Abstract

The relationship between DSM V diagnosis and relationship anxiety, relationship avoidance, and attachment security was studied in 120 child and adolescent patients enrolled in a partial hospitalization psychiatric program.  Many statistically significant correlations were found between demographics (age, race, gender) and diagnosis; as well as relationship anxiety and attachment security.  Some DSM V diagnoses were also found to have significant correlations with relationship anxiety and attachment security.

Introduction

This aim of this study was to assess for a correlation (Pearson R) and significance of correlation (Chi-Square, P) between mental health diagnosis per DSM 5 diagnosis and attachment style (via relationship anxiety and relationship avoidance) in children and adolescents.

Previous research (Palitsky et al., 2013) concluded that in an adult population "insecure attachment styles were associated with greater reporting of suicidal ideation, (suicide) attempts, and all mental disorder categories analyzed" as well as that "secure attachment styles were associated with a decreased likelihood of reporting suicidal ideation, attempt, and any anxiety disorder."   Specific to adolescents and young adults, Brennan and Shaver (1998) found that in "a nonclinical group of 1407 individuals... results indicated substantial overlap between attachment and personality disorder measures.  Brennan and Shaver also noted that "the one personality disorder factor that is unrelated to attachment appears akin to psychopathy."

This relationship has also been studied by Fonagy (et al., 1996), using the Adult Attachment Interview (AAI) and DSM (3rd ed., rev.) diagnosis, who found that in nonpsychotic inpatients Axis I diagnosis anxiety was associated with unresolved attachment status and that AAI scales were able to discriminate between depression and eating disorder.  Fonagy also found that Borderline Personality Disorder was linked to the experience of severe trauma and lack of resolution (in attachment forming).  The study also noted that "results suggest that individuals rated as dismissing on the AAI are more likely to show improvements in psychotherapy."

Figure 1 (Fraley 2012)

The theory of attachment was developed by John Bowlby who was "primarily focused on the infant-caregiver relationship, he believed that attachment characterized human experience from 'the cradle to the grave' (Fraley, 2018)."  Following the work of Hazan and Shaver, Kelly Brennan conducted further research suggesting "that there are two fundamental dimensions with respect to adult attachment patterns" (Fraley, 2018), attachment-related anxiety and attachment-related avoidance.  These dimensions result in four attachment styles depicted in Figure 1.

The Experience in Close Relationships Questionnaire (ECR) is a 36-question assessment to measure avoidance and anxiety in relationships, thereby assessing for attachment style (Fraley 2012, 2018).  Higher scores in Avoidant Attachment reflect greater avoidance in close relationships, whereas higher scores in Anxious Attachment reflect greater anxiety in close relationships.  Attachment styles using the ECR and ECR-R have been studied previously (Fraley, 2012) where the later reduces the 36-question assessment to 19 questions.  The survey was further reduced in Shaver's 2017 presentation to 4 questions and administered to the attendees of the presentation.

Farley (2012) noted that "there is no 'natural' or 'correct' way to assign people to attachment categories or styles."  However, Bateman and Fonagy (2017, p.46) noted that 75% of patients who met criteria for BPD (Borderline Personality Disorder) fell into the rarely used adult attachment subgroup of "fearfully preoccupied with respect to trauma."  This suggests that there is a correlation between diagnostic criteria (and thereby diagnosis) and attachment style – with specific regard to personality disorders.  As such, the author hypothesizes that there is a statistically significant relationship between DSM V diagnosis and both relationship anxiety and relationship avoidance, thereby indicating a significant relationship between diagnosis and attachment style.  The null, then, being that there is not a statistically significant relationship between DMS V diagnosis and either relationship avoidance or relationship anxiety.

Procedure and Methods

A modified version of the ECR/ECR-R, as presented by P. Shaver (2017)^[1], was administered to children and adolescents enrolled in a Partial Hospitalization Program and Intensive Outpatient Program at an outpatient psychiatric hospital.  Children were asked to read and complete the survey in Figure 2 below. 

Figure 2

Rate each of the following relationship styles according to how you think each description describes your general relationship style using a 1-to-7 scale where 1 = not at all like me, 4 = somewhat like or unsure, and 7 = very much like me.

_____ A) It's easy for me to become emotionally close to others. I am comfortable depending on them and having them depend on me. I don't worry about being alone or having others not accept me.

_____ B) I am uncomfortable getting close to others. I want emotionally close relationships, but I find it difficult to trust others completely, or to depend on them. I worry that I will be hurt if I allow myself to become too close to others.

_____ C) I want to be completely emotionally intimate with others, but I often find that others are reluctant to get as close as I would like. I am uncomfortable being without close relationships, but I sometimes worry that others don't value me as much as I value them.

_____ D) I am comfortable without close emotional relationships. It is very important to me to feel independent and self-sufficient, and I prefer not to depend on others or have others depend on me.

In order to facilitate accurate responses with consideration of age, developmental level, and cognitive functioning level of participants, all participants were given the opportunity to ask questions about or for clarification of the above statements.  However, clarification was not provided unless explicitly solicited from administering clinician by the participant.

Responses to the above (A,B,C,D) were used to compute an X-coordinate (representing attachment anxiety) and a Y-coordinate (representing attachment avoidance) where  X = (B + C) - (A+ D) and Y = (B + D) - (A + C).  Coordinates contained in Quadrant I represent a fearful attachment style.  Coordinates contained in Quadrant II represent a dismissive attachment style.  Coordinates contained in Quadrant III represent a secure attachment style.  Coordinates contained in Quadrant IV represent a preoccupied attachment style.

The R Programming Language (Maechler et al., 2013) was used to compute statistical correlation (Pearson r) and confidence intervals (p) for identified variables listed in results below.  The programming script used to compute the results from the raw data is open source and available online.^[2] 

Samples were collected during an 9 month period (from December 2017 until August 2018).  Diagnosis were assigned by attending psychiatrist or psychiatric nurse practitioner per DMS V criteria.  Of note, some diagnosis change throughout the course of treatment as new information is obtained and symptomology and case conceptualization are better understood.  However, for the purposes of this research the only diagnosis included were those give upon admission to the program by attending psychiatrist or nurse practitioner.  Also of note.  rule out (R/O) diagnosis were not included in the collected data.  However, information regarding rule in diagnosis (R/I), not otherwise specified (NOS), and otherwise specified (OS) was included in the data.  

Specifiers such as "with psychotic features" were not distinguished from "without psychotic features" in diagnosis such as Bipolar Disorder (BP) and Major Depressive Disorder (MDD) where such specifiers are noted by diagnosing physician.  In the case of psychotic features, those diagnosis were categorized by their primary diagnose (e.g. MDD, BP) rather than as a Schizophrenia Spectrum Disorder (including Schizophrenia, Schizoid Personality Disorder, Schizophreniform Disorder, Schizoaffective Disorder, and Schizotypal Disorder).

Participants

A sample of 120 child and adolescent patients (mean age 13.98 years) actively enrolled in a partial hospitalization and intensive outpatient programs were sampled with the above assessment.  All participants were receiving mental health services inclusive of group and family counseling as well as psychiatric medication management.  Sample racial, gender, and diagnostic demographics are noted in Table 1.  Patients who were admitted to the treatment program, sampled, and later readmitted to program after discharge did not complete subsequent surveys.  

Table 1

Results

Several notable demographic statistics can be observed from this data.  65% of patients sampled were of Caucasian, 70.8% were female gender, 73% were diagnosed with MDD, 45% were diagnosed with ADHD, and 41% were diagnosed with PTSD or trauma related stressors.

In this sample, the mean relationship anxiety score was –0.6 while the mean relationship avoidance score was 0.9833.  Given these results, the mean of the population sampled is graphed into the avoidant insecure attachment quadrant of attachment styles.  Of the population sampled, 20% (N = 24) met criteria for having secure attachments (negative values for both relationship anxiety and relationship avoidance) while 80% (N = 96) did not meet criteria for the secure attachment quadrant.  These results differ notably from Brennan et. al (1998) who's assessment of "thousands" of adults found a mean attachment (relationship) anxiety of 1.3 and a mean attachment (relationship) avoidance of 3.15.  It should be noted that in both populations, relationship / attachment avoidance was more than double relationship / attachment anxiety.

Table 2 shows the results of significant (p < 0.05) correlations (r) between variables (age, race, gender, and diagnosis) and relationship anxiety and / or relationship avoidance; as well as the strength of correlation (p).  Of note, there were no variables with a correlation (p) of less than 0.05 with both relationship anxiety and relationship avoidance which confirms the null hypothesis that there is not a significant relationship between diagnosis and attachment style.  However, several diagnoses did have statistically significant (p < 0.05) correlations with attachment style when attachments styles were categorically simplified to secure (negative values for both relationship anxiety and relationship avoidance) and insecure (a non-negative numeral for either relationship anxiety or relationship avoidance).

Table 2

Notice should be given to the sample size of each diagnostic category.  While the overall sample size was large (N = 120), some diagnoses have very small sample sizes (N < 10); as noted in Table 1.

There were no variables that had a significant relationship with both relationship anxiety and relationship avoidant, suggesting that there is not a significant relationship between diagnosis and attachment style; that is, between secure, preoccupied, fearful, and dismissive attachment styles.  However, when attachment categories were simplified to secure (negative value for both relationship avoidance and relationship anxiety) or not secure; there were some significant correlations found.

Age (N = 120) had a positive correlation with relationship avoidance (r = 0.27, p = 0.0026) with greater than 99% confidence.  Male gender (N=34) was also found to have a negative correlation with relationship anxiety (r = -0.28. p = 0.0023) and a positive correlation with secure attachment (r = 0.29, p = 0.0015), both with greater than 99% confidence.  Female gender (N = 85) was found to have a positive correlation (r = 0.25, p = 0.0054) and a negative correlation with secure attachment (r = -0.28, p = 0.0024) with greater than 99% confidence.

Major Depressive Disorder (MDD, N = 73) was found to have a positive correlation (r = 0.25, p = 0.0062) with relationship anxiety at greater than 99% confidence.  Attention Deficit Hyperactivity Disorder (ADHD, N = 45) had a positive correlation (r = 0.22, p = 0.0181) with relationship anxiety at greater than 98% confidence.  Oppositional Defiant Disorder (ODD, N = 24) also had a positive correlation with relationship anxiety (r = 0.23, p = 0.0339) at greater than 95% confidence.

Some other significant correlations were observed where both category variables contained large samples (N > 20).  African American Race (N = 32) had a negative correlation with Oppositional Defiant Disorder (r = -0.22, p = 0.0174).  Male gender had positive correlations with ADHD (r = 0.35, p = 0.0000) and ODD (r = 0.19, p = 0.0336).  Similarly, female gender had negative correlations with ADHD (r = -0.34, p = 0.002) and ODD (r = -0.54, p = 0.0450).  Bipolar Disorder was found to have a negative correlation (r = -0.54, p = 0.0000) with MDD, but MDD was also found to have a positive correlation with Generalized Anxiety Disorder (r = 0.24, p = 0.0081).

Discussion

Studying attachment styles and their correlation to mental health diagnosis may provide additional insight to allow for more intentionally chosen and mindfully endorsed clinical interventions and clinician education and training.  Randomly Controlled Trials of empirically supported treatments (Target, 2014) demonstrated that 5 – 10% of patients demonstrate deterioration in treatment; thus, demonstrating a continued need for exploration and re-evaluation of how treatment modalities are informed and implemented.

The above data collected, in addition to and in conjunction with diagnostic and attachment style statistics, can directly inform programmatic approaches, training opportunities offered by the facility and education team, as well as treatment specialization and certification among clinical staff to directly addressing and related to milieu presenting to program.

While no diagnostic categories had significant correlations to both relationship anxiety and relationship avoidance in this sample, six categories did have significant correlations to relationship anxiety alone, which may suggest that the measure used is a more accurate indicator of relationship anxiety than overall attachment style.

Similarly, five diagnostic categories had significant correlations with secure attachments which may suggest that, while the correlation between diagnosis and attachment style is not significant when comparing all four attachment styles (secure, preoccupied, avoidant, and fearful), when the attachment categories are reduced either secure or insecure there is a stronger correlation in the data.

Many statistically significant correlations (see Table 2) were within a –0.3 / 0.3 margin, which suggests that while the correlations were statistically significant, the correlation is relatively weak.  However, there were many occasions where there were particularly strong relationships outside of the –0.3 / 0.3 margin.  The correlation (p) between male gender and ADHD diagnosis was 0.35 with a p-value of less than 0.0000; while ADHD diagnosis had a correlation of –0.34 with female gender (r = 0.0002).  The correlation (p) between male gender and Intermittent Explosive Disorder diagnosis was 0.32 with a p-value of 0.0004, and was nearly mirrored by female gender as well (p = -0.31, r = 0.0054).  

Additionally, the correlation (p) between Autism Spectrum Disorder (N=7) diagnosis and Obsessive Compulsive Disorder diagnosis was 0.37 with a p-value of 0.0000.  The correlation (p) between Bipolar Disorder and Major Depressive Disorder was –0.54 (r = 0.0000) and the correlation (p) between ADHD and Intermittent Explosive Disorder was 0.32 (r = 0.0003).

There are several areas for improvement upon this study, allowing for more strictly controlled variables and more precise outcomes.  Firstly, the overall sample size remains relatively small considering the vast variance between the categories examined (diagnosis, relationship anxiety, relationship avoidance, age, gender, race, etc...).  While the overall sample size was considerable, certain diagnostic categories were underrepresented when compared to the general population.  The overall task of mental health diagnostics with children is undoubtedly nebulous task, however it has noted (see Bateman & Fonagy, 2016, p.149 and Laurenssen ,2013) that there is a bias among mental health professionals against assigning certain diagnoses where "fears are rightly expressed about pejorative overtones, judgmental attitudes, blaming the patient, attacking the very 'soul' of the individual, and stigmatizing the patient for life."

Secondly, multiple, more comprehensive measures of attachment style (relationship anxiety and relationship avoidance) have been published such as the original 36 question ECR and 19 question ECR-R which may provide more detailed and accurate results.  In addition, the results were provided by self-report which may produce skewed results for reasons including, but not limited to: patient capacity for reading comprehension, patient capacity for insight and awareness of their relationship / attachment style, and attachment styles themselves affecting reported measures (e.g. persons with avoidant attachments underreporting both anxiety and avoidance in relationships).

Third, diagnostic criteria itself may be a confounding variable in multiple ways.  For example, MDD Criteria A3 denotes "significant weight loss or gain" and Criteria A4 denotes "insomnia or hypersomnia" (APA, p.94, 2016), any of which would present in varying symptomatic (and thereby phenomenological) ways, yet would all be categorized under a positive MDD diagnosis.  Similarly, Post Traumatic Stress Disorder (PTSD) Criteria C explicitly describes "persistent avoidance" while Criteria E denotes "Marked alterations in arousal and reactivity" characterized by irritability, recklessness, exaggerated startle response, problems with concentration, hypervigilance, or sleep disturbance" (APA, p.143-149, 2016).  Further still, diagnostic criteria for some diagnosis are subsumed by other diagnosis or even required a prerequisite criterion (see Conduct Disorder (APA, p.221-224, 2016) and Oppositional Defiant Disorder (313.81, F91.3); as well as Disruptive Mood Dysregulation Disorder (296.99, F34.8), Cyclothymic Disorder (301.13, F34.0), and Bipolar Disorder (APA, p.65-75, 2016)).

Presentation in treatment is also a dynamic process.  That is, as more information comes forth during treatment, diagnostic status may be altered, or all together changed from the assigned diagnosis during admission to treatment; with specific consideration to 'rule-in' and 'rule-out' diagnoses.

Similarly, with specific reference to personality disorders, the National Institute of Mental Health estimated in 2017 that the prevalence of personality disorders among adolescents in the general population was between 10 and 15%; specifically, Borderline Personality Disorder was estimated at 1.4%.  However, in the studied sample of patients actively receiving mental health services Borderline Personality Disorder was the only noted personality disorder and accounted for only 2.5% of the documented diagnoses.  The disparity suggests that there is indeed either a great avoidance of or overall underreporting of some mental health diagnoses – specifically personality disorders.  This largely confounds the ability of researchers to accurately studied correlations between diagnostic criteria and other factors which could largely improve the efficacy of mental health services.

Conclusion

There were no statistically significant correlations found between DSM V diagnosis and both relationship anxiety and relationship avoidance using a shortened version of the ECR.  This suggests that there is not a statistically significant correlation between mental health diagnosis and attachment style (secure, avoidant, preoccupied, fearful).  However, six of twenty sampled diagnoses did have statistically significant correlations with relationship anxiety alone using the identified measure.  Also, five of twenty sampled diagnoses had statistically significant correlations with secure attachment (versus insecure attachment) which suggests that the simplified attachment categories are more accurately measured.

In several cases where a statistically significant correlations were found, the sample size for that particular diagnosis was very low (N<10).  There also seem to be inconsistent and largely varying diagnostic practices and criteria which indicate varying attachment styles and degrees for relationship anxiety and avoidance within the same diagnosis or spectrum of diagnoses.

The relationship between diagnosis and attachment style warrants further study and exploration as many diagnoses manifest phenomenologically parallel to insecure attachment styles (preoccupied, fearful, and avoidant) and comorbid social and interpersonal dysfunction.

References

American Psychiatric Association. (2016). Desk reference to the diagnostic criteria from

            DSM-5. Arlington: American Psychiatric Association.

Bateman, A., & Fonagy, P. (2016). Mentalization based treatment for personality disorders: A

practical guide. Oxford: Oxford University Press.

Brennan, K.A., Clark, C.L., & Shaver, P.R. (1998). Self-report measurement of adult romantic

attachment: An integrative overview. In J.A. Simpson & W.S Rholes (Eds.), Attachment

theory and close relationships (pp. 46-76). New York:  Guilford Press

Brennan, K. A., & Shaver, P. R. (1998). Attachment Styles and Personality Disorders: Their

Connections to Each Other and to Parental Divorce, Parental Death, and Perceptions

of Parental Caregiving. Journal of Personality, 66(5), 835-878.

doi:10.1111/1467-6494.00034

Fonagy, P., Leigh, T., Steele, M., Steele, H., Kennedy, R., Mattoon, G., . . . Gerber, A. (1996).

The relation of attachment status, psychiatric classification, and response to

psychotherapy. Journal of Consulting and Clinical Psychology, 64(1), 22-31.

doi:10.1037/0022-006x.64.1.22

Fraley, R. C. (2012, November). Information on the Experiences in Close Relationships-

Revised (ECR-R) Adult Attachment Questionnaire. Retrieved March 28, 2018, from

http://internal.psychology.illinois.edu/~rcfraley/measures/ecrr.htm

Fraley, R. Chris (2018). Adult Attachment Theory and Research: A brief overview. Retrieved

June 6, 2018, from http://labs.psychology.illinois.edu/~rcfraley/attachment.htm

Laurenssen, E. M., Hutsebaut, J., Feenstra, D. J., Busschbach, J. J., & Luyten, P. (2013).

Diagnosis of personality disorders in adolescents: A study among psychologists. Child

and Adolescent Psychiatry and Mental Health,7(1), 3. doi:10.1186/1753-2000-7-3

 Maechler, M., Rousseeuw, P., Struyf, A., Hubert, M., Hornik,  K.(2013).  cluster: Cluster

Analysis Basics and Extensions. R package  version 1.14.4.

National Institute of Mental Health. (2017, November). Personality Disorders. Retrieved June

22, 2018, from https://www.nimh.nih.gov/health/statistics/personality-disorders.shtml

Palitsky, D., Mota, N., Afifi, T. O., Downs, A. C., & Sareen, J. (2013). The Association

Between Adult Attachment Style, Mental Disorders, and Suicidality. The Journal of

Nervous and Mental Disease, 201(7), 579-586. doi:10.1097/nmd.0b013e31829829ab

Shaver, P.R. (2017, October). American Academy of Psychotherapists Institute and

Conference:  Plenary. Santa Fe, NM

Target, M., PhD. (2014, May). Attachment and metnalization: Common factors in effective

therapeutic practice. Retrieved June 13, 2018, from

https://www.ranzcp.org/Files/Resources/Conference-presentations/Mary-Target-

attachment-and-mentalisation.aspx