Virtual Reality for AnxIety Disorders - Randomized Controlled Trial

INTRODUCTION

Social anxiety disorder (SAD) is a common anxiety disorder characterized by excessive fear of being scrutinized or criticized by others leading to avoidance of social situations. According to the ICD-10 classification of mental and behavioral disorders, engaging in feared situations is accompanied by autonomic symptoms of anxiety such as sweating, trembling or increased heartrate (HR). The lifetime prevalence of SAD ranges between 8.4% and 12.1% and the 12 month prevalence ranges between 4.2% and 7.1%.

SAD is related to reduced health-related quality of life and is also associated with substantial psychiatric comorbidity including other anxiety disorders, mood disorders and substance use disorders. Epidemiological studies show that SAD most often precedes depression and that SAD is related to a substantially and consistently increase in risk of subsequent depression. Similarly, symptoms of social anxiety often precedes alcohol dependence.

SAD is an adolescent-onset disorder with a long recovery period. Despite the prolonged recovery, few individuals with SAD seek treatment for their disorder. Only about one-third of lifetime cases report ever seeking treatment for SAD. Not seeking treatment may be related to the nature of the disorder itself. Individuals with SAD avoid treatment because the treatment itself constitutes a social situation that provokes anxiety.

The treatment of choice for social anxiety is cognitive behavioral therapy (CBT). Treatment is conducted both individually and in group-settings. Exposure therapy is central to CBT and is very effective in fear reduction.

In vivo exposure is effective when treating SAD, but conducting in vivo exposure in session can be challenging because relevant social situations might be difficult to obtain and control. In addition, finding the right setting for exposure can be time consuming and costly.

Exposure in virtual reality has several advantages compared to in vivo exposure. Virtual reality provides readily available environments for exposure, such as a meeting room with a group of people waiting for the patient to give a presentation. Furthermore, exposure in VR is highly controllable and can be modified to fit the needs of the patient. Finally, exposure takes place in confidentiality within the safety of the therapy room and thus the threshold for initiating exposure might be lower than for in vivo exposure.

AIMS AND HYPOTHESES

Primary hypothesis:

At post treatment the investigators expect that CBT including exposure therapy using individually tailored VR-content and a system to track anxiety levels (CBT-ExpVR) will result in lower levels of social anxiety than CBT with exposure in vivo (CBT-Exp). The outcome on social anxiety will be measured using the total score on Social Interaction Anxiety Scale (SIAS).

Secondary hypotheses:

At 6 and 12 months follow-up, the investigators expect that VR-treatment will result in lower levels of social anxiety than in vivo-treatment.

Post treatment and at 6 and 12 months follow up the investigators expect that VR-treatment will result in lower levels of depression and higher levels of self-rated health than in vivo-treatment.

The dropout rate the investigators expect will be lower for the VR-treatment compared to the in vivo-treatment.

In addition to the evaluation of effect, a health economic evaluation will be made from a societal perspective.

METHODS

Design:

The trial is a randomized controlled, assessor-blinded, parallel-group superiority trail. The study is conducted at Centre for Telepsychiatry in the Mental Health Services in the Region of Southern Denmark

Participants and recruitment:

Participants are referred to the trial's website where they are provided with written information about the study and are invited to complete online questionnaires screening for social anxiety symptoms and symptoms of depression. The questionnaire consists of Social Interaction Anxiety Scale and Major Depression Inventory and questions on current treatment and use of medication. Cut-off score for inclusion is <22 on SIAS and cut-off score for exclusion on MDI is >29. The online questionnaire might be supplemented by a phone call to inquire further information on current treatment and medication. No information will be obtained from patient records.

Eligible participants are invited to an assessment at Centre for Telepsychiatry. The assessment will be carried out using the short version of the Present State Examination. Participants who meet the inclusion criteria will be offered to participate in the study. Before the pre-treatment assessment, included patients will be asked to complete an informed consent form.

Randomization:

Participants are block randomized (1:1) using random block size. The block sizes will not be disclosed, to ensure concealment. Computer generated random numbers using the platform Sealed Envelope (www.sealedeenvelope.com) will be used to generate the allocation sequence. The allocation sequence is handled by a data manager from the Patient data Explorative Network (OPEN) and is unavailable to those who enroll and assign participants.

Blinding:

Psychologist responsible for the diagnostic interview are blind to treatment assignment. Participants are blinded to their treatment assignment until their first treatment session. Participants will only be blind to treatment assignment at the pre-test but not at later assessments. A clinician blind to treatment assignment will administer the assessment taking place at pre-test, mid-test, and post-test.

Sample size:

A recent study reported approximately 10-point drop in SIAS for standard treatment involving CBT and 20-point drop for CBT with virtual exposure with standard deviations around 15 points. A drop of 13 points on SIAS is considered reliable change.

If 35 participants are recruited for each group, this will lead to a statistical power of 0.80 comparing the VR-treatment to in vivo-treatment at the 0.05 significance level. To consider a 20% dropout the investigators plan to invite 90 participants in total.

Fidelity:

The interventions are manualized to increase treatment fidelity. To ensure that the treatment is delivered consistently and reliably in accordance with the manual the therapist will after each treatment session answer a self-report questionnaire on specific treatment targets for each session.

Data collection and management:

Data collected from the participants using self-report measures and data reported by therapist are collected using REDCap (Research Electronic Data Capture). To ensure confidentiality assigned researchers and the data manager at OPEN will be the only people with access to data at REDCap. Informed consent forms will be scanned and stored in REDCap . Data is stored at OPEN's server located in the Regions of Southern Denmark. Subjective Units of Distress Scale, Heart Rate, Electrodermal activity, and estimated anxiety are collected using iMotions, and data is saved on a secure folder on SharePoint.

Statistical analyses:

The primary statistical analysis will be carried out as intention to treat (ITT). The investigators will use linear mixed models to analyze the data. Separate analysis will be performed for each outcome variable. The investigators will use a two-level model with observations nested within participants. The fixed effects will be time, intervention, and the interaction between time and intervention as well as the baseline score. In addition to the primary analysis, a per-protocol analysis will be carried out on those participants completing at least 50% of the exposure sessions.

A sensitivity analysis will be carried out where missing data will be handled by multiple imputation (m=100). Imputations will be based on baseline characteristics and secondary outcome measures using chained equations. The moderating effect of the working alliance, depressive symptoms, alcohol and drug use on the treatment outcome will be explored as subgroup analysis with continuous moderators by including them as covariates interacting with treatment and time.

Model validation in the linear mixed model will be performed by inspection of QQ-plots of residuals and BLUPs (best linear unbiased predictors) to assess normality, and plotting residuals against fitted values to check homoscedasticity. If assumptions are violated, analysis will be performed after log-transformation. If assumptions do not hold on log-scale, bootstrapping will be applied.

Monitoring:

Cybersickness similar to motion sickness may occur in the VR-setting. Cybersickness will be monitored using the Simulator Sickness Questionnaire. Adverse events will be registered by the therapists.

Dissemination policy:

Results will be disseminated regardless of the magnitude or direction of effect. Both positive, negative, and inconclusive results will be made public and both beneficial and harmful effects of adaptive virtual reality exposure therapy will be reported. Dissemination will happen in scientific journals, at scientific conferences, as well as via www.clinicaltrials.gov. Authorship will be determined according to the Vancouver Guidelines