Child and Parent Variables Associated with Treatment Response in Narcissistic Youths: The Role of Self-Blame and Shame

Abstract

Introduction

this retrospective study aimed at exploring the impact of parental and youth narcissism on service utilization and response to psychotherapy.

Method

thirty-six 9–13 years-old narcissistic youths receiving a combined treatment (psychodynamic psychotherapy + parental counselling) were independently assessed for child pathological narcissism, parental narcissism, child’s treatment attitude and psychiatric/ psychosocial services utilization.

Results

Parent and child narcissistic characteristics were negatively correlated with the use of OPD services. Child self-blame and lack of empathy predicted a lesser OPD utilization. Child pathological narcissism was significantly correlated with the specific treatment response pattern composed of the set of attitudes consistently observed in psychotherapy with narcissistic youths. Sixty percent of the variance in treatment response was accounted for by four narcissistic characteristics of the child: devaluation of others, avoidance of vulnerability, boredom, and self-blame. Post hoc analyses highlighted the role of the child’s self-blame, which mediated the relation between parental narcissism and service use while acting as a moderator in reversing the relation between parent narcissism and treatment response. Relationships between self-blame, shame-proneness and guilt, and limitations of the study are discussed.

Conclusion

Prospective studies with a larger sample are needed to confirm the association between self-blame and response to treatment in narcissistic youths.

INTRODUCTION

Clinical studies have consistently stressed how challenging is the treatment of narcissistic disorders in children and adolescents (Bleiberg, 2001; Kernberg et al., 2000). Narcissistic youths seem prone to drop out from treatment or act out if their behaviour is inappropriately managed within the therapeutic frame (Kernberg et al., 2000). They elicit in their therapists “the helplessness and vulnerability they cannot tolerate in themselves” (Bleiberg, 2001). Clinically, they appear oppositional and self absorbed, demanding of others but giving little in return. They behave as if they are entitled to do what they want, avoid making any effort to realize their academic potential and tend to control others through dramatic tantrums. Their grandiose self-experience is organized around over-evaluation of their achievements, fantasies of invulnerability and a feeling of uniqueness. Dealing with a frail self-esteem, they remain hypersensitive to criticism and defeat. They blame others for their longstanding problems of adjustment in school and at home which are in fact related to their lack of empathy, their feelings of envy and their unrelieved need for admiration (Guilé, 1996).

Recent research reported high stabilities of pathological narcissistic behaviours in boys and girls from early adolescence into adulthood (Crawford et al., 2001). However, the role treatment can play in preventing those youths from structuring their psychopathology into a Personality Disorder in adulthood is still unknown. Research on factors influencing the treatment outcome of these disturbances in children and adolescents is mostly confined to case-report studies underscoring their flickering compliance (Bleiberg, 2001; Kernberg et al., 2000; Guilé, 1996). Extensive research on paediatric Axis I disorders has already demonstrated that both the child psychopathology and the parental characteristics predict treatment response (Weisz et al., 2002; Verhulst et al., 1997). In keeping with these findings, this retrospective study aimed at exploring the impact of parental and youth narcissism on service utilization and response to psychotherapy.

METHODS

Sample

Participants were pre-adolescents referred to an outpatient specialized clinic for youths at risk of developing Personality Disorder and attending the first step of the OPD multiphasic treatment plan. Attendees participated in a one school-year individual psychodynamic psychotherapy whereas parents were given parental counselling. Only subjects who received one yearlong treatment at the time of the study were included in the study. The sample included 36 Caucasian subjects ranging in age from 9 to 13 years (mean: 10.92 +1.34 years). The Gender proportion was 75% boys and 25% girls. Two thirds of the children lived with their 2 parents. DSM-IV Axis I diagnosis composition was 52.8% disruptive disorders equally composed of ADHD and ODD, 27.8% anxiety disorders (mainly separation anxiety disorder), and 5.6% adjustment disorders. A remaining 13.9% presented a combined type with disruptive and anxiety disorders. ADHD children were receiving 20 to 35 mg Methylphenidate daily. No subject had been hospitalized.

Measures

Based on a review of the clinical literature (Guilé, 1996), four variables have been described which have a relationship with the treatment of narcissistic youths: child and parent narcissism; and two variables which correspond to the response to treatment: treatment attitude and service use.

1. Child pathological narcissistic behaviour

The Retro-Diagnostic Interview for Narcissism (Retro-DIN) is an adaptation, for reviewing children’s medical charts, of a largely-used research semi-structured interview (Gunderson et al., 1990). This adaption has been shown to be a specific and sensitive instrument for identifying narcissistic behaviour in adults. The Retro-DIN regroups 34 narcissistic traits into 5 sections (grandiosity, interpersonal relationships, reactiveness, mood states and adaptation) paralleling the adult version (see Appendix 1). By summing up the section scores one can derive a continuous score (0–13) as a quantitative index of pathological narcissistic behaviour of the child. The total score is then defined as a dimensional personality trait and not as a categorical diagnosis. Interrater-reliabilities were calculated for this study and ranged from .62 to 1.0. Only the interrator reliability coefficient for denial of dependency needs and idealization scored below .70 (k = .62, CI=0–1.00). Sections’ alpha coefficients were respectively: grandiosity (I): .73, relationships (II): .71, reactiveness (III): .70, mood states (IV): .58, adaptation (V): .52 and Total items score: .79.

2. Parental Narcissism

Parental narcissism (PN) includes a set of traits reported in the clinical literature: presence of narcissistic traits in one of the parents, overidealization, and over-investment of partial characteristics of the child’s performance and behaviour instead of an integrated investment of the child (Kernberg et al., 2000; Guilé, 1996; Segal, 1996).

3. Treatment attitude

This variable encompasses a set of behaviours consistently observed in psychodynamic psychotherapy with narcissistic youths (Kernberg et al., 2000; Guilé, 1996) including frequent use of sadistic themes in play, mirroring play, hypersensitivity to the therapist’s interventions, reluctance to engage in pretend play.

4. Service use

The third variable regroups psychosocial and health services received at school, in Community Services(CS) and in Child Psychiatry Out-patients Departments (OPD).

Interrater-reliabilities (using interclass correlation coefficients (ICC) calculated for the last three variables were good: parental narcissism (ICC=.97, CI=.92–.99), child’s treatment attitude (ICC=.94, CI=.84–.98), psychosocial and health services received at school (ICC=.90, CI=.72–.96), CS (ICC= .77, CI=.39–.92) and OPD (ICC =.76, CI=.36–.91).

Procedure

Axis I diagnoses were assigned by an experienced psychiatrist (JMG) according to DSM-IV criteria. Mental retardation, psychotic and major affective disorders were excluded. Two independent raters, blind to the child’s identity and diagnosis, scored narcissistic pathology through a chart review process using the Retro-DIN. Interrater reliability was then calculated. As a second step, two other independent raters, blind to the child’s diagnosis and Retro-DIN score, assessed the three other variables mentioned in the measures section. Interrater reliability was also established for each of the measurements.

Data analysis

Preliminary analyses including chi-square and t tests were conducted to determine whether demographic variables were related to the tested variables. Correlation analyses and multiple regression techniques were then used to examine the relationships between child and parent narcissistic variables and the two variables targeting treatment response as criterion variables. With respect to the model proposed by Baron and Kenny (1986), post hoc analyses were performed to determine whether the independent variables play a moderator or mediator role within the interaction. Based on a model including one co-variate which yields a R² of .50, an alpha set at .05 and power at .80, an increment of .10 would be obtained with a sample size of 35 subjects.

RESULTS

Characteristics of the sample and preliminary analyses

Scores obtained with the Retro-DIN on the pathological narcissistic behaviour of the child confirmed that the sample was mainly composed of narcissistic subjects (mean total score = 11.03 ± 2.16). Skewness and Kurtosis coefficients were not significant. Preliminary analyses revealed no effect of age, gender, family structure, Axis I diagnoses or medication (methylphenidate) on the tested variables.

Parental and child narcissism

Parental narcissism (PN) was positively correlated with the child’s anhedonia (r=.44, p<.01), denial of dependency (r=.40, p<.05), devaluation (r=.49, p<.01) and with the entire DIN Mood section (r=.34, p<.05). PN was inversely correlated with the child self-destructiveness (r=−.38, p<.05).

PN was inversely correlated with OPD services utilization (r=−.38, p<.05). Child narcissistic behaviour was inversely correlated with OPD use (see Table 1). Child self-destructiveness was the only characteristic to be positively correlated with OPD use (r=.36, p< .05). In addition, the suicidal reactions section (III, 21), composed of suicidal threats when the child is facing criticism or defeat, was correlated with support at school (r=.36; p<.05) whereas avoidance of vulnerability was inversely correlated (r=−.51, p<.01). No correlation was found between child and parent narcissism, and community services.

TABLE 1.

Child Retro-DIN’s

sections & traits	Use of psychiatric services	Treatment attitude
I Grandiosity	−.33*
I.1exaggeration of talents	−.34*
I.3 avoidance of vulnerability		.49**
I.4 grandiose fantasies	−.35*
II Interpersonal relationships		.36*
II.6 lack of empathy	−.42*
II.11 devaluation		.50**
III Reactiveness		.42*
IV Mood states		.70***
IV.25 boredom		.67**
IV.26 irritability		.61**
IV.27 self-destructiveness	.36*
IV.28 self-blame	−.58**	−.37*
Total score		.58**

Pair-wise Correlations Between Child Narcissism and Response to Treatment

^*p < .05,

^**p < .01,

^***p < .001

Only the significant correlations are reported

Multiple regression analyses were then performed with the child narcissistic characteristics as predictors and OPD use as the criterion variable. Stepwise analyses revealed that only lack of empathy (β=−.34, t=−2.58; p<.05) and self-blame (β=− .52, t=−3.99; p<.001) contributed to the variance in OPD use (R² =.44, R²adj=.41, F (2,33) =13.17, p<.001). Parental narcissism (PN) emerged also as a predictor of service use (R²adj=.12, F(1,34)=5.68, p<.05). Post hoc probing for mediating/moderating role of the different variables was performed. Self-blame (SB) remained significant whereas the predictive power of PN was no longer found (PN: β=−.15, t=−.81; NS; PN x SB: β=−.18, t=.71; NS). Thus, child self-blame appeared to mediate the interaction between parental narcissism and services use.

Regression analyses conducted with parent and child narcissism as predictors and treatment-attitude as criterion confirmed that child narcissism was a potent predictor of treatment attitude contributing to 30% of the variance (R²=.34, R²adj=.29, F(2,26)=7.26, p<.005). Stepwise regression performed on the child narcissism correlated variables yielded a 4 variables equation (R²=.67, R²adj=.62, F(4,26)=13.11, p< .001) including boredom (β=.38, t=2.76; p<.05), self-blame (SB) (β=−.39, t=3.30; p<.001), devaluation (β=.28, t=2.19; p<.05) and avoidance of vulnerability (β=.28, t=2.16; p<05). More than 60% of the variance in treatment attitude was accounted for by a combination of these four narcissistic characteristics of the child. Interestingly the two affect variables (boredom and self-blame) accounted for about 50% (R²adj=49, F 2,28)=5.76, p<001). Post hoc analyses revealed a significant interaction between self-blame (SB) and parental narcissism (PN) (SB: β=1.05, t=3.55; p<001; PN: β=60, t=2.89; p=007; SB x PN: β=. t=.40; p=002) which pleaded for a moderating role of the child self-blame on the interaction between PN and treatment attitude.

DISCUSSION

Parent and child narcissism and the child’s response to treatment

This study highlights the impact of parent and child narcissistic traits on the child’s response to treatment in several ways. The services utilization and interpersonal relationships profile of the children seemed to mirror those of their parents. First, although only gleaned from the child’s chart, the parental narcissism correlated both with the child’s anhedonia and with two defense mechanisms observed in the child’s interactions with others: denial of dependency and devaluation. Second, both the child’s and the parent’s narcissism showed an inverse correlation with outpatient service utilization.

Third, as one would predict intuitively, those youth who expressed suicidal reactions were able to garner services from the school, whereas those who were vulnerable and avoidant did not access such services. From a clinical perspective, this might suggest that the fear of being vulnerable prevents the child from seeking services that might threaten his precarious sense of self-esteem, and thus presents a challenge to the school preventive care system intent on identifying and helping such youth. Fourth, two clinical features on the child’s part, lack of empathy and self blame, accounted for 40% of the variance in OPD service use, and parental narcissism for another 12% of that variance. The above suggests that a significant inherent trait of such youth, an underdeveloped sense of empathy, might prevent them from perceiving a potential treatment benefit on their interpersonal relationships, and thus lead to treatment avoidance.

Fifth, this study has confirmed the association between the child narcissistic psychopathology and the previously reported set of attitudes which is observed in psychodynamic psychotherapy with narcissistic youths i.e., mirroring and sadistic themes in play, hypersensitivity to the therapist’s interventions, and reluctance to engage in pretend play (Kernberg et al., 2000; Guilé, 1996). Furthermore, the results suggest that four characteristics of the child’s narcissism are of particular predictive value. They include: the avoidance of vulnerability; the devaluation of others; the presence of boredom, and self-blame. Therapists working with youth manifesting those characteristics can anticipate the eventual therapeutic appearance of such issues as the development of sadistic play themes, hypersensitivity to the therapist’s interventions and reluctance to engage in pretend play.

Our results are thus consistent with previous studies stressing the impact of parental characteristics on the treatment of childhood disorders (Weisz et al. 2002; Brent et al., 1998). The co-occurrence of a mental disorder of the same kind in the parent and in the adolescent has been identified in depression as a predictor of adverse outcome (Brent et al., 1998). In our study, the parental narcissism (PN), including a specific pattern of parent- child interaction (Kernberg et al., 2000; Guilé, 1996, Segal, 1996), emerged as a predictor of psychiatric services utilization and treatment attitude. The more narcissistically woven the parent-child relationship is, the less the psychiatric services are used, and the stronger is the child’s reluctance to engage in psychotherapy. It suggests that the members of the narcissistic parent/ child dyad specifically do not want to disrupt the idealized parent/child relationship, and will thus not risk a psychotherapy process.

Self-blame, shame and self-esteem in psychotherapy

Self-blame implies the capacity to attribute the source of a problem to the self rather than to others. As the results showed, self-blame moderated the impact of parental narcissism on the child’s treatment attitude for those youth who were already engaged in the treatment process, but curtailed the treatment for those who were not yet engaged.

From a research perspective on narcissism, self-blame might be a more promising construct than seemed to be self-esteem in the area of narcissistic psychopathology. In narcissistic disordered youths, self-esteem appears to be an ambiguous construct which does not necessarily co-vary with narcissism. Watson and colleagues (1996) showed that in a community sample of 459 young adults, the relation between narcissism and self-esteem depends on the level of social adjustment. Moreover it seems that the correlation between self-esteem and narcissism depends largely on the instruments used to measure self-esteem (Watson et al., 1996; Hibbard, 1993). Other constructs related to pathological narcissism, as shame or self-blame mentioned earlier, might be more relevant and specific. Self-blaming transactions have been reported as a key feature in psychodynamic psychotherapy of narcissistic patients (Zaslav, 1998). They denote “transactions that in psychotherapy mobilize shame”. Zaslav underlined shame-proneness as a patient’s particular vulnerability requiring, when present, specific adaptation of the psychotherapy technique. In a university sample of 701 youths (mean age=21), Hibbard (1993) showed that shame was significantly correlated with pathological narcissism. Gaze aversion which might been related to shame-proneness, has been also pinpointed as a component of the child narcissistic picture (Kernberg et al., 2000). Building on this association between narcissism and shame, it could be hypothesized that shame-proneness plays a significant role in treatment participation. It could be stated that the proposal of treatment elicits shame in the patient and then prevents him/her from assenting to treatment. This perception that the treatment is shame-provoking constitutes a barrier to treatment participation in the sense of Kazdin and Wassell (1999). These authors stressed the importance of such factors in demonstrating that barriers to treatment participation were significantly correlated with therapeutic change.

While applying these notions to childhood and adolescence, one should take a developmental stance as Kagan posited (2001). Within the context of the emotional development of the child, shame refers to the recognition of the violation of a standard of conduct. Guilt which appears later, is related to the recognition that the violation of a personal standard could have been inhibited and its cause should be attributed to self rather than to others. In keeping with our results, it could be also hypothesized that once in treatment, the youth who is able to blame himself, experiences guilt to some extent and achieves a better collaboration in psychotherapy if the therapist provides a warm and empathetic support. The set of treatment attitudes observed in psychotherapy with narcissistic youths and reported above, could correspond to this process. From a clinical and psychodynamic perspective, these attitudes represent both a form of resistance to treatment, and thus negative attributes, and by contrast, essential characteristics of a treatment process which need to be manifest, addressed and resolved, and thus positive. Their absence could result in unresolved conflict and thus an unsuccessful treatment.

Our findings are not conclusive due to several limitations of this study. First, chart review, rather than direct clinical observation, guided characterization of parent/child interaction and determination of parental narcissism. As well, the study was retrospective and the sample size was small and restricted to narcissistic children enrolled in an OPD treatment setting. Future studies should consider a prospective design using measures of parental and child narcissism, and include information concerning the child’s mood, the therapeutic alliance and behavioural outcomes.

APPENDIX 1

RETRO-DIN ADAPTED FOR CHART REVIEW
Section I: Grandiosity	Section III: Reactiveness
1- Exaggeration of talentsa	19- Hypersensitivity to criticism
2- Belief in invulnerability	20- Hypersensitivity to defeat
3- Avoidance of vulnerability	21- Suicidal reactions
4- Grandiose fantasiesa	22- Aggressive reactions
5- Denial of dependency needs	23- Reactions to others’ envy
6- Uniquenessa	Section IV: Mood States
7- Superiority	24- Anhedonia
8- Boastful behaviour	25- Boredom
Section II: Interpersonal Relationships	26- Irritability
9- Need for admirationa	27- Self-destructiveness
10- Idealization	28- Self-blame
11- Devaluation/contempt	Section V: Social/Moral Adaptation
12- Envya	29- Academic and athletic achievements
13- Entitlementa	30- Superficial values and interests
14- Arrogant/haughty behavioura	31- Disregard for values/rules
15- Self-centered behaviour	32- Antisocial reactions
16- Lack of empathya	33- Recurrent antisocial behavioursb
17- Lack of reciprocity	34- Psychosexual development
18- Exploitivenessa

^aCriterion used in DSM-IV for narcissistic personality disorder.

^bPresence of this characteristic weighs against a diagnosis of narcissistic personality disorder.