Scandinavian Journal of Psychology, 2015, 56, 53–61
DOI: 10.1111/sjop.12174
Health and Disability Associations between attachment-related symptoms and later psychological problems among international adoptees: Results from the FinAdo study € €A € 5 and HELENA LAPINLEIMU6 MARKO ELOVAINIO,1,2 HANNA RAASKA,3 JARI SINKKONEN,4 SANNA MAKIP A 1
National Institute for Health and Welfare, Helsinki, Finland Institute for Behavioural Siences, University of Helsinki, Finland 3 University of Helsinki, Helsinki University Hospital, Finland 4 University of Turku and Save the Children, Finland 5 Capacitas Familia, Helsinki, Finland 6 Department of Pediatrics, Turku University Hospital and University of Turku, Finland 2
Elovainio, M., Raaska, H., Sinkkonen, J., M€akip€a€a, S. & Lapinleimu, H. (2015). Associations between attachment-related symptoms and later psychological problems among international adoptees: results from the FinAdo study. Scandinavian Journal of Psychology, 56, 53–61. We examined the associations between attachment-related symptoms (symptoms of reactive attachment disorder (RAD), symptoms of disinhibited social engagement disorder (DSED), and clinging) and later psychological problems among international adoptees. The study population comprised internationally adopted children (591 boys and 768 girls, 6–15 years) from the ongoing Finnish Adoption (FinAdo) study. Data were gathered with self-administered questionnaires both from adoptive parents and from adoptees aged over 9 years. Attachment-related symptoms were measured using of a short (8-item) questionnaire and later behavioral/emotional problems were assessed using the Child Behavior Checklist (CBCL) and the Five to Fifteen (FTF) scale for attention-deficit/hyperactivity disorder (ADHD) symptoms. RAD and DSED symptom subscales were associated with an increased risk of emotional and behavioral problems and ADHD. Especially the mixed type of attachment-related symptoms was strongly associated with later emotional and behavioral problems. Key words: Attachment, adoption, survey. Marko Elovainio, National Institute for Health and Welfare, PO Box 30, 00271 Helsinki, Finland. E-mail: marko.elovainio@thl.fi
INTRODUCTION Children raised in institutions often exhibit typical and recognizable behavioral patterns termed in earlier literature as “excessive need for adult attention,” “indiscriminate friendliness” (O’Connor, Marvin, Rutter, Olrick & Britner, 2003; O’Connor & Rutter, 2000), or lack of comfort-seeking even when distressed (O’Connor & Rutter, 2000). Several studies have shown that a key etiological factor in all of these patterns is the absence of a consistent caregiver. In the previous diagnostic manual (DSM-IV from 1994), these patterns were included in reactive attachment disorder (RAD). RAD was defined as aberrant social behavior divided into two basic types, inhibited and disinhibited. Children defined as suffering from the inhibited subtype typically resist initiating or responding to social interactions. They are emotionally withdrawn, do not seek comfort from others, and their reactions especially in distress are typically inhibited, vigilant, or ambivalent. In the current diagnostic manuals (DSM-V and ICD-10), RAD is defined based on this inhibited subtype. In the DSM-V, the disinhibited attachment type is currently termed the disinhibited social engagement disorder (DSED). Children suffering from this disorder are typically non-selective or indiscriminate in their attachment behavior. The child will readily go off with strangers, show limited differentiation among adults, and fail to check back with the parent in stressful situations. He or she has difficulties in forming close, confiding relationships. The symptoms of attachment disorder typically present
© 2014 Scandinavian Psychological Associations and John Wiley & Sons Ltd
with the closest attachment figures, for example, the parents (Boris & Zeanah, 2005; Rutter, 1998; Rutter et al., 2007; Rutter, Kreppner & O’Connor, 2001; Rutter & Sonuga-Barke, 2010). In addition to these two disorders, also a mixed behavioral pattern with features from both disorders has been reported (Gleason, Fox, Drury et al., 2011; Smyke, Dumitrescu & Zeanah, 2002; Zeanah, Scheeringa, Boris, Heller, Smyke & Trapani, 2004; Zeanah, Smyke & Dumitrescu, 2002), and in the DC:0-3R it is included in the deprivation/maltreatment disorder as a subtype of its own (Zero to Three, 2005). Although clinging is mentioned in ICD-10 as one symptom of indiscriminate RAD subtype in toddlers, few previous studies have taken this symptom into account. Excessively clingy behavior is often described by adoptive parents, but its significance with respect to other attachment-related symptoms is far from clear. In this study, we refer to symptoms of all of these disorders (RAD, DSED, and clinging) as attachment-related symptoms. Disordered attachment-related behavior patterns have been suggested to be linked to severe maltreatment and caregiving neglect (O’Connor & Rutter, 2000). Children with an institutional background and also children adopted from foster care have been shown to have a higher risk of attachment disturbances (Bruce, Tarullo & Gunnar, 2009; Smyke et al., 2002). Length of time in institutional care has also been reported to be associated with attachment-related symptoms (Bruce et al., 2009). Specifically, the symptoms of RAD (previously inhibited
54 M. Elovainio et al. subtype) have been associated with caregiving quality features, such as lack of sensitivity, flat affect, and detachment, in previous studies (Gleason et al., 2011). Relatively little research has focused on the developmental patterns of attachment-related symptoms, on the relationship between such symptoms and other childhood psychiatric disorders, or on the potential long-term consequences of attachmentrelated symptoms. Severe attachment-related symptoms have, however, been associated with subsequent functional impairment. RAD (previously inhibited subtype) has been shown to be associated with depressive symptoms (Gleason et al., 2011), and DSED (previously disinhibited RAD) with symptoms resembling attention-deficit/hyperactivity disorder (ADHD) (Bruce et al., 2009; Gleason et al., 2011). In addition, in severely deprived Romanian adoptees, DSED was associated with inattention, overactivity, cognitive impairment, and “quasi-autistic” features, forming deprivation-specific patterns (Kumsta, Kreppner, Rutter et al., 2010). McGoron and others (2012) showed that attachment security was associated with symptoms of multiple forms of psychopathology, including ADHD and both internalizing and externalizing symptoms. Attachment behaviors in young children have been proposed to reflect a child’s sense of predictability and protection by his/her caregiver. These perceptions have been revealed to be associated with enhanced emotion and behavioral regulation abilities, socially engaging behaviors, and social competence. Problems in these domains are linked to RAD and socially indiscriminant behavior as well as to internalizing and externalizing symptoms (DeKlyen & Greenberg, 2008). In this study, we tested whether attachment-related symptoms were associated with later behavioral/emotional problems (measured using the Child Behavior Checklist (CBCL)) and ADHD symptoms (measured using the Five to Fifteen (FTF) scale) (Achenbach & Rescorla, 2001; Kadesjo et al., 2004). Based on the previous literature, we hypothesized that (1) RAD symptoms would be associated with internalizing problems, (2) DSED would be associated with externalizing problems and ADHD, (3) the mixed subtype would associated with externalizing problems, internalizing problems, and ADHD, and (4) clinging would be associated with internalizing problems.
METHODS Study population This study is a part of the ongoing Finnish Adoption (FinAdo) study, including all internationally adopted children (591 boys and 768 girls, response rate 55.7%) through the three legalized adoption organizations in Finland during 1985–2007. Data were gathered with self-administered questionnaires both from the adoptive parents and from adoptees aged over 9 years (in 2007–2009). In the current study, a subsample of children aged 6–15 years with no missing values in attachment-related symptom measures was used (N = 853).
Scand J Psychol 56 (2015) Parents were asked to assess the following behavioral symptoms: (1) (2) (3) (4) (5) (6) (7) (8)
child readily goes off with a stranger; child does not check back with parent even in a stressful situation; child seeks comfort from parents when distressed; child is scared and wary and does not calm down when parent is comforting or soothing him/her; child withdraws from contact with others; child appears apathetic and hopeless; child is excessively clingy to both parents; and child is excessively clingy to one of the parents.
For each item, parents were asked to evaluate the severity of the symptom on a scale ranging from 0 to 2: She or he had “No / never such behavior = 0,” “Some or sometimes such behavior = 1,” “A lot or often such behavior = 2.” The instructions were: “Did you child behave in ways described in the following items when he/she arrived to Finland for the first time?”
Behavioral and emotional problems We used the Child Behavior Checklist (CBCL) to evaluate behavioral and emotional problems. In addition to the total score, we grouped symptoms into internalizing versus externalizing, which is the most commonly used distinction. Internalizing symptoms reflect problems mainly within the self, such as anxiety, depression, somatic complaints without medical cause, and withdrawal from social contacts. Externalizing symptoms represent conflicts with other people and their expectations for children’s behavior (Achenbach & Rescorla, 2001). Each item is rated as 0 = “not true,” 1 = “somewhat or sometimes true,” and 2 = “very true or often true.” The CBCL has previously been used in internationally adopted children and has shown good psychometric properties (Xing Tan, Marfi & Dedrick, 2010).
ADHD ADHD symptoms of the adoptees were evaluated by using part of the Five to Fifteen (FTF) parental questionnaire (Kadesjo et al., 2004; Trillingsgaard, Damm, Sommer et al., 2004). Overall, FTF comprises 181 statements related to behavioral or developmental problems. The items are grouped into eight domains (memory, learning, language, executive functions, motor skills, perception, social skills, and emotional/ behavioral problems). The domain of executive functions consists of 18 items reflecting ADHD symptoms, based on DSM-IV criteria. Symptoms can be grouped into inattentive and hyperactive-impulsive subtypes. Each item has three rating alternatives (0 = “does not apply,” 1 = “applies sometimes or to some extent,” 2 = “definitely applies”). Cronbach’s alpha of the FTF dimensions ranged from 0.97 to 0.99.
Attachment The Kinship Center Attachment Questionnaire (KCAQ) is a quantitative screening instrument (20 items, response format from 0 = never/rarely to 6 = almost always) of child attachment completed by the caregiver. It is appropriate for use with children younger than 6 years and has shown application in tracking difficulties in attachment over time (Kappenberg & Halpern, 2006). The total score comprises four factors: 1 = Positive adjustment/development (six items), 2 = Negative behavior (six items), 3 = Emotional reactivity (four items), 4 = Distancing from caregiver support (four items). In addition it has been suggested that the KCAQ can discriminate between children referred for mental health services and children from a non-clinical sample with a strong effect of group membership on KCAQ total score (Cohen’s d = 0.79) (Kappenberg & Halpern, 2006).
Measurements Attachment-related symptoms. The attachment-related symptoms measure was developed based on the Disturbances of Attachment Interview (Smyke et al., 2002) and clinically relevant symptoms suggesting RAD and DSED and clinging from DSM-V and ICD-10. © 2014 Scandinavian Psychological Associations and John Wiley & Sons Ltd
Covariates The parental questionnaire of the FinAdo study included questions about the child’s background and life circumstances in the birth country. The
Results from the FinAdo study 55
Scand J Psychol 56 (2015) age at adoption, age at time of evaluation, and sex of the child were recorded. The country of birth was classified into four classes based on the continent of birth: Asia (China, Thailand, Philippines, India, Vietnam, Sri Lanka, Pakistan, Nepal, Taiwan, Southern Korea, Israel), Africa (Southern Africa, Ethiopia, Mozambique), America (Colombia, Guatemala, the Dominican Republic, USA, Canada), and Eastern Europe (Russia, Romania, Estonia, Poland). For the statistical analysis, the quality and number of different placements before adoption were classified into four classes: 1 = “only foster home,” 2 = “only one orphanage,” 3 = “2–3 placements before adoption,” and 4 = “4 or more placements before adoption.” The adoptive family’s socioeconomic status (SES) was classified into three classes. The child’s physical health was screened by ascertaining any medical diagnoses that the child had received in Finnish medical evaluations conducted by a doctor. Disabilities were defined as deafness or hearing impairment of both ears, blindness or poor vision in both eyes, intellectual impairment, mental retardation, and autism.
Statistical analysis Because the attachment-related symptoms scale is not a widely used and established scale, its factorial validity was tested with exploratory factor analyses (varimax rotation) (Maharee-Lawler, Rodwell & Noblet, 2010) and structural equation modeling (J€oreskog, 1993). First, we tested the factor structure and number of dimensions using exploratory factor analyses with eigenvalue 1 as a criterion for the appropriate number of factors. Second, we tested the final structure of the RAD measure using structural equation modeling. We used the v2 test, the goodness-of-fit index (GFI), which indicates the amount of variance and covariance explained by the model (values over 0.90 indicating acceptable fit), the adjusted goodness-of-fit index (AGFI, which adjusts the GFI for degrees of freedom in the model, with values above 0.90 indicating a good fit), and the RMSEA criterion (Kelloway, 1998). The comparative fit index (CFI) was used as a measure of comparative fit, with values exceeding 0.90 indicating a good fit. Parsimonious fit was measured by consistent Akaike information criterion (CAIC). Testing the final structure was done in two steps. A one-factor model was first estimated, where all remaining items loaded on the same underlying dimension (null model). Then a second-order model was estimated that represented the theoretical model, with first-order factors loading on the general attachment-related symptoms factor. The associations between the attachment-related symptoms measure and outcomes were examined by linear and logistic regression analyses. First, we calculated an unadjusted model and then we adjusted for age, gender, age at adoption, continent of birth, number and quality of pre-adoption placements, and family SES. We used SAS for Windows version 9.2 and LISREL 9 for the statistical analyses.
RESULTS Characteristics of subjects are shown in Table 1. The mean age of participants (girls 54%) was 2.4 years at the time of adoption and 8.5 years at the time of data collection. Most of the children came from Asia. Of the potential confounders, age at adoption ( p < 0.001), gender ( p = 0.02), number of placement ( p = 0.01), country of origin ( p < 0.001), and physical health status ( p < 0.001) were associated with attachment-related symptoms. The exploratory factor analyses with varimax rotation suggested that the three-factor solution offered the best fit to the data in both samples following the eigenvalue > 1. In the threefactor solution, all items had the strongest loading to their corresponding factors: inhibited (RAD), disinhibited (DSED), and clinging (Table 2). The final structural validity of the attachment-related symptoms scale was tested with confirmatory factor analysis. We © 2014 Scandinavian Psychological Associations and John Wiley & Sons Ltd
Table 1. Characteristics of the study population N = 853 Mean (SD) Age 8.5 (2.9) Age at adoption 2.4 (1.3) Sex Girl Boy Continent of birth Asia Africa America Eastern Europe Number and quality of pre-adoption placements One foster home One orphanage Many placements Parental socioeconomic status (SES) High Middle Low Attention-deficit/hyperactivity disorder symptoms (FTF) Hyperactivity (mean range 0–2) 0.5 (0.9) Attention problems (mean range 0–6) 0.4 (0.9) Behavioral control problems (mean range 0–8) 0.7 (0.9) Psychological/behavioral symptoms (CBCL) Internalizing (range 0–32) 4.9 (5.2) Externalizing (range 0–55) 8.2 (3.2) Total symptoms score (range 0–115) 24.3 (22.8)
N (%)
461 (54) 390 (46) 435 92 136 187
(51) (11) (16) (22)
72 (9) 465 (55) 315 (36) 483 (60) 176 (21) 149 (19)
Table 2. Exploratory factor analysis of attachment related symptoms for international adoptees (n = 850). Varimax rotation Factor loadings Items Child readily goes off with a stranger Child does not check back with parent even in a stressful situation Child seeks comfort from parents when distressed Child is scared and wary and does not calm down when parent comforts or soothes him/her Child withdraws from contact with others Child is apathetic and hopeless Child is excessively clingy to both parents Child is excessively clingy to one of the parents Eigenvalues
Disinhibited
Inhibited
Clinging
0.797
0.025
0.068
0.793
0.210
0.097
0.669
0.393
0.049
0.278
0.596
0.183
0.216
0.734
0.101
0.041 0.133
0.738 0.138
0.065 0.826
0.132
0.283
0.745
2.58
1.32
1.01
tested the overall fitness of the factor structure with definite loading structure and whether the items corresponded to the underlying theoretical construct. Table 3 shows the summary of the fit-indices for the two competitive models of the
56 M. Elovainio et al.
Scand J Psychol 56 (2015)
dimensions were associated with both of the psychological variables, whereas clinging was only weakly associated with ADHD and not associated with behavioral problems. Second, we constructed the final attachment-related symptoms measure with four categories: (1) no symptoms (n = 348, 41%); (2) only disinhibited symptoms (n = 153, 19%); (3) only inhibited symptoms (n = 137, 18%); and (4) (both disinhibited and inhibited symptoms) mixed subtype (n = 214, 25%), and tested whether this measure was associated with internalizing/externalizing behavioral problems or ADHD symptoms (Tables 4 and 5). Compared with children with no symptoms, those with any of the attachment-related symptoms had more internalizing or externalizing behavioral problems and more ADHD symptoms. These associations were robust and remained despite adjustments for age, age at adoption, number and quality of pre-adoption placements, and family SES. Children who belonged to the mixed subtype category had much more symptoms and problems than those in the other categories. Because the items measuring attachment-related symptoms may be restricted to subjects aged under 6 years, we replicated the analyses using those under 6 years of age upon arrival to Finland. No changes in the results were found (all p-values < 0.001, except between attachment symptoms and attention problems, p = 0.230). As additional analyses, we tested the association between our attachment-related symptoms measure and KCAQ. The results in the subpopulation of 6 years of age or under presented in Table 6 showed that those with attachment-related symptoms had significantly more attachment problems than those who did not have such symptoms. Again, those who belonged to the mixed subtype had the most problems.
Table 3. Summary of goodness-of-fit for the models of confirmatory factor analyses in international adoptees (n = 850) Fit-indices
Null model
Measurement model
v2 df v2 difference df p-value GFI NFI RMSEA CAIC
851.70 20
340.09 17 511.61 3 0.000 0.91 0.91 0.149 487.23
0.81 0.81 0.223 975.66
Note: Null model: A one-factor model where all items load on a single factor; Measurement model 1: A model with three first-order factors: disinhibited, inhibited, and clinging.
confirmatory factor analysis. The three-factor model fits the data significantly better than the null model and reached an acceptable level of fit (GFI > 0.90, AFGI > 0.90, RMSEA < 0.20). NFI and CAIC (as measures of comparative and parsimonious fit) favored the three-factor model (Table 3). The loadings of items ranged from 1.00 to 0.69. The model with two factors (only clingy behavior and others) produced a worse fit than the three-factor model v2 (19) = 731.33, CFI = 0.83, RMSEA = 0.21, CAIC = 863.03. We tested the associations between attachment-related symptoms and internalizing/externalizing psychological problems and ADHD symptoms using two analyses. First, we tested using SEM whether the attachment-related symptom dimensions were associated with psychological problems and ADHD symptoms (Figs. 1 and 2). Both inhibited (RAD) and disinhibited (DSED)
Disinhibited
0.66
Inhibited
-0.22
ADHD
0.14
Clingy behavior
Fig. 1. Associations between RAD and ADHD systems. © 2014 Scandinavian Psychological Associations and John Wiley & Sons Ltd
Results from the FinAdo study 57
Scand J Psychol 56 (2015)
Disinhibited
0.28
Inhibited
0.33
CBCL
-0.00
Clingy behavior
Fig. 2. Associations between RAD and CBCL.
DISCUSSION According to our results, both RAD and DSED were associated with behavioral problems and ADHD among international adoptees. Children with symptoms suggesting inhibited attachmentrelated symptoms (RAD) had higher internalizing, externalizing, and total problem scores in CBCL, whereas children with disinhibited symptoms (DSED) only had higher externalizing and total problem scores than children with no attachment-related symptoms. Thus, children with inhibited attachment-related symptoms may be considered to have a worse outcome than children with indiscriminate features. The clinical significance of the two subtypes of attachment-related symptoms is not completely clear. According to Rutter, Kreppner and Sonuga-Barke (2009), the inhibited form responds better to enhanced caregiving, and thus, has a more favorable prognosis. Withdrawal may also be an adaptive reaction to situations in which the child receives care from multiple and changing caregivers. In our study, children belonging to the mixed subtype category had much more symptoms and problems in all areas than children in the other, more clear-cut categories. These results are in line with the idea that a constant behavioural pattern or behavioral strategy, here inhibited or disinhibited forms of attachment-related symptoms, functions at some level, and the mixed subtype simply reflects a failure to find a coherent way of responding in difficult circumstances (Gleason et al., 2011). Our results concerning this mixed subtype are, however, preliminary, and we are aware that theorizing about this issue has been controversial (Rutter et al., 2007). In previous studies, indiscriminate behavior has been associated with ADHD symptoms overall (Bruce et al., 2009; Gleason et al., 2011; Kumsta et al., 2010). In our study, both disinhibited © 2014 Scandinavian Psychological Associations and John Wiley & Sons Ltd
(RAD) and inhibited (DSED) symptoms were associated with hyperactivity, but neither of them was connected to attention problems. Recently, growing interest has emerged in finding diagnostic hallmarks between “genuine” ADHD and ADHD symptoms among children from deprived environments. Our results indicate that the presentation of ADHD symptoms associated with attachment-related symptoms may rather be of the hyperactive subtype, rather than being linked to attention problems. Relatively few studies have explored psychopathology with regard to the inhibited form of attachment-related symptoms. Our study shows that also inhibited children were more hyperactive and impulsive – but not inattentive – which may present a pattern of generalized regulatory difficulties. This finding is in line with a previous study among adult ADHD patients; those who had the hyperactive-impulsive subtype presented affective modulation difficulties, whereas those with the inattentive subtype did not (Conzelmann, Mucha, Jacob et al., 2009). Compared with results from previous cross-cultural studies (Viola, Garrido & Rescoria, 2011), the CBCL scores obtained here could be considered relatively low, especially in view of the fact that we were focusing on a risk group. Thus, some cultural differences in reporting symptoms may exist. However, although the levels may be different the associations between various constructs may be similar, and thus, CBCL is probably a good measure of behavioral problems also in Finland. Theoretical and clinical developments in attachment-related psychopathology have been rapid, and while DSM-IV distinguished inhibited and disinhibited types of attachment disorder, the latest revision (DSM-V) distinguishes two separate disorders: reactive attachment disorder (RAD; formerly the inhibited type)
58 M. Elovainio et al.
Scand J Psychol 56 (2015)
Table 4. Association between attachment-related symptom dimensions and behavioral problems measured with CBCL (n = 859). Results from multivariate regression analyses Internalizing Adjusted for age, gender, age at adoption, continent of birth, number and quality of pre-adoption placements, and family SES
Unadjusted model Mean No symptoms Inhibited Disinhibited Mixed subtype
3.46 5.88 3.87 7.14
95% CI 2.84 4.87 2.94 6.38
4.06 6.90 4.79 7.89
p-value
Mean
< 0.001
3.62 5.87 3.77 6.97
95% CI 3.00 4.83 2.82 6.17
p-value 4.25 6.89 4.69 7.75
< 0.001
Externalizing Adjusted for age, gender, age at adoption, continent of birth, number and quality of pre-adoption placements, and family SES
Unadjusted model Mean No symptoms Inhibited Disinhibited Mixed subtype
5.00 8.14 7.82 13.17
95% CI 3.95 6.38 6.22 11.86
6.06 9.91 9.43 14.5
p-value
Mean
< 0.001
5.87 7.85 7.54 12.55
95% CI 4.83 6.14 6.00 11.23
p-value 6.91 9.56 9.08 13.86
< 0.001
Total CBCL Adjusted for age, gender, age at adoption, continent of birth, number and quality of pre-adoption placements, and family SES
Unadjusted model Mean No symptoms Inhibited Disinhibited Mixed subtype
15.10 25.12 23.00 39.00
95% CI 12.57 20.91 19.16 35.85
17.63 29.32 26.83 42.15
p-value
Mean
< 0.001
17.28 24.34 22.25 37.53
and disinhibited social engagement disorder (DSED; formerly the disinhibited type). Of further note is that this proposed DSED is no longer considered to be a reactive attachment disorder, but rather a disorder of social behavior. The survey measure used in this study, the FinAdo attachment-related symptoms scale, provided the differentiation needed for this new diagnostic conceptualization. Our results suggest that the FinAdo attachment-related symptoms scale is a valid survey measure of attachment-related symptoms. In addition to being capable of separating two subtypes of attachment-related symptoms, it offers a tool to evaluate a child’s clingy behavior. The three factors of inhibited symptoms (RAD, 3 items), disinhibited symptoms (DSED, 3 items), and clingy behavior (2 items) made a discriminate effect with associations with criteria measuring psychopathology previously associated with attachment-related symptoms (Gleason et al., 2011). The first two items in the complete scale (readily goes off with a stranger, lack of checking back with parent even in a stressful situation) are commonly described as typical clinical features suggesting disinhibited behavior (Boris & Zeanah, 2005; Rutter et al., 2007; Rutter & Sonuga-Barke, 2010). The third © 2014 Scandinavian Psychological Associations and John Wiley & Sons Ltd
95% CI 14.77 20.12 18.53 34.35
p-value 19.79 28.48 25.98 40.71
< 0.001
item evaluating a child’s behavior when distressed was formulated here focusing on the child’s tendency to turn to parents, rather than evaluating the child’s general reactions to distress or comforting (“not seeking comfort from parents when distressed”). Consequently, the item loaded on the same factor with the first two items suggesting disinhibited behavior. Knowing that a general lack of comfort-seeking is a symptom of inhibited attachment-related symptoms (RAD) (Boris & Zeanah, 2005; Rutter et al., 2007), our finding suggests that either type of attachment disorder can explain a child’s aberrant behavior in distress depending on which kind of behavior is being evaluated. It may indicate a lack of a close confiding relationship with a privileged person (lack of seeking comfort from parent), as in an indiscriminate subtype of attachment-related symptoms, or a disturbance in emotion regulation in distress (general lack of comfort-seeking), suggesting disinhibited attachment-related symptoms. The next three items focused on the child’s emotional reactivity, especially in relation to others (child is scared and wary and does not calm down when a parent comforts or soothes him/her, child withdraws from contact with others, child is apathetic and
Results from the FinAdo study 59
Scand J Psychol 56 (2015)
Table 5. Association between attachment-related symptom dimensions and AHDH symptoms measured with FTF (n = 859). Results from multivariate regression analyses Hyperactivity Adjusted for age, gender, age at adoption, continent of birth, number and quality of pre-adoption placements, and family SES
Unadjusted model Mean No symptoms Inhibited Disinhibited Mixed subtype
0.34 0.48 0.55 0.83
95% CI 0.28 0.40 0.47 0.76
0.39 0.56 0.64 0.90
p-value
Mean
DOI: 10.1111/sjop.12174
Health and Disability Associations between attachment-related symptoms and later psychological problems among international adoptees: Results from the FinAdo study € €A € 5 and HELENA LAPINLEIMU6 MARKO ELOVAINIO,1,2 HANNA RAASKA,3 JARI SINKKONEN,4 SANNA MAKIP A 1
National Institute for Health and Welfare, Helsinki, Finland Institute for Behavioural Siences, University of Helsinki, Finland 3 University of Helsinki, Helsinki University Hospital, Finland 4 University of Turku and Save the Children, Finland 5 Capacitas Familia, Helsinki, Finland 6 Department of Pediatrics, Turku University Hospital and University of Turku, Finland 2
Elovainio, M., Raaska, H., Sinkkonen, J., M€akip€a€a, S. & Lapinleimu, H. (2015). Associations between attachment-related symptoms and later psychological problems among international adoptees: results from the FinAdo study. Scandinavian Journal of Psychology, 56, 53–61. We examined the associations between attachment-related symptoms (symptoms of reactive attachment disorder (RAD), symptoms of disinhibited social engagement disorder (DSED), and clinging) and later psychological problems among international adoptees. The study population comprised internationally adopted children (591 boys and 768 girls, 6–15 years) from the ongoing Finnish Adoption (FinAdo) study. Data were gathered with self-administered questionnaires both from adoptive parents and from adoptees aged over 9 years. Attachment-related symptoms were measured using of a short (8-item) questionnaire and later behavioral/emotional problems were assessed using the Child Behavior Checklist (CBCL) and the Five to Fifteen (FTF) scale for attention-deficit/hyperactivity disorder (ADHD) symptoms. RAD and DSED symptom subscales were associated with an increased risk of emotional and behavioral problems and ADHD. Especially the mixed type of attachment-related symptoms was strongly associated with later emotional and behavioral problems. Key words: Attachment, adoption, survey. Marko Elovainio, National Institute for Health and Welfare, PO Box 30, 00271 Helsinki, Finland. E-mail: marko.elovainio@thl.fi
INTRODUCTION Children raised in institutions often exhibit typical and recognizable behavioral patterns termed in earlier literature as “excessive need for adult attention,” “indiscriminate friendliness” (O’Connor, Marvin, Rutter, Olrick & Britner, 2003; O’Connor & Rutter, 2000), or lack of comfort-seeking even when distressed (O’Connor & Rutter, 2000). Several studies have shown that a key etiological factor in all of these patterns is the absence of a consistent caregiver. In the previous diagnostic manual (DSM-IV from 1994), these patterns were included in reactive attachment disorder (RAD). RAD was defined as aberrant social behavior divided into two basic types, inhibited and disinhibited. Children defined as suffering from the inhibited subtype typically resist initiating or responding to social interactions. They are emotionally withdrawn, do not seek comfort from others, and their reactions especially in distress are typically inhibited, vigilant, or ambivalent. In the current diagnostic manuals (DSM-V and ICD-10), RAD is defined based on this inhibited subtype. In the DSM-V, the disinhibited attachment type is currently termed the disinhibited social engagement disorder (DSED). Children suffering from this disorder are typically non-selective or indiscriminate in their attachment behavior. The child will readily go off with strangers, show limited differentiation among adults, and fail to check back with the parent in stressful situations. He or she has difficulties in forming close, confiding relationships. The symptoms of attachment disorder typically present
© 2014 Scandinavian Psychological Associations and John Wiley & Sons Ltd
with the closest attachment figures, for example, the parents (Boris & Zeanah, 2005; Rutter, 1998; Rutter et al., 2007; Rutter, Kreppner & O’Connor, 2001; Rutter & Sonuga-Barke, 2010). In addition to these two disorders, also a mixed behavioral pattern with features from both disorders has been reported (Gleason, Fox, Drury et al., 2011; Smyke, Dumitrescu & Zeanah, 2002; Zeanah, Scheeringa, Boris, Heller, Smyke & Trapani, 2004; Zeanah, Smyke & Dumitrescu, 2002), and in the DC:0-3R it is included in the deprivation/maltreatment disorder as a subtype of its own (Zero to Three, 2005). Although clinging is mentioned in ICD-10 as one symptom of indiscriminate RAD subtype in toddlers, few previous studies have taken this symptom into account. Excessively clingy behavior is often described by adoptive parents, but its significance with respect to other attachment-related symptoms is far from clear. In this study, we refer to symptoms of all of these disorders (RAD, DSED, and clinging) as attachment-related symptoms. Disordered attachment-related behavior patterns have been suggested to be linked to severe maltreatment and caregiving neglect (O’Connor & Rutter, 2000). Children with an institutional background and also children adopted from foster care have been shown to have a higher risk of attachment disturbances (Bruce, Tarullo & Gunnar, 2009; Smyke et al., 2002). Length of time in institutional care has also been reported to be associated with attachment-related symptoms (Bruce et al., 2009). Specifically, the symptoms of RAD (previously inhibited
54 M. Elovainio et al. subtype) have been associated with caregiving quality features, such as lack of sensitivity, flat affect, and detachment, in previous studies (Gleason et al., 2011). Relatively little research has focused on the developmental patterns of attachment-related symptoms, on the relationship between such symptoms and other childhood psychiatric disorders, or on the potential long-term consequences of attachmentrelated symptoms. Severe attachment-related symptoms have, however, been associated with subsequent functional impairment. RAD (previously inhibited subtype) has been shown to be associated with depressive symptoms (Gleason et al., 2011), and DSED (previously disinhibited RAD) with symptoms resembling attention-deficit/hyperactivity disorder (ADHD) (Bruce et al., 2009; Gleason et al., 2011). In addition, in severely deprived Romanian adoptees, DSED was associated with inattention, overactivity, cognitive impairment, and “quasi-autistic” features, forming deprivation-specific patterns (Kumsta, Kreppner, Rutter et al., 2010). McGoron and others (2012) showed that attachment security was associated with symptoms of multiple forms of psychopathology, including ADHD and both internalizing and externalizing symptoms. Attachment behaviors in young children have been proposed to reflect a child’s sense of predictability and protection by his/her caregiver. These perceptions have been revealed to be associated with enhanced emotion and behavioral regulation abilities, socially engaging behaviors, and social competence. Problems in these domains are linked to RAD and socially indiscriminant behavior as well as to internalizing and externalizing symptoms (DeKlyen & Greenberg, 2008). In this study, we tested whether attachment-related symptoms were associated with later behavioral/emotional problems (measured using the Child Behavior Checklist (CBCL)) and ADHD symptoms (measured using the Five to Fifteen (FTF) scale) (Achenbach & Rescorla, 2001; Kadesjo et al., 2004). Based on the previous literature, we hypothesized that (1) RAD symptoms would be associated with internalizing problems, (2) DSED would be associated with externalizing problems and ADHD, (3) the mixed subtype would associated with externalizing problems, internalizing problems, and ADHD, and (4) clinging would be associated with internalizing problems.
METHODS Study population This study is a part of the ongoing Finnish Adoption (FinAdo) study, including all internationally adopted children (591 boys and 768 girls, response rate 55.7%) through the three legalized adoption organizations in Finland during 1985–2007. Data were gathered with self-administered questionnaires both from the adoptive parents and from adoptees aged over 9 years (in 2007–2009). In the current study, a subsample of children aged 6–15 years with no missing values in attachment-related symptom measures was used (N = 853).
Scand J Psychol 56 (2015) Parents were asked to assess the following behavioral symptoms: (1) (2) (3) (4) (5) (6) (7) (8)
child readily goes off with a stranger; child does not check back with parent even in a stressful situation; child seeks comfort from parents when distressed; child is scared and wary and does not calm down when parent is comforting or soothing him/her; child withdraws from contact with others; child appears apathetic and hopeless; child is excessively clingy to both parents; and child is excessively clingy to one of the parents.
For each item, parents were asked to evaluate the severity of the symptom on a scale ranging from 0 to 2: She or he had “No / never such behavior = 0,” “Some or sometimes such behavior = 1,” “A lot or often such behavior = 2.” The instructions were: “Did you child behave in ways described in the following items when he/she arrived to Finland for the first time?”
Behavioral and emotional problems We used the Child Behavior Checklist (CBCL) to evaluate behavioral and emotional problems. In addition to the total score, we grouped symptoms into internalizing versus externalizing, which is the most commonly used distinction. Internalizing symptoms reflect problems mainly within the self, such as anxiety, depression, somatic complaints without medical cause, and withdrawal from social contacts. Externalizing symptoms represent conflicts with other people and their expectations for children’s behavior (Achenbach & Rescorla, 2001). Each item is rated as 0 = “not true,” 1 = “somewhat or sometimes true,” and 2 = “very true or often true.” The CBCL has previously been used in internationally adopted children and has shown good psychometric properties (Xing Tan, Marfi & Dedrick, 2010).
ADHD ADHD symptoms of the adoptees were evaluated by using part of the Five to Fifteen (FTF) parental questionnaire (Kadesjo et al., 2004; Trillingsgaard, Damm, Sommer et al., 2004). Overall, FTF comprises 181 statements related to behavioral or developmental problems. The items are grouped into eight domains (memory, learning, language, executive functions, motor skills, perception, social skills, and emotional/ behavioral problems). The domain of executive functions consists of 18 items reflecting ADHD symptoms, based on DSM-IV criteria. Symptoms can be grouped into inattentive and hyperactive-impulsive subtypes. Each item has three rating alternatives (0 = “does not apply,” 1 = “applies sometimes or to some extent,” 2 = “definitely applies”). Cronbach’s alpha of the FTF dimensions ranged from 0.97 to 0.99.
Attachment The Kinship Center Attachment Questionnaire (KCAQ) is a quantitative screening instrument (20 items, response format from 0 = never/rarely to 6 = almost always) of child attachment completed by the caregiver. It is appropriate for use with children younger than 6 years and has shown application in tracking difficulties in attachment over time (Kappenberg & Halpern, 2006). The total score comprises four factors: 1 = Positive adjustment/development (six items), 2 = Negative behavior (six items), 3 = Emotional reactivity (four items), 4 = Distancing from caregiver support (four items). In addition it has been suggested that the KCAQ can discriminate between children referred for mental health services and children from a non-clinical sample with a strong effect of group membership on KCAQ total score (Cohen’s d = 0.79) (Kappenberg & Halpern, 2006).
Measurements Attachment-related symptoms. The attachment-related symptoms measure was developed based on the Disturbances of Attachment Interview (Smyke et al., 2002) and clinically relevant symptoms suggesting RAD and DSED and clinging from DSM-V and ICD-10. © 2014 Scandinavian Psychological Associations and John Wiley & Sons Ltd
Covariates The parental questionnaire of the FinAdo study included questions about the child’s background and life circumstances in the birth country. The
Results from the FinAdo study 55
Scand J Psychol 56 (2015) age at adoption, age at time of evaluation, and sex of the child were recorded. The country of birth was classified into four classes based on the continent of birth: Asia (China, Thailand, Philippines, India, Vietnam, Sri Lanka, Pakistan, Nepal, Taiwan, Southern Korea, Israel), Africa (Southern Africa, Ethiopia, Mozambique), America (Colombia, Guatemala, the Dominican Republic, USA, Canada), and Eastern Europe (Russia, Romania, Estonia, Poland). For the statistical analysis, the quality and number of different placements before adoption were classified into four classes: 1 = “only foster home,” 2 = “only one orphanage,” 3 = “2–3 placements before adoption,” and 4 = “4 or more placements before adoption.” The adoptive family’s socioeconomic status (SES) was classified into three classes. The child’s physical health was screened by ascertaining any medical diagnoses that the child had received in Finnish medical evaluations conducted by a doctor. Disabilities were defined as deafness or hearing impairment of both ears, blindness or poor vision in both eyes, intellectual impairment, mental retardation, and autism.
Statistical analysis Because the attachment-related symptoms scale is not a widely used and established scale, its factorial validity was tested with exploratory factor analyses (varimax rotation) (Maharee-Lawler, Rodwell & Noblet, 2010) and structural equation modeling (J€oreskog, 1993). First, we tested the factor structure and number of dimensions using exploratory factor analyses with eigenvalue 1 as a criterion for the appropriate number of factors. Second, we tested the final structure of the RAD measure using structural equation modeling. We used the v2 test, the goodness-of-fit index (GFI), which indicates the amount of variance and covariance explained by the model (values over 0.90 indicating acceptable fit), the adjusted goodness-of-fit index (AGFI, which adjusts the GFI for degrees of freedom in the model, with values above 0.90 indicating a good fit), and the RMSEA criterion (Kelloway, 1998). The comparative fit index (CFI) was used as a measure of comparative fit, with values exceeding 0.90 indicating a good fit. Parsimonious fit was measured by consistent Akaike information criterion (CAIC). Testing the final structure was done in two steps. A one-factor model was first estimated, where all remaining items loaded on the same underlying dimension (null model). Then a second-order model was estimated that represented the theoretical model, with first-order factors loading on the general attachment-related symptoms factor. The associations between the attachment-related symptoms measure and outcomes were examined by linear and logistic regression analyses. First, we calculated an unadjusted model and then we adjusted for age, gender, age at adoption, continent of birth, number and quality of pre-adoption placements, and family SES. We used SAS for Windows version 9.2 and LISREL 9 for the statistical analyses.
RESULTS Characteristics of subjects are shown in Table 1. The mean age of participants (girls 54%) was 2.4 years at the time of adoption and 8.5 years at the time of data collection. Most of the children came from Asia. Of the potential confounders, age at adoption ( p < 0.001), gender ( p = 0.02), number of placement ( p = 0.01), country of origin ( p < 0.001), and physical health status ( p < 0.001) were associated with attachment-related symptoms. The exploratory factor analyses with varimax rotation suggested that the three-factor solution offered the best fit to the data in both samples following the eigenvalue > 1. In the threefactor solution, all items had the strongest loading to their corresponding factors: inhibited (RAD), disinhibited (DSED), and clinging (Table 2). The final structural validity of the attachment-related symptoms scale was tested with confirmatory factor analysis. We © 2014 Scandinavian Psychological Associations and John Wiley & Sons Ltd
Table 1. Characteristics of the study population N = 853 Mean (SD) Age 8.5 (2.9) Age at adoption 2.4 (1.3) Sex Girl Boy Continent of birth Asia Africa America Eastern Europe Number and quality of pre-adoption placements One foster home One orphanage Many placements Parental socioeconomic status (SES) High Middle Low Attention-deficit/hyperactivity disorder symptoms (FTF) Hyperactivity (mean range 0–2) 0.5 (0.9) Attention problems (mean range 0–6) 0.4 (0.9) Behavioral control problems (mean range 0–8) 0.7 (0.9) Psychological/behavioral symptoms (CBCL) Internalizing (range 0–32) 4.9 (5.2) Externalizing (range 0–55) 8.2 (3.2) Total symptoms score (range 0–115) 24.3 (22.8)
N (%)
461 (54) 390 (46) 435 92 136 187
(51) (11) (16) (22)
72 (9) 465 (55) 315 (36) 483 (60) 176 (21) 149 (19)
Table 2. Exploratory factor analysis of attachment related symptoms for international adoptees (n = 850). Varimax rotation Factor loadings Items Child readily goes off with a stranger Child does not check back with parent even in a stressful situation Child seeks comfort from parents when distressed Child is scared and wary and does not calm down when parent comforts or soothes him/her Child withdraws from contact with others Child is apathetic and hopeless Child is excessively clingy to both parents Child is excessively clingy to one of the parents Eigenvalues
Disinhibited
Inhibited
Clinging
0.797
0.025
0.068
0.793
0.210
0.097
0.669
0.393
0.049
0.278
0.596
0.183
0.216
0.734
0.101
0.041 0.133
0.738 0.138
0.065 0.826
0.132
0.283
0.745
2.58
1.32
1.01
tested the overall fitness of the factor structure with definite loading structure and whether the items corresponded to the underlying theoretical construct. Table 3 shows the summary of the fit-indices for the two competitive models of the
56 M. Elovainio et al.
Scand J Psychol 56 (2015)
dimensions were associated with both of the psychological variables, whereas clinging was only weakly associated with ADHD and not associated with behavioral problems. Second, we constructed the final attachment-related symptoms measure with four categories: (1) no symptoms (n = 348, 41%); (2) only disinhibited symptoms (n = 153, 19%); (3) only inhibited symptoms (n = 137, 18%); and (4) (both disinhibited and inhibited symptoms) mixed subtype (n = 214, 25%), and tested whether this measure was associated with internalizing/externalizing behavioral problems or ADHD symptoms (Tables 4 and 5). Compared with children with no symptoms, those with any of the attachment-related symptoms had more internalizing or externalizing behavioral problems and more ADHD symptoms. These associations were robust and remained despite adjustments for age, age at adoption, number and quality of pre-adoption placements, and family SES. Children who belonged to the mixed subtype category had much more symptoms and problems than those in the other categories. Because the items measuring attachment-related symptoms may be restricted to subjects aged under 6 years, we replicated the analyses using those under 6 years of age upon arrival to Finland. No changes in the results were found (all p-values < 0.001, except between attachment symptoms and attention problems, p = 0.230). As additional analyses, we tested the association between our attachment-related symptoms measure and KCAQ. The results in the subpopulation of 6 years of age or under presented in Table 6 showed that those with attachment-related symptoms had significantly more attachment problems than those who did not have such symptoms. Again, those who belonged to the mixed subtype had the most problems.
Table 3. Summary of goodness-of-fit for the models of confirmatory factor analyses in international adoptees (n = 850) Fit-indices
Null model
Measurement model
v2 df v2 difference df p-value GFI NFI RMSEA CAIC
851.70 20
340.09 17 511.61 3 0.000 0.91 0.91 0.149 487.23
0.81 0.81 0.223 975.66
Note: Null model: A one-factor model where all items load on a single factor; Measurement model 1: A model with three first-order factors: disinhibited, inhibited, and clinging.
confirmatory factor analysis. The three-factor model fits the data significantly better than the null model and reached an acceptable level of fit (GFI > 0.90, AFGI > 0.90, RMSEA < 0.20). NFI and CAIC (as measures of comparative and parsimonious fit) favored the three-factor model (Table 3). The loadings of items ranged from 1.00 to 0.69. The model with two factors (only clingy behavior and others) produced a worse fit than the three-factor model v2 (19) = 731.33, CFI = 0.83, RMSEA = 0.21, CAIC = 863.03. We tested the associations between attachment-related symptoms and internalizing/externalizing psychological problems and ADHD symptoms using two analyses. First, we tested using SEM whether the attachment-related symptom dimensions were associated with psychological problems and ADHD symptoms (Figs. 1 and 2). Both inhibited (RAD) and disinhibited (DSED)
Disinhibited
0.66
Inhibited
-0.22
ADHD
0.14
Clingy behavior
Fig. 1. Associations between RAD and ADHD systems. © 2014 Scandinavian Psychological Associations and John Wiley & Sons Ltd
Results from the FinAdo study 57
Scand J Psychol 56 (2015)
Disinhibited
0.28
Inhibited
0.33
CBCL
-0.00
Clingy behavior
Fig. 2. Associations between RAD and CBCL.
DISCUSSION According to our results, both RAD and DSED were associated with behavioral problems and ADHD among international adoptees. Children with symptoms suggesting inhibited attachmentrelated symptoms (RAD) had higher internalizing, externalizing, and total problem scores in CBCL, whereas children with disinhibited symptoms (DSED) only had higher externalizing and total problem scores than children with no attachment-related symptoms. Thus, children with inhibited attachment-related symptoms may be considered to have a worse outcome than children with indiscriminate features. The clinical significance of the two subtypes of attachment-related symptoms is not completely clear. According to Rutter, Kreppner and Sonuga-Barke (2009), the inhibited form responds better to enhanced caregiving, and thus, has a more favorable prognosis. Withdrawal may also be an adaptive reaction to situations in which the child receives care from multiple and changing caregivers. In our study, children belonging to the mixed subtype category had much more symptoms and problems in all areas than children in the other, more clear-cut categories. These results are in line with the idea that a constant behavioural pattern or behavioral strategy, here inhibited or disinhibited forms of attachment-related symptoms, functions at some level, and the mixed subtype simply reflects a failure to find a coherent way of responding in difficult circumstances (Gleason et al., 2011). Our results concerning this mixed subtype are, however, preliminary, and we are aware that theorizing about this issue has been controversial (Rutter et al., 2007). In previous studies, indiscriminate behavior has been associated with ADHD symptoms overall (Bruce et al., 2009; Gleason et al., 2011; Kumsta et al., 2010). In our study, both disinhibited © 2014 Scandinavian Psychological Associations and John Wiley & Sons Ltd
(RAD) and inhibited (DSED) symptoms were associated with hyperactivity, but neither of them was connected to attention problems. Recently, growing interest has emerged in finding diagnostic hallmarks between “genuine” ADHD and ADHD symptoms among children from deprived environments. Our results indicate that the presentation of ADHD symptoms associated with attachment-related symptoms may rather be of the hyperactive subtype, rather than being linked to attention problems. Relatively few studies have explored psychopathology with regard to the inhibited form of attachment-related symptoms. Our study shows that also inhibited children were more hyperactive and impulsive – but not inattentive – which may present a pattern of generalized regulatory difficulties. This finding is in line with a previous study among adult ADHD patients; those who had the hyperactive-impulsive subtype presented affective modulation difficulties, whereas those with the inattentive subtype did not (Conzelmann, Mucha, Jacob et al., 2009). Compared with results from previous cross-cultural studies (Viola, Garrido & Rescoria, 2011), the CBCL scores obtained here could be considered relatively low, especially in view of the fact that we were focusing on a risk group. Thus, some cultural differences in reporting symptoms may exist. However, although the levels may be different the associations between various constructs may be similar, and thus, CBCL is probably a good measure of behavioral problems also in Finland. Theoretical and clinical developments in attachment-related psychopathology have been rapid, and while DSM-IV distinguished inhibited and disinhibited types of attachment disorder, the latest revision (DSM-V) distinguishes two separate disorders: reactive attachment disorder (RAD; formerly the inhibited type)
58 M. Elovainio et al.
Scand J Psychol 56 (2015)
Table 4. Association between attachment-related symptom dimensions and behavioral problems measured with CBCL (n = 859). Results from multivariate regression analyses Internalizing Adjusted for age, gender, age at adoption, continent of birth, number and quality of pre-adoption placements, and family SES
Unadjusted model Mean No symptoms Inhibited Disinhibited Mixed subtype
3.46 5.88 3.87 7.14
95% CI 2.84 4.87 2.94 6.38
4.06 6.90 4.79 7.89
p-value
Mean
< 0.001
3.62 5.87 3.77 6.97
95% CI 3.00 4.83 2.82 6.17
p-value 4.25 6.89 4.69 7.75
< 0.001
Externalizing Adjusted for age, gender, age at adoption, continent of birth, number and quality of pre-adoption placements, and family SES
Unadjusted model Mean No symptoms Inhibited Disinhibited Mixed subtype
5.00 8.14 7.82 13.17
95% CI 3.95 6.38 6.22 11.86
6.06 9.91 9.43 14.5
p-value
Mean
< 0.001
5.87 7.85 7.54 12.55
95% CI 4.83 6.14 6.00 11.23
p-value 6.91 9.56 9.08 13.86
< 0.001
Total CBCL Adjusted for age, gender, age at adoption, continent of birth, number and quality of pre-adoption placements, and family SES
Unadjusted model Mean No symptoms Inhibited Disinhibited Mixed subtype
15.10 25.12 23.00 39.00
95% CI 12.57 20.91 19.16 35.85
17.63 29.32 26.83 42.15
p-value
Mean
< 0.001
17.28 24.34 22.25 37.53
and disinhibited social engagement disorder (DSED; formerly the disinhibited type). Of further note is that this proposed DSED is no longer considered to be a reactive attachment disorder, but rather a disorder of social behavior. The survey measure used in this study, the FinAdo attachment-related symptoms scale, provided the differentiation needed for this new diagnostic conceptualization. Our results suggest that the FinAdo attachment-related symptoms scale is a valid survey measure of attachment-related symptoms. In addition to being capable of separating two subtypes of attachment-related symptoms, it offers a tool to evaluate a child’s clingy behavior. The three factors of inhibited symptoms (RAD, 3 items), disinhibited symptoms (DSED, 3 items), and clingy behavior (2 items) made a discriminate effect with associations with criteria measuring psychopathology previously associated with attachment-related symptoms (Gleason et al., 2011). The first two items in the complete scale (readily goes off with a stranger, lack of checking back with parent even in a stressful situation) are commonly described as typical clinical features suggesting disinhibited behavior (Boris & Zeanah, 2005; Rutter et al., 2007; Rutter & Sonuga-Barke, 2010). The third © 2014 Scandinavian Psychological Associations and John Wiley & Sons Ltd
95% CI 14.77 20.12 18.53 34.35
p-value 19.79 28.48 25.98 40.71
< 0.001
item evaluating a child’s behavior when distressed was formulated here focusing on the child’s tendency to turn to parents, rather than evaluating the child’s general reactions to distress or comforting (“not seeking comfort from parents when distressed”). Consequently, the item loaded on the same factor with the first two items suggesting disinhibited behavior. Knowing that a general lack of comfort-seeking is a symptom of inhibited attachment-related symptoms (RAD) (Boris & Zeanah, 2005; Rutter et al., 2007), our finding suggests that either type of attachment disorder can explain a child’s aberrant behavior in distress depending on which kind of behavior is being evaluated. It may indicate a lack of a close confiding relationship with a privileged person (lack of seeking comfort from parent), as in an indiscriminate subtype of attachment-related symptoms, or a disturbance in emotion regulation in distress (general lack of comfort-seeking), suggesting disinhibited attachment-related symptoms. The next three items focused on the child’s emotional reactivity, especially in relation to others (child is scared and wary and does not calm down when a parent comforts or soothes him/her, child withdraws from contact with others, child is apathetic and
Results from the FinAdo study 59
Scand J Psychol 56 (2015)
Table 5. Association between attachment-related symptom dimensions and AHDH symptoms measured with FTF (n = 859). Results from multivariate regression analyses Hyperactivity Adjusted for age, gender, age at adoption, continent of birth, number and quality of pre-adoption placements, and family SES
Unadjusted model Mean No symptoms Inhibited Disinhibited Mixed subtype
0.34 0.48 0.55 0.83
95% CI 0.28 0.40 0.47 0.76
0.39 0.56 0.64 0.90
p-value
Mean
