Cognitive Bias Modification (CBM) interventions directly target cognitive mechanisms associated with vulnerability to anxiety disorders, such as attention and interpretation biases (see (1
)). CBM affects cognition via repeated practice on tasks that require rapid processing (for reviews of CBM, see ((3
)). To date, two independent randomized controlled trials (RCT) have demonstrated preliminary efficacy of CBM targeting attention bias (CBM-A) in individuals with Social Anxiety Disorder (SAD) (6
). Both trials utilized the same protocol, which comprised a dot probe task completed in eight, twice-weekly computer sessions. Between-group effect sizes (d = 0.35 to 1.59) were comparable to existing psychosocial and pharmacological treatments for SAD, and symptom reduction was maintained at a four month follow-up.
Researchers have also developed CBM tasks to target interpretation bias (CBM-I). Although no RCTs have been conducted in samples meeting diagnostic criteria for SAD, one randomized trial examined a CBM-I treatment in highly socially anxious individuals (8
). After eight twice-weekly sessions of a CBM-I intended to reinforce benign interpretations and extinguish threat interpretations of ambiguous sentences, CBM-I successfully decreased social anxiety symptoms compared to the placebo condition.
In order to maximize clinical impact, a logical next step is the combination of CBM-A and CBM-I treatments. Recently, a small (n = 13) open trial (9
) tested a combined CBM-A-I treatment and the feasibility of delivering CBM in an outpatient psychology office. In their sample of patients with SAD or Generalized Anxiety Disorder, CBM yielded significant effects on state and trait anxiety. Although these results are encouraging, they must be interpreted in the context of several major design limitations (e.g., lack of control group, small sample, completer analyses).
Positive results are promising given the lack of clinician contact, brevity, and ease of administration of CBM. However, most previous studies did not explicitly assess credibility or patient acceptability and satisfaction. Regarding credibility, researchers have asked participants at the end of the study whether they believed they received a treatment or placebo. Only 6% of participants correctly guessed that they received a treatment following CBM-A (7
), compared to 46% following CBM-I (8
). While such findings quell any concerns about demand effects, they suggest that CBM may not be a very credible treatment, particularly CBM-A.
Regarding acceptability, a recent open trial of CBM-A suggests that this type of treatment is acceptable to anxious children and their parents based on a participant acceptability questionnaire(10
). However, acceptability has yet to be systematically examined in adults with SAD, which is crucial for eventual update of treatments in the real world. Brosan and colleagues’ open trial of a combined CBM-A-I protocol informally asked patients about their experiences. Patients found CBM-I to be helpful in increasing their awareness of their negative thinking patterns. However, patient comments suggest that CBM-A may have limited acceptability due to its “boring” nature (p.5).
Attrition rates from previous CBM studies (0% to 8%) do not provide ideal data regarding acceptability because each study was conducted in an anxiety clinic and provided compensation for session attendance. We do not know the tolerability of these procedures when participants are recruited directly from the community and are not paid for completing sessions. These data, as well as the previously mentioned potential problems with credibility, suggest that a more systematic evaluation of credibility and patient acceptability is warranted.
Thus, the current RCT tested an Attention and Interpretation Modification (AIM) CBM treatment for SAD and examined its efficacy, credibility, and acceptability. We aimed to examine the potential for AIM to have real world utility. First, we combined the CBM-A and CBM-I treatments in order to maximize clinical impact. We also provided participants with a brief treatment rationale stating that the computer programs were designed to help participants develop healthier mental habits. Second, we recruited patients from the general community and primary care clinics, and we allowed patients to be engaged in concurrent treatment in order to obtain a less selected sample. Third, we did not compensate participants for session attendance in order to obtain more informative attrition rates. Finally, we included established measures of credibility and acceptability. We expected participants to report less social anxiety symptoms following AIM compared to a control condition. To extend previous RCT's assessment of efficacy, we also included a behavioral measure of social anxiety (i.e., impromptu speech). Based on our experience and preliminary findings from Brosan and colleagues (9
), we expected participants receiving AIM to rate the CBM-A task as less acceptable compared to the CBM-I task.