بدرفتاری عاطفی و اختلال خوردن در نوجوانان: تست نقش میانجی تنظیم احساسات

Abstract The present study aimed to determine if emotion regulation mediates the relationship between emotional maltreatment and disordered eating behavior in adolescents. Participants were 222 secondary school pupils (aged 14–18 years) from a state high school in the UK. Standardized questionnaire measures were used to gather self-report data on emotional abuse and emotional neglect, functional and dysfunctional emotion regulation strategies and disordered eating behavior. Results showed that disordered eating was associated with emotional abuse, dysfunctional emotion regulation and being female. Multiple mediation analysis found an indirect relationship between emotional abuse and disordered eating through dysfunctional emotion regulation. Interestingly, emotional neglect predicted lower levels of functional emotion regulation. The findings support previous research showing emotion regulation to mediate the relationship between childhood abuse and disordered eating in adults and a differential effect of abuse and neglect on emotion regulation. Longitudinal studies are required to confirm the direction of relationships; however these data suggest that dysfunctional emotion regulation is a significant variable in the development of disordered eating and may be a useful target for intervention.

Results Prevalence of emotional maltreatment and disordered eating Table 1 shows the scores on all scales for males, females, and the whole sample. Participants were considered to display a history of emotional maltreatment if they scored within either of the top two categories of the childhood trauma scale subscales (within the moderate-severe or severe-extreme range). Twenty four participants (10.8%) fell within these ranges for emotional abuse, and 45 participants (20.3%) for emotional neglect. These means are very similar to those found in community samples and prevalence rates broadly consistent (Baker & Maiorino, 2010), though the rate of emotional abuse was somewhat lower in the present sample. The rates were also similar to those in a previous study in a Canadian high school sample (Wolfe, Scott, Wekerle, & Pittman, 2001) although Wolfe and colleagues found higher rates of abuse than neglect. There was a reasonably high level of co-occurrence of abuse and neglect: of the 24 participants whose scores indicated history of emotional abuse, 14 had a history of emotional neglect indicated. Therefore, 31.1% of those with history of emotional neglect also had history of emotional abuse indicated, and 58.3% of those with history of emotional abuse had scores suggesting a history of emotional neglect. Using the recommended score of ≥20 on the EAT-26, the scores of 17 participants (7.7%) were indicative of them being at risk of an eating disorder. Table 1. Descriptive data for maltreatment, emotion regulation and disordered eating for males, females, and total sample. Measure Males Females Total sample Mean (SD) Mean (SD) Mean (SD) CTQ Emotional abuse 8.16 (4.02) 8.55 (3.74) 8.33 (3.90) Emotional neglect 10.67 (4.26) 10.52 (4.22) 10.60 (4.23) REQ Functional 2.66 (.58) 2.80 (.50) 2.72 (.55) Dysfunctional 2.02 (.58) 2.06 (.58) 2.04 (.58) EAT-26 5.56 (9.66) 9.63 (11.44) 7.37 (10.66) Note: CTQ = Childhood Trauma Questionnaire; REQ = Regulation of Emotions Questionnaire; EAT-26 = Eating Attitudes Test-26 total score. Table options Gender differences in scores across measures For comparisons of scores between males and females, log transformed data were used. Childhood emotional maltreatment scores did not differ significantly between males and females (t(220) = 1.14, p = .25 for emotional abuse; t(220) = .18, p = .86 for emotional neglect). Males and females did not differ in the use of dysfunctional emotion regulation strategies (t(220) = .63, p = .53). Females scored more highly on the use of functional emotion regulation strategies, though this difference was not quite at a significant level (t(220) = 1.93, p = .06). Females scored more highly on EAT-26 total score (t(220) = 4.60, p < .001). Correlations between emotional maltreatment, emotion regulation and disordered eating Pearson correlation coefficients using log transformed scores are reported in Table 2. Emotional abuse and emotional neglect were both significantly positively correlated with EAT-26 total scores (both p < .001). Emotional abuse and emotional neglect were both significantly positively correlated with dysfunctional emotion regulation and negatively correlated with functional emotion regulation (all p < .001). Dysfunctional emotion regulation was correlated with the disordered eating score (p < .001) but functional emotion regulation was not (p = .82). A medium sized correlation between emotional abuse and emotional neglect supported the decision to treat these as separate constructs in the subsequent analysis. Table 2. Pearson's correlation coefficients between emotional abuse and neglect, emotion regulation and disordered eating (transformed data) (N = 222). Variables 1. 2. 3. 4. 5. CTQ EA – CTQ EN .47** – REQ functional −.25** −.44** – REQ dysfunctional .63** .28** −.15* – EAT total .44** .23** −.02 .43** – Note: CTQ = Childhood Trauma Questionnaire, EA = emotional abuse, EN = emotional neglect; REQ = Regulation of Emotions Questionnaire, Functional = functional emotion regulation, Dysfunctional = dysfunctional emotion regulation; EAT = Eating Attitudes Test-26 total score. * p < .05. ** p < .01 (2-tailed). Table options Emotion regulation as a mediator of the relationship between maltreatment and disordered eating To test whether emotion regulation mediated the relationship between emotional maltreatment and disordered eating, analyses were carried out using the Mediate macro, with disordered eating (EAT-26 total score) as the outcome variable, emotional neglect and abuse as separate independent variables, and both functional and dysfunctional emotion regulation scores as mediators. Transformed scores were used within this analysis. As males and females differed in disordered eating scores, gender was added as a covariate. The output from the mediation analysis is presented in the following order in the text and in Table 3: each emotion regulation score was regressed onto emotional abuse, neglect and gender (i.e., to test pathways from the independent variables to mediators) then the disordered eating score was regressed onto emotion regulation and emotional abuse, neglect, and gender (i.e., to test the pathway from independent variables and mediators to dependent variable). Confidence intervals for indirect effects were then examined to determine whether these effects were significant. Table 3. Regression models for predicting mediators from independent variables and dependent variable from independent variables and mediators. Outcome and predictors B coefficient SE t p Outcome = REQ functional Constant .7103 .0411 17.28 <.0001 Emotional abuse −.0344 .0385 .89 .37 Emotional neglect −.2187 .0366 5.98 <.001 Gender −.0244 .0114 2.13 .03 Outcome = REQ dysfunctional Constant −.0709 .0430 1.65 .10 Emotional abuse .4260 .0402 10.59 <.001 Emotional neglect −.0142 .0382 .37 .71 Gender .0015 .0120 .13 .90 Outcome = EAT Constant −.1579 .2680 −.59 .56 REQ functional .3956 .2862 1.38 .17 REQ dysfunctional .9408 .2738 3.44 <.001 Emotional abuse .6273 .2004 3.13 .002 Emotional neglect .2079 .1668 1.25 .21 Gender −.2182 .0489 4.47 <.001 Note: SE = standard error; REQ = Regulation of Emotions Questionnaire, Functional = functional emotion regulation, Dysfunctional = dysfunctional emotion regulation; EAT = Eating Attitudes Test-26 total score. Table options The first part of Table 3 shows that functional emotion regulation was predicted by gender and emotional neglect, such that being female was associated with higher levels of functional emotion regulation and scoring more highly on neglect was associated with lower functional emotion regulation scores. Emotional abuse did not predict functional emotion regulation. The second part of Table 3 shows that dysfunctional emotion regulation was predicted only by emotional abuse, where higher abuse scores were associated with greater dysfunctional emotion regulation scores. The third section of Table 3 shows the coefficients for disordered eating scores regressed onto emotional abuse and neglect, functional and dysfunctional emotion regulation, and gender. Within this model, dysfunctional emotion regulation, emotional abuse, and gender were significant predictors, where higher levels of disordered eating were associated with being female and greater levels of abuse, and dysfunctional emotion regulation. This model for predicting disordered eating was significant (F(5,216) = 19.48, p < .001), accounting for 31.1% of the variance in disordered eating scores. Table 4 shows the bias accelerated bootstrap confidence intervals for the indirect effects from emotional abuse and neglect to disordered eating through functional and dysfunctional emotion regulation. The only confidence intervals not including a zero value were for the indirect effect from emotional abuse to disordered eating through dysfunctional emotion regulation. Because emotional abuse was still accounting for a significant amount of variance in disordered eating when disordered eating was regressed onto emotional maltreatment and emotion regulation (Table 3), it was concluded that dysfunctional emotion regulation did not account for all the variance in the relationship between the two variables. Table 4. Bootstrap confidence intervals for the indirect effects of emotional maltreatment on disordered eating through emotion regulation. Variables Effect SE BCA bootstrap CI Lower Upper IV = Emotional abuse REQ functional −.0136 .0256 −.0946 .0168 REQ dysfunctional .4008 .1271 .1689 .6736 IV = Emotional neglect REQ functional −.0865 .0584 −.2109 .0191 REQ dysfunctional −.0134 .0388 −.0936 .0638 Note: SE = standard error; BCA = bias corrected and accelerated, CI = confidence intervals. Confidence intervals are based on 5,000 samples.