emerging evidence that mentalisation deficits exist among individuals in both chronic and early course psychotic illness (Kettle, O’Brien-Simpson & Allen, 2008; Vogely, 2007). A recent case study (Brent, 2009) illustrated the application of mentalisation based approach to treatment to a patient with a psychotic disorder with promising results (Mentalisation Based Therapy: MBT, Bateman & Fonagy, 2006). MBT is undemanding in terms of training and therefore practical to implement into the MDT; however, it remains to be seen whether this therapy will prove effective in clinical trials for psychosis. Furthermore, it will be important to ascertain whether skills developed in therapy will translate into the client’s life outside of the therapy environment.

It is anticipated that future research will strengthen and develop the current study’s findings. Preliminary evidence suggests that the DERS is a useful measure of emotion regulation in psychosis samples; however, further work to establish reliability and validity of the measure in this population is warranted. Experienced Sampling Methods (ESM: Larson & Csikszentmihalyi, 1983) have also been used to study emotion regulation in psychosis samples by recording daily fluctuations in mood (Myin-Germeys, van Os, Schwartz, Stone & Delespaul, 2001), and could be used in studies investigating attachment and the therapeutic alliance in this client group.

Previous articleOn the Importance of Moral Imagination
Next articlePeter Breggin Interviews MIA blogger Dr. Michael Cornwall
Kermit Cole
Kermit Cole, MFT, founding editor of Mad in America, works in Santa Fe, New Mexico as a couples and family therapist. Inspired by Open Dialogue, he works as part of a team and consults with couples and families that have members identified as patients. His work in residential treatment — largely with severely traumatized and/or "psychotic" clients — led to an appreciation of the power and beauty of systemic philosophy and practice, as the alternative to the prevailing focus on individual pathology. A former film-maker, he has undergraduate and master's degrees in psychology from Harvard University, as well as an MFT degree from the Council for Relationships in Philadelphia. He is a doctoral candidate with the Taos Institute and the Free University of Brussels. You can reach him at [email protected].


  1. Doctors don’t realise nowadays the importance of the therapeutic alliance. Only 50 years ago they knew all about it. The personal touch has all but disappeared from medicine and it is a pitty and it is getting worse. Our GP’s surgery is turning into a computerised something-no human touch left. Psychiatrists haven’t got a clue about emotions and anxieties. Sad really.

    Report comment

    • I don’t think it sad, I think what you say about psychiatrists is an indictment. They are charged with looking after people in extreme mental distress and according to you, and this tallies with my personal experience, they haven’t got a clue about emotions and anxieties. Yet they have very nice salaries, with all the benefits of paid holidays and pensions.

      I’m a gardener, I wouldn’t get away with knowing nothing about plants or the basics of how to mow a lawn. I’d be unemployable if I was.

      In this model psychiatrists need retraining or sacking.

      Report comment

  2. “The sensitive caregiver recognises changes in emotional states of the infant based on subtle shifts in expression and body language. Signs of distress are attended to and managed before they become overwhelming. Similarly, a sophisticated mental health worker may process details of the client’s posture, gaze and tone of voice, and use this rich information to inform the therapy (Meares, 2005). Through this, the client begins to recognise changes in his/her own internal state and becomes more efficient at managing dysregulating symptomatology.”

    Or in other words a caring and understanding person can help you calm down and get over the problems you accrued in early life.

    And: “key workers who are attuned to their own feelings and internal states may be more likely to facilitate the recognition and labelling of internal states in their patients, ultimately enhancing emotion regulation.

    Or in other words, someone who knows them self is more able to understand others and this is often the key to feeling cared for and understood.

    However, “This ability to understand the internal states of oneself and others is known as ‘mentalisation.’” – Really? – I thought it was called being sensitive and understanding.

    Report comment