This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Mental Health, is properly cited. The complete bibliographic information, a link to the original publication on http://mental.jmir.org/, as well as this copyright and license information must be included.
Prior research has demonstrated the efficacy of internet-based cognitive behavioral therapy (ICBT) for social anxiety disorder (SAD). However, it is unclear how shame influences the efficacy of this treatment.
This study aimed to investigate the role shame played in the ICBT treatment process for participants with SAD.
A total of 104 Chinese participants (73 females; age: mean 24.92, SD 4.59 years) were randomly assigned to self-help ICBT, guided ICBT, or wait list control groups. For the guided ICBT group, half of the participants were assigned to the group at a time due to resource constraints. This led to a time difference among the three groups. Participants were assessed before and immediately after the intervention using the Social Interaction Anxiety Scale (SIAS), Social Phobia Scale (SPS), and Experience of Shame Scale (ESS).
Participants’ social anxiety symptoms (self-help: differences between pre- and posttreatment SIAS=−12.71; Cohen
The findings of this study suggest that participants’ engagement in the exposure module in ICBT alleviates social anxiety symptoms by reducing the levels of shame proneness. Our study provides a new perspective for understanding the role of shame in the treatment of social anxiety. The possible mechanisms of the mediation effect and clinical implications are discussed.
Chinese Clinical Trial Registry ChiCTR1900021952; http://www.chictr.org.cn/showproj.aspx?proj=36977
Internet-based cognitive behavioral therapy (ICBT) entails similar content to conventional in-person cognitive behavioral therapy (CBT), which has been proven to have treatment effects equivalent to pharmacological treatments [
Shame shares many similarities with social anxiety, including self-directed attention; fear of negative evaluations from others; and regarding oneself as unwelcome, unattractive, or worthless in others’ view [
Some empirical studies have investigated the relationship between shame and social anxiety through self-reported scales [
In this study, we investigated the following questions among a sample of Chinese individuals with SAD: (1) Is shame proneness significantly reduced over the course of treatment using a Chinese version of the ICBT? and (2) If so, which modules in the ICBT influence the levels of shame proneness? We hypothesized that the levels of shame proneness would be reduced over the course of ICBT treatment and that shame proneness would mediate the relationship between ICBT modules and social anxiety symptoms. If successful, this investigation will further elucidate the treatment of SAD and contribute new insights into the development of more detailed and targeted ICBT programs.
This research was an 8-week clinical trial. Participants were recruited from 2015 to 2017 in two different stages: a pilot study, which consisted of only the self-help and the wait list control (WLC) groups, and a controlled trial with 3 groups. All data came from a larger program of ICBT. This study was approved by the local ethics committee and registered in Peking University. The trial registration number is ChiCTR1900021952.
We used a community sample in the study. Participants were recruited through different internet platforms, and they were informed about the basic information, aim, and procedure of the study. Individuals who were interested in the study were required to finish several self-reported questionnaires on the website (N=1479). In addition, they were invited to participate in the Chinese version of the Mini International Neuropsychiatric Interview (MINI; N=784) [
The main inclusion criteria were as follows: participants who were older than 18 years and who met the diagnostic criteria of SAD in the Structural Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV) Axis I Disorders. Their Social Interaction Anxiety Scale (SIAS) score was higher than 22, with Social Phobia Scale (SPS) score higher than 33. They did not take any antipsychotic drugs or undergo other psychological treatments in the last year, and they did not meet the diagnostic criteria of schizophrenia, bipolar disorder, and high suicidal tendency. Participants had to agree that they could finish the 8-week ICBT program and the posttreatment measurements. Detailed information of the screening process and the eligibility criteria are shown in
Flowchart of this study.
A total of 104 participants, including 31 males and 73 females, aged 18 to 45 years (mean 24.92, SD 4.59 years) met the criteria and agreed to attend the treatment program.
The 25-item Experience of Shame Scale was composed by Qian et al [
SIAS is a 19-item scale, originally composed by Mattick and Clarke and revised into a Chinese version [
The Chinese version of the Beck Depression Inventory (BDI) scale is widely used in the measurement of depressive symptoms, with high reliability (standard Cronbach alpha=.890 and split-half reliability=0.879) and validity [
The MINI [
The primary outcome measures were changes in the ESS score and the relationships of ESS and ICBT, whereas others were recorded as secondary outcomes.
The ICBT program is an internet-based self-help cognitive behavioral intervention course and was first developed at the University of Bern [
The 8-week courses can roughly be divided into 5 parts. First, motivation arousing, which guides the participants to think about and write down why they want to change and what life would be like if social anxiety symptoms reduce. Relaxation training would also be introduced to participants in this module. Second, psychoeducation, which explains the relevant theories of SAD, the concepts of negative thoughts, safety behaviors, self-focus attention, and their relationships, helping participants gradually construct the case formulation of their own. Third, cognitive construct, which instructs participants to identify and re-examine their nonadaptive negative thoughts and to take notes on the
Overall, two forms of the ICBT intervention were included in the study: the self-help ICBT and guided ICBT. A total of 3 therapists were included in the program, all of whom were masters-level graduate students in clinical psychology, who had undergone formal CBT training and had at least 1 year of experience of individual counseling, and they were supervised by a licensed clinical psychologist on a weekly basis. Each therapist assists a certain number of participants when needed in the guided group. The assistance of the therapists consisted of a weekly email to each patient, aiming at motivating and reinforcing their usage of the ICBT program. Furthermore, therapists answered participants’ questions about the ICBT program. Therapists also needed to know the basic information of their patients and their progress in the program, the last time of their visit, and the homework record. Approximately 15 min were needed to prepare and reply to the email per patient for each week. The program had an independent network platform for therapists, and they can check the login information and relative data of all participants on the platform (such as their homework and the time they spent on each module).
After the screening process, the participants would first sign the digital informed consent form via internet and were provided with the instructions of the program. After which they would be divided into 3 groups: guided group, self-help group, and WLC group. Each individual needed to fill out the SPS, SIAS, and ESS scales before and immediately after completing the ICBT program (or 2 months later for the WLC group).
All analyses were conducted using SPSS version 20 (IBM Corp). First, differences among various groups in demographic and pretreatment clinical variables were tested using chi-square and one-way analysis of variance (ANOVA) tests. Repeated measures of ANOVA were also conducted to verify the participants’ improvement after ICBT.
For further analysis, we introduced a variable, residual gain (RG), to indicate the intervention changes. A linear regression model was fitted to find the specific modules of ICBT, which have an impact on RG of shame proneness. Afterward, we conducted a mediation analysis to investigate the role of shame proneness in the ICBT treatment. The results related to ESS (shame proneness) were regarded as the primary outcome.
The descriptive statistics of all variable scores are shown in
The descriptive statistics of all variables before and after the treatment.
Sociodemographics | Intervention (n=80) | Chi-square ( |
|||||||||||||
|
Self-help (n=47) | Guided (n=33) | Wait list (n=24) |
|
|
|
|||||||||
|
Values |
|
Values |
|
Values |
|
|
|
|
||||||
|
0.9 | ||||||||||||||
|
Female | 34 (72) | N/A | 22 (67) | N/A | 17 (71) | N/A | N/A | N/A |
|
|||||
Age (years), mean (SD) | 25.91 (4.25) | N/A | 24.73 (5.40) | N/A | 23.25 (3.59) | N/A | 2.81 (2,101) | .07 |
|
||||||
|
0.9 | ||||||||||||||
|
Low/middle | 31 (66) | N/A | 22 (67) | N/A | 16 (67) | N/A | N/A | N/A |
|
|||||
|
High | 16 (34) | N/A | 11 (33) | N/A | 8 (33) | N/A | N/A | N/A |
|
|||||
|
0.9 | ||||||||||||||
|
SADb | 23 (49) | N/A | 15 (45) | N/A | 9 (38) | N/A | N/A | N/A |
|
|||||
|
SAD+MDDc | 9 (19) | N/A | 9 (27) | N/A | 6 (25) | N/A | N/A | N/A |
|
|||||
|
SAD+ADd | 8 (17) | N/A | 5 (15) | N/A | 5 (20) | N/A | N/A | N/A |
|
|||||
|
SAD+MDD+AD | 7 (15) | N/A | 4 (12) | N/A | 4 (16) | N/A | N/A | N/A |
|
|||||
|
N/A | 0.80 | N/A | 0.88 | N/A | 0.25 | N/A | N/A | N/A | ||||||
|
Pretreatment | 75.45 (10.53) | N/A | 74.61 (12.94) | N/A | 76.29 (12.48) | N/A | 0.14 (2,101) | .87 |
|
|||||
|
Posttreatment | 68.11 (13.92) | N/A | 64.64 (14.97) | N/A | 76.92 (13.13) | N/A | 5.47 (2,101) | .006 |
|
|||||
|
N/A | 1.01 | N/A | 1.20 | N/A | 0.21 | N/A | N/A | N/A | ||||||
|
Pretreatment | 66.62 (10.62) | N/A | 70.67 (9.36) | N/A | 66.04 (10.61) | N/A | 1.96 (2,101) | .15 |
|
|||||
|
Posttreatment | 53.91 (14.80) | N/A | 51.21 (13.25) | N/A | 66.50 (13.11) | N/A | 9.27 (2,101) | <.001 |
|
|||||
|
N/A | 0.89 | N/A | 0.96 | N/A | 0.25 | N/A | N/A | N/A | ||||||
|
Pretreatment | 56.02 (14.16) | N/A | 53.48 (13.87) | N/A | 55.42 (13.74) | N/A | 0.33 (2,101) | .72 |
|
|||||
|
Posttreatment | 44.89 (17.54) | N/A | 40.03 (14.99) | N/A | 56.29 (16.90) | N/A | 6.83 (2,101) | .002 |
|
aN/A: not applicable.
bSAD: social anxiety disorder.
cMDD: major depressive disorder.
dAD: other anxiety disorders.
The dropout rate difference between the self-help (32.86%) and guided (52.86%) ICBT groups was significant, with higher dropout rate in the guided group (χ²1=5.7;
We used repeated measures of ANOVA to access whether ICBT can reduce participants’ shame proneness. The results showed that the interaction effect of group and time on ESS (
In our study, RG (post-pre) was used as an improvement index of the ICBT treatment. RG is calculated as follows: Z2−(Z1×r12), in which Z2 means the Z score of posttreatment, Z1 is the pretreatment Z score, and r12 refers to the Pearson correlation of pre- and posttreatment scores [
To investigate the relationship between different levels of shame and ICBT, we analyzed the Pearson correlation between the pretest ESS score (ESS-pre) and the RG of social anxiety (RG-SIAS and RG-SPS). The results did not show significant correlations (RG-SIAS:
Thus, we further investigated which module of ICBT had an effect on the decrease of shame. Using the feedback system of the network platform, we analyzed the Pearson correlation between each module’s involvement (using frequency and time as the indexes) and pretest ESS score and RG-ESS (the RG of ESS). The results showed that there was no significant correlation between pretest ESS and any involvement index; however, the frequency of relaxation training (
Furthermore, we used ENTERING method to perform a linear regression analysis on gender, age, group, the frequency of relaxation training, the total number of words writing in graded exposure and the average number of words writing in graded exposure as well as systematic problems. This linear regression analysis determined whether these parameters had influenced the score of RG-ESS. The results showed that the goodness-of-fit was the highest when the regression model included only gender and the average number of words of graded exposure as the predictive variables (
The regression coefficients of the regression model.
Variable | β | SE | Tolerance | Variance inflation factor | ||
Constant | N/Aa | N/A | 1.10 | .27 | N/A | N/A |
Gender | .21 | 0.175 | 1.97 | .05 | 0.98 | 1.02 |
The average number of words of graded exposure | −.33 | 0.002 | −3.13 | .002 | 0.98 | 1.02 |
aN/A: not applicable.
To further investigate the relationship among ICBT, shame proneness, and social anxiety, we did a mediation analysis. We used the average number of words of the exposure module as the predictive variable, the RG of ESS (the decrease of shame) as the mediation variable, and the RG of SIAS and SPS (the improvement of social anxiety) as dependent variables. The results of our analysis revealed evidence of a significant indirect effect of the average number of words of the exposure module on the improvement of social anxiety symptoms via their decrease of shame (for SIAS: β=−.0049; SE 0.0016; 95% CI −0.0085 to −0.0019 and for SPS: β=−.0039; SE 0.0015; 95% CI −0.0075 to −0.0012). Two graphical depictions of the model were seen in
The mediation effect of shame decrease on exposure to Social Interaction Anxiety Scale.
The mediation effect of shame improvement on exposure to Social Phobia Scale.
We conducted a group (self-help, guided, and WLC) × time (pre-/posttreatment) repeated measures of ANOVA analysis to access the treatment effect of ICBT. The results showed that the interaction effects of group and time on SIAS (
We used the RG of SIAS, SPS, and ESS as dependent variables and conducted independent
Our study used the Chinese version of the ICBT program to investigate whether shame can be significantly reduced during the treatment of SAD and which modules of ICBT exert an influence on the decrease of shame proneness. This study found significant reductions in participants’ shame proneness and social anxiety scores over the course of ICBT treatment in both the self-help and guided groups. Our study also suggested that gender (being female) and level of involvement in the exposure module (ie, higher average word count in completed homework assignments) were the only two significant predictors of reductions in shame proneness.
Furthermore, shame proneness fully mediated the relationship between the participants’ average word count in the exposure module and change in social anxiety scores. That is, greater engagement in the exposure module led to greater improvements in shame proneness, which, in turn, led to greater improvements in social anxiety symptoms.
As for the dropout rate and adherence, our results showed that the dropout rate of the guided group was significantly higher than that of the self-help ICBT group, and shame proneness was not a moderator in the relationship between the form of ICBT and treatment adherence.
The effectiveness of both the self-help and guided ICBT on SAD showed in this study is consistent with previous research [
The relationship between shame and social anxiety has long been debated. According to the psychoevolutionary model [
In addition, our study indicated that shame proneness played a mediation role in the relationship between the participants’ average word count in the exposure module of the ICBT and change in social anxiety scores. Some previous evidence might explain the mediational model. Many studies have confirmed that early negative experiences (such as emotional neglect and abuse) have an influence on feelings of shame and social anxiety, which are subsequently internalized, causing more stable shameful-based schemas [
Furthermore, our study showed a higher dropout rate in the guided ICBT group compared with the self-help group. This might be because of more perceived burden of participants in the guided group, who thought of the email support as another homework. Haug et al [
There are several limitations worth noting in this study. First, our study did not investigate the follow-up effect of ICBT on shame proneness and social anxiety symptoms. Future research is needed to explore the long-term effects of ICBT on these constructs and their interaction. Second, it remains possible that the sequence of the interventions may have contributed to the mediation effect. ICBT is a continuous therapy with 8 different modules, in which exposure is the last one. Participants’ motivation and involvement of the exposure might be influenced by previous modules, which we were not able to differentiate in this study. Finally, in our study, the exposure was implemented as a one-time intervention, whereas the measurement of shame proneness was measured as change over the course of the 8-week intervention, which may also confound the effects of other aspects of the intervention. Future dismantling studies are needed to separate these influences and further verify this mediation effect by using only the exposure intervention rather than the entire ICBT package.
In accordance with the theories mentioned earlier, our results suggest that shame proneness is an important factor in treating SAD and can be reduced through engagement in a web-based, self-guided exposure treatment. To our knowledge, this is the first study to investigate the mediation effect of shame proneness in the relationship between ICBT (particularly the exposure component of ICBT) and social anxiety symptoms. Our results suggest that among all the ICBT modules we investigated, only the completion of the exposure component significantly improved social anxiety symptoms by reducing the level of shame proneness. In short, this investigation further elucidates a process-based approach to alleviate shame and social anxiety and contribute insights into the development of more tailored exposure-based ICBT programs.
analysis of variance
Beck Depression Inventory
cognitive behavioral therapy
Diagnostic and Statistical Manual of Mental Disorders, 4th edition
Experience of Shame Scale
internet-based cognitive behavioral therapy
mean deviation
Mini International Neuropsychiatric Interview
posterior cingulate cortex
residual gain
social anxiety disorder
Social Interaction Anxiety Scale
Social Phobia Scale
wait list control
This work was supported by grants to SL from the National Key R&D Program of China (2017YFB1002503) and the National Natural Science Foundation of China (31571127). The authors would like to acknowledge Lingyu Lin for her kind guidance in the writing of the manuscript.
None declared.