Measuring Attachment disorders in school-aged children: Validation of the ETRADQ (Early TRAuma-related Disorders Questionnaire): Final data

• In Event: Poster Session 11
  In Poster Session: PS 11 Topic - Developmental Psychopathology

Abstract

Introduction: Attachment Disorders (AD), namely Reactive Attachment Disorder (RAD) and Disinhibited Social Engagement Disorder (DSED), are psychiatric disorders that children may develop as a result of severe early adversity (e.g., maltreatment, institutionalization, repeated change of caregivers). Although measures of AD are currently available, they present several weaknesses: most focus only on DSED, none has been through a complete validation process and none are based on the new diagnostic criteria of DSM-5 (APA, 2013). Consequently, we developed the ETRADQ (Early TRAuma-related Disorders Questionnaire; Monette, submitted), a new questionnaire measuring both RAD and DSED symptoms and based on DSM-5 criteria. The objective of this study is to present the ETRADQ and data supporting its reliability and validity.

Method
Participants. The sample consists of 823 primary caregivers of 6-12 years old children from the community (n = 578) or at risk of an AD (adopted, foster care or residential care children, n = 275).

Measures. Primary caregivers completed the following questionnaires: 1. Sociodemographic questionnaire; 2. ETRADQ (42 items); and 3. Relationship problem questionnaire (RPQ, Minnis et al., 2002), a screening questionnaire to assess RAD (6 items) and DSED (4 items), based on DSM-IV and ICD-10 criteria. Primary caregivers were also questioned using the RADA interview (Lehmann et al., 2020), a recently updated semi-structured interview assessing RAD (11 items) and DSED (9 items) based on the DSM-5 criteria, originating from the CAPA-RAD (Minnis et al., 2009).

Results and discussion: Confirmatory factor analysis (see Figure 1) with WLSMV estimator on ETRADQ items indicates that the best fitting model includes two second-order factors: 1) a RAD scale (with 3 subscales: Low selective attachment, Low social and emotional responsiveness, Emotional unpredictability) and a DSED scale (with 2 subscales: Interaction with unfamiliar adults, Social disinhibition). Fit indices indicate an excellent fit of the model (X2 = 2102.827, df = 774, RMSEA = .046, CFI = .959, TLI = .957, SRMR = .061). The RAD and DSED scales are moderately correlated (r = .60). Because the item measuring DSED criterion A3 saturated on the RAD factor, instead of the DSED factor, it was removed. Other researchers have obtained similar results (Dobrova-Krol et al., 2010, Pears et al., 2010), which calls into question the relevance of the DSED criterion A3.
The ETRADQ RAD scale (sum of 25 items) correlates in the expected direction with the RPQ RAD scale (r = .83) and the RADA interview RAD scale (r = .75). The ETRAD-Q DSED scale (sum of 16 items) correlates in the expected direction with the RPQ DSED scale (r = .89) and the RADA interview DSED scale (r = .78). Internal consistency of both RAD and DSED ETRAD-Q scales are very high (McDonald Omega: > .88). Test-retest correlations are also very high (r >.90) for both scales. Finally, at-risk groups (adopted, foster care, residential care children) all showed higher scores on ETRADQ scales, compared to children from the general population. These results indicate that the ETRADQ has excellent psychometric properties.

Authors

• Sebastien Monette, Université du Québec à Montréal (UQAM)
  Presenting Author

• Chantal Cyr, Université du Québec à Montréal (UQAM)
  Non-Presenting Author

• Miguel M. Terradas, Univerisité de Sherbrooke
  Non-Presenting Author

• Sophie Couture, Université de Sherbrooke
  Non-Presenting Author