Internet-based treatment for social phobia: a randomized controlled trial

Internet-Based Treatment for Social Phobia: A Randomized Controlled Trial m

Thomas Berger, Ele´ onore Hohl, and Franz Caspar University of Bern

In this study conducted in the French-speaking part of Switzerland, 52 individuals with social phobia were randomly assigned either to an Internet-based cognitive–behavioral treatment with minimal contact with therapists via e-mail or to a waiting-list control group. Significant differences between the two groups were found at posttreatment on all primary outcome measures (social anxiety measures) and on two of the secondary outcome measures (general symptomatology, therapy goal attainment). On average, within-groups effect sizes were large for the primary outcomes (Cohen’s d 5 0.82) and for secondary outcomes (Cohen’s d 5 1.04). Moreover, subjects in the treatment group fulfilled the criteria of clinically significant improvement significantly more often than subjects in the control group on all measured dimensions (58% vs. 20%). Users’ acceptance of the program was high. The results from the present study lend further support to the hypothesis that Internet-delivered interventions with minimal therapist contact are a promising treatment approach to social phobia. & 2009 Wiley Periodicals, Inc. J Clin Psychol 65: 1021–1035, 2009. Keywords: Internet-based treatment; social phobia; randomized controlled trial; cognitive–behavioral therapy; computer

Internet-based psychotherapeutic interventions are increasingly tested in controlled trials, and have been developed for various mental health problems, including anxiety and mood disorders (Andersson, Bergstro¨m, Carlbring, & Lindefors, 2005; Andersson, Cuijpers, Carlbring, & Linderfors, 2007; Barak, Hen, Boniel-Nissim, & Shapira, 2008; Spek et al., 2007). Most of this growing body of evidence comes from studies evaluating minimal-contact treatments, in which the therapist is actively involved, though to a lesser degree than in traditional therapy (Newman, Erickson, Przeworski, & Dzus, 2003). In Web-based approaches, minimal contact typically This study was funded by the Swiss National Science Foundation (SNF 100011–112345). We would like to thank Bertrand Auckenthaler, Michelle Bischoff, Me´lissa David, Robert Richardson and Isabelle Schmutz for their therapeutic work. Roxane Fumeaux and Bertrand Auckenthaler contributed greatly to the development of the Internet-based program. Correspondence concerning this article should be addressed to: Thomas Berger, University of Bern, Gesellschaftsstrasse 49, CH-3012 Bern, Switzerland; e-mail: [email protected]

JOURNAL OF CLINICAL PSYCHOLOGY, Vol. 65(10), 1021--1035 (2009) & 2009 Wiley Periodicals, Inc. Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jclp.20603

1022

Journal of Clinical Psychology, October 2009

means that therapists assist and support patients via e-mail. The main component, however, is that patients work their way through a self-help guide adapted for the World Wide Web. The existing self-help guides are more or less interactive, including text and/or multimedia, and are almost exclusively grounded in evidence-based, cognitive–behavioral approaches. The advantages of Web-based approaches, such as increased accessibility to and affordability of evidence-based treatments (Bauer, Golkaramnay, & Kordy, 2005; Caspar, 2004; Rochlen, Zack, & Speyer, 2004), may be of particular importance in highly prevalent disorders such as social phobia (Degonda & Angst, 1993; Magee, Eaton, & Wittchen, 1996). Although there are effective psychosocial treatments for social phobia (Acarturk, Cuijpers, van Straten, & de Graaf, in press; Heimberg, 2001), far from all individuals suffering from social anxiety seek and eventually find help (Kessler, 2003). Possible reasons for low treatment-seeking rates may not only lie in the costs of a treatment and in the limited access to appropriate mental health professionals (e.g., because of a shortage of competent therapists, long waiting lists, or the fact that people live in more rural areas), but may also be related to the disorder-specific fear of seeking and meeting a therapist (King & Poulos, 1998; Newman et al., 2003). Internet-based treatments may thus lower the threshold for seeking help (Budman, 2000). Following earlier work on Internet-based interventions for fear of public speaking (Botella, Hofmann, & Moscovitch, 2004), Internet-based treatments for social phobia have been tested in a series of efficacy trials independently conducted by a Swedish and an Australian group (Andersson et al., 2006; Carlbring, Furmark, Steczko, Ekselius, & Andersson, 2006; Carlbring et al., 2007; Tillfors et al., 2008; Titov, Andrews, Choi, Schwencke, & Mahoney, 2008; Titov, Andrews, & Schwencke, 2008; Titov, Andrews, Schwencke, Drobny, & Einstein, 2008). Overall, both groups found significant improvements for treated subjects in contrast to controls on most of the social anxiety measures and secondary outcome dimensions, with large within-groups effect sizes (Cohen’s d40.8), and medium to large betweengroups effect sizes (Cohen’s d40.5). In the Swedish trials, comparable effect sizes were found in three of the trials, although the protocols included either just a minimal-contact treatment (i.e., an Internet-delivered self-help program plus therapist contact via e-mail; Carlbring et al., 2006), a minimal-contact treatment plus two group exposure sessions in real life (Andersson et al., 2006), or a minimalcontact treatment that also included weekly telephone calls (Carlbring et al., 2007). The only obvious difference was found for attrition rates. Attrition was substantively higher in the minimal-contact treatment with only e-mail support (Carlbring et al., 2006), suggesting that the compliance to treatment can be increased with closer contact with a therapist, a finding that was also seen in the meta-analysis by Spek et al. (2007). This is also supported by a recent trial of the Australian group testing the effects of an Internet-based self-help guide for social phobia with or without additional therapist e-mails. The attrition rate was substantially higher in the pure self-help condition, and treatment effects were significantly higher in the minimalcontact condition in which regular e-mail contact with a therapist was included (Titov et al., 2008). Such findings fit well with research on self-help through written materials for social phobia, showing that in contrast to pure self-help, self-help augmented by therapist assistance resulted in marked improvements in symptoms of social phobia (Rapee, Abbott, Baillie, & Gaston, 2007). To our knowledge, we are the third research group that independently evaluated an Internet-delivered minimal-contact treatment for social phobia. Because the Journal of Clinical Psychology

DOI: 10.1002/jclp

Internet-Based Treatment for Social Phobia

1023

robustness of the empirical support for Internet-based treatments can still be questioned (e.g., Nguyen, Carrieri-Kohlman, Rankin, Slaughter, & Stulbarg, 2004), such independent replications are important. In addition, although the Swedish, the Australian, and our Swiss self-help packages all build upon cognitive–behavioral (CBT) approaches (e.g., Clark & Wells, 2003), they differ in the way the material is presented. For instance, whereas the Swedish solution is largely text-based and has also been termed an Internet-based bibliotherapy (e.g., Carlbring et al., 2006) or as Net-bibliosystem CBT (Marks, Cavanagh, & Gega, 2007), our program uses many of the interactive and multimedia features the Internet has to offer. The current study therefore also addresses the question as to whether more sophisticated technical solutions are feasible and effective. Method Recruitment and Selection of Participants Participants were recruited by means of advertisements in regional newspapers. Several persons found our Web page via search engines or links from other Web sites; this page presented general information about social phobia and its treatment, an outline of the study, a link to 24-hour emergency phone numbers, and an application form. We completed the first selection of participants by evaluating several self-report questionnaires. Only subjects that exceeded cutoff scores on the Social Phobia Scale (SPS; Mattick & Clark, 1998) or the Social Interaction Anxiety Scale (SIAS; Mattick & Clark, 1998) were included in the study. We referred subjects who exceeded a score of 1 on the suicide item of the Beck Depression Inventory (Beck & Steer, 1987), or who reported parasuicidal behavior in the Screening Questionnaire of the Structural Clinical Interview for DSM-IV, Axis-II (SCID-II; First, Gibbon, Spitzer, Williams, & Benjamin, 1997) to local psychiatrists or psychotherapists. In addition, we excluded subjects who reported more than 8 of the 15 symptoms of the borderline personality disorder module in the Screening Questionnaire of the SCID-II. Subsequently, participants who signed an informed consent form were interviewed, either in person or over the phone, using the Structured Clinical Interview for DSM-IV (SCID-I and SCID-II; First, Spitzer, Gibbon, & Williams, 1995; First et al., 1997). To be included in the study, participants had to meet the following additional criteria: (a) fulfill the DSM-IV criteria for social phobia according to the SCID-I; (b) follow no other psychological treatment for the duration of the study; (c) have access to a computer with an Internet connection; (d) be at least 18 years old, but no older than 45 (We only included participants in this age range because at the beginning of the project, Internet use was much more common among younger people. We only wanted to include participants who were naturally familiar with the Internet to limit uncontrolled variance.); and (e) not currently meet diagnostic criteria for psychosis or borderline personality disorder. According to the SCID-I, social phobia had to be considered to be the primary diagnosis if other comorbid disorders were present. We excluded 6 participants with very complex conditions and a treatment history going back more than 25 years. Even though they were not suicidal at baseline, we decided to refer these 6 patients to more intensive face-to-face treatments. Fifty-two participants met all inclusion criteria and were randomly assigned to the treatment group or waiting-list control group (31 to the treatment group; 21 to the control group). Figure 1 shows the participant flow. Journal of Clinical Psychology

DOI: 10.1002/jclp

1024

Journal of Clinical Psychology, October 2009

Figure 1. Participant flow.

Outcome Measures The following social anxiety scales constituted the primary outcome measures of the study: the self-report version of the Liebowitz Social Anxiety Scale (LSAS-SR; Baker, Heinrich, Kim, & Hofmann, 2002; Liebowitz, 1987), the Social Phobia Scale (SPS), and Social Interaction Anxiety Scale (SIAS; Mattick & Clark, 1998). In addition, the following secondary measures were used to assess depression, general symptomatology, interpersonal problems, and therapy goal attainment: Beck Depression Inventory (BDI; Beck & Steer, 1987), Symptom Check List (SCL-90-R; Derogatis, 1983), Inventory of Interpersonal Problems (IIP; Horowitz, Strauss, & Kordy, 2000), and the revised Goal Attainment Scale (GAS-R; Kiresuk & Lund, 1979). All outcome measures were sent by post and administered on paper. Goal attainment scaling involved the following steps: (a) participants were asked to specify the three most important therapy goals at pretreatment; (b) on a continuum of possible outcomes between 2 (worst expected outcome) and 14 (best expected outcome), pretreatment performance was set at 0; (c) criteria for scoring at each level were specified; and (d) participants were asked to evaluate the extent of goal attainment at posttreatment. The average score over all three goals was used as outcome measure. Procedure Participants, intention to treat, and attrition. We used a weighted randomization procedure (Altman, 1991), such that 60% were assigned to the treatment condition Journal of Clinical Psychology

DOI: 10.1002/jclp

Internet-Based Treatment for Social Phobia

1025

and 40% to the waiting-list control group. According to a computer-generated randomization scheme, 31 of the 52 subjects included in the randomization were assigned to the treatment group and 21 to the control group. After randomization, 5 participants (3 in the treatment group and 2 in the control group) dropped out during the course of the study and did not complete postassessment (9.6%). According to an intention-to-treat paradigm, we brought the pretreatment data forward for the noncompleters to replace missing values. Of the 3 participants in the treatment group who did not complete postassessment, one female participant did not write any message to the therapist (although the male therapist wrote five messages dedicated to motivating her), and she only completed approximately 15% of the self-help guide. The two other participants who did not fill out the postassessment were very active for 5 and 7 weeks, respectively, during which they both wrote six messages to the therapist and accomplished approximately 40% and 50%, respectively, of the self-help guide. When asked for their reasons for not continuing with the program and not completing the postassessment measures, neither participant provided us with any reason, and one wrote she did not have enough time to fill out the questionnaires. The 28 participants in the treatment group who filled out the postassessment measures completed on average approximately 85% of the self-help guide. Of these 16 (57.1%) completed the whole self-help guide (5 sessions), 6 (21.4%) completed approximately 80% of the self-help guide (4 sessions), 5 (17.9%) completed about 60% (3 sessions), and one participant (3.6%) only completed the first session. Most participants were Swiss citizens (46; 88%). Five participants lived in France (10%) and one in Belgium (2%). Of the total sample, 26.9% had a current psychiatric Axis I disorder in addition to the social phobia (e.g., major depression, panic disorder), and 38.5% had a current avoidant personality disorder. According to the Social Interaction Anxiety Scale (SIAS), 65% of the participants could have been classified as suffering from generalized social anxiety disorder (SIASZ42; Hart, Fresco, Turk, & Heimberg, 2000). See Table 1 for more demographic information. Intervention. The intervention consisted of a 10-week Internet-based program that included an interactive self-help guide, a module for establishing regular textbased contact with a therapist, a continuous monitoring and feedback system of patients’ responses, as well as collaborative online group elements, offering participants the opportunity to share their experiences with the other participants. We used SSL (Secure Sockets Layer) encryption to secure all Internet-based communication, and participants were identified using anonymous login names and passwords. The self-help guide is based on the established cognitive–behavioral approach by Clark and Wells (1995; Stangier, Heidenreich, & Peitz, 2003), which was adapted for the Internet. It is interactive in the sense that participants are able to freely navigate through the Web pages and repeat exercises and sessions; the program is also responsive to what the participants have entered in previous modules. The self-help package consists of a total of 57 Web sites that are divided into five sessions. In the first session, Clark and Wells’s (1995) model for social phobia is outlined. Various components of the model are explained and participants are asked to develop their own individualized model of their social anxiety step by step. They are asked to enter social situations they fear, cognitive and physical symptoms, safety behaviors and mental images that may be related to increased self-focused attention. Subsequently, the relationship between these components is explained with graphical Journal of Clinical Psychology

DOI: 10.1002/jclp

1026

Journal of Clinical Psychology, October 2009

Table 1 Demographic Description of the Participants Treatment (n 5 31) Variable

n

Gender Women Men Age Mean age (standard deviation) Min-Max Marital status Married/living together with child Married/living together without child Divorced with child Divorced without child Single without children Highest educational level Nine-year compulsory school Vocational school Secondary school College/University (not completed) College/University (completed) Treatment history Sought help before Earlier psychological treatment

18 13 28.1 21–43

%

Control (n 5 21) n

58.1 41.9 5.4

11 10 30.0 19–43

%

52.4 47.6 5.1

Total (n 5 52) n

29 23 28.9 19–43

%

55.8 44.2 5.3

6 9 0 1 15

19.4 29.0 0.0 3.2 48.4

2 5 1 0 13

9.5 23.8 4.8 0.0 61.9

8 14 1 1 28

15.4 26.9 1.9 1.9 53.9

0 1 5 10 15

0.0 3.2 16.1 32.3 48.4

1 4 5 4 7

4.8 19.1 23.8 19.1 33.3

1 5 10 14 22

1.9 9.6 19.2 26.9 42.3

9 2

29.0 6.5

9 7

42.9 33.3

18 9

34.6 17.3

animations where individual inputs are integrated. At the end of this session, important aspects are summarized and participants may print out an extensive summary of the model. The goal of the second session is to have participants experience the impact of increased self-focused attention, as well as the impact of safety behaviors on anxiety and social performance. This is realized by first asking participants to judge the intensity of the expected anxiety for several social situations such as reading a text in front of an audience or asking a question in a lecture room filled with 50 people. Movies of anxiety-provoking situations are then presented on the computer screen and participants are instructed, for instance, to read a text in front of an audience present on the screen. They do this twice. First they increase self-focused attention and activate safety behaviors, and the second time they focus their attention on the task and avoid using any safety behavior. At the end of each exercise, participants are asked to report where their attention was focused, what safety behaviors they used, and how intense their anxiety was. Finally, participants get a feedback on the association between, for example, self-focused attention and the intensity of anxiety. Our data suggest that such exercises were appropriate to demonstrate to participants the theoretically expected positive relationship between self-focused attention and the experienced intensity of anxiety. For instance, in 255 exercises accomplished by the subjects, a correlation of r(253) 5 0.81 ( po.001) was found between the subjects’ ratings of self-focused attention and the experienced intensity of anxiety. In the third session, users are informed about the importance of confronting social situations in reality and are encouraged to plan and engage in in vivo exposures. A diary for behavioral experiments is available where participants can enter the exposures they plan to carry out, as well as expectations regarding the anxiety level, Journal of Clinical Psychology

DOI: 10.1002/jclp

Internet-Based Treatment for Social Phobia

1027

physical symptoms, possible safety behaviors, and other variables. After the exposure, participants are invited to enter their observations regarding the outcome of their experiment (e.g., the experienced intensity of anxiety) and to compare these with their expectations. The fourth session targets the identification and modification of negative cognitions. First, negative cognitions entered in previous sessions (e.g., during the development of an individual model of social anxiety in Session 1) are challenged with questions typically used in cognitive approaches (e.g., in the situation ‘‘speaking freely in front of an audience’’ you are thinking ‘‘I will be at a loss for words, the others will think I am ridiculous,’’ what evidence do you have in support of your belief. Please enter the evidence.), and participants are instructed to find more balanced and helpful thoughts. Finally, a diary of negative thoughts is introduced where participants are asked to enter their usual negative cognitions, and to challenge these by finding more balanced cognitions. Each Web site and each session builds upon the previous one, and users only gain access to the next site if the previous sites/tasks have been completed. However, as many tasks and exercises are repeated, working with the self-help guide is far from a sequential process. Repetition is what the fifth and last session targets. It emphasizes the importance of practice and involves repeating the exercises and tasks introduced in the previous sessions. Participants have the opportunity to share their experiences with the other participants in various forums. For instance, after experiencing and playing with self-focused attention in Session 2, participants are encouraged to write about how they managed to focus their attention on the task, or about problems they had in directing their attention outwards. Participants are also encouraged to react to the contributions of others. This sharing of experiences takes place throughout all sessions, and participants may decide if they want to anonymously publish their inputs, such as their feared situations, their negative and helpful cognitions, or the behavioral experiments they realized. These collaborative elements are a real advantage of the Internet, and they may be a promising new way of bringing into play some of the therapeutic factors typical of group therapies within an individual setting (Berger & Caspar, 2008). Similar to most Internet-based treatments, the present program can be best described as minimal-contact therapy (Andersson et al., 2007; Newman et al., 2003). When patients log into the program for the first time, they are redirected to the contact module, where a therapist has introduced her- or himself and informed the patients that they may contact her/him whenever they want to. Therapists responded within at least 3 days after receiving a question from a patient. In addition, therapists were instructed to write a short motivating message to the clients once a week. In the present study, the mean number of messages written by the patients was 5.5 (Minimum: 0; Maximum: 16) with an average number of 189 words per message.

Therapists. The therapists were six psychologists with a master’s degree in clinical psychology (four women, two men). Four of them were in the first year of a CBT psychotherapy training program. Two therapists were following a postgraduate course in clinical psychology and psychotherapy. In our Web-based solution, therapists have their own Internet-based platform where they have an overview of all their patients and an easy and quick access to the Web-based environment of each patient. There they can follow the work and inputs of their clients. Journal of Clinical Psychology

DOI: 10.1002/jclp

1028

Journal of Clinical Psychology, October 2009

Statistical analysis. Primary and secondary outcome measures were first analyzed with a multivariate analysis of variance (MANOVA) on the pre- and postchange scores. Then, univariate analyses of variance (ANOVAs) on the change scores were calculated for each measure. Calculations of within- and between-groups effect sizes (Cohen’s d ) were based on the pooled standard deviation. For the Goal Attainment Scale (GAS-R), we used the standard deviation at postassessment to calculate withingroups effect sizes because pretreatment scores were set to 0. Clinical significance was defined as (a) a statistically reliable change according to the Reliable Change Index (RCI), and (2) being within two standard deviations of a nonclinical group at postassessment (Jacobson & Truax, 1991). Normative data were taken from French data sets when available; otherwise German norms were used. Significance testing of between-groups differences in terms of clinically significant change was conducted with chi-square tests. All analyses were conducted on an intent-to-treat basis. Results Main Outcome Results on the social anxiety scales and the secondary measures are presented in Table 2. Although the control group had slightly higher scores on almost all measures at pretreatment, the two groups did not differ significantly on any of the measures, t(51) 5 .07–1.71, p 5 .95–.19. The three social anxiety measures and also the secondary measures were highly and significantly intercorrelated at each point of assessment, primary measures from r(50) 5 .47 to r(50) 5 .76; secondary measures from r(50) 5 .33 to r(50) 5 .80; all pso.05). Table 2 Outcome Measures at Pre- and Postassessment, Including Effect Sizes Effect size (Cohen’s d ) Pretreatment score Mean (SD) Liebowitz Social Anxiety Scale Treatment group 68.7 (16.9) Control group 75.0 (17.4) Social Phobia Scale Treatment group 35.6 (14.2) Control group 35.1 (10.8) Social Interaction Anxiety Scale Treatment group 43.2 (10.9) Control group 47.4 (11.3) Global Severity Index (SCL-90-R) Treatment group 0.90 (0.33) Control group 0.95 (0.49) Beck Depression Inventory Treatment group 16.6 (6.2) Control group 17.9 (10.4) Inventory of Interpersonal Problems Treatment group 1.74 (0.40) Control group 1.74 (0.47) Goal Attainment Scaling Treatment group Control group

Posttreatment score Mean (SD)

Within group

Between group

52.7 (21.9) 70.7 (17.2)

0.82 0.25

0.89

23.5 (13.2) 30.3 (10.8)

0.88 0.44

0.55

34.1 (12.9) 44.0 (10.1)

0.76 0.32

0.84

1.05 0.42

0.54

1.03 0.44

0.44

1.40 (0.65) 1.56 (0.51)

0.63 0.37

0.27

1.61 (1.12) 0.88 (1.07)

1.44 0.82

0.66

0.52 (0.39) 0.75 (0.47) 8.6 (9.1) 13.1 (11.5)

Journal of Clinical Psychology

DOI: 10.1002/jclp

Internet-Based Treatment for Social Phobia

1029

The participants’ primary and secondary measures were analyzed with a multivariate analysis of variance (MANOVA) on the change scores, followed by univariate ANOVAs on the change scores for each measure. For the three social anxiety measures, the MANOVA revealed a significant between-groups effect at posttreatment, F(3,48) 5 3.1, p 5 .035. Univariate ANOVAs showed a highly significant between-groups effect for the LSAS-SR, F(1,50) 5 8.8, p 5 .005, and a significant effect for the SPS, F(1,50) 5 5.9, p 5 .019, and the SIAS, F(1,50) 5 4.1, p 5 .048. For the secondary outcome measures, the overall MANOVA did not indicate significant between-groups differences, F(4,47) 5 1.7, p 5 .178. Even though the MANOVA showed no significant difference among the groups on the dependent variables, we were interested in determining if the groups differed on any of the secondary outcome measures for exploratory purposes (Huberty & Morris, 1989). Subsequent ANOVAs revealed a significant effect for the SCL-90, F(1,50) 5 4.4, p 5 .041, and the GAS-R, F(1,50) 5 5.2, p 5 .027. No significant differences were found for the BDI, F(1,50) 5 2.0, p 5 .162, and the IIP improvement scores, F(1,50) 5 2.4, p 5 .131. Effect Sizes Effect sizes (Cohen’s d) were calculated both within and between groups, and are presented in Table 2. The mean within-groups effect size was high at d 5 0.82 for the primary outcome measures (LSAS-R, SPS, SIAS) and at d 5 1.04 for the secondary outcome measures (SCL-90-R, BDI, IIP, GAS-R). The between-groups effect sizes for primary outcome measures varied between d 5 0.55 (SPS) and 0.89 (LSAS-SR). Clinical Significance Data on clinically significant improvement are presented in Table 3. Consistently across all measures, about 58% of the participants in the treatment group fulfilled the criteria of clinically significant improvement compared to only 20% of the participants in the control group. All between-groups differences were significant. To examine possible moderators of outcome, we tested associations between clinically significant change and several baseline characteristics of the participants in the treatment group. A chi-square test revealed no significant difference between participants with an additional Axis I diagnosis compared to those with only a social phobia diagnosis, w2 (1) 5 0.49, p 5 0.64. Furthermore, no association was found with regard to an additional avoidant personality disorder, w2 (1) 5 2.41, p 5 0.18, participant’s sex, w2 (1) 5 0.68, p 5 0.45, or any other sociodemographic variable. However, the small group size reduced power to detect moderators of treatment outcome. Treatment Satisfaction Participants reported a high level of satisfaction with the Internet-based treatment. More specifically, 63% (n 5 17) of the participants were completely satisfied, 22% (n 5 6) were moderately satisfied, and 11% (n 5 3) were not satisfied. One (3.7%) of the responding participants rated the self-help modules as being too difficult and only one participant (3.7%) indicated that she did not immediately understand the specific purpose of the various self-help sessions. In addition 96% (n 5 26) of the subjects reported that their computer knowledge was totally appropriate to benefit from the program. Journal of Clinical Psychology

DOI: 10.1002/jclp

1030

Journal of Clinical Psychology, October 2009

Table 3 Percentage and Number of Participants Recovering From Treatment as Defined by Jacobson and Truax (1991) Treatment Measure

%

Liebowitz Social Anxiety Scale Recovered 54.8 Not recovered 45.2 Social Phobia Scale Recovered 58.1 Not recovered 41.9 Social Interaction Anxiety Scale Recovered 58.1 Not recovered 41.9 Global Severity Index (SCL-90) Recovered 58.1 Not recovered 41.9 Beck Depression Inventory Recovered 58.1 Not recovered 41.9 Inventory of Interpersonal Problems Recovered 58.1 Not recovered 41.9

Control n

%

n

w2

(1)

17.1 17 14

0.0 100.0

0 21

18 13

19.0 81.0

4 17

18 13

23.8 76.2

5 16

18 13

19.0 81.0

4 17

18 13

28.6 71.4

6 15

18 13

19.0 81.0

4 17

7.8 6.0 7.8 4.4 7.8

po.05. po.01.

Discussion The results from the present study give further support to the notion that Internetdelivered CBT with minimal therapist contact is a promising treatment approach in the treatment of social phobia. Compared to a waiting-list control group, the Internet-based treatment showed statistically significant improvements on all social anxiety measures (LSAS-SR; SPS; SIAS). The between-groups effect size for primary outcomes was at Cohens’ d 5 0.76. This could be compared with the effects reported in a recent meta-analysis by Acarturk et al. (2008), in which a subgroup analysis of 35 studies with waiting-list comparisons revealed effect sizes of d 5 0.86 on social anxiety measures for traditional cognitive behavioral therapy, cognitive therapy, social skills training, relaxation, and/or exposure. All statistical analyses used intent-to-treat principles in which pretreatment data were carried forward for noncompleters to replace missing values. However, there were few drop-outs in our study, with 90% of the participants completing the questionnaires at postassessment. The few missing data suggest that other methods to replace missing values would have been very unlikely to yield different results. Based on the intent-to-treat sample, the Internet-delivered treatment yielded large within-groups effect sizes on measures of social anxiety (Cohens’ d 5 0.82), and on measures of general symptomatology (SCL-90), depression (BDI), interpersonal problems (IIP), and therapy goal attainment (GAS-R; Cohen’ d 5 1.04). Moderate within-groups improvements were also observed across all measures in the waitinglist control group. The mean within-groups effect size in the control condition was at d 5 0.34 for primary outcomes and at d 5 0.51 for secondary outcomes. This attenuated between-groups differences, especially with regard to the secondary outcome measures. For instance, though within-groups effect sizes were at d 5 1.03 Journal of Clinical Psychology

DOI: 10.1002/jclp

Internet-Based Treatment for Social Phobia

1031

for the BDI and at d 5 0.63 for the IIP, between-groups effect sizes of d 5 0.44 and d 5 0.27, respectively, were found. Unfortunately, this study was underpowered to statistically detect this level of effects, which could be an explanation for the lack of significance between the two conditions on these two measures. However, the effects in the control group underline the importance of controlling the treatment condition for time and assessment effects. We suppose that in addition to a regression to the mean, the Structured Clinical Interview and the goal attainment scaling, which were conducted either by phone or in person after the preassessment, may have acted as a therapeutic intervention. Even so, it must be acknowledged that the waiting-list control group design does not allow any conclusion regarding the specificity of the intervention. Comparisons with other Internet-based interventions and face-to-face treatments will be the logical next step. This has been done in two recent trials on panic disorder. In both studies, the authors found equivalent outcomes for Internet delivered and face-to-face treatment (Carlbring et al., 2005; Kiropoulos et al., 2008). In this study, we followed the recommendation of using multiple outcome measures, including both universal scales and one individualized measure that assess progress in terms of specific goals set for a given patient (Kiesler, 1966). With regard to the Goal Attainment Scale (GAS-R), the treatment group showed significantly more goal attainment than the control group. On average, treated participants graded their goal attainment at postassessment close to a score of 2. Thus, treated participants indicated having taken concrete steps to attain their therapy goals while still not reaching the best possible outcome. The analyses of the proportion of participants fulfilling the criteria of clinically significant improvement revealed statistically significant differences between the treatment and control group on all measures. About 58% of the treated participants were classified as clinically significantly changed at posttreatment. This finding was very consistent across all measures. Not many studies evaluating psychological faceto-face treatments for social phobia provide information on clinically significant change, making it difficult to interpret our results. A fair estimate of the proportion of patients recovering during face-to-face CBT treatments is 65% (Rodebaugh, Holaway, & Heimberg, 2004). Thus, the percentage of participants fulfilling the criteria for clinical significant improvement was in the range of that found in evidence-based face-to-face approaches. However, the fact that about one third of patients were not recovering from treatment clearly indicates that there is still significant room for improvement in the therapeutic approach. There are several limitations to these interpretations and to the present study that should be addressed. It is unlikely that our sample was representative of the general population of social phobics, or that we evaluated the usual population that seeks or is referred to traditional face-to-face treatments. For instance, as in many clinical trials of Internet-based interventions, the sample was considerably better educated than the general population. In addition, due to the fact that we only included participants between 18 and 45 years of age, our participants were younger than those included in many other studies. This limits the generalizability of our results and raises the question whether less educated and older people would also benefit from such an Internet-based intervention. Also, a selection bias could be present because the sample was selected from individuals who had expressed interest in our Internet-based treatment. In addition, our extensive screening and diagnostic procedure for the recruitement of patients may have led to a selection of participants who were especially motivated for treatment. Most of the individuals who were initially interested in our treatment were lost because they did not return the Journal of Clinical Psychology

DOI: 10.1002/jclp

1032

Journal of Clinical Psychology, October 2009

comprehensive package of pretest questionnaires (Figure 1). The considerable initial investment may have represented a barrier for individuals who were hesitant about participating in our treatment, especially in a context where no personal contact was involved in the first screening step. However, our sample was severely impaired at pretreatment according to the social anxiety measures. For example, according to the Social Interaction Anxiety Scale (SIAS), approximately 65% of our participants could have been classified as suffering from generalized social anxiety disorder (SIASZ42; Hart, Fresco, Turk, & Heimberg, 2000), and the mean pretreatment score on the SIAS was above the mean score of a clinical sample of individuals suffering from social anxiety disorder as reported by Mattick and Clark (1998). Overall, with regard to the SPS, SIAS, and BDI, the pretreatment means of the participants in the present study were in the range of baseline means reported in recent efficacy and effectiveness trials on social phobia (Clark et al., 2003; Clark et al., 2006; Lincoln et al., 2003; Stangier, Heidenreich, Peitz, Lauterbach & Clark, 2003). However, at least part of our sample may be representative of a new and severely impaired population of social phobics that typically seek information about their disorder on the Internet. As evidenced by Erwina, Turk, Heimberg, Frescoa, and Hantula (2004) there seems to exist quite a large population of individuals with severe social anxiety who found a "home" on the Internet, who are not well represented in traditional clinical and epidemiological studies of social anxiety disorder, and who are likely not to receive adequate psychotherapy. There is thus a fundamental question, which sample is actually respresentative of social phobia patients. There are good reasons to also question the representativeness of patients seeking face-to-face treatment. It is interesting to note that our findings are very similar to the results of the Internet-based social phobia trials conducted in Sweden and Australia (e.g., Andersson et al., 2006; Titov et al., 2008). For instance, if one compares the scores on the social anxiety measures (SPS, SIA,S and LSAS-SR) of the current trial with the results of the Andersson et al. (2006) study, pre- and posttest scores were almost identical on all three measures (all discrepancies between mean scores were within half a standard deviation of a particular measure). To some extent, this comparable effectiveness is surprising. Although the treatment conditions in both studies were Internet-delivered, and although both self-help guides were derived from evidencebased cognitive-behavioral approaches, they differed in the way the material was presented. Whereas the Swedish self-help guide was a largely text-based, Internetdelivered book in PDF format, we invested a lot in the development of a sophisticated technical solution with many interactive and multimedia features. However, even though the participants have been satisfied with our solution and did not report technical problems with this rather complex program, the results of our trial do not suggest a superiority of the more interactive program that uses the advantages of the Internet in comparison to the rather simple ebook solution. Experimental comparisons of Internet-based protocols delivered in different formats could shed more light on how Internet-based treatments should best be presented. Indeed, the reason for the comparable effectiveness of the Swedish and Swiss program could be due to various other factors (e.g., the content of the self-help guide or the amount and quality of contacts with therapists). Finally, among the limitations of this study is the sole reliance on self-report measures. Behavioral tests, and clinician-administered scales at postassessment would have strengthened the results. In several Swedish studies, Internet-based programs have been found to be effective with regard to interview measures such as Journal of Clinical Psychology

DOI: 10.1002/jclp

Internet-Based Treatment for Social Phobia

1033

the SCID (Carlbring et al., 2005, 2006). In addition, follow-up of the results at 6 months will provide information about the sustainability of our Internet-based approach to social phobia. In conclusion, this trial suggests that guided Internet-delivered treatment of social anxiety disorder is a promising treatment option. Patient preferences, clinician preferences, and cost effectiveness will be important to consider in the future dissemination of the treatment approach. References Acarturk, C., Cuijpers, P., van Straten, A., & de Graaf, R. (2008). Psychological treatment of social anxiety disorder: A meta-analysis. Psychological Medicine, 39, 241–254. Altman, D.G. (1991). Practical statistics for medical research. London: Chapman and Hall. Andersson, G., Bergstro¨m, J., Carlbring, P., & Lindefors, N. (2005). The use of the internet in the treatment of anxiety disorders. Current Opinion in Psychiatry, 18, 73–77. Andersson, G., Carlbring, P., Holmstro¨m, A., Sparthan, E., Furmark, T., Nilsson-Ihrfelt, E., et al. (2006). Internet-based self-help with therapist feedback and in-vivo group exposure for social phobia: A randomized controlled trial. Journal of Consulting and Clinical Psychology, 74, 677–687. Andersson, G., Cuijpers, P., Carlbring, P., & Linderfors, N. (2007). Effects of internetdelivered cognitive behaviour therapy for anxiety and mood disorders. Review Series Psychiatry, 9, 9–14. Baker, S.L., Heinrich, N., Kim, H.-J., & Hofmann, S.G. (2002). The Liebowitz Social Anxiety Scale as a self-report instrument: A preliminary psychometric analysis. Behaviour Research and Therapy, 50, 701–715. Barak, A., Hen, L., Boniel-Nissim, M., & Shapira, N. (2008). A comprehensive review and a meta-analysis of the effectiveness of internet-based psychotherapeutic interventions. Journal of Technology in Human Services, 26, 109–160. Bauer, S., Golkaramnay, V., & Kordy, H. (2005). E-mental-health: Neue Medien in der psychosozialen Versorgung [E-mental-health: New medias in psychosocial services]. Psychotherapeut, 50, 7–15. Beck, A.T., & Steer, R.A. (1987). Beck Depression Inventory–manual. San Antonio: The Psychological Association. Berger, T., & Caspar, F. (2008). Von anderen Patienten lernen: Konsequenzen neuer Entwicklungen im Internet fu¨r webbasierte, psychotherapeutische Angebote [Learning from other patients: Consequences of new developments in the internet for web-based psychotherapeutic offers]. Psychotherapeut, 53(2), 130–137. Botella, C., Hofmann, S.G., & Moscovitch, D.A. (2004). A self-applied, internet-based intervention for fear of public speaking. Journal of Clinical Psychology, 60(8), 821–830. Budman, S.H. (2000). Behavioral health care dot-com and beyond: Computer-mediated communication in mental health and substance abuse treatment. American Psychologist, 55(11), 1290–1300. Carlbring, P., Furmark, T., Steczko, J., Ekselius, L., & Andersson, G. (2006). An open study of internet-based bibliotherapy with minimal therapist contact via email for social phobia. Clinical Psychologist, 10, 30–38. Carlbring, P., Gunnarsdottir, M., Hedensjo¨, L., Andersson, G., Ekselius, L., & Furmark, T. (2007). Treatment of social phobia: A randomized trial of internet delivered cognitive behaviour therapy and telephone support. British Journal of Psychiatry, 190, 123–128. Carlbring, P., Nilsson-Ihrfelt, E., Waara, J., Kollenstam, C., Buhrman, M., Kaldo, V., et al. (2005). Treatment of panic disorder: Live therapy vs. self-help via the internet. Behaviour Research and Therapy, 43, 1321–1333. Journal of Clinical Psychology

DOI: 10.1002/jclp

1034

Journal of Clinical Psychology, October 2009

Caspar, F. (2004). Technological developments and applications in clinical psychology and psychotherapy: Introduction. Journal of Clinical Psychology, 60, 221–238. Clark, D.M., Ehlers, A., Hackmann, A., McManus, F., Fennell, M., Grey, N., et al. (2006). Cognitive therapy versus exposure and applied relaxation in social phobia: A randomized controlled trial. Journal of Consulting and Clinical Psychology, 74, 568–578. Clark, D.M., Ehlers, A., McManus, F., Hackmann, A., Fennell, M., Campbell, H., et al. (2003). Cognitive therapy versus fluoxetine in generalized social phobia: A randomized placebo-controlled trial. Journal of Consulting and Clinical Psychology, 71, 1058–1067. Clark, D.M., & Wells, A. (1995). A cognitive model of social phobia. In R.G. Heimberg, M. Liebowitz, D. Hope, & F. Schneider (Eds.), Social phobia: Diagnosis, assessment and treatment (pp. 69–93). New York: Guilford Press. Degonda, M., & Angst, T. (1993). The Zurich Study. Social phobia and agoraphobia. European Archives of Psychiatry and Clinical Neuroscience, 243, 95–102. Derogatis, L.R. (1983). Symptom Checklist 90-Revised/SCL-90-R: Administration, scoring, and procedures manual. Baltimore, MD: Clinical Psychometric Research. Erwina, B.A., Turk, C.L., Heimberg, R.G., Frescoa, D.M., & Hantula, D.A. (2004). The internet: Home to a severe population of individuals with social anxiety disorder? Journal of Anxiety Disorders, 18, 629–646. First, M.B., Gibbon, M., Spitzer, R.L., Williams, J.B.W., & Benjamin, L.S. (1997). Scid-II: Structured Clinical Interview for DSM-IV Axis II Disorders. Washington, DC: American Psychiatric Press. First, M.B., Spitzer, R.L., Gibbon, M., & Williams, J.B.W. (1995). Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I). Washington, DC: American Psychiatric Press. Hart, T.A., Fresco, D.M., Turk, C.L., & Heimberg, R.G. (2000). Using self-report measures to identify social anxiety disorder and its subtypes: A receiver operating characteristic analysis. Unpublished manuscript. Heimberg, R.G. (2001). Current status of psychotherapeutic interventions for social phobia. Journal of Clinical Psychiatry, 62, 136–142. Horowitz, L.M., Strauss, B., & Kordy, H. (2000). Inventar zur Erfassung Interpersonaler Probleme–Deutsche Version (IIP-D) [Inventory of Interpersonal Problems–German version]. Weinheim: Beltz Verlag. Huberty, C.J., & Morris, J.D. (1989). Multivariate analysis versus multiple univariate analyses. Psychological Bulletin, 105, 302–308. Jacobson, N.S., & Truax, P. (1991). Clinical significance: A statistical approach to defining meaningful change in psychotherapy research. Journal of Consulting and Clinical Psychology, 59, 12–19. Kessler, R.C. (2003). The impairments caused by social phobia in the general population: Implications for intervention. Acta Psychiatrica Scandinavica, 108, 19–27. Kiesler, D.J. (1966). Some myths of psychotherapy research and the search for a paradigm. Psychological Bulletin, 65, 110–136. King, S.A., & Poulos, S.T. (1998). Using the internet to treat generalized social phobia and avoidant personality disorder. CyberPsychology and Behaviour, 1, 29–36. Kiresuk, T., & Lund, S. (1979). Goal attainment scaling: Research, evaluation and utilization. In H.C. Schulberger & F. Parker (Eds.), Program evaluation in health fields (vol. 2, pp. 214–237). New York: Human Science Press. Kiropoulos, L.A., Klein, B., Austin, D.W., Gilson, K., Pier, C., Mitchell, J., et al. (2008). Is internet-based CBT for panic disorder and agoraphobia as effective as face-to-face CBT? Journal of Anxiety Disorders, 22, 1273–1284. Liebowitz, M.J. (1987). Social phobia. Modern Problems of Pharmacopsychiatry, 22, 141–173. Journal of Clinical Psychology

DOI: 10.1002/jclp

Internet-Based Treatment for Social Phobia

1035

Lincoln, T.M., Rief, W., Hahlweg, K., Frank, M., von Witzleben, I., Schroeder, B., et al. (2003). Effectiveness of an empirically supported treatment for social phobia in the field. Behaviour Research and Therapy, 41, 1251–1269. Magee, W.J., Eaton, W.W., & Wittchen, H.U. (1996). Agoraphobia, simple phobia, and social phobia in the national comorbidity survey. Archives of General Psychiatry, 53, 159–168. Marks, I.M., Cavanagh, K., & Gega, L. (2007). Hands-on help: Computer-aided psychotherapy. Hove: Psychology Press. Mattick, R.P., & Clark, J.C. (1998). Development and validation of measures of social phobia scrutiny fear and social interaction anxiety. Behavior Research and Therapy, 36, 455–470. Newman, M.G., Erickson, T., Przeworski, A., & Dzus, E. (2003). Self-help and minimalcontact therapies for anxiety disorders: Is human contact necessary for therapeutic efficacy? Journal of Clinical Psychology, 59(3), 259–274. Nguyen, H.Q., Carrieri-Kohlman, V., Rankin, S.H., Slaughter, R., & Stulbarg, M. (2004). Internet-based patient education and support interventions: A review of evaluation studies and directions for future research. Computers in Biology and Medicine, 34, 95–112. Rapee, M., Abbott, M.J., Baillie, A.J., & Gaston, J.E. (2007). Treatment of social phobia through pure self-help and therapist-augmented self-help. British Journal of Psychiatry, 191, 246–252. Rochlen, A.B., Zack, J.S., & Speyer, C. (2004). Online therapy: Review of relevant definitions, debates and current empirical support. Journal of Clinical Psychology, 60(3), 269–283. Rodebaugh, T.L., Holaway, R.M., & Heimberg, R.G. (2004). The treatment of social anxiety disorder. Clinical Psychology Review, 24, 883–908. Spek, V., Cuijpers, P., Nyklicek, I., Riper, H., Keyzer, J., & Pop, V. (2007). Internet-based cognitive behaviour therapy for symptoms of depression and anxiety: A meta-analysis. Psychological Medicine, 37(3), 319–328. Stangier, U., Heidenreich, T., & Peitz, M. (2003). Soziale Phobien. Ein kognitivverhaltenstherapeutisches Behandlungsmanual [Social phobia. A cognitive-behavioral treatment manual]. Weinheim: Beltz. Stangier, U., Heidenreich, T., Peitz, M., Lauterbach, W., & Clark, D.M. (2003). Cognitive therapy for social phobia: Individual versus group treatment. Behaviour Research and Therapy, 41, 991–1007. Tillfors, M., Carlbring, P., Furmark, T., Lewenhaupt, S., Spak, M., Eriksson, A., et al. (2008). Treating university students with social phobia and public speaking fears: Internet delivered self-help with or without live group exposure sessions. Depression and Anxiety, 25(8), 708–717. Titov, N., Andrews, G., Schwencke, G., Drobny, J., & Einstein, D. (2008). Shyness 1: Distance treatment of social phobia over the internet. Australian and New Zealand Journal of Psychiatry, 42(7), 585–594. Titov, N., Andrews, G., & Schwencke, G. (2008). Shyness 2: Treating social phobia online: Replication and extension. Australian and New Zealand Journal of Psychiatry, 42(7), 595–605. Titov, N., Andrews, G., Choi, I., Schwencke, G., & Mahoney, A. (2008). Shyness 3: An RCT of guided versus unguided internet based CBT for social phobia. Australian and New Zealand Journal of Psychiatry, 42(12), 1030–1040.

Journal of Clinical Psychology

DOI: 10.1002/jclp