Skip to main content

Increased attachment security is related to early therapy drop-out in substance use disorders



Previous research work suggests a positive association between secure attachment and increased therapy adherence (TA) in different patient groups. However, there is still a strong need for research focusing on the influence of attachment on TA in substance use disorder (SUD) treatment. Hence, this study attempts to investigate the predictive value of different attachment patterns concerning TA in SUD inpatients.


122 (34 female) SUD inpatients completed the Attachment Style Questionnaire (ASQ) during the entry phase of therapeutic community treatment. After 6 weeks, subjects who remained in therapy (n = 47) completed the ASQ for a second time. In line with the literature, agglomerative Cluster Analysis suggested a two-cluster solution (Cluster I: increased secure attachment pattern; Cluster II: increased insecure attachment pattern). Notably, inpatients in Cluster I were more likely to drop out of treatment within the first 6 weeks (p < .001). Furthermore, subjects showed less “Confidence in Self and Others” (p < .05) after 6 weeks of treatment. Our findings indicate a negative predictive value of increased attachment security for TA in SUD inpatients. This finding probably mirrors a more realistic kind of self-assessment. More generally, the importance of considering attachment styles in SUD treatment is underlined.


Substance use disorders (SUD) have been linked with insecure attachment in several studies [1]. This well-established connection is explained by several characteristics observed in individuals displaying an insecure attachment pattern: dysfunctional attributes such as the increased susceptibility and physiological response to stress, as well as the underuse of social support (in conjunction with inadequate help seeking strategies), lead to a predisposition for the frequent use of psychoactive substances as a futile attempt to regulate affects [2,3,4].

Research focusing on the relationship between attachment styles and treatment adherence (TA) is relatively sparse, these studies have been conducted mostly on somatic patients treated for diabetes and lupus [5, 6]: the results indicate that insecure attachment, especially dismissing and avoidant styles, correspond to worse treatment outcome and compliance. Insecure attachment might lead to increased difficulties in the relationship between patients and caregivers, due to patients’ inability to accept help from others. A study on the role of attachment in TA in SUD inpatients has found anxious-preoccupied attachment to be a predictor for treatment retention [7]. Moreover, several studies also observed a negative association between TA and parameters such as comorbid personality disorder, cognitive deficits and age [8, 9].

In previous research, we observed a higher amount of insecure attachment in different SUD patient groups in comparison to healthy controls [10]. As an extension to this study [10], we intended to examine the role of attachment styles for TA by investigating this group of SUD inpatients for a second time after 6 weeks. Based on former results we hypothesized increased TA in more securely attached SUD patients.

Main text


Participants and procedure

The initial sample consisted of 122 inpatients diagnosed with AUD (F10.2x) (n = 66) or poly drug use disorder (PUD) (F19.2x) (n = 57) by a licensed psychiatrist according to ICD 10 [11]. As our previous research indicated that patients diagnosed with AUD or PUD do not differ in their attachment style and their personality organization [10], we decided not to further differentiate between AUD and PUD inpatients in the present study. After 6 weeks, the 47 participants remaining in treatment were investigated for a second time. All subjects were abstinent while in treatment. Further inclusion criteria were: no current psychotic symptoms and full command of the German language. Assessment took place at therapeutic facilities of the “Grüner Kreis” society, an Austrian association dedicated to the treatment and rehabilitation of people with SUD.

Assessment of clinical characteristics and attachment dimensions

Participants where asked about age, gender and level of education. Current medication and comorbid psychiatric diagnoses were assessed through the facilities’ database.

The Attachment Style Questionnaire (ASQ) [12] is a 40-item questionnaire which measures five dimensions of adult attachment: “Confidence in Self and Others” (15 items), “Discomfort with Closeness” (8 items), “Relationships as Secondary” (7 items), “Need for Approval” (4 items) and “Preoccupation with Relationships” (6 items). Items are rated on a 6-point Likert scale. While higher scores in “Confidence in Self and Others” indicate a more secure attachment, higher scores in the other four scales indicate more insecure attachment facets.

Statistical analysis

We conducted agglomerative hierarchical cluster analysis including a squared Euclidian distance interval and Ward’s method to establish a parameter for the different attachment attitudes within the inpatient sample. Furthermore, we performed general linear model multivariate and step-wise hierarchical logistic regression analysis for group comparison and the prediction of treatment drop-out. Additional group comparisons were conducted by means of χ2 tests. The alpha-level was set to .05.


Cluster analysis of attachment dimensions

All ASQ sub-scales showed satisfying internal consistencies (see Table 1), with the exception of “Discomfort with Closeness”, which was therefore not taken into account for further interpretation.

Table 1 Differences between Cluster I (n = 81) and Cluster II (n = 41) in attachment dimensions

In line with de Rick and Vanheule [13], a two-cluster solution could be accepted after performing agglomerative cluster-analysis using all ASQ subscales. As shown in Table 1 Cluster I, “increased secure attachment pattern”, was comprised of inpatients with increased scores in “Confidence in Self and Others” (which equals more secure attachment) and decreased scores in almost all other dimensions representing facets of insecure attachment [11]. Conversely, Cluster II, “increased insecure attachment pattern” consisted of inpatients with lower scores in “Confidence in Self and Others”, in combination with increased scores in most insecure attachment dimensions, with the exception of “Preoccupation with Relationships”.

Furthermore, Cluster I inpatients exhibited significantly more comorbid disorders than Cluster II inpatients (\(\chi^{2}{_{(1)}}\) = 13.49, p < .001). In comparison to healthy controls [9], Bonferroni corrected post hoc analysis revealed that Cluster I inpatients showed lower scores in “Confidence in Self and Others” (p < .001) and higher scores in “Relationships as Secondary” (p < .001), but lower scores in “Need for Approval” (p < .001) and “Preoccupation with Relationships” (p < .001). Cluster II inpatients exhibited lower scores in “Confidence in Self and Others” (p < .001) and higher scores in “Relationships as Secondary” (p < .01), but did not show any differences (p > .05) in “Need for Approval” or “Preoccupation with Relationships” in comparison to control subjects.

Demographics and clinical characteristics

Characteristics for both groups, inpatients who remained in treatment for at least 6 weeks (TR; n = 47) as well as inpatients who dropped out of treatment during the first 6 weeks (TD; n = 75), are displayed in Table 2. Both groups consisted of more male (TR: 83%; TD: 65%) than female participants (\(\chi^{2}{_{(1)}}\) = 4.48, p < .05), while comorbid psychiatric disorders were less common in TR inpatients (Affective Disorder: 15%, Anxiety Disorder: 2%, Schizophrenia: 4%, Personality Disorder 4%) than in TD (Affective Disorder: 43%, Anxiety Disorder: 7%, Schizophrenia: 3%, Personality Disorder: 1%), (\(\chi^{2}{_{(5)}}\) = 13.44, p < .05).

Table 2 Group differences in demographic and clinical characteristics

Finally, we found that TR (77%) were more likely than TD (49%) to belong to Cluster II (increased insecure attachment pattern) (\(\chi^{2}{_{(1)}}\) = 10.44, p < .001).

Attachment predictors of therapy adherence

Within TR inpatients, “Confidence in Self and Others” decreased during the first 6 weeks of treatment (F(1,46) = 300.89, p < .05, η2 = .13). No relevant differences were found for any of the other dimensions (p > .05).

In hierarchical regression analyses (see Table 3), the final regression model accounts for 18% of the estimated variance in TA: sex was entered as a control variable at Step 1 (B = 0.95, Nagelkerke R2 = .05, p < .05), comorbidity at Step 2 (B = − 1.05, Nagelkerke R2 = .12, p < .01) and attachment security at Step 3 (B = − 1.10, Nagelkerke R2 = .18, p < .01). Therefore, attachment security added approximately 6% of the variance to the predictive validity of the regression model.

Table 3 Predictors of treatment adherence after 6 weeks (hierarchical logistic regression modelling; n = 122)


The primary goal of this study was to explore the impact of attachment on TA, for the entry phase of treatment (6 weeks), in a group of SUD inpatients. An Agglomerative Cluster analysis was conducted whereby a two-cluster solution, more secure (Cluster I) versus more insecure (Cluster II) attachment patterns, could be acknowledged. In contrast to our assumptions, Cluster I subjects were more likely to drop out during the entry phase of treatment and also more likely to be diagnosed with a comorbid disorder than participants with a more insecure attachment pattern (Cluster II). Hierarchical regression analysis revealed that the influence of the attachment security concealed the effects of the control variables sex and comorbidity on TA. Moreover, the remaining patients showed a significant decrease in the ASQ subscale “Confidence in Self and Others” at the second point of measurement. No other significant changes in attachment dimensions, assessed at the beginning of treatment, were found after 6 weeks.

The surprising results regarding the influence of attachment on TA, which contradicts previous findings, might be partially explained by a lack of self-reflection within TD inpatients. In this study, a self-report measure of adult attachment was applied, which might be susceptible to distorted self-images and lacks the capacity to detect repressed information [14]. This problem might be more prominent in SUD patients than with patients with solely somatic condition [15]. By further considering the increased rate of comorbidity within the TD and the more secure attached group, it seems possible that the self-reported high attachment security more likely reflects an unrealistic (or idealized) image of the self in relationships with others, characterized by primitive defense mechanism like splitting and denial [16]. Furthermore, our findings could mirror a specific mechanism taking place specifically within therapeutic communities, which represents a threat to narcissistically biased self-images: The high amount of group cohesion within the community might lead to strong cognitive dissonances in vulnerable individuals, which in turn leads to early treatment drop-out [17]. A reduction in narcissism might also provide an explanation for a decreased “Confidence in Self and Others” score after 6 weeks of treatment. This result fits with the general approach of therapeutic community treatment to support patients in developing more conscious attitudes towards interpersonal deficits during their stay within the community [18]. Either way, the outcome of a decreasing “Confidence in Self and Others” might mirror an initial euphoria combined with the experience of being sober after such a long time of severe substance misuse. This exhilaration possibly vanishes after some time spent in the therapeutic community.


In conclusion, our highly preliminary results indicate that caregivers should be especially attentive with patients exhibiting a high amount of comorbidity in combination with a seemingly secure attachment attitude, especially if attachment was captured via standardized self-report measures. However, in line with the literature [5,6,7], this study confirms the importance of considering the attachment dimension for the treatment of SUD.


  • A rather small sample size and the use of a self-report measure for adult attachment styles.

  • The focus was only on the first short treatment period of 6 weeks. For the description of long-term treatment effects within a therapeutic community setting further research might focus on an enhanced time-span.

  • The study examined a convenience sample. It is necessary to replicate these highly preliminary findings in future studies using randomized controlled trials.

  • For a more in depth understanding of our results, further research might especially focus on attachment based therapeutic interventions on TA.



Attachment Styles Questionnaire


alcohol use disorder


poly drug use disorder


substance use disorders


patients, who dropped out of treatment


patients, who remained in treatment


therapy adherence


  1. Unterrainer HF, Hiebler-Ragger M, Koschutnig K, Fuchshuber J, Tscheschner S, Url M, Wagner-Skacel J, Reininghaus EZ, Papousek I, Weiss E, Fink A. Addiction as an attachment disorder: white matter impairment is linked to increased negative affective states in poly-drug use. Front Hum Neurosci. 2017;11:208.

    Article  PubMed  PubMed Central  Google Scholar 

  2. Flores PJ. Addiction as an attachment disorder: implications for group therapy. Int J Group Psychother. 2001;51:63–81.

    Article  CAS  PubMed  Google Scholar 

  3. Maunder RG, Hunter JJ. Attachment and psychosomatic medicine: developmental contributions to stress and disease. Psychosom Med. 2001;63:556–67.

    Article  CAS  PubMed  Google Scholar 

  4. Schindler A, Bröning S. A review on attachment and adolescent substance abuse: empirical evidence and implications for prevention and treatment. Subst Abus. 2015;36:304–13.

    Article  PubMed  Google Scholar 

  5. Bennett JK, Fuertes JN, Keitel M, Phillips R. The role of patient attachment and working alliance on patient adherence, satisfaction, and health-related quality of life in lupus treatment. Patient Educ Couns. 2011;85:53–9.

    Article  PubMed  Google Scholar 

  6. Ciechanowski PS, Katon WJ, Russo JE, Walker EA. The patient–provider relationship: attachment theory and adherence to treatment in diabetes. Am J Psychiatry. 2001;158:29–35.

    Article  CAS  PubMed  Google Scholar 

  7. Fowler JC, Groat M, Ulanday M. Attachment style and treatment completion among psychiatric inpatients with substance use disorders. Am J Addict. 2013;22:14–7.

    Article  PubMed  Google Scholar 

  8. Brorson HH, Ajo Arnevik E, Rand-Hendriksen K, Duckert F. Drop-out from addiction treatment: a systematic review of risk factors. Clin Psychol Rev. 2013;33:1010–24.

    Article  PubMed  Google Scholar 

  9. Oslin DW, Slaymaker VJ, Blow FC, Owen PL, Colleran C. Treatment outcomes for alcohol dependence among middle-aged and older adults. Addict Behav. 2005;30:1431–6.

    Article  PubMed  Google Scholar 

  10. Hiebler-Ragger M, Unterrainer HF, Rinner A, Kapfhammer HP. Insecure attachment styles and increased borderline personality organization in substance use disorders. Psychopathology. 2016;49:341–4.

    Article  PubMed  Google Scholar 

  11. World Health Organization. International classification of diseases, 10th version (ICD-10). Geneva: World Health Organization; 2015.

    Google Scholar 

  12. Hexel M. Validierung der deutschen Version des Attachment Style Questionnaire (ASQ) bei Personen mit und ohne psychiatrische Diagnosen [Validation of the German Version of the Attachment Style Questionnaire (ASQ) in participants with and without psychiatric diagnosis]. Z Klin Psychol Psychother. 2004;33:79–90.

    Article  Google Scholar 

  13. de Rick A, Vanheule S. Attachment styles in alcoholic inpatients. Eur Addict Res. 2007;13:101–8.

    Article  PubMed  Google Scholar 

  14. Ravitz P, Maunder R, Hunter J, Sthankiya B, Lancee W. Adult attachment measures: a 25-year review. J Psychosom Res. 2010;69:419–32.

    Article  PubMed  Google Scholar 

  15. Savov S, Atanassov N. Deficits of affect mentalization in patients with drug addiction: theoretical and clinical aspects. ISRN Addict. 2013.

    PubMed  PubMed Central  Google Scholar 

  16. Kernberg O. Borderline personality organization. J Am Psychoanal Assoc. 1967;15:641–85.

    Article  CAS  PubMed  Google Scholar 

  17. Chiesa M, Fonagy P, Holmes J, Drahorad C. Residential versus community treatment of personality disorders: a comparative study of three treatment programs. Am J Psychiatry. 2004;161:1463–70.

    Article  PubMed  Google Scholar 

  18. De Leon G. The therapeutic community: theory, model and method. New York: Springer Publishing Company; 2000.

    Google Scholar 

Download references

Authors’ contributions

JF, MHR, and HU conceptualized the work. JF and AR acquired the data. JF and MHR analyzed and interpreted the data. JF, KR and MHR drafted and revised the article. HU and HK supervised the analysis and interpretation of data as well as the drafting and revising of the article. All authors read and approved the final manuscript.


We would like to acknowledge the work of Nikolas Bonatos for making helpful and invaluable critical comments to the manuscript.

Competing interests

The authors declare that they have no competing interests.

Availability of data and methods

All data generated or analysed during this study are included in this published article.

Consent for publication

Not applicable.

Ethics approval and consent to participate

The study was approved by the ethics committee of the Medical University of Graz, Austria and all participants provided written consent to participate.


There was no funding for this study.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Author information

Authors and Affiliations


Corresponding author

Correspondence to H. F. Unterrainer.

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (, which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Fuchshuber, J., Hiebler-Ragger, M., Ragger, K. et al. Increased attachment security is related to early therapy drop-out in substance use disorders. BMC Res Notes 11, 141 (2018).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI:


  • Attachment styles
  • Cluster analysis
  • Substance use disorder
  • Therapeutic community
  • Treatment adherence