Adult attachment representation moderates psychotherapy treatment efficacy in clinically depressed inpatients

Journal of Affective Disorders 195 (2016) 163–171

Contents lists available at ScienceDirect

Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad

Research paper

Adult attachment representation moderates psychotherapy treatment efficacy in clinically depressed inpatients I. Reiner a,n, M.J. Bakermans-Kranenburg b, M.H. Van IJzendoorn b, E. Fremmer-Bombik c, M. Beutel a a

University Medical Center Mainz, Department of Psychosomatic Medicine and Psychotherapy, Untere Zahlbacher Str. 8, 55131 Mainz, Germany Center for Child and Family Studies, Rommert Casimir Institute for Developmental Psychopathology, Leiden University, P.O. Box 9555, 2300 RB Leiden, The Netherlands c Clinic for Child Psychiatry, Psychosomatic and Psychotherapy, Universitätsstraße 84, 93053 Regensburg, Germany b

art ic l e i nf o

a b s t r a c t

Article history: Received 20 September 2015 Received in revised form 29 January 2016 Accepted 6 February 2016 Available online 9 February 2016

Background: We explored in a sample of clinically depressed patients the influence of attachment security and unresolved trauma on psychotherapeutic outcome as well as changes in attachment representation through psychotherapeutic intervention. Methods: The sample consisted of 85 women (aged 19–52), 43 clinically depressed patients from a psychosomatic inpatient unit, and 42 healthy control subjects matched for age and education. Average length of hospitalization in the patient group was eight weeks. Attachment representations were assessed with the Adult Attachment Interview at the time of admission (baseline) and at discharge. Depressive symptoms were measured using the PHQ-9 at T1 and T2. Results: Insecure attachment representations were overrepresented in depressed patients. Treatment effects were moderated by baseline attachment representation: patients with higher attachment security scores at admission benefited more from the inpatient treatment and were less depressed at time of discharge than less secure patients (η2 ¼.07). Generally, attachment security increased (η2 ¼.19) and depressive symptoms decreased (η2 ¼ .23) after inpatient psychotherapy treatment in the patient group. No significant effects for unresolved symptoms were found. Limitations: The study is not a randomized controlled study, but used a quasi-experimental matched control group design with female subjects only. Conclusions: Our results suggest that attachment representations play a major role in both the development and treatment of clinical depression. Baseline attachment security may influence psychotherapeutic outcome, perhaps through relational factors such as therapeutic working alliance. Inpatient psychotherapy may also need to address psychological issues associated with depression such as attachment insecurity. & 2016 Elsevier B.V. All rights reserved.

Keywords: Adult attachment Depression Psychotherapy Coherence AAI In-patients

1. Introduction Childhood trauma and negative attachment experiences have been found to be major risk factors for depression in adulthood. With depressed adult patients, psychotherapy aims at symptom reduction and increase of psychosocial functioning. Empirically, however, little is known about the impact and modification of attachment representation in the process of inpatient psychotherapy. The objective of the present study is to test whether attachment representation and unresolved attachment trauma moderate the psychotherapeutic impact on depressive symptom n

Corresponding author. E-mail address: [email protected] (I. Reiner).

http://dx.doi.org/10.1016/j.jad.2016.02.024 0165-0327/& 2016 Elsevier B.V. All rights reserved.

levels. Furthermore, we examine whether psychotherapy improves the representation of early attachment experiences in the direction of attachment security, and whether it decreases the level of unresolved trauma symptoms. With a life-time prevalence between 10% and 15% in the general population major depressive disorder is one of the most frequent and debilitating diseases worldwide, with highly adverse impact on the subjects' quality of life in domains such as psychosocial functioning, work productivity, relationship quality, and mortality risk (Lepine and Briley, 2011). Women are at higher risk to suffer from depression than men, with female-male prevalence ratios around 2:1 (Burt and Stein, 2001). Causes of depression are multifaceted and based on a complex interplay of genetic and environmental influences (Karg et al., 2011; Sullivan et al., 2000), yet childhood trauma, insecure attachment and maladaptive

164

I. Reiner et al. / Journal of Affective Disorders 195 (2016) 163–171

relationship patterns have been found to be significant risk factors in the development and perpetuation of depression (Kendler et al., 2004; Wiersma et al., 2009). John Bowlby, the founder of attachment theory, stressed the importance of early attachment experiences on mental health and conceptualized attachment as a lifespan development concept in order to explain “the many forms of emotional distress and personality disturbance, including anxiety, anger, depression, and emotional detachment, to which unwilling separation and loss give rise” (Bowlby, 1979). In childhood and in adulthood, theory as well as research underlines that differences in internal working models of attachment mirror differences in emotion regulation: While secure individuals are (mostly) effective in emotional regulation when feeling distressed, insecure individuals cope less well with distress. In particular, insecure-dismissing adults tend to deny negative emotions and attachment needs, and insecurepreoccupied adults are overwhelmed by them with maximizing attachment signals as assessed with the Adult Attachment Interview, the AAI (Hesse, 2008; Main et al., 1985). Adults with unresolved attachment representation fail to regulate emotions effectively when talking about an experienced loss or trauma in the AAI (Bakermans-Kranenburg and Van IJzendoorn, 2009; Hesse, 2008). In line with the expectation, unresolved symptoms and unresolved attachments are strongly overrepresented in abused individuals and those suffering from PTSD, but-surprisingly-not in depressed samples. Depressive symptoms seem related to attachment insecurity in a non-systematic manner; they were found to be associated with insecure-dismissing as well as insecurepreoccupied attachment representations (Bakermans-Kranenburg and Van IJzendoorn, 2009). It should be noted that when adult attachment style was assessed via self-report questionnaires in terms of a specific relationship quality or schema, insecure attachment style was strongly linked with depressive symptoms (Marganska et al., 2013; Roelofs et al., 2011), but attachment style and representation are considered different, independent dimensions of adult attachment (Crowell et al., 2008; Roisman et al., 2007). In the current study we focus on attachment representations. Psychotherapy with its “purpose of assisting people to modify their behaviors, cognitions, emotions, and/or other personal characteristics in directions that the participants deem desirable” (Norcross, 1990, page 218) is highly relevant for the clinical application of attachment: Bowlby (1988) stressed the dynamic nature of internal working models, suggesting that the attachment system displays significant continuity but is also open to modification and change, for instance through psychotherapy, that may help the individual to create a more coherent autobiographical narrative. The significance of coherence for mental health has been already noted by Freud (1997), who observed that in patients autobiographical narratives “connections-even the ostensible onesare for the most part incoherent, and the sequence of different events is uncertain" (Freud, 1997, p.10). Empirically, psychotherapy improves narrative coherence about emotionally relevant events (Adler et al., 2013). Fonagy and colleagues (Fonagy et al., 2002, 1991b; Fonagy and Target, 2005) extended attachment theory and modern psychoanalytic theory to include the concept of mentalization (Fonagy et al., 1998). Mentalization is strongly linked with attachment and they develop hand in hand (Bowlby, 1988): parents' mindful and sensitive parenting increases both secure attachment and mentalization in a child (Bowlby, 1979, 1988; Fonagy et al., 1991a; Meins et al., 2001; Slade, 2005). Research in clinical adult samples support the link between insecure attachment and poor mentalization, and mentalization-based treatment (MBT) has been applied with some success in the treatment of Borderline Personality disorders (Bateman and Fonagy, 2009). Also

other diagnostic patient groups having experienced adverse and/ or traumatic interactions with attachment figures may benefit from mentalization-focused psychotherapeutic approaches as mentalization may “buffer” the effect of adverse childhood experiences on mental health (Outcalt et al., 2016), possibly through reflecting on and thus altering internal working models (Bowlby, 1979; Fonagy and Bateman, 2006). Internal working models can be viewed as relationship or interpersonal schemata (Bretherton and Munholland, 2008), and, although differing in treatment concepts and tools, evidencebased psychotherapies have in common that they aim at interpersonal problems and the revision of maladaptive relationship patterns. In the treatment of nonpsychotic depressive disorders, patients with adverse childhood experiences and trauma have been shown to benefit at least as much from psychotherapy as from pharmacotherapy (Nemeroff et al., 2003). Most patients with major depressive disorder can be treated in an outpatient setting, but severe and complex cases do not respond to outpatient treatment (Cuijpers et al., 2014) and may need inpatient treatment (Harter et al., 2010). Psychosomatic inpatient treatment programs in Germany are characterized by a strong focus on psychotherapy as the main treatment modality, with outcomes of overall medium to large effect sizes (Liebherz and Rabung, 2013). Specifically for clinically depressed females, Franz et al. (2015) reported in a naturalistic multi-center study including 15 German psychosomatic hospital units robust positive treatment outcomes and invited researchers to further explore predictors of therapeutic effects. Psychotherapy studies including the AAI indicate that symptom reduction is linked to changes in attachment narratives: in a sample of women suffering from PTSD, unresolved symptoms in the AAI were strongly associated with PTSD avoidant symptoms, and addressing imaginal exposure therapy led to both decreases of unresolved symptoms in the AAI and of PTSD symptoms (StovallMcClough and Cloitre, 2006). Harari et al. (2009) distinguished in a sample of Dutch veterans with and without PTSD between deployment-related and non-deployment related unresolved symptoms and found deployment-related, but not non-deployment related unresolved symptoms associated with PTSD symptom severity. In a sample of borderline patients comparing transferencefocused treatment, dialectical behavior therapy and supportive psychotherapy, transference-focused treatment focusing on attachment themes led to significant increase in attachment security as well as attachment-related metacognition and mentalization capacities (reflective functioning) one year after treatment (Levy et al., 2006), however, no changes in unresolved symptoms were reported. From these reports, it remains unclear whether unresolved symptoms in the AAI are directly linked to clinical symptoms and how symptom improvement is related to changes in unresolved symptomatology. In clinically depressed mothers, parent-child psychotherapy led to both higher reflective functioning in the AAI as well as more positive relationships to their toddlers (Toth et al., 2008, 2006). Psychotherapy may impact the patient's attachment narratives and therefore move toward a more secure attachment orientation: The therapist as attachment figure assists the patient with his emotional regulation and reflection on past and current attachment relationships, including the one with the therapist – provided that the patient can use the therapist as “secure base” and form a working alliance with the therapist (Bowlby, 1979, 1988; Byng‐Hall, 1995; Dozier and Tyrrell, 1998). Several studies on patients' attachment and therapeutic alliance employing attachment-related self-report assessments reported that patients' greater attachment security is linked with better therapeutic working alliance, while attachment insecurity (anxious and avoidant attachment styles) was associated with lower working

I. Reiner et al. / Journal of Affective Disorders 195 (2016) 163–171

alliance (Bernecker et al., 2014; Diener and Monroe, 2011). So far only one study (Fonagy et al., 1996) systematically involved the AAI in a (psychoanalytically oriented) inpatient psychotherapy setting with nonpsychotic patients suffering from different mental disorders (e.g. affective, personality, eating disorders). Baseline insecure-dismissing patients showed the highest degree of symptom improvement. Furthermore, depressive and dysthymic patients were at baseline mostly preoccupied and unresolved in the AAI (Fonagy et al., 1996). It remained unclear, how symptom reduction was linked to changes in the AAI or how initial attachment status influenced treatment process variables. In sum, these studies emphasize that attachment theory may facilitate the understanding of psychopathological development and might be also a useful tool to explore differences in treatment process and outcome. The evidence-base is, however, rather poor and more treatment studies are urgently needed. In a sample of depressed females with a carefully matched mentally healthy control group, we explored the impact of and changes in attachment representations assessed with the AAI before and after psychotherapy inpatient treatment. Based on previous studies, we expected that in our depressed sample insecure attachment would be overrepresented at baseline. We also expected a significant decrease in depressive symptomatology after inpatient treatment. Assuming that attachment security is related to better therapeutic working alliance and that unresolved trauma may hinder treatment processes, we expected a) symptom reduction after treatment to be moderated by higher baseline attachment security at time of admission; and b) symptom reduction after treatment to be moderated by lower unresolved symptoms at time of admission. Next, we hypothesized that through addressing interpersonal problems, losses, traumata as well as adverse early attachment experiences in psychotherapy inpatient treatment c) attachment security would increase and unresolved symptoms would decrease. Finally, we explored if potential increases in attachment security and decreases in unresolved symptoms were associated with decreases in depressive symptoms.

2. Methods 2.1. Participants and procedure The total sample consisted of 85 women aged 19–52 (M¼30.1 years, SD ¼9.0 years). 58% (n¼ 49) of the participants had a high school degree, 38% (n ¼33) secondary education and 4% (n ¼3) lower secondary education. All participants were Caucasian and fluent in German language. The clinical sample consisted of 43 depressed patients, who were recruited from the inpatient unit of a Department of Psychosomatic Medicine and Psychotherapy in Rhineland-Palatinate, Germany. The control group involved 42 healthy control subjects who were recruited by flyers and posters and were matched for age and education. Inclusion criteria for the patient group were the diagnosis of Depression and/or Dysthymia using the German version of the Structured Clinical Interview for DSM-IV Disorders; SCID-I and SCID-II; (Wittchen et al., 1997). Patients with borderline, antisocial or narcissistic personality disorders as well as patients with psychotic disorders or eating disorders or substance abuse were not eligible. In the control group, women who were currently in psychotherapy or had a mental disorder diagnosis (using SCID-I and SCID-II) were excluded from further study participation. Moreover, pregnant and breastfeeding women and subjects who suffered from adrenocorticotrophic, gynecological and neurological diseases were excluded. The patient group had the following main diagnoses: Major depression (n ¼42), dysthymia (n ¼1). Comorbid conditions were anxiety disorders (n¼ 18), dysthymia (n¼ 8), and depersonalization

165

disorder (n ¼4). Data collection for the patient group was performed at admission (T1) and discharge (T2). Treatment duration varied between 5 and 12 weeks (M ¼8.3, SD ¼1.7). All 85 study participants filled out self-report questionnaires on depressive symptoms (PHQ-9) at T1 and T2. Two patients terminated treatment prematurely and one subject from the control group dropped out, therefore their adult attachment representations were not available at T2. Participants from the control group were matched for age (7 3 years), education (school degree) and time interval between the two assessments. The study was approved by the Ethics Committee of the State Board of Physicians of Rhineland-Palatinate (Germany). All participants provided their written informed consent to participate in this study. 2.2. Treatment Depressed patients were treated with a multimodal psychotherapy approach in line with the “Mainz Model” (Beutel et al., 2008). Treatment included two psychodynamic individual sessions and two group therapy sessions per week as well as body-oriented, art therapy, and other treatment elements (relaxation techniques, psychoeducational groups) to manage depressive symptoms and activate and work through underlying trauma and conflict. Admission of new group members to group treatment follows the “slow-open” principle: When a patient is discharged after eight to twelve weeks of treatment, a newly admitted patient takes his or her place. The therapeutic community on the ward provides ample interpersonal experiences (e.g. self-assertion, cooperation, conflict, initiative): While we did not use differential approaches dependent on individual patterns of attachment, providing safe opportunities for trustful attachment experiences was an overall clinical goal. Thus, not only single psychotherapeutic sessions, but also our clinical setting itself stimulates the attachment system - patients intensely interact with each other (exchange and exposure of problems to other group member in group therapy sessions) and spend time together outside of therapeutic sessions (e.g. sharing kitchen, lounge, and sleeping in twobed-rooms). Under these circumstances, interpersonal communication and conflict styles as well as relationship patterns linked to early attachment experiences are activated and new attachment bonds are made. Thus, both current and past attachment relationships and conflicts are discussed with the therapist, and continual presence of the nursing staff provides a safe and sensitive environment assisting the patient's emotional regulation. Clinical staff does not carry white coats. Regular team meetings serve the important function of drawing together the patients' experiences in the treatment modules and the multiple interactions on the ward. Here, patients' attachment experiences and conflicts are discussed, a psychodynamic and attachment-oriented understanding of each patient is formulated, determining specific treatment goals, and reviewing the progress of patients. 2.3. Adult attachment representations The Adult Attachment Interview (AAI; Hesse, 2008) is a semistructured and widely-used instrument that measures adult attachment on the representational level. The interview consists of 18 questions concerning attachment related experiences in early and middle childhood, including the relationship to the primary caregivers (usually both parents) and other attachment experiences, such as separations and losses. The AAI takes about 60 min to administer, and it has to be electronically recorded and transcribed verbatim afterwards. Transcriptions of AAIs were coded according to the ‘AAI Scoring and Classification System’ (Main

166

I. Reiner et al. / Journal of Affective Disorders 195 (2016) 163–171

et al., 2002) that is often considered the ‘gold standard’ in the field (Bakermans-Kranenburg and Van IJzendoorn, 2009). There are four major AAI Classifications: “Secure”, “Insecure-Dismissing” and “Insecure-Preoccupied”. These three classifications describe groups that portray continuously organized state of mind with respect to attachment. The fourth category “Unresolved” (‘U’)-category stands for unresolved/disorganized state of mind with respect to loss and/or abuse. In case an “unresolved”-assignment is considered, it is made in conjunction with a forced, best-fitting alternative organized state of mind category. The prototypical secure adult is ‘free to evaluate’ (and therefore often indicated by ‘F’) his attachment experiences and has good access to childhood memories. Secure adults are able to reflect on their own as well as on other's feelings and thoughts, and talk openly about negative feelings and hurt without excessive anger or avoidance of the interview topics. Insecure-dismissing individuals cannot talk openly about negative experiences. They have trouble remembering childhood and often idealize one or both parents. Dismissing adults struggle to communicate their feelings and to regulate negative emotions that might be raised during the interview by denying them. Insecure-preoccupied adults are emotionally strongly entangled (and therefore often indicated by ‘E’) with their attachment experiences and usually discuss childhood memories extensively. Typically, they are characterized by emotional confusion, or ambivalence and anger with regards to the relationship with their caregiver. With respect to the “Unresolved” -category, the determining factor for assigning this category is not the pure incidence of loss or abuse but rather an incomplete and disoriented (and therefore unresolved) processing and representing of those experiences. According to the Main et al. (2002) classification system, an unresolved state of mind with respect to loss/ abuse becomes apparent in a cognitive disorientation or disorganization during the discussion of those traumatic experiences, e.g. in extreme attention to the details of the traumatic event or indications that the interviewee felt unrealistically responsible or guilty for the loss or trauma. Main et al. (2002) provide guidelines for scoring unresolved attachment category by means of a 1–9 rating scale for unresolved trauma. If the score reaches 5 or above, the ‘U’ category as final classification is assigned. Dismissing, preoccupied and unresolved/disorganized attachment status have been linked to psychopathology (Dozier et al., 2008). The central indicator of attachment security in the AAI is narrative coherence. The “coherence-of-transcript scale” is highly relevant for the final classification - secure adults are characterized by a coherent description of their attachment related experiences, independent of parental behavior. The concept of coherence refers to the four maxims of Grice (1975), that is (1) quality (be truthful and have evidence of for what you say), (2) quantity (be complete but succinct), (3) relevance (be relevant to the question) and (4) manner (be clear and orderly). A transcript in which the parent is described with highly positive adjectives but the interviewee is unable to provide examples that support those adjectives would score low on coherence as the maxim of quality is violated. In another example, the interviewee may describe her parent with negative adjectives. When asked to give examples, she might be able specify situations in which she felt her parent treated her poorly. However, if she still feels angry about the parent's behavior, she may address the parent during the AAI as if he was in the room and may blame him for present difficulties. Also in this example, coherence is low as the maxims relevance and manner may be violated. In the present study we used the coherence-of-transcript scale (range 1–9) as a marker of attachment security for statistical analyses (Reiner and Spangler, 2010). The unresolved trauma scale was used as dimensional marker for unresolved trauma (Bakermans-Kranenburg et al., 2011). All interviews were rated by two reliable and trained coders unaware of the subjects' other data.

Rater agreement was assessed for attachment classification of 32 of the 168 interviews with a Cohen's Kappa ¼.81 (88% exact agreement), indicating very good agreement. Agreement for the coherence-of-transcript scale (in the following indicated as “attachment security”) and the unresolved trauma scale (in the following indicated as “unresolved symptoms”) was assessed yielding satisfactory agreement (r (32) ¼ .73 and r (32) ¼.71, respectively). 2.4. Depressive symptoms Depressive symptoms were measured with the depression module of the Patient Health Questionnaire, the PHQ-9 (Lowe et al., 2004). The nine-items of the PHQ-9 correspond to the DSM criteria of major depression as “0” (not at all) to “3” (nearly every day). Internal consistency of the PHQ-9 scale was high for the total sample (Cronbach's α ¼ 0.93) and acceptable for both subsamples (depressed sample: Cronbach's α ¼0.75, healthy control group: Cronbach's α ¼ 0.67). Scores range from 0 up to a maximum of 27. Scores of 15 and 20 or higher represent moderately severe and severe depression (Kroenke et al., 2010). 2.5. Statistical analyses Statistical analyses were performed using IBM SPSS Statistics 20 (IBM, New York, NY). To investigate relations between groups (patient/ control; attachment groups), chi square tests were used. Depressive symptoms (PHQ-9), dimensional attachment security and unresolved symptoms were treated as continuous scales. Pearson correlation coefficients were computed to indicate the strength and direction of relation between two variables. To compare groups with respect to continuous scales, analyses of variance were carried out. To examine changes in depressive symptoms, attachment security and unresolved symptoms, repeated measures analyses of variance were conducted. Lastly to explore associations between changes from admission to discharge, additional variables were composed subtracting time 1 from time 2 scores (e.g. admission attachment security minus discharge attachment security). A two-sided p-value of 0.05 or less was considered as statistically significant and effect sizes were computed.

3. Results 3.1. Preliminary analyses Participant age and duration of treatment (inpatient group) were not associated with attachment classification, F(3,81) ¼ 0.30, p¼ .82, η2 ¼ .01, nor with dimensional attachment security, r(85) ¼ 0.04, p¼ .70, or unresolved symptoms, r(85) ¼0.07, p ¼.51. As expected, insecure attachment groups were significantly overrepresented in the clinical sample at time of admission (χ2(3, N ¼85) ¼33.21, p o.001, φ ¼ 0.63), see Table 1a. Attachment security was significantly lower in the patient than in the control group, F(1,83)¼ 41.94, p o.001, η2 ¼ .34 (Patient M ¼3.31, SD ¼1.44 vs. Control M ¼5.62, SD ¼1.82). Unresolved symptoms were significantly higher in the patient than in the control group, F(1,83) ¼ 14.68, p o.001, η2 ¼ .15 (Patient M¼4.14, SD ¼2.31 vs. Control M¼2.57, SD ¼1.32). As evident from Tables 1a, b and 3, differences between patients and controls with regard to distribution of attachment classifications, attachment security and unresolved symptoms, respectively, remained significant at T2 (discharge): χ2(3, N ¼82) ¼10.59 p¼ .017, φ ¼ 0.34; Attachment security: F (1,80) ¼10.59, p ¼.002, η2 ¼.12 (Patient M¼4.38, SD ¼1.60 vs. Control M ¼5.50, SD ¼1.52): Unresolved symptoms: F(1,80) ¼ 10.93, p ¼.001, η2 ¼.12 (Patient M ¼3.79, SD ¼1.90 vs. Control M¼2.59, SD ¼1.34).

I. Reiner et al. / Journal of Affective Disorders 195 (2016) 163–171

Table 1 Attachment classifications (four- and three way) of depressed patients and healthy controls at T1 admission (aþ c) and T2 discharge (b þd): distribution in number of subjects. a)

Attachment classification (T1-admission) Secure

Patients Controls Total b) Patients Controls Total c) Patients Controls Total d) Patients Controls Total

Dismissing

5 8 27 10 32 18 Attachment classification Secure Dismissing 18 4 26 9 44 13 Attachment classification Secure Dismissing 10 9 29 10 32 18 Attachment classification Secure Dismissing 23 6 28 9 44 13

Preoccupied

Total Unresolved

11 19 2 3 13 18 (T2-discharge) Preoccupied Unresolved 6 13 2 4 8 17 – three way (T1-admission) Preoccupied Total 24 43 3 42 13 85 – three way (T2-discharge) Preoccupied Total 12 41 4 41 8 82

43 42 85 Total 41 41 82

Note. Assessment times of controls correspond to assessment times of patients (matched time intervals).

As evident from Table 1a and b, 13 patients switched from an insecure (dismissing or preoccupied) or unresolved state of mind to a secure state of mind. In total, changes in attachment classifications were evident in 15 patients (37%): two more patients with unresolved status turned to an insecure-dismissing and insecurepreoccupied attachment status, respectively, without major signs of lack of resolution of trauma. In the control group, changes of attachment status were evident in six subjects (15%): One subject classified as preoccupied at T1 was unresolved at T2, three subjects turned from dismissing (T1) to secure (T2), two subjects turned from secure (T1) to dismissing (T2) and one subject from secure (T1) to preoccupied (T2). At Time 1, the three drop-outs were classified as unresolved and preoccupied (patient group) and secure (control group). Table 1c and d shows distributions of three way attachment classifications, when unresolved cases were forced into an organized state of mind attachment classification as the second-best choice. Again, insecure attachment groups were significantly overrepresented in the clinical sample at time of admission (χ2(2, N ¼85) ¼ 25.63, p o.001, φ ¼0.55), but significant differences between patients and controls with regard to distribution of attachment groups were absent at T2 (discharge): χ2(2, N ¼82) ¼ 5.10 p ¼.07, φ ¼0.25. As presented in Table 2, the measured constructs – depressive symptoms, attachment security, and unresolved symptoms – showed moderate to strong correlations between the two points of assessment (T1 and T2). Only one further significant correlation was found: Attachment security and unresolved symptoms correlated negatively in the patient group at time of discharge. 3.2. Decrease in depressive symptoms and influence of baseline attachment security To examine changes in depressive symptoms a repeated-measure ANOVA was performed, with diagnostic group as a betweensubjects factor and time (admission and discharge) as a withinsubject factor, revealing a significant decrease of depressive symptoms due to changes in the patient group F(1,83) ¼24.60, p o.001, η2 ¼.23 (see also Table 3). To examine the influence of attachment security at the time of

167

admission on decreases in depressive symptoms, two groups were created based on a median split on AAI coherence (¼ attachment security) scores at baseline: one captured subjects with lower attachment security (4.6 and lower, total n¼ 47 – patients n ¼34, controls n¼ 13) while the other one included subjects with higher attachment security scores (equal or higher than 4.7, total n ¼38 – patients n ¼9, controls n ¼29). A repeated-measure ANOVA was performed with patient vs control group and attachment security (lower vs. higher attachment security) as between-subjects factors and time (admission and discharge) as a within-subject factor, revealing a significant two-way interaction between group (patient vs. control) and attachment security (high vs. low) on changes of depressive symptoms F(1,81)¼ 5.84, p¼ .018, η2 ¼.07. Fig. 1a and b illustrates that the decrease in depressive symptoms in depressed patients was moderated by baseline attachment security: The decrease in depressive symptoms (PHQ-9) between T1 and T2 was significantly larger in patients with relatively high attachment security, (T1: M¼ 17.22 SD ¼4.38 vs T2 M ¼8.44 SD ¼4.63) than in patients with low baseline attachment security (T1: M¼ 14.21 SD ¼ 4.66 vs T2 M ¼10.73 SD ¼5.06). In the control group, mean differences in depressive symptoms (PHQ-9) between T1 and T2 were not significant: Control subjects with relatively high baseline attachment security (T1: M¼2.31 SD ¼2.09 vs T2 M ¼2.21 SD ¼1.76) did not differ significantly in terms of depressive symptoms from control subjects with low baseline attachment security (T1: M ¼1.54 SD ¼ 1.99 vs T2 M ¼1.85 SD ¼ 1.72). 3.3. Influence of baseline unresolved trauma on decrease in depressive symptoms To examine the influence of unresolved trauma at time of admission on decreases in depressive symptoms, two groups were created based on a median split on their AAI unresolved trauma score: subjects with lower unresolved symptoms (3.0 and lower, total n ¼44 – patients n¼ 16, controls n ¼28) were distinguished from subjects with higher unresolved symptoms (equal and higher than 3.1, total n ¼41 – patients n ¼27, controls n ¼16). Again, a repeated-measure ANOVA was performed, with patient vs control group and unresolved trauma (low vs. high unresolved symptoms) as between-subjects factors and time (admission and discharge) as a within-subject factor, revealing no significant effects, F(1,81) ¼ 0.56, p ¼.458, η2 ¼ .01. This means that baseline unresolved symptoms did not moderate the decrease of depressive symptoms in depressed patients after psychotherapeutic treatment. 3.4. Increase of attachment security and decrease of unresolved symptoms after psychotherapy To examine changes in attachment security, we performed a repeated-measure ANOVA with patient vs control group as a between-subjects factor and time (admission and discharge) as a within-subject factor, that revealed a significant increase in attachment security in the patient group F(1,80)¼ 19.12, p o.001, η2 ¼ .19. To examine changes in unresolved symptoms, we conducted a repeated-measure ANOVA with patient vs control group as a between-subjects factor and time (admission and discharge) as a within-subject factor, revealing no significant decreases in unresolved symptoms, F(1,80)¼1.07, p ¼.303, η2 ¼ .01 (see also Table 3). 3.5. Associations between changes in depressive symptoms, attachment security and unresolved symptoms In order to explore if decreases in depressive symptoms were associated with increases in attachment security and decreases in

168

I. Reiner et al. / Journal of Affective Disorders 195 (2016) 163–171

Table 2 Correlations between depressive symptoms (PHQ-9), attachment security (coherence) and unresolved symptoms (unresolved state of mind scale) at T1 (admission) and T2 (discharge) for patient and control group.

Depressive symptoms T1 Attachment security T1 Unresolved symptoms T1 Depressive symptoms T2 Attachment security T2 Unresolved symptoms T2

Depressive symptoms T1

Attachment security T1

Unresolved symptoms T1

Depressive symptoms T2

Attachment security T2

Unresolved symptoms T2

–

.23

.08

.33n

.28

.02 nn

 .04

.05

–

 .06

 .16

.59

.07

 .05

–

.13

 .27

.70nn

.38n

.04

.19

–

 .07

.06

.15

.87

nn

 .11

.06

–

 .31n

.20

 .25

.61nn

.07

 .25

–

Note. Patient, n (T1) ¼ 43, n (T2) ¼ 41; control – underlinded, n (T1) ¼ 42, n (T2) ¼ 41. Assessment times of controls correspond to assessment times of patients (matched time intervals). n

pr .05. p r.01.

nn

Table 3 Depressive symptoms (PHQ-9), attachment security (coherence) and unresolved symptoms (unresolved trauma score) at admission (T1) and discharge (T2): means and standard deviations. Depressive symptoms (PHQ9)

Attachment secur- Unresolved ity (coherence) symptoms (Uscale)

M

M

SD

Patients T1 n ¼43 14.84a T2 n ¼41 10.25b Control T1 T2

n ¼42 2.10 n ¼41 2.10

SD

M

SD

4.71 5.01

3.31a 4.38b

1.44 1.60

4.11 3.79

2.35 1.33

2.01 1.74

5.62 5.50

1.82 1.52

2.57 2.59

1.34 1.34

Note. a, b: different letters indicate statistically significant differences (DUNCAN, p r.01). Assessment times of controls correspond to assessment times of patients (matched time intervals).

unresolved symptoms respectively, time 1 minus time 2 variables were created and correlated. As shown in Table 4, decreases in depressive symptoms were not significantly associated with increases of attachment security or unresolved symptoms. Correlations in the control groups did also not reach statistical significance.

Table 4 Correlations between changes in depressive symptoms (time 1 minus time 2), attachment security (time 1 minus time 2), and unresolved symptoms (time 1 minus time 2) for patient and control group.

Depressive symptoms T1-T2 Attachment security T1-T2 Unresolved symptoms T1-T2

Depressive symptoms T1-T2

Attachment security T1-T2

Unresolved symptoms T1-T2

–

.01

.00

 .09

–

.01

 .27

.19

–

Note. Patient, n¼41; control – underlinded, n¼ 41. Assessment times of controls correspond to assessment times of patients (matched time intervals).

4. Discussion We explored the influence of attachment security and unresolved symptoms on psychotherapeutic outcome and changes in attachment representation through psychotherapeutic intervention in a sample of depressed females. As expected and reported previously, insecure attachment representations were significantly overrepresented in depressed patients compared to age-matched controls. However, contrary to prior findings in depressed samples (Bakermans-Kranenburg and Van IJzendoorn, 2009; Dozier et al., 2008), with overrepresentation of especially insecure-dismissing

a) Low baseline coherence

b) High baseline coherence

20 18 Depressive Symptoms PHQ-9

16

n=9 n = 34

14 12 Control

10 Patient

8 6 4

n = 13

n = 29

2 0 Admission T1

Discharge T2

Note. N = 85; patient group n= 43, control group n = 42. Fig. 1. Decrease of depressive symptoms (PHQ-9 score, means and standard errors) in patient and control group by low (a) and high (b) baseline attachment security.

I. Reiner et al. / Journal of Affective Disorders 195 (2016) 163–171

and insecure-preoccupied and not of unresolved subjects, our depressed group had primarily unresolved and insecure-preoccupied attachment classifications at time of admission. Our distribution of AAI groups was comparable to the Fonagy et al. (1996) sample of unipolar depressed and dysthymic patients from an inpatient unit, with strong overrepresentations of preoccupied and unresolved classifications. It is important to note that depressed patients from our psychosomatic unit represent clinically severe and complex cases, typically with a long medical history. Although none of our subjects suffered from clinical PTSD, many of them experienced severe attachment traumata such as losses, abuse and early separation from parents. When understanding the role of attachment representation in the development and treatment of depression, considering “subtypes” of depression may be a useful approachsevere and chronic cases of unipolar depression may be associated with particularly preoccupied or unresolved attachment status, whereas milder forms of depression might not be associated with attachment insecurity (Bakermans-Kranenburg and Van IJzendoorn, 2009). Furthermore, the overrepresentation of preoccupied and unresolved attachment statuses in our depressed group might indicate a higher risk for developing chronic and severe depression. Alternatively or in addition, the relatively low rate of insecure-dismissing patients might be a result of a reluctance to utilize support and seek treatment in depressed individuals with dismissing attachment representations. Previous studies have shown that dismissing persons have difficulties to seek help and disclose negative feelings (Dozier and Lee, 1995; Larose and Bernier, 2001; Reiner and Spangler, 2013). However caution is advised when interpreting cross-sectional associations between clinical depression and attachment status: attachment representation and coherence can be also influenced by depression and related restricted cognitive resources or altered emotion regulation capacities. Depressive symptomatology in the patient group significantly decreased after inpatient treatment. Psychotherapy has proven to be effective in the treatment of depression (Cuijpers et al., 2008). Regarding inpatient psychodynamically oriented psychotherapy in German psychosomatic units like the one from the present study, Franz et al. (2015) report large effect sizes for the treatment of depressive disorders from a naturalistic, multicenter intervention study. Our study revealed that attachment security at time of admission moderated therapeutic outcome: depressed patients with higher initial attachment security displayed a significantly steeper decrease of depressive symptoms at discharge and benefitted from therapy more than patients with low attachment security. We cautiously suggest that patients with higher attachment security may engage easier in the therapeutic setting and in the therapeutic relationship, being more comfortable with opening up towards therapeutic interventions in both single and group sessions. Independent of the severity of depression, patients with very low security in the AAI may struggle more with establishing a positive relationship to the therapist and a therapeutic working alliance, which has been repeatedly found to be one of the most consistent predictors for psychotherapy outcome (Lambert and Barley, 2001). We speculate that highly insecure patients, who struggle to form a relationship to their therapist, would first have to feel a sense of security in the (therapeutic) relationship in order to attain depressive symptom reduction as therapeutic goal more efficiently. Because we did not assess therapeutic working alliance and other variables influencing the patient-therapist relationship (e.g. therapist characteristics), our interpretation remains speculative, but our findings offer grounded hypotheses for further, randomized psychotherapy experiments controlling for initial attachment status and therapeutic working alliance. Contrary to our expectations, unresolved symptoms were not

169

linked to depressive symptom reduction and did not significantly decrease after treatment, indicating that – in our broad treatment concept for clinically depressed patients - unresolved symptoms do not directly influence or alter the outcome of inpatient psychotherapy. There has been evidence that imaginal exposure therapy impacts unresolved symptoms in the AAI (StovallMcClough and Cloitre, 2006). Trauma-focused in- or outpatient psychotherapy targeting at the resolution of traumatic experiences more than at depressive symptoms and symptom-associated difficulties in relationships may lead to significant decreases in unresolved symptoms in the AAI. Confirming our hypothesis, inpatient psychotherapy led to increases in attachment security. Theoretically, attachment insecurity in the AAI goes along with unprocessed adverse attachment experiences in childhood, which might have contributed to the pathogenesis of depressive symptoms. An increase in attachment security through psychotherapy seems intuitive, in particular because our psychodynamic treatment concept aims to create a more coherent narrative of past adverse experiences and better mentalization of current relationships (Beutel et al., 2008; Fonagy et al., 2002). Yet, inpatient psychotherapy is multimodal and includes many different components, so it is impossible to know what the effective component really is and which components influence each other: Although depression was the cause for hospitalization and depressive symptom relief considered an effective outcome, inpatient psychotherapy does not only target at depressive symptoms, but addresses also interpersonal problems most likely linked to depression. Thus, in psychotherapeutic sessions, “knowledge” and awareness of own attachment representations might be acquired by patients. By working through difficulties in present relationships, they may build an understanding that at least some of these difficulties may be rooted in early attachment experiences. Of course, psychotherapy cannot alter adverse or traumatic experiences with attachment figures, however, our study shows that psychotherapy helps to process and integrate emotionally difficult experiences, which is reflected in more coherent attachment representations. The following example of a 28-year old female depressed “patient X” from our study may illustrate this: At time of admission, she described the relationship to her parents as follows: “The relationship to my mother was scary. She was so unpredictable – o yeah she still is – and always screaming at us and I always was like ‘buah, just let her voice turn down’. (…) My father left me high and dry. He explained to us why my mother was loud towards us but didn't prevent it! But he must have sensed it! Nobody can tell me that this man came home thinking ‘uh what a harmonic atmosphere' (...). When being asked how her attachment experiences influenced her today, patient X said: “Well, my childhood is not the yardstick for all things. There was always this ‘be always strong, be always there for others, do everything perfectly, survive in this world, be someone special’. (…) I always feel inferior to people who went to college. My mother always wanted me to go to college. But I didn't want to!”. The AAI from patient X was classified “insecure-preoccupied” at time of admission – she was very emotionally entangled with her attachment figures and angry with both father and mother. Her speech was incoherent (coherence-score¼3) as can be seen particularly in violations of relevance and manner (e.g. talking about the present: o yeah, she still is; angry speech: Nobody can tell me that this man came home (…); unintroduced speech: There was always this ‘be always strong…’). By the end of the inpatient treatment, patient X described the relationship to her parents as follows in the AAI: “My mother was choleric. She hardly spoke but when she did, she started screaming and yelling for the most time. She was very angry after a short time. For instance, when I skinned my elbows and I cried, she

170

I. Reiner et al. / Journal of Affective Disorders 195 (2016) 163–171

yelled at me ‘why are you so stupid to fall down like this’. (…). Then, my father practically translated that and mediated and said: ‘Your mother yells at you, because it hurts her that you are hurt’. Well, that's why I think he was trying to mediate – to explain to me why my mother behaved so strangely. What I – I can only tell for myself – never understood, I never questioned that, I was five years old, these things were just normal for me.” When being asked how her attachment experiences influenced her today, she said: “I learned to tough it out. By growing up under such weird circumstances I have developed sensitive antennas – When I enter a room, I sense immediately, who's feeling well and who's not feeling well. I think people like being around me because of that. But I have also learned to suppress feelings, em, mostly my own feelings. Because it was bad to show your feelings and I, em di– didn't admit them. That's why I couldn't cry anymore. And at some point of time this, this doesn't work anymore, then your body shows some kind of reaction, at least my body reacted and I became depressed and burned out.”. At time of discharge, patient's X AAI was classified “secure” – although she still describes unfavorable experiences with her angry mother and difficulties with her father's efforts to intervene, she describes them in a balanced and contained way. Her speech was overall moderately coherent (coherence-score ¼6), there were no major violations of quality, quantity, relevance and manner. Compared to her AAI taken at admission, she developed a sense of forgiveness for her father's imperfections and her mother's temper. She understood how early childhood experiences influenced her present life – in good and in bad terms – and how they might be linked to her depressive illness. The present example illustrates not only well how psychotherapeutic treatment may change representation about attachment experience, but also how attachment security (coherence) is related with mentalization: While statements in the admission AAI did not mirror highly developed mentalization skills, patient X shows high reflective functioning in her AAI taken at discharge (Fonagy et al., 1998): she shows that she is aware of the nature of her and her father's mental states (e.g.: ‘Well, that's why I think he was trying to mediate – to explain to me why my mother behaved so strangely’) and is able to recognize diverse perspectives and developmental aspects of mental states (e.g.: ‘What I – I can only tell for myself – never understood, I never questioned that, I was five years old’). She takes into account how others may perceive her (‘I think people like being around me because of that’) and puts effort to tease out mental states underlying her own behavior (e.g.: ‘That's why I couldn't cry anymore. And at some point of time this, this doesn't work anymore, then your body shows some kind of reaction, at least my body reacted and I became depressed and burned out’). Also, the intensive psychotherapeutic care (e.g. the possibility to approach therapeutically trained nursing staff at all times) may also lead to an increase in feeling a “general sense of attachment security”. Increases in attachment security may contribute to decreases in depression or vice versa – our results shows no significant correlation between changes regarding attachment and changes regarding depressive symptoms and suggest that both change independently. However, there might be latent treatment variables that have not been investigated in the present study. Yet, our findings propose that psychotherapy in psychosomatic settings promotes attachment security and not only depressive symptom reduction. This may prevent patients from (future) painful relationship experiences and therefore a depressive relapse. The following limitations of the present study need to be acknowledged: our sample consisted of female inpatients and matched comparisons only and, although our study design included two times of assessment with a carefully matched mentally

healthy control group, it remains a quasi-experimental design and hypothesis generating exploratory study. Our therapy concept included many different components and it remains unclear which of those mostly affected attachment representations and depressive symptoms. Several possible confounders influencing treatment processes and underlying the patient-therapist relationship such as psychopharmacological treatment or therapist's personality or attachment representation could not be controlled. Although the PHQ-9 is a reliable instrument to detect depression outcome and changes over time (Lowe et al., 2004), severity of depressive symptoms has been assessed via self-report, which might be biased by subject's characteristics such as medical history, age or attachment status (Dozier and Lee, 1995; Eaton et al., 2000). In order to better understand efficacy factors in the treatment of depression and the causal role of attachment representation, further empirical examinations involving male subjects, outpatient psychotherapeutic settings and randomized controlled study designs are required. Our study provides some grounded hypotheses on the role of attachment security in moderating therapeutic efficacy in depressed patients such that secure attachment representations is suggested to facilitate the efficacy of the inpatient treatment.

Acknowledgments We would like to express our special thanks to all patients and control subjects for their participation in this study. We also thank the project coordinators Johanna Ottemeyer and Gotje Trojan. Marinus van IJzendoorn and Marian Bakermans-Kranenburg are members of the Leiden Consortium on Individual Development and were supported by the Dutch Ministry of Education, Culture, and Science and the Netherlands Organization for Scientific Research (Gravitation program NWO grant number 024.001.003, SPINOZA, VICI grant 453-099-003). MJBK was also funded by the European Research Council (AdG 669249).

References Adler, J.M., Harmeling, L.H., Walder-Biesanz, I., 2013. Narrative meaning making is associated with sudden gains in psychotherapy clients' mental health under routine clinical conditions. J. Consult. Clin. Psychol. 81, 839. Bakermans-Kranenburg, M.J., Van IJzendoorn, M.H., 2009. The first 10,000 adult attachment interviews: distributions of adult attachment representations in clinical and non-clinical groups. Attach. Hum. Dev. 11 (3), 223–263. http://dx. doi.org/10.1080/14616730902814762. Bakermans-Kranenburg, M.J., van IJzendoorn, M.H., Caspers, K., Philibert, R., 2011. DRD4 genotype moderates the impact of parental problems on unresolved loss or trauma. Attach. Hum. Dev. 13 (3), 253–269. http://dx.doi.org/10.1080/ 14616734.2011.562415 936575953 [pii]. Bateman, A., Fonagy, P., 2009. Randomized controlled trial of outpatient mentalization-based treatment versus structured clinical management for borderline personality disorder. Am. J. Psychiatry 166, 1355–1364. Bernecker, S.L., Levy, K.N., Ellison, W.D., 2014. A meta-analysis of the relation between patient adult attachment style and the working alliance. Psychother. Res. 24, 12–24. Beutel, M.E., Michal, M., Subic-Wrana, C., 2008. Psychoanalytically-oriented inpatient psychotherapy of somatoform disorders. J. Am. Acad. Psychoanal. Dyn. Psychiatry 36 (1), 125–142. http://dx.doi.org/10.1521/jaap.2008.36.1.125. Bowlby, J., 1979. The Making and Breaking of Affectional Bonds. Tavistock Publications, London. Bowlby, J., 1988. A Secure Base. Parent-Child Attachment and Healthy Human Development. Basic Books, New York. Bretherton, I., Munholland, K.A., 2008. Internal Working Models in Attachment Relationships: Elaborating a Central Construct in Attachment Theory. In: Cassidy, J., Shaver, P. (Eds.), Handbook of Attachment Theory and Research. Guilford, New York, pp. 102–131. Burt, V.K., Stein, K., 2001. Epidemiology of depression throughout the female life cycle. J. Clin. Psychiatry 63, 9–15. Byng-Hall, J., 1995. Creating a secure family base: some implications of attachment theory for family therapy. Fam. Process 34, 45–58. Crowell, J.A., Fraley, R.C., Shaver, P.R., 2008. Measurement of Individual Differences in Adolescent and Adult Attachment. Cuijpers, P., van Straten, A., Andersson, G., van Oppen, P., 2008. Psychotherapy for depression in adults: a meta-analysis of comparative outcome studies. J. Consult. Clin. Psychol. 76 (6), 909–922. http://dx.doi.org/10.1037/a0013075. Cuijpers, P., Karyotaki, E., Weitz, E., Andersson, G., Hollon, S.D., van Straten, A., 2014. The effects of psychotherapies for major depression in adults on remission,

I. Reiner et al. / Journal of Affective Disorders 195 (2016) 163–171

recovery and improvement: a meta-analysis. J. Affect. Disord. 159, 118–126. http://dx.doi.org/10.1016/j.jad.2014.02.026. Diener, M.J., Monroe, J.M., 2011. The relationship between adult attachment style and therapeutic alliance in individual psychotherapy: a meta-analytic review. Psychotherapy 48, 237. Dozier, M., Lee, S.W., 1995. Discrepancies between self-report and other-report of psychiatric symptomatology – effects of dismissing attachment strategies. Dev. Psychopathol. 7, 217–226. Dozier, M., Tyrrell, C., 1998. The Role of Attachment in Therapeutic Relationships. In: Simpson, J.A., Rholes, W.S. (Eds.), Attachment Theory and Close Relationships. Guilford Press, NewYork, pp. 221–248. Dozier, M., Stovall-McClough, K.C., Albus, K.E., 2008. Attachment and Psychopathology in Adulthood. In: Cassidy, J., Shaver, P. (Eds.), Handbook of Attachment. Theory, Research, and Clinical Applications, 2nd ed. Guilford, New York, pp. 718–744. Eaton, W.W., Neufeld, K., Chen, L.-S., Cai, G., 2000. A comparison of self-report and clinical diagnostic interviews for depression: diagnostic interview schedule and schedules for clinical assessment in neuropsychiatry in the Baltimore epidemiologic catchment area follow-up. Arch. Gen. Psychiatry 57, 217–222. Fonagy, P., Target, M., 2005. Commentary: bridging the transmission gap: an end to an important mystery of attachment research? Attach. Hum. Dev. 7, 333–343. Fonagy, P., Bateman, A.W., 2006. Mechanisms of change in mentalization‐based treatment of BPD. J. Clin. Psychol. 62, 411–430. Fonagy, P., Steele, H., Steele, M., 1991a. Maternal representations of attachment during pregnancy predict the organization of infant-mother attachment at one year of age. Child. Dev. 62, 891–905. Fonagy, P., Steele, M., Steele, H., Moran, G.S., 1991b. The capacity for understanding mental states: the reflective self in parent and child and its significance for security of attachment. Infant Ment. Health J. 12, 201–218. Fonagy, P., Gergely, G., Jurist, E.L., Target, M., 2002. Affect Regulation, Mentalization and the Development of the Self. Other Press, New York. 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. J. Consult. Clin. Psychol. 64, 22–31. Fonagy, P., Target, M., Steele, H., Steele, M., 1998. Reflective-Functioning Manual, version 5.0, for Application to Adult Attachment Interviews. University College London, London. Franz, M., Seidler, D., Jenett, D., Schafer, R., Wutzler, U., Kammerer, W., Deters, M.A., Schauenburg, H., Becher-Dortschy, N., Frommhold, K., Keller, W., Gosda, J., Tress, W., Zeller, A., Beutel, M.E., Langenbach, M., Hoffmann, T., Haberger, R., Hartkamp, N., 2015. [Psychodynamically oriented inpatient psychotherapy of depressive disorders-First findings from a naturalistic, multicenter study]. Z. Psychosom. Med. Psychother. 61 (1), 19–35. http://dx.doi.org/10.13109/ zptm.2015.61.1.19. Freud, S., 1997. Dora: An Analysis of a Case of Hysteria. Simon and Schuster, New York. Grice, H.P., 1975. Logic and Conversation. In: Cole, P., Morgan, J. (Eds.), Syntax and Semantics. Academic Press, New York. Harari, D., Bakermans-Kranenburg, M.J., de Kloet, C.S., Geuze, E., Vermetten, E., Westenberg, H.G.M., van Ijzendoorn, M.H., 2009. Attachment representations in Dutch veterans with and without deployment-related PTSD. Attach. Hum. Dev. 11 (6), 515–536. http://dx.doi.org/10.1080/14616730903282480. Harter, M., Klesse, C., Bermejo, I., Schneider, F., Berger, M., 2010. Unipolar depression: diagnostic and therapeutic recommendations from the current S3/national clinical practice guideline. Dtsch. Arzteblatt Int. 107 (40), 700–708. http: //dx.doi.org/10.3238/arztebl.2010.0700. Hesse, E., 2008. The Adult Attachment Interview: Protocol, Method of Analysis, and Empirical Studies. In: Cassidy, J., Shaver, P.R. (Eds.), Handbook of Attachment: Theory, Research, and Clinical Applications. Guilford Press, New York, pp. 552–598. Karg, K., Burmeister, M., Shedden, K., Sen, S., 2011. The serotonin transporter promoter variant (5-HTTLPR), stress, and depression meta-analysis revisited: evidence of genetic moderation. Arch. Gen. Psychiatry 68 (5), 444–454. http://dx. doi.org/10.1001/archgenpsychiatry.2010.189. Kendler, K.S., Kuhn, J.W., Prescott, C.A., 2004. Childhood sexual abuse, stressful life events and risk for major depression in women. Psychol. Med. 34 (8), 1475–1482. Kroenke, K., Spitzer, R.L., Williams, J.B., Lowe, B., 2010. The patient health questionnaire somatic, anxiety, and depressive symptom scales: a systematic review. Gen. Hosp. Psychiatry 32 (4), 345–359. http://dx.doi.org/10.1016/j. genhosppsych.2010.03.006. Lambert, M.J., Barley, D.E., 2001. Research summary on the therapeutic relationship and psychotherapy outcome. Psychotherapy 38 (4), 357–361. http://dx.doi.org/ 10.1037//0033-3204.38.4.357.

171

Larose, S., Bernier, A., 2001. Social support processes: mediators of attachment state of mind and adjustment in late adolescence. Attach. Hum. Dev. 3 (1), 96–120. http://dx.doi.org/10.1080/14616730010024762. Lepine, J.P., Briley, M., 2011. The increasing burden of depression. Neuropsychiatr. Dis. Treat. 7 (Suppl. 1), S3–S7. http://dx.doi.org/10.2147/NDT.S19617. Levy, K.N., Meehan, K.B., Kelly, K.M., Reynoso, J.S., Weber, M., Clarkin, J.F., Kernberg, O.F., 2006. Change in attachment patterns and reflective function in a randomized control trial of transference-focused psychotherapy for borderline personality disorder. J. Consult. Clin. Psychol. 74 (6), 1027–1040. http://dx.doi.org/ 10.1037/0022-006X.74.6.1027. Liebherz, S., Rabung, S., 2013. [Effectiveness of psychotherapeutic Hospital treatment in German speaking countries: a meta-analysis]. Psychother. Psychosom. Med. Psychol. 63 (9–10), 355–364. http://dx.doi.org/10.1055/s-0033-1337909. Lowe, B., Kroenke, K., Herzog, W., Grafe, K., 2004. Measuring depression outcome with a brief self-report instrument: sensitivity to change of the patient health questionnaire (PHQ-9). J. Affect. Disord. 81 (1), 61–66. http://dx.doi.org/ 10.1016/S0165-0327(03)00198-8. Main, M., Kaplan, N., Cassidy, J., 1985. Security in infancy, childhood, and adulthood-a move to the level of representation. Monogr. Soc. Res. Child. Dev. 50 (1– 2), 66–104. http://dx.doi.org/10.2307/3333827. Main, M., Goldwyn, R., Hesse, E., 2002. Adult Attachment Scoring and Classification Systems – Version 7.1. Unpublished Manuscript. University of California at Berkeley. Marganska, A., Gallagher, M., Miranda, R., 2013. Adult attachment, emotion dysregulation, and symptoms of depression and generalized anxiety disorder. Am. J. Orthopsychiatry 83, 131–141. Meins, E., Fernyhough, C., Fradley, E., Tuckey, M., 2001. Rethinking maternal sensitivity: mothers' comments on infants' mental processes predict security of attachment at 12 months. J. Child Psychol. Psychiatry 42, 637–648. Nemeroff, C.B., Heim, C.M., Thase, M.E., Klein, D.N., Rush, A.J., Schatzberg, A.F., Keller, M.B., 2003. Differential responses to psychotherapy versus pharmacotherapy in patients with chronic forms of major depression and childhood trauma. Proc. Natl. Acad. Sci. USA 100 (24), 14293–14296. http://dx.doi.org/ 10.1073/pnas.2336126100. Norcross, J.C., 1990. An Eclectic Definition of Psychotherapy. In: Zeig, J.K., Munion, W.M. (Eds.), What is Psychotherapy?. Contemporary Perspectives Jossey-Bass, San Francisco, pp. 218–220. Outcalt, J., Dimaggio, G., Popolo, R., Buck, K., Chaudoin-Patzoldt, K.A., Kukla, M., Olesek, K.L., Lysaker, P.H., 2016. Metacognition moderates the relationship of disturbances in attachment with severity of borderline personality disorder among persons in treatment of substance use disorders. Compr. Psychiatry 64, 22–28. Reiner, I., Spangler, G., 2010. Adult attachment and gene polymorphisms of the dopamine D4 receptor and serotonin transporter (5-HTT). Attach. Hum. Dev. 12 (3), 209–229. http://dx.doi.org/10.1080/14616731003759674. Reiner, I., Spangler, G., 2013. Representations of early attachment experiences and personality in adulthood. J. Adult Dev. 20 (1), 38–45. http://dx.doi.org/10.1007/ s10804-013-9154-x. Roelofs, J., Lee, C., Ruijten, T., Lobbestael, J., 2011. The mediating role of early maladaptive schemas in the relation between quality of attachment relationships and symptoms of depression in adolescents. Behav. Cogn. Psychother. 39, 471–479. Roisman, G.I., Holland, A., Fortuna, K., Fraley, R.C., Clausell, E., Clarke, A., 2007. The adult attachment interview and self-reports of attachment style: an empirical rapprochement. J. Personal. Soc. Psychol. 92, 678–697. Slade, A., 2005. Parental reflective functioning: an introduction. Attach. Hum. Dev. 7, 269–281. Stovall-McClough, K.C., Cloitre, M., 2006. Unresolved attachment, PTSD, and dissociation in women with childhood abuse histories. J. Consult. Clin. Psychol. 74 (2), 219–228. http://dx.doi.org/10.1037/0022-006X.74.2.219. Sullivan, P.F., Neale, M.C., Kendler, K.S., 2000. Genetic epidemiology of major depression: review and meta-analysis. Am. J. Psychiatry 157 (10), 1552–1562. Toth, S.L., Rogosch, F., Cicchetti, D., 2008. Attachment-Theory-Informed Intervention and Reflective Functioning in Depressed Mothers. Clinical Applications of the Adult Attachment Interview. Guilford Press, New York, NY, pp. 154–172. Toth, S.L., Rogosch, F.A., Manly, J.T., Cicchetti, D., 2006. The efficacy of toddler-parent psychotherapy to reorganize attachment in the young offspring of mothers with major depressive disorder: a randomized preventive trial. J. Consult. Clin. Psychol. 74 (6), 1006–1016. http://dx.doi.org/10.1037/0022-006X.74.6.1006. Wiersma, J.E., Hovens, J.G., van Oppen, P., Giltay, E.J., van Schaik, D.J., Beekman, A.T., Penninx, B.W., 2009. The importance of childhood trauma and childhood life events for chronicity of depression in adults. J. Clin. Psychiatry 70 (7), 983–989. Wittchen, H.U., Zaudig, M., Fydrich, T., 1997. Strukturiertes Klinisches Interview für DSM-IV. Achse I und II. Handanweisung. Hogrefe, Göttingen.