Brozina abela 2005 klinische pschologie PDF

Preview

Summary

Pathoplastische model
Een temperament kan hier een moderator zijn voor een stoornis zonder dat het een directe verband heeft. En hoe mensen hierop reageren is afhankelijk van je temperament. Sociale interacties hebben hierop een invloed. Je temperament zal bepalen hoe een stoornis zal ontwikke...

Description

ARTICLE IN PRESS

Behaviour Research and Therapy 44 (2006) 1337–1346 www.elsevier.com/locate/brat

Behavioural inhibition, anxious symptoms, and depressive symptoms: A short-term prospective examination of a diathesis-stress model Karen Brozina , John R.Z. Abela Department of Psychology, McGill University, 1205 Dr. Penfield Avenue, Montreal, Que., Canada H3A 1B1 Received 12 July 2005; received in revised form 25 August 2005; accepted 15 September 2005

Abstract The purpose of the current study was to examine the relationship between behavioural inhibition (BI) and anxious symptoms within a diathesis-stress framework, using a short-term prospective design. In addition, we examined whether BI acts as a specific vulnerability to anxious symptoms, or as a common vulnerability to both anxious and depressive symptoms. At time 1, 384 children in grades 3 through 6 completed self-report measures of BI, anxious symptoms, and depressive symptoms. Six weeks later, they completed self-report measures of hassles, anxious symptoms, and depressive symptoms. Results demonstrated that children with high BI who experienced high levels of hassles during the 6-week follow-up interval showed increases in anxious symptoms, but not depressive symptoms. r 2005 Elsevier Ltd. All rights reserved. Keywords: Behavioural inhibition; Depression; Anxiety; Diathesis-stress; Specificity

Introduction

(e.g., Garcia Coll, Kagan, & Reznick, 1984). ), is considered to be a measurable childhood temperament incorporating aspects of shyness, fearfulness, and social withdrawal (Hirshfeld-Becker, Biederman, & Rosenbaum, 2004). Further, BI is proposed to confer vulnerability to anxiety within a diathesis-stress framework, such that it ‘‘interacts with developmental experience and life events to determine ultimate likelihood of onset of anxiety disorder’’ (Biederman, Rosenbaum, Chaloff, & Kagan, 1995, p. 76). A rich history of research has examined hypotheses related to the role of BI in the development of childhood anxiety (for a review, see Biederman et al., 1995; for critical reviews, see Oosterlaan, 2001; Turner, Beidel, & Wolff, 1996). . In a sample of children aged 

Corresponding author. Tel.: +1 514 398 6127; fax: +1 514 398 4896. E-mail address: [email protected] (K. Brozina).

0005-7967/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.brat.2005.09.010

ARTICLE IN PRESS 1338

K. Brozina, J.R.Z. Abela / Behaviour Research and Therapy 44 (2006) 1337–1346

2–7 years with a parent undergoing psychiatric treatment for one of panic disorder and agoraphobia (PDAG), major depressive disorder (MDD), both PDAG and MDD, or neither PDAG nor MDD, the children of parents diagnosed with PDAG were more likely to have high BI than the children of parents without a diagnosis of PDAG (Rosenbaum et al., 1988). . In a second study using the above-mentioned sample (Biederman et al., 1990), children with high BI demonstrated significantly higher rates of multiple anxiety disorders than control children. In a small longitudinal cohort of 721-year-old children who had been classified as having either high or low BI as toddlers, the parents of children with high BI demonstrated higher rates of anxiety disorders than did either the parents of children with low BI or the parents of control children (Rosenbaum et al., 1991). . In a study examining both of the above-mentioned samples, researchers found that parents of inhibited children with multiple anxiety disorders had higher rates of multiple anxiety disorders themselves, as compared to both the parents of inhibited children without anxiety disorders and the parents of not-inhibited children without anxiety disorders (Rosenbaum et al., 1992). Taken together, the results of these studies suggest a strong association between childhood BI and anxiety. Recent research extends the investigation of this association to older children. In three large samples of adolescents ranging in age from 11 to 18 years (Muris, Meesters, & Spinder, 2003; Muris, Merckelbach, Schmidt, Gadet, & Bogie, 2001; Muris, Merckelbach, Wessel, & van de Ven, 1999), researchers found that participants who classified themselves as high on a measure of BI reported higher levels of anxious symptoms than those who classified themselves as middle or low on BI. In addition, adolescents with high BI were more likely than the other adolescents to report anxiety disorders in the sub-clinical range (Muris et al., 1999).

(e.g., Garber & Hollon, 1991). One study using a large sample of high-school students found that retrospective reports of childhood BI predicted the onset of social phobia 4 years later (Hayward, Killen, Kraemer, & Taylor, 1998). However, other longitudinal research has produced mixed results. For example, one cohort of toddlers was followed through to the ages of 721 (Biederman et al., 1990), 11 (Biederman et al., 1993), and 13 (Schwartz, Snidman, & Kagan, 1999) years. Within this sample, as compared to children with low BI as toddlers, children with high BI as toddlers had significantly higher rates of (a) separation anxiety at age 11, but not at ages 721 or 13; (b) avoidant disorder at age 11, but not at age 721; (c) phobic disorders at age 721, but not at ages 11 or 13; (d) social anxiety at age 13, but not at age 11; and (e) multiple anxiety disorders at age 11, but not at age 721.

(Biederman et al., 1995).

Despite all of the speculation that BI confers vulnerability to anxiety within a diathesis-stress framework, previous research has examined BI within a maineffect model. A second possible explanation for previous inconsistencies in the longitudinal examination of BI is that early studies were based on the assumption that BI is a distal factor in the development of anxiety. This assumption presumes that BI is a highly stable individual difference, and as such, studies classified BI in toddlers and then assessed anxiety in middle to late childhood (e.g., Biederman et al., 1990, 1993; Schwartz et al., 1999). However, a review of the relevant literature suggests that even among extreme groups, BI is not necessarily an enduring attribute (Turner et al., 1996). For example, in examining repeated BI classifications in toddlers through to the age of 721 years, only 12 of 22 high BI children remained inhibited, and only 9 of 19 low

ARTICLE IN PRESS K. Brozina, J.R.Z. Abela / Behaviour Research and Therapy 44 (2006) 1337–1346

1339

BI children remained uninhibited (Hirshfeld et al., 1992). In a study comparing inhibited and uninhibited children to controls (Kerr, Lambert, Stattin, & Klackenberg-Larsson, 1994), researchers found that BI ratings were much more stable for the extreme groups from the ages of 22 months to 712 years. In addition, they found that BI at age 7 predicted BI at age 16, but only for inhibited girls. Within normative samples, the stability of BI has been found to be only modest (e.g., Reznick, Gibbons, Johnson, & McDonough, 1989; Scarpa, Raine, Venables, & Mednick, 1995). Taken together, these results suggest that BI is a variable characteristic. Therefore, previous inconsistencies may be due to the fact that some inhibited toddlers became less inhibited through development, and some non-inhibited toddlers became more inhibited through development. We propose that BI can act as a proximal factor in the development of anxiety. . To our knowledge, no study has examined whether BI predicts increases in anxious symptoms in children across a short time-interval. Although most research on the psychopathological correlates of BI have focused on anxiety, Biederman et al. (1995) have speculated that BI may also be antecedent to depression. Adolescent self-reports of BI are not only positively correlated with anxious, but also with depressive symptoms (Muris et al., 1999, 2001, 2003). In a prospective study (Hayward et al., 1998), the fearfulness component of BI was associated with an increased risk for the development of depression in adolescents. In a long-term longitudinal study (Caspi, Moffitt, Newman, & Silva, 1996), young adults who had been classified as inhibited at the age of 3 years, were more likely to receive a diagnosis of depression than those who had been classified as undercontrolled (similar to uninhibited), or well adjusted at the age of 3 years. One possible explanation for the association between BI and depression is that BI confers vulnerability to both anxiety and depression. This would be congruent with the common aetiology hypothesis (e.g., Frances, Widiger, & Fyer, 1990; Merikangas, 1990), which proposes that anxiety and depression co-occur at high rates because they share aetiological factors. A second explanation is that the association between BI and depression is spurious, and is ultimately due to the strong association that anxiety has separately with: (1) BI, and (2) depression. Results from a study by Muris et al. (2001) provide preliminary support for this second alternative. However, because their data was cross-sectional, prospective studies are required to determine which alternative is the better explanation of the association between BI and depressive symptoms.

Children in grades 3 through 6 completed self-report measures of BI, anxious symptoms, and depressive symptoms. Six weeks later, these children completed self-report measures of stress, anxious symptoms, and depressive symptoms. Consistent with the proposition that BI is a vulnerability factor for anxiety, we hypothesized that BI would predict changes in anxious symptoms over the 6-week follow-up. Consistent with a diathesis-stress framework, we hypothesized that children with high BI who experienced high levels of stress would show greater increases in anxious symptoms than other children. Finally, in line with the common aetiology hypothesis, we hypothesized that: (a) BI would also predict changes in depressive symptoms; and (b) children with high levels of BI who experienced high levels of stress would show greater increases in depressive symptoms than other children. Method Participants Participants were recruited from 10 schools in the English Montreal School Board. Information letters and consent forms were sent home with students in grades 3 through 6. Consent was obtained actively, such that only students who returned consent forms participated in the study. At Time 1, 429 students completed all of the measures. At Time 2, 384 of these students completed all of the measures. The age of the final sample ranged from 8 to 13 years (M ¼ 10.6, SD ¼ 1.2 years) and 47% were males. Participants were asked to identify their ethnic backgrounds; 65% identified themselves as white, 14% as South Asian, 7% as East Asian,

ARTICLE IN PRESS 1340

K. Brozina, J.R.Z. Abela / Behaviour Research and Therapy 44 (2006) 1337–1346

6% as black, 3% as Middle Eastern, 3% as Hispanic, and 1% as Native American. Although data on socialeconomic status were not available, participating schools were located in economically deprived regions of the city, middle class regions, and affluent suburbs, suggesting that a wide range of social classes were represented. There were no significant differences in gender, ethnicity, or total scores on the Time 1 measures between the children who completed only the first assessment and children who completed both assessments. Children completing only the first assessment were significantly younger (M ¼ 10.1 years, F(1,427) ¼ 6.96, po.01), as one of the participating grade 3 classes cut the second assessment short in order to attend a school function. As a result, only 3 of 27 students in this class completed all of the Time 2 questionnaires. Procedure At Time 1, students completed the (MASC; March, Parker, (CDI; Kovacs, 1985), and the Sullivan, Stallings, & Conners, 1997), (BIS; Muris et al., 1999). Six weeks later (Time 2), students completed the MASC, the CDI, and the (CHAS; Kanner, Feldman, Weinberger, & Ford, 1987). Students completed all of the questionnaires in a classroom with a teacher, and 2 research assistants present. Research assistants read the questions aloud and answered any questions the participants had. Measures MASC The MASC (March et al., 1997) is a 39-item self-report measure of anxious symptoms designed for youth aged 8–19. Children were asked to indicate on a 4-point Likert scale how often each item has been true for them in the last week. Items are scored from 0 to 3. Total scores range from 0 to 117, with higher scores indicating higher anxiety. In an examination of its psychometric properties, the MASC demonstrated good reliability and validity (March et al., 1997). CDI The CDI (Kovacs, 1985) is a 27-item self-report measure of the cognitive, affective, and behavioural symptoms of depression. For each item, children are asked to consider how they have been thinking or feeling in the last week, and to indicate which of three statements best describes them. Each of the three options varies in symptom severity, and is scored from 0 to 2 with higher scores indicating greater symptom severity. In the current study, CDI item 9 concerning suicidal ideation was omitted from the scale at the request of the English Montreal School Board. Thus, total scores could range from 0 to 52. In an examination of its psychometric properties, the CDI demonstrated good reliability and validity (Kovacs, 1985). BIS The BIS (Muris et al., 1999) is a 4-item, self-report measure designed to assess four aspects of BI (see Gest, 1997): shyness, communication/talkativeness, fearfulness, and smiling/laughter. Children are asked to indicate on a 4-point Likert scale how often each item is true for them. After recoding reversed items, scores are summed and can range from 4 to 16, with higher scores indicating greater degrees of BI (i.e., apprehensive, shy, difficulty initiating social interactions with unfamiliar people). Previous studies using the BIS demonstrate moderate to high internal consistency (Muris et al., 1999, 2001). An important consideration in assessing both anxiety and BI through self-report measures is that the association between BI and anxiety may be tautological in nature. In other words, having anxiety may influence respondents’ self-reports of BI. The creators of the BIS attempted to address this issue by examining the association between self-reports and parent-reports. The correlation was r ¼ 0.49, po.001 (Muris et al., 2003), suggesting that there is external validation (i.e., parent agreement) for self-reports of BI. In addition, although significant, there was only a modest association between the BIS and the Time 1 MASC (r ¼ .23) in the current study. Although the direct correspondence between self-report and observational methods of assessing BI has yet to be established, many previous studies (e.g., Gest, 1997; Mick & Telch, 1998; Muris et al., 1999, 2001, 2003;

ARTICLE IN PRESS K. Brozina, J.R.Z. Abela / Behaviour Research and Therapy 44 (2006) 1337–1346

1341

Neal, Edelman, & Glachan, 2002; Reznick, Hegeman, Kaufman, Woods, & Jacobs, 1992) provide support for the utility and validity of such measures. In addition, previous research has found that parental ratings on the BIS are associated both with laboratory observations of BI (van Brakel, Muris, & Bogels, 2004), and with children’s ratings on the BIS (Muris et al., 2003). An examination of the scale’s convergent validity suggests that the BIS is positively correlated with measures of shyness and fearfulness, while it is negatively correlated with measures of extraversion and regulation (van Brakel & Muris, in press). Although past studies have not examined the discriminant validity of the BIS, in the current study there was no significant association between the BIS and Time 1 CDI (r ¼ .07). CHAS The CHAS (Kanner et al., 1987) contains a list of 39 daily hassles that children may experience in the areas of family, school, friends, and play (e.g., your parents were fighting; kids at school teased you; your teacher was mad at you because of your behaviour; you were punished for something you didn’t do). ’’. The items were developed based on children’s responses to a semi-structured interview about stress in their lives. On the CHAS, children are asked to indicate whether or not each hassle has occurred in their lives over the previous 6 weeks, and if it has occurred, how frequently it has occurred (0—Never to 4—All the time). The CHAS has previously been found to have high internal consistency (Kanner et al., 1987) and moderate test–retest reliability (Abela, Zuroff, Adams, & Hankin, in press). Results Descriptive data Table 1 displays the means, standard deviations, internal consistencies, and intercorrelations of all measures used in the study. . Overview of analyses To examine the diathesis-stress model, hierarchical multiple regression analysis was conducted (Cohen, Cohen, West, & Aiken, 2003). In examining anxious symptoms, Time 2 MASC score was the dependent variable. Time 1 MASC scores were entered into the equation first, in order to control for differences among participants in initial levels of anxious symptoms. CHAS scores were entered into the second step to account

Table 1 Intercorrelations, means, standard deviations, and internal consistencies of all measures Time 1 MASC Time 2 MASC Time 1 CDI Time 2 CDI BIS CHAS Mean Standard deviation Cronbach’s Alpha

.70*** .25*** .22*** .23*** .24*** 48.23 15.75 .87

Time 2 MASC

.18*** .30*** .29*** .31*** 45.96 15.78 .89

Time 1 CDI

.69*** .07 .43*** 9.32 6.96 .85

Time 2 CDI

.06 .52*** 8.58 7.18 .89

BIS

CHAS

.05 10.23 2.82 .60

83.10 23.91 .92

Note: MASC ¼ Multidimensional anxiety scale for children, CDI ¼ Children’s depression inventory, BIS ¼ Behavioural inhibition scale, CHAS ¼ Children’s hassles scale.  po:05,  po:01,  po:001.

ARTICLE IN PRESS 1342

K. Brozina, J.R.Z. Abela / Behaviour Research and Therapy 44 (2006) 1337–1346

for stress. BIS scores were entered into the third step. Finally, the BIS  CHAS interaction was entered, allowing us to examine whether the diathesis-stress interaction had an effect above and beyond the main effects of BIS and CHAS. In order to control for multicollinearity among main effect variables and interaction terms, all variables were standardized before interactions were calculated, and analyses were run. Based on previous research indicating that girls report higher levels of BI and anxious symptoms than boys (e.g., Muris et al., 1999), we examined the effect of gender by regressing Time 2 MASC onto Time 1 MASC, CHAS, BIS, gender and all possible two- and three-way interactions. Neither the BIS  gender, nor the BIS  CHAS  gender interaction was significant. As previous studies using the BIS have been in slightly older samples of children, we similarly tested for grade effects, and found none. In addition, the assumption of homogeneity of covariance (Joiner, 1994) was examined and met in the analysis, suggesting that the diathesisstress interaction does not have varying effects based on initial levels of anxious symptoms. Behavioural inhibition and anxious symptoms Results for the analysis are presented in Table 2—Panel a.

To illustrate the form of the diathesis-stress interaction, we followed the procedure recommended by Cohen et al. (2003) and computed residual change scores by inserting specific values for predictor variables (i.e., one standard deviation above and below the mean) into the regression equation. As can be seen in Fig. 1, children with high BI who experienced high levels of hassles showed increases in anxious symptoms across the 6-week follow-up. In contrast, children with high BI who experienced low levels of hassles and children with low BI (regardless of levels of hassles) showed decreases in anxious symptoms across the follow-up. Behavioural inhibition and depressive symptoms Similar analyses were conducted to examine whether BI predicts increases in depressive symptoms. The dependent variable was Time 2 CDI ...