CBT and Interpretation Bias Modification for Generalized Anxiety Disorder: Examining the Roles of Intolerance of Uncertainty and Interpretation Bias in Symptom Reduction Eleanor Donegan A Thesis In the Department of Psychology Presented in Partial Fulfillment of the Requirements For the Degree of Doctor of Philosophy (Psychology) at Concordia University Montréal, Québec, Canada November 2016 © Eleanor Donegan, 2016 CONCORDIA UNIVERSITY School of Graduate Studies This is to certify that the thesis prepared By: Eleanor Donegan Entitled: CBT and Interpretation Bias Modification for Generalized Anxiety Disorder: Examining the Roles of Intolerance of Uncertainty and Interpretation Bias in Symptom Reduction and submitted in partial fulfilment of the requirements for the degree of Doctor of Philosophy (Psychology) complies with the regulations of the University and meets the accepted standards with respect to originality and quality. Signed by the final examining committee: ________________________________ Chair Dr. Virginia Penhune ________________________________ External Examiner Dr. Bethany Teachman ________________________________ External to Program Dr. Sylvia Kairouz ________________________________ Examiner Dr. Jean-Philippe Gouin ________________________________ Co-Supervisor Dr. Adam Radomsky ________________________________ Co-Supervisor Dr. Michel Dugas Approved by: __________________________________________ Dr. Karen Li, Graduate Program Director _________________, 2016 ___________________________________________ Dr. André Roy, Dean of Faculty of Arts and Sciences ABSTRACT CBT and Interpretation Bias Modification for Generalized Anxiety Disorder: Examining the Roles of Intolerance of Uncertainty and Interpretation Bias in Symptom Reduction Eleanor Donegan, Ph.D. Concordia University, 2016 Generalized Anxiety Disorder (GAD) is characterized by excessive worry and anxiety (APA, 2013). Although cognitive behaviour therapy (CBT) is efficacious, 20-50% of individuals with GAD continue to meet diagnostic criteria following treatment (Hanrahan et al., 2013). To improve outcomes, it is essential that we develop a better understanding of the factors involved in symptom reduction and ensure that these factors are targeted effectively. Two factors that are associated with excessive worry and anxiety are intolerance of uncertainty (IU) and negative interpretation bias (Ladouceur et al., 2000; Rosen & Knaüper, 2009), both of which have been proposed to play a role in the maintenance of GAD symptoms (Hayes & Hirsch, 2007; Koerner & Dugas, 2006). In this program of research, the first goal was to examine the impact of an IU- focused CBT (Dugas & Ladouceur, 2000) on IU and negative interpretation bias and to determine whether these factors played a role in symptom reduction. In Study 1, 80 adults completed CBT for GAD. By post-treatment, CBT was associated with reductions in GAD symptoms, IU and negative interpretation bias. Moreover, reductions in negative interpretation bias predicted reductions in GAD symptoms and this effect was partially mediated by reductions in IU. Cognitive bias modification programs (CBM-I) have also been developed to target interpretation bias, primarily among socially anxious individuals, and have been proposed as low-cost alternatives to CBT (Amir & Taylor, 2012). The second goal in this program of research was to validate a new CBM-I program designed to target interpretation bias in GAD worry domains. In Study 2, participants who completed CBM-I (n = 16) exhibited greater reductions in negative interpretation bias than participants in an interpretation control condition (n = 14). However, CBM-I training did not lead to anticipated reductions in worry or anxiety. Overall, this program of research provided further support for an IU-focused CBT and insight iii into change processes during treatment. Although the CBM-I program examined here cannot yet be recommended as a stand-alone intervention, other clinical uses of CBM-I are discussed, including the possibility of implementing CBM-I as an adjunct intervention to enhance the efficacy of CBT. iv ACKNOWLEDGEMENTS The studies described in this dissertation were conducted with financial support from a grant from the Canadian Institutes of Health Research, as well as doctoral research awards from Les Fonds de la Recherche en Santé du Québec and Concordia University. I have been extremely grateful for this financial support which made the research described here possible. This dissertation would also not have been possible without the help and support of a number of people. I would like to thank my doctoral thesis supervisor, Dr. Michel Dugas, for his guidance over the years. Like most clinical programs in psychology, our program is designed to encourage a scientist-practitioner approach and I have appreciated the way in which he has embodied and modelled this approach in both his work as a researcher and as a clinician. I am also grateful to Dr. Adam Radomsky, who welcomed me into his laboratory and agreed to act as co-supervisor on my dissertation. In addition, I would like to thank Jean-Philippe Gouin, my internal committee member, for his feedback on my proposed program of research. I am extremely grateful to the members of the Anxiety Disorders Laboratory at Concordia University, including Kristin Anderson, Sonya Deschênes, Elizabeth Hebert, and former members Avital Ogniewicz and Kathryn Sexton. During my years of graduate training, I had the opportunity to work alongside these enthusiastic, intelligent, creative and often hilarious women who provided helpful feedback, encouragement, laughter and friendship. Our discussions over the years during potlucks, dinners, book clubs and lab hallway conversations were a highlight of my time as a graduate student. I would also like to thank the undergraduate thesis students and volunteers at the Anxiety Disorders Laboratory, as well as Amélie Cossette, Claudie Bax-D’Auteuil and the research assistants at l’Université de Sherbrooke and l’Université du Québec à Trois-Rivières who helped with components of my dissertation studies. Finally, I would like to thank the members of my family for their unconditional support and encouragement. Your unfailing generosity, advice, patience and faith in me made this work possible and I am truly grateful for it. v CONTRIBUTION OF AUTHORS The following thesis comprises two manuscripts: Study 1 (Chapter 2) Donegan, E., Dugas, M. J., Turcotte, J., Dao, T-V., & Savard, P. (2016). Cognitive predictors of symptom change during CBT for GAD: Examining the role of intolerance of uncertainty and interpretation bias. Manuscript in preparation for publication. Study 2 (Chapter 4) Donegan, E., & Dugas, M.J. (2016). Validation of a Multi-Session Cognitive Bias Modification (CBM-I) Training Program among Individuals with Elevated Worry and Anxiety. Manuscript in preparation for publication. I was responsible for choosing the overall focus of this program of research, as well as the focus of the specific studies included here, in consultation with Dr. Michel Dugas. The analyses presented in Study1 formed part of a larger 5-year clinical trial of CBT for GAD at l’Hôpital du Sacré-Coeur de Montréal (PI Dr. Michel Dugas). I was principally responsible for the selection of the specific research questions within this larger clinical trial, selection of relevant measures, statistical analyses, interpretation of findings, and manuscript preparation. Day-to-day coordination of clinical activities was performed by Céline Doucet and treatment was provided by a team of hospital-based clinicians, including Julie Turcotte, Isabelle Geninet, Thu- Van Dao, Pascale Harvey, and Pierre Savard. In Study 2, I was principally responsible for all aspects of study design and implementation, including selection of stimuli, coordination of computer programming, participant recruitment (via the Psychology Participant Pool at Concordia University and poster advertisements in Montreal and surrounding areas), participant scheduling, data collection, statistical analyses, interpretation of findings, and manuscript preparation. I also provided training and supervision to the research assistants who helped with data collection, including Amélie Cossette (l’Université du Québec à Trois-Rivières) and Claudie Bax-D’Auteuil (l’Université de Sherbrooke). My committee members, including Dr. Adam Radomsky and Dr. Jean-Philippe Gouin, provided recommendations and approved my program of research during a dissertation proposal meeting. Additional feedback on study design and statistical analyses was also provided by Dr. Roisin O’Conner during my dissertation proposal meeting. vi TABLE OF CONTENTS List of Figures ……………………………………………………….................. viii List of Tables ……………………………………………………….................. ix Chapter 1 General Introduction………………………………………........ 1 Chapter 2 Cognitive predictors of symptom change during CBT for GAD: Examining the role of intolerance of uncertainty and interpretation bias ……………………………………………… 10 Method…………………………………………………………. 13 Results………………………………………………………….. 17 Discussion……………………………………………………… 24 Chapter 3 Bridge………………………………………………………….. 27 Chapter 4 Validation of a Multi-Session Cognitive Bias Modification (CBM-I) Training Program among Individuals with Elevated Worry and Anxiety ……………………………………………. 29 Method…………………………………………………………. 32 Results………………………………………………………….. 41 Discussion……………………………………………………… 48 Chapter 5 General Discussion…………………………………………….. 54 References ……………………………………………………….................. 66 Appendix A Advertisement for Participant Recruitment (Study 1)…………. 80 Appendix B Information and Consent for Participation (Study 1)………….. 81 Appendix C Advertisement for Participation Recruitment (Study 2)……….. 88 Appendix D Information and Consent for Participation (Study 2)..…………. 89 Appendix E Computerized CBM-I/ICC Instructions (Study 2) .……………. 91 Appendix F Sample WSAP CBM-I and ICC Word and Sentence Pairs (Study 2) 92 Appendix G Scrambled Sentence Task Instructions (Study 2).....……………. 96 Appendix H Sample Scrambled Sentence Task Items (Study 2)...……………. 98 vii LIST OF FIGURES Figure 1 Mediation models (Study 1)……………………………………… 21 Figure 2 Participant flow through study (Study 2)........................………… 38 Figure 3 Sample trial in the CBM-I condition (Study 2)........................…... 40 viii LIST OF TABLES Table 1 Pre and Post-treatment Means in Threat Interpretations, Intolerance of Uncertainty and Symptoms during CBT for GAD (N = 80) …………………..…………………………… 19 Table 2 Correlations between Residualized Change Scores of Threat Interpretations, Intolerance of Uncertainty, and Symptoms during CBT for GAD (N = 80)……………………..……………. 20 Table 3 Summary of Mediation Analyses during CBT for GAD (N = 80) 23 Table 4 Comparison of Pre-Training Sample Characteristics in the CBM-I (n = 16) and ICC (n = 14) Conditions …………………. 42 Table 5 Means and Standard Deviations of Cognitive and Symptom Variables in a Completers-Only Sample .………………………. 44 Table 6 Mixed Effects ANOVA Analyses: Comparing Change in CBM-I (n = 15) and ICC (n = 14) Conditions from Pre- to Post-Training……………………………….…………………… 45 ix 1 CHAPTER 1 GENERAL INTRODUCTION Generalized Anxiety Disorder (GAD) first appeared in the anxiety disorders section of the Diagnostic and Statistical Manual of Mental Disorders for DSM-III in 1980 (American Psychiatric Association, 1980). Initially described as a relatively mild condition, epidemiological studies in the 1990’s began to provide evidence that GAD was in fact associated with considerable impairment (Wittchen, Zhao, Kessler, & Eaton, 1994). GAD is now recognized as a chronic condition characterized by excessive and uncontrollable worry and anxiety, as well as associated symptoms (e.g., restlessness, muscle tension) (APA, 2013). In addition to being one of the most commonly-diagnosed anxiety disorders, with lifetime prevalence rates ranging from 4% to 6%, GAD is also often co-morbid with other anxiety and mood disorders (Kessler et al., 2005). Even when not co-morbid with other conditions, the level of impairment associated with GAD rivals that of Major Depressive Disorder (Kessler, Dupont, Berglund, & Wittchen, 1999). Cognitive behaviour therapy (CBT) has been recognized as the gold standard psychological treatment for anxiety disorders for at least the past two decades, with meta- analyses indicating that CBT is associated with moderate-to-large effect sizes (e.g., Butler, Chapman, Forman, & Beck, 2006; Hofmann & Smits, 2008). CBT protocols designed to target GAD-specific vulnerabilities have also been demonstrated to be effective. Early studies found, for instance, that CBT was more effective for GAD than no treatment or non-directive therapy (Borkovec & Costello, 1993; Gould, Otto, Pollack, & Yap, 1997) and was associated with mean pre-to-post effect sizes in at least the moderate range (e.g., ESμ = 0.70; Gould et al., 1997). The benefits of CBT were also found to be long-lasting, with gains being maintained up to 12 months following treatment (Borkovec & Costello, 1993; Gould et al., 1997). More recent meta-analyses suggest that gains may be maintained up to 8 or even 10 years following treatment (e.g., Durham, Chambers, MacDonald, Power, & Major, 2003). Despite this treatment success, however, optimism about the overall utility of CBT protocols for GAD has been somewhat muted by the finding that 20% to 50% of individuals continue to meet diagnostic criteria for GAD following treatment (e.g., Fisher, 2006; Hanrahan, Field, Jones, & Davey, 2013). Moreover, the proportion of individuals who experience significant residual symptoms following CBT has been larger than for other anxiety disorders (Gould, Safren, O’Neill, Washington, &