ترس، اجتناب و علائم فیزیولوژیکی در طول درمان شناختی رفتاری برای اختلال اضطراب اجتماعی

Abstract We examined fear, avoidance and physiological symptoms during cognitive-behavioral therapy (CBT) for social anxiety disorder (SAD). Participants were 177 individuals with generalized SAD who underwent a 14-week group CBT as part of a randomized controlled treatment trial. Participants filled out self-report measures of SAD symptoms at pre-treatment, week 4 of treatment, week 8 of treatment, and week 14 of treatment (post-treatment). Cross-lagged Structural Equation Modeling indicated that during the first 8 weeks of treatment avoidance predicted subsequent fear above and beyond previous fear, but fear did not predict subsequent avoidance beyond previous avoidance. However, during the last 6 weeks of treatment both fear and avoidance predicted changes in each other. In addition, changes in physiological symptoms occurred independently of changes in fear and avoidance. Our findings suggest that changes in avoidance spark the cycle of change in treatment of SAD, but the cycle may continue to maintain itself through reciprocal relationships between avoidance and fear. In addition, physiological symptoms may change through distinct processes that are independent from those involved in changes of fear and avoidance.

Results Missing data Our sample included 177 individuals with fear, avoidance, and physiological symptoms measured at four time-points during treatment. Thus the total data set included 2124 data points with 336 (15.8%) containing missing data. These data were missing completely at random (Little's MCAR test χ2 = 58.60, df = 69, p = 0.81, n.s.). Missing data were imputed using Bayesian multiple imputation, and all analyses were conducted on the imputed data set. SEM analyses Table 1 presents the correlation matrix between all variables in the model, as well as means and standard deviations. We estimated a model with all synchronous effects, stability effects, and cross-lagged effects. This model had a good fit with the data (Satorra–Bentler scaled χ2 = 39.49, df = 27, p = 0.06; TLI = 0.97; CFI = 0.99; RMSEA = 0.05). In the next step, we removed all non-significant pathways according to the procedures outlined by Bentler and Moojaart (1989). The resulting model included all synchronous effects and stability effects, as well as the cross-lagged pathways between avoidance at pre-treatment and fear at week 4 (β = 0.19, SE = 0.06, p < 0.05), avoidance at week 4 and fear at week 8 (β = 0.28, SE = 0.05, p < 0.001), avoidance at week 8 and fear at week 14 (β = 0.23, SE = 0.06, p < 0.001), and fear at week 8 and avoidance at week 14 (β = 0.22, SE = 0.10, p < 0.01). This model had an excellent fit with the data (Satorra–Bentler scaled χ2 = 52.61, df = 41, p = 0.11; TLI = 0.98; CFI = 0.99; RMSEA = 0.04). This final model is presented in Fig. 1. Table 1. Correlation matrix between all variables in the model. 1 2 3 4 5 6 7 8 9 10 11 12 1. Fear 0 – 2. Fear 4 0.53*** – 3. Fear 8 0.43*** 0.62*** – 4. Fear 14 0.35*** 0.44*** 0.75*** – 5. Avoid 0 0.76*** 0.48*** 0.41*** 0.35*** – 6. Avoid 4 0.42*** 0.69*** 0.56*** 0.43*** 0.51*** – 7. Avoid 8 0.26*** 0.44*** 0.71*** 0.65*** 0.37*** 0.59*** – 8. Avoid 14 0.21** 0.37*** 0.59*** 0.81*** 0.29*** 0.44*** 0.66*** – 9. Phys 0 0.24** 0.09 0.11 0.02 0.24** 0.16* 0.02 0.01 – 10. Phys 4 0.11 0.21** 0.14 0.08 0.15 0.24** 0.13 0.10 0.47*** – 11. Phys 8 0.00 0.11 0.31*** 0.28*** 0.10 0.10 0.32*** 0.28*** 0.34*** 0.47*** – 12. Phys 14 0.06 0.14 0.31*** 0.41*** 0.11 0.24** 0.31*** 0.44*** 0.19* 0.32*** 0.51*** – M 16.52 14.85 12.28 10.01 15.84 13.21 10.46 8.09 6.35 4.87 3.31 2.93 SD 4.17 3.78 3.95 4.30 5.02 5.41 4.65 5.14 3.09 2.89 2.40 2.66 Note. Fear = Fear subscale of the Brief Social Phobia Scale; Avoid = Avoidance subscale of the Brief Social Phobia Scale; Phys = Psychophysiological symptoms subscale of the Brief Social Phobia Scale; 0 = week 0 (pre-treatment); 4 = week 4; 8 = week 8; 14 = week 14 (post-treatment). *p < 0.05, **p < 0.01, ***p < 0.001. Table options Final SEM model. Note. Numbers represent regression weights (β) and pertain to ... Fig. 1. Final SEM model. Note. Numbers represent regression weights (β) and pertain to statistically significant parameters. Synchronous effects (i.e. correlations between subscales at a given time point) are not presented for weeks 4, 8, and 14 for purposes of clarity. Correlations at week 4 are 0.61, 0.20 and 0.23 (for fear–avoidance, avoidance–physiological symptoms, and fear–physiological symptoms respectively). Correlations at week 8 are 0.60, 0.37 and 0.36 (for fear–avoidance, avoidance–physiological symptoms, and fear–physiological symptoms respectively). Correlations at week 14 are 0.66, 0.31 and 0.26 (for fear–avoidance, avoidance–physiological symptoms, and fear–physiological symptoms respectively). Figure options We also examined whether this model held for all three treatment conditions. When all pathways were constrained to equality across the conditions, the resulting Satorra–Bentler scaled χ2 was 217.49 (df = 173, p < 0.01). When the pathways were allowed to differ between conditions, the Satorra–Bentler scaled χ2 was 155.09 (df = 123, p < 0.05). The Satorra–Bentler scaled χ2 difference was 62.41 (df = 50) which is below the critical value of 67.51. This indicates that the difference between models is non-significant and that allowing the pathways to vary does not result in a better fitting model. Therefore, the final model has a similar fit in all conditions. Synchronous effects We compared synchronous effects at each time point using Meng, Rosenthal, and Rubin's (1992) formula. We found that at each time point, fear–avoidance correlations (r = 0.60–0.76) were significantly greater than correlations between fear and physiological symptoms (r = 0.23–0.36), and correlations between avoidance and physiological symptoms (r = 0.20–0.37; all ps < 0.05). Mediation We examined whether the effect of avoidance at week 4 on avoidance at week 14 was mediated by fear at week 8. We followed the guidelines of Preacher and Hayes (2004) and conducted the mediation analysis using a bootstrapping procedure (5000 samples). We found evidence of a significant indirect effect (95% confidence interval = 0.18–0.35) indicating that fear significantly mediated the relationship between avoidance at week 4 and avoidance at week 14. The index of mediation ( Preacher & Hayes, 2008; Preacher & Kelley, 2011), a standardized effect size for the indirect effect, was 0.28, indicating that avoidance at week 14 increases by 0.28 standard deviations for every 1 SD increase in avoidance at week 4 indirectly through fear at week 8. Expectancy We examined whether expectancy ratings predicted post-treatment BSPS scores while controlling for pretreatment BSPS scores using regression analyses. We found no significant relationship between expectancy ratings and post-treatment SAD severity (β = −0.11, t = −1.18, p = 0.24). Similarly, no significant relationship was found between expectancy ratings and improvement during treatment (β = −0.11, t = −1.53, p = 0.13).