Web-based Interpretation Training For Anxiety

The study aims to develop a web-based Cognitive Bias Modification infrastructure to train interpretations, and evaluate the usability, acceptability, and feasibility of the program to reduce anxiety symptoms.

Approximately half of the U.S. population experiences serious mental health problems during their lifetime, including 29% with anxiety pathology severe enough to qualify for an anxiety disorder diagnosis [1]. Critically, more than two thirds of individuals struggling with a mental illness do not receive treatment [1]. With this level of mental illness burden, it is clear that treating people one-on-one in an office setting will never meet the existing needs [2]. There are many barriers to treatment, including costs [3], difficulties accessing evidence-based treatments in many regions [4], and associated stigma [5]. Thus, there is a pressing need to consider alternative, larger scale approaches to delivering mental health services. Cognitive Bias Modification (CBM) interventions hold considerable promise as a way to meet these needs, especially for anxiety difficulties [6]. These computer-based programs are designed to alter biased ways of thinking, such as a tendency toward negative interpretations, which cause and maintain anxiety [6]. Because these programs do not require therapist contact and can be administered on any computer with an Internet connection, CBM holds promise as a cost-effective method that can be disseminated widely. However, while CBM for interpretation bias (CBM-I) has established efficacy when administered in-person in the laboratory [7], it now needs to be tested with broader populations using a web-based infrastructure to examine: a) whether the program will be effective in a web environment, b) whether the program continues to engage the targeted mechanism (i.e., interpretation bias), c) the feasibility of this delivery method, and d) the modifications needed to adapt the program for the web (in particular, to prime anxiety-linked negative thinking in an online environment, we test the effect of adding a guided anxious imagery exercise to prime feared outcomes prior to each training session). Together, the current proposal will develop an infrastructure to pilot test the effectiveness of web-based CBM-I for anxiety symptoms. CBM-I training will target moderate to severe anxiety symptoms, a widespread problem area with considerable occupational and social impairment [8]. Participants will be visitors to Project Implicit Mental Health (PIMH), an existing website directed by the Principal Investigator that allows visitors to assess their cognitive biases tied to mental health concerns. Consistent with the RFA's priorities, this approach encourages efficiencies by capitalizing on the existing PIMH site and its heavy traffic. Further, the site's large number of visitors and use of automated assessments will make it efficient to assess baseline demographic characteristics and interpretation bias as moderators of CBM-I effects that can be tested in future trials. Aim 1: Develop and evaluate usability and acceptability of web-based CBM-I for anxiety symptoms. Aim 1 will build the web-based interpretation bias training program using the PIMH infrastructure. We will pilot the program on a small test group of moderate to highly anxious participants (N=15) who will complete questionnaires and semi-structured interviews to provide feedback on the programs' usability and acceptability. Further, an advisory board (N=8) of anxiety researchers, clinicians, and experts in CBM and web-based research will provide feedback on the program and study protocol. Using a "deployment-focused" approach, this feedback from experts and end-users will be used to iteratively modify the program for the trial planned for Aims 2 and 3. Thus, even at this initial pilot stage, we will measure the targeted outcome (anxiety symptoms) and mechanism (interpretation bias) to determine whether modifications to enhance target engagement are needed. Note, within Research Domain Criteria (RDoC), this outcome falls under the Potential Threat/Anxiety construct within the Negative Valence System, and the targeted mechanism (interpretation bias) falls under the Response Selection, Inhibition construct within the Cognitive (effortful) control system. Both the outcome and mechanism will be objectively measured using multiple units of analysis (e.g., behavior and self-report). Further, mechanisms underlying the guided anxious imagery prime's effects will be measured by assessing subjective distress, imagery vividness, and activation of feared outcomes following the manipulation. This prime was selected in part because of its potential to be disseminated widely in future trials, given it does not require human contact. Aim 2: Test target engagement, feasibility and effectiveness of web-based CBM-I. Aim 3: Evaluate the impact of an anxious prime on web-based CBM-I for anxiety symptoms. Aims 2 and 3 will test the feasibility of an 8-session web-based interpretation training program among individuals with moderate to severe anxiety symptoms (based on screening at the PIMH site). Participants will be randomly assigned to positive CBM-I (90% positive scenario training), 50% positive/50% negative CBM-I, or a no scenario control condition. Half the participants in each of these 3 conditions will receive an anxious imagery prime prior to each training session, and half will receive a neutral imagery prime, resulting in a 3 training condition x 2 prime design (N=210; target of n=35 per condition). Feasibility will be determined by analyses of recruitment, attrition, acceptance of randomization, adherence to and appropriateness of the measurement model, caseness, extent of missing data, and safety. Additionally, target engagement (change in interpretation bias) and preliminary tests of effectiveness at reducing anxiety symptoms will be evaluated.

Positive Cognitive Bias Modification - Interpretation training paired with a preceding Anxious Imagery Prime

Positive Cognitive Bias Modification - Interpretation training paired with a preceding Neutral Imagery Prime

50/50 Cognitive Bias Modification - Interpretation (half positive & half negative scenarios) paired with a preceding Anxious Imagery Prime

50/50 Cognitive Bias Modification - Interpretation (half positive & half negative scenarios) paired with a preceding Neutral Imagery Prime

Training involves presenting participants with brief scenarios that introduce an ambiguous potential threat. Critically, the ambiguity regarding how the situation is resolved remains until the last word of the scenario, which is presented as a word fragment that the participant must solve, which will then assign a benign (rather than threatening) meaning to the scenario.

This condition follows the same design and content as the Positive Training conditions, except the word fragments differ; rather than consistently resolving the scenario in a positive direction, half of the scenarios will end with a negative (anxiety-congruent) word fragment, and half will end with a positive word fragment.

We will test whether priming anxious concerns and feared negative outcomes on the web via the addition of a guided anxious imagery prime at the start of each session will enhance effects.

At the start of each session, participants will complete a neutral control imagery exercise where they do a guided imagery exercise imagining upcoming mundane tasks, like brushing one's teeth.

To measure interpretation bias, participants will read ambiguous scenarios with titles, after which they will see the titles of each scenario, followed by 2 disambiguated interpretations of the scenario: 1 positive and 1 negative. Participants will rate each disambiguated interpretation based on how similar in meaning it is to the original scenario.

Baseline, and after sessions 3 (~2 weeks following baseline), 6 (~10 days after session 3), and 8 (~7 days after session 6) & at 2-month follow-up. Measure will be completed immediately following that day's training session.

This 5-item measure of anxiety symptom severity and impairment has good psychometric properties, shows treatment sensitivity, and is valid in community and clinical samples.

Baseline, and after sessions 2,3,4,5,6,7,8 (sessions will be spaced ~3-4 days apart) & at 2-month follow-up. Measure will be completed immediately following that day's training session. Session 2 is expected to occur within 3-8 days of the baseline.

This questionnaire assesses users' opinions about the intervention, including the extent participants found the intervention helpful for their anxiety, their satisfaction with the length of the intervention, how engaged they were, and their experience of being randomly assigned to condition.

To assess interpretation change following the CBM-I trial, participants are presented with ambiguous events and then asked to rate three alternative explanations for why the event might have occurred. One option is always negative, whereas the other responses are either neutral and/or positive.

Baseline, and after sessions 3 (~2 weeks following baseline), 6 (~10 days after session 3), and 8 (~7 days after session 6) & at 2-month follow-up. Measure will be completed immediately following that day's training session.

Impact of Anxious Imagery Prime (change over time is exploratory but should show anxious/neutral prime differences; e.g., in subjective distress)

Following the anxious or neutral imagery prime, participants report subjective distress (peak anxiety), vividness of the image, and estimate the probability that the situation they imagined will turn out well or badly, and provide a catastrophe judgment (i.e., how manageable it would be if this situation did turn out badly). All items rated on a 1-5 scale.

After sessions 1,2,3,4,5,6,7,8 (sessions will be spaced ~3-4 days apart). Measure will be completed following that day's imagery prime (which immediately precedes the training session).

This 7-item measure of anxiety symptoms has good psychometric properties, shows treatment sensitivity, and is valid in community and clinical samples.

This 16-item questionnaire was chosen to evaluate change in quality of life in five domains: recreation, relationships, well-being, personal fulfillment and development, and social activities.

Baseline, and after sessions 3 (~2 weeks following baseline), 6 (~10 days after session 3), and 8 (~7 days after session 6) & at 2-month follow-up. Measure will be completed immediately following that day's training session.

Inclusion Criteria: The target population will be adults age 18 and over who score in the moderate to extremely severe anxiety range (i.e., 10 or higher) on the Depression, Anxiety, Stress Scales - Short Form: Anxiety Subscale and have regular access to the Internet. Exclusion Criteria: None listed.

Steinman SA, Portnow S, Billingsley AL, Zhang D, Teachman BA. Threat and benign interpretation bias might not be a unidimensional construct. Cogn Emot. 2020 Jun;34(4):783-792. doi: 10.1080/02699931.2019.1682973. Epub 2019 Oct 25.