Linguistic Analysis of Communication in a Therapist- Assisted Internet-Based Cognitive Behaviour Therapy Program for Individuals with Generalized Anxiety Disorder

Abstract

Background: Therapist-assisted Internet-based Cognitive Behaviour Therapy (TAICBT)

is an empirically supported method of treating generalized anxiety disorder (GAD). In

addition to the cognitive and behavioural elements found within face-to-face Cognitive

Behaviour Therapy (CBT), TAICBT includes elements of expressive writing, through

email communication with a therapist. Expressive writing has been associated with

positive outcomes; furthermore, certain linguistic dimensions in expressive writing have

been associated with psychological and physical health benefits, such as less frequent

use of negative emotion words, and greater use of positive emotion words. No research

to date has analyzed linguistic dimensions in client communication over the course of a

TAICBT program. Objective: The purpose of the present study was to understand how

clients with symptoms of GAD are communicating with their online therapists, and

explore whether their communication style was related to successful participation.

Research Questions: Six research questions were posed for analyses: Is there a

relationship between client online communication (i.e., word count and percentage of

specific word categories) and symptom measures (e.g., general anxiety, work and social

adjustment)? Does communication change over the course of 12 modules? Do

participants who have completed all 12 modules, compared to those that completed

fewer than 12 modules, differ in baseline communications submitted to their therapists?

Do program completers who are classified as symptomatic versus asymptomatic on the

GAD-7 after program completion differ in overall communication submitted to their

therapists? For program completers, do change scores on symptom measures from preto

post-treatment correlate with overall communication? For program completers, do

scores on therapeutic alliance, treatment satisfaction, and comfort with emotional expressivity correlate with overall communication? Method: Data was obtained from a

previously-conducted study. Specifically, written communication submitted to therapists

by 107 adult clients with symptoms of GAD during the course of a 12 module TAICBT

program was analyzed along with generalized anxiety, worry, and work and social

adjustment scores completed pre-, mid- and post-treatment. Analyses: Clients’ written

communication was screened for typographical errors and analyzed by the Linguistic

Inquiry and Word Count Software (Pennebaker, Chung, Ireland, Gonzales, & Booth,

2007). Results: Significant relationships emerged between negative emotion words,

anxiety words, insight words, and past tense words, with symptom measures of anxiety

and worry; these categories also changed over time. A higher number of negative

emotion and anxiety words differentiated clients that dropped out of the program earlier

versus those that dropped later, and demonstrated a negative relationship with

therapeutic alliance and treatment satisfaction. Discussion: Results suggest there is

utility in paying attention to linguistic categories as they can be used as indicators of

client progress and well-being. Findings improve our understanding of the client

experience of participating in TAICBT, with some linguistic trends being consistent with

what clinicians would expect to see from a client's successful participation in therapy.

Other findings offer new insight into communication that may be particularly relevant to

clients with symptoms of GAD. Findings confirm that clients are engaging with the

program and opening up to their therapists. In addition to generating new knowledge

about the process of TAICBT, the potential mechanisms of change, and potential

predictors of success within TAICBT, the current findings could also facilitate

improvement in the way TAICBT is delivered in the future.