درمان اختلال اضطراب اجتماعی با استفاده از محیط های مجازی اینترنتی در زندگی دوم

Abstract Over 80% of people with social anxiety disorder (SAD) do not receive any type of treatment, despite the existence of effective evidence-based treatments. Barriers to treatment include lack of trained therapists (particularly in nonmetropolitan areas), logistical difficulties (e.g., cost, time, transportation), concerns regarding social stigma, and fear of negative evaluation from health care providers. Interventions conducted through electronic communication media, such as the Internet, have the potential to reach individuals who otherwise would not have access to evidence-based treatments. Second Life is an online virtual world that holds great promise in the widespread delivery of evidence-based treatments. We assessed the feasibility, acceptability, and initial efficacy of an acceptance-based behavior therapy in Second Life to treat adults with generalized SAD. Participants (n = 14) received 12 sessions of weekly therapy and were assessed at pretreatment, midtreatment, posttreatment, and follow-up. Participants and therapists rated the treatment program as acceptable and feasible, despite frequently encountered technical difficulties. Analyses showed significant pretreatment to follow-up improvements in social anxiety symptoms, depression, disability, and quality of life, with effect sizes comparable to previously published results of studies delivering in-person cognitive behavior therapy for SAD. Implications and future directions are discussed.

Results Both intent-to-treat (ITT) and completer-only analyses were conducted and results were equivalent; therefore, only the former are reported. Two participants who dropped out of treatment completed final assessments at the time of discontinuation. Multiple imputation (Rubin, 1987) using SPSS was conducted to account for missing data. Because this was a pilot study to explore if implementing CBT using Second Life is even feasible, we were mindful of balancing concerns over Type I error with those regarding Type II error; we therefore elected not to adjust alpha to control for experiment-wise error. Acceptability and Feasibility All participants, including dropouts, completed a posttreatment patient satisfaction survey, and reported satisfaction with their treatment (93% were very or mostly satisfied) and their therapists (100% were completely or mostly satisfied). Most patients believed that the treatment decreased their fears in social situations (86% strongly agreed or agreed) and decreased their avoidance of social situations (72% strongly agreed or agreed). Ninety-three percent of patients would recommend the treatment to a friend. Most patients reported that receiving treatment through Second Life was easy (79% reported very or fairly easy). The dropout rate was 14%, with 12 out of 14 participants completing treatment. One participant dropped out after the third session to avoid in-session exposure exercises, and the other participant dropped out after the ninth session to avoid in-session exposure exercises and because she was uncomfortable communicating with a therapist she could not see. All three therapists rated treatment in Second Life as “fairly feasible” (5 = very feasible, 4 = fairly feasible, 3 = neutral, 2 = fairly unfeasible, 1 = very unfeasible). Therapists reported insignificant (i.e., did not affect quality of treatment session) or no technical difficulties for 61% of sessions, moderate technical difficulties for 27% of sessions, and major to severe technical difficulties for 12% of sessions. Inability to transmit or receive sound through the headset on the part of the patient, therapist, or role-player was the most commonly reported technical problem, occurring in 21% of sessions; sometimes the microphone settings or headset volume was not adjusted properly, and other times the reason for the problem was unknown. Poor sound quality was experienced in 17% of sessions, with voices garbled or cutting out in the middle of conversation, or “echoing,” which occurs when one hears the sound of his or her own voice back through the headset. Due to technical difficulties with sound in Second Life, therapists reported using the telephone (concurrently with Second Life) for all or part of the time in 27% of sessions. In 6% of sessions, patients experienced hardware or Internet connection problems, such as being disconnected from Second Life, having a slow Internet connection, having their computer freeze, etc. Therapists reported that the moderate to severe technical difficulties impacted treatment by resulting in less time for discussion and exposure exercises, interrupted flow of the session, and difficulty in hearing what the other party was saying. The mean presence score was 4.71 (SD = 1.94), with 43% reporting a score of 6 or 7. Most patients anecdotally reported feeling anxious during in-session exposures, suggesting a level of psychological presence. However, correlations were small for the relationship between presence and residual change (pretreatment to follow-up) in social anxiety symptoms as measured by the Brief FNE, r(14) = .17, p = .56, the SPAI-SP, r(14) = .09, p = .76, and the LSAS-Total, r(14) = .05, p = .86; note that a validated measure of presence was not used. Treatment Outcome Measures Repeated measures ANOVAs were conducted for the primary outcome measures across pretreatment, midtreatment, posttreatment, and 12-week follow-up (see Table 2). The ANOVA results were significant with large effect sizes for all measures of social anxiety: SPAI-SP, Wilks's lambda = 0.32, F(3, 11) = 7.92, p < .01; LSAS-Fear, Wilks's lambda = 0.28, F(3, 11) = 9.52, p < .01; LSAS-Avoidance, Wilks's lambda = 0.33, F(3, 11) = 7.44, p < .01; Brief-FNE, Wilks's lambda = 0.24, F(3, 11) = 11.49, p < .01; and CGI-Severity, Wilks's lambda = 0.19, F(3, 11) = 15.22, p < .01. Tukey's post hoc tests showed decreases in social anxiety from pretreatment to follow-up for all social anxiety measures. At follow-up, the majority of participants (n = 8; 57%) no longer met DSM-IV criteria for SAD, and clinician ratings on the CGI-Improvement scale were as follows: 21% of participants were given a rating of 1 (very much improved), 43% a rating of 2 (much improved), and 21% a rating of 3 (minimally improved). Using the criteria suggested by Jacobson and Truax (1991), 36% of participants demonstrated clinically significant improvement, having a reliable change index greater than 1.96 and a SPAI-SP score below 106.5 ( Osman et al., 1996), which is closer to a non-SAD population. Treatment expectancy after Session 1 was not significantly correlated with residual change in social anxiety symptoms for the SPAI-SP, r(14) = .44, p = .11, Brief FNE, r(14) = -.16, p = .60, nor LSAS-Total, r(14) = .371, p = .19. Table 2. Means, Standard Deviations, and Effect Sizes for Intent-to-Treat Sample Measure M SD Effect Size (Cohen's d) SPAI-SP Pre-treatment 134.60 26.45 Mid-treatment 112.95 26.37 Post-treatment 96.96 38.93 1.42 Follow-up 94.90 29.35 1.50 LSAS-Total Pre-treatment 81.36 23.85 Mid-treatment 68.57 24.62 Post-treatment 55.86 27.22 1.07 Follow-up 52.50 24.20 1.21 LSAS-Fear Pre-treatment 41.79 11.37 Mid-treatment 35.93 11.41 Post-treatment 29.07 12.16 1.12 Follow-up 28.36 10.85 1.18 LSAS-Avoidance Pre-treatment 39.57 13.56 Mid-treatment 32.64 14.01 Post-treatment 26.79 16.19 0.94 Follow-up 24.14 14.24 1.14 Brief-FNE Pre-treatment 48.00 7.71 Mid-treatment 44.07 9.65 Post-treatment 38.64 10.79 1.21 Follow-up 38.07 9.24 1.29 BDI Pre-treatment 14.29 10.74 Mid-treatment 8.64 7.41 Post-treatment 6.64 8.61 0.71 Follow-up 7.29 7.65 0.65 CGI-Severity Pre-treatment 5.07 1.14 Mid-treatment 3.86 1.29 Post-treatment 3.29 1.64 1.56 Follow-up 2.93 1.49 1.88 QOLI Pre-treatment -0.93 1.93 Mid-treatment -0.09 1.41 Post-treatment 0.72 1.62 0.85 Follow-up 0.64 1.25 0.81 SDS-Total Pre-treatment 20.57 5.71 Mid-treatment 18.21 7.17 Post-treatment 11.86 6.95 -1.53 Follow-up 13.86 8.47 -1.18 AAQ-II Pre-treatment 30.50 10.11 Mid-treatment 28.00 6.48 Post-treatment 23.21 6.89 0.72 Follow-up 24.17 7.95 0.63 Table options The ANOVA on the BDI-II was also significant, with a medium effective size, Wilks's lambda = 0.38, F(3, 11) = 5.95, p = .01, with post hoc tests revealing decreases in depression from pretreatment to follow-up. Psychosocial functioning and disability also improved, with significant ANOVA results and large effect sizes for the SDS, Wilks's lambda = 0.12, F(3, 11) = 26.18, p < .01, and the QOLI, Wilks's lambda = 0.18, F(3, 11) = 17.08, p < .01, with post hoc tests showing improvements from pretreatment to follow-up. Significant results and a medium effect size was found for the AAQ-II, Wilks's lambda = 0.22, F(3, 11) = 10.62, p < .01, with post hoc tests showing decreases in experiential avoidance from pretreatment to follow-up.