Transcript Processes of Change

Acceptance and Commitment Therapy:
A Transdiagnostic Model of Behavior
Change
Jason B. Luoma, Ph.D.,
Steven C. Hayes, Ph.D.
University of Nevada, Reno
Frank W. Bond, Ph.D.,
Goldsmiths College, University of London
Akihiko Masuda, M.A.,
University of Nevada, Reno
Why pay attention to transdiagnostic
processes of change?
• Without transdiagnostic processes of change, behavioral
technologies are likely to gather into an ever expanding pile with
no means for simplification or possibly advancement (there are
currently how many hundred DSM diagnoses…)
• It is unlikely that scientists and practitioners will be able to
maximize the efficacy of our interventions if the most proximal
psychological processes are not understood.
• If processes that cut across diagnostic categories can be found,
clinician training might be made more efficient and effective
DSM is based on topography
Topographically-defined clusters of behavior/symptoms
may not tell us much about etiology or maintenance of
these patterns
Behaviors that appear different in form may have similar
functions
Examples:
• Eating to avoid feeling lonely (BED/Bulimia?)
• Not getting out of bed to avoid feeling lonely (Depression?)
Common function? avoidance
Implication - Behaviors that look different may
actually be the same when viewed functionally.
DSM is based on:
The Assumption of Healthy Normality
•By their nature humans are psychologically
healthy
•Abnormality is a disease or syndrome driven
by unusual pathological processes
•We need to understand these processes and
change them
ACT: Human Suffering is
Ubiquitous and Normal
Lots of data - high rates of serious suicidal ideation,
high lifetime prevalence of “disorders”, prejudice,
divorce, abuse, etc.
Hypothesis: Normal human psychological processes,
particularly side effects of language, result in much
suffering (Hayes, Barnes-Holmes, Roche, 2001)
“Disorder-specific” processes can be exacerbated by
normal language processes
Example: ACT and Psychotic Symptoms




Can ACT help with what a “disorder specific”
pathological process?
Bach & Hayes (2002): 80 S’s hospitalized with
hallucinations and/or delusions randomized to
either ACT or TAU
4 hours of ACT; all but one session inpatient
Recently replicated by Gaudiano and Herbert
(2004) with similar results
Impact on Rehospitalization
1.0
ACT
.9
.8
.7
.6
Treatment as Usual
40
80
Days After Initial
Release
120
Processes of Change:
Symptoms
100
ACT
75
50
Control
25
Pre
Post
Phase
Processes of Change:
Believability
80
Control
60
ACT
40
Post
Pre
Phase
Relational Frame Theory
Stimulus Equivalence: An Example of
the Core Verbal Process
Lemon
Derived Function
(e.g., taste)
Very early on (<14 months old or so), a human will
begin to derive...
Function (e.g.,
taste)
These Three Relations Are the Basis
for Suffering
When frames of coordination (previous slide), time or
contingency, and comparative frames become part of a
person’s repertoire, problem solving is made possible, but
also:
• Comparison to an ideal
• Worry about imagined futures
• Social comparison / prejudice / stigma
• Self-loathing
• Social inhibition (e.g., fear of negative evaluation)
Because of Relational Frames
Self-knowledge of painful events is painful
The actual abuse
causes emotional
pain
Abuse
Later, just describing or
thinking about the event
causes emotional pain, so
thinking about it is
avoided
Description
Description
Implications of Relational Frames
At least two destructive processes result naturally from language:
Experiential Avoidance
•The tendency to attempt to alter the form, frequency, or situational
sensitivity of historically produced negative private experience
(emotions, thoughts, bodily sensations) even when attempts to do so
cause psychological and behavioral harm
•Based originally on natural processes of language but is amplified by
the culture
Cognitive Fusion/Literality
The domination of derived functions (i.e., those based on language) over
other response functions even when this process creates psychological
and behavioral harm
Two Processes Aimed at the Root Cause
Acceptance involves
•Encouraging the direct moment-to-moment contact with previously avoided
private events (that functionally need not be avoided) as they are directly
experienced to be, not as they “say they are”
•E.g., interoceptive exposure; Gestalt exercises; challenging a control agenda
Cognitive defusion involves
•a change in the normal use of language and cognition such that the ongoing
process of thinking is more evident and the normal functions of the products of
thinking are broadened.
•Similar to mindfulness techniques (as seen in MBCT, DBT) clients are taught to
observe thoughts without becoming entangled in them; a thought is understood, but
it is also heard as a sound, seen as a habit, or dispassionately observed as an
automatic verbal relation
ACT Outcomes to Date
 At least 31 completed studies (25 published),
including 11 randomized controlled trials
 Problems: pain, anxiety, psychosis, depression,
eating disorders, conduct disorder, prejudice,
substance abuse, smoking, stress, burnout, school
performance, stigma, OCD, diabetes
 Variable in lengths and emphases
 Always better than control; often has performed
better than active treatment comparators
ACT Mediational Results
• Diabetes - ACT compared to diabetes education - diabetes-related
acceptance shown to be a mediator of self-management behaviors
(Gregg, 2004)
• Smoking Cessation - ACT compared to nicotine patch smoking-related acceptance shown to be mediator of smoking
cessation outcomes (Gifford, Kohlenberg, Hayes et al., 2004)
• Workplace stress - ACT compared to Innovation Promotion and
waitlist - general acceptance (AAQ) mediated general mental health
outcomes (Bond & Bunce, 2000)
• Counselor Stigma and Burnout - ACT compared to multicultural
training and education - believability of stigmatizing thoughts
mediated outcomes on burnout and frequency of stigma (Hayes et al.,
2004).
Process of Change Outcomes
Believability of problem-relevant thoughts is reduced by ACT
•
•
•
•
depression (Zettle & Hayes, 1986)
psychosis (Bach & Hayes, 2003; Gaudiano, 2004)
polysubstance abuse (relative to control; Bissett, 2001)
counselor stigma and burnout (Hayes et al., 2004)
Acceptance is increased by ACT
•
•
•
•
•
•
•
chronic pain (McCracken, Vowles, & Eccleston, in press)
diabetes self-management (Gregg, 2004)
mathematics anxiety (Zettle, 2003)
parents of autistic children (Blackledge, 2004)
self-stigma in substance abuse (Kohlenberg, Luoma, et al., 2004)
smoking cessation (relative to control; Gifford et al., 2004)
workplace stress (Bond & Bunce, 2000)
Experimental Psychopathology
Studies
• Positive results comparing defusion vs. control instructions on
reducing discomfort and believability of negative self-relevant
thoughts (Masuda et al., 2004)
• in 2 cold pressor/1 analogue pain task experiments, individuals given
an acceptance-based rationale were able to tolerate higher levels of
pain than those given a control rationale (Gutierrez, Luciano, & Fink,
2004; Hayes et al., 1999; Takahashi et al., 2002)
• in 2 experiments studying tolerance of CO2 enriched air, participants
(normals/panic disordered) given an acceptance based rationale
reported less distress and were more willing to try the task again
(Eifert & Heffner, 2003; Levitt, Brown, Orsillo, & Barlow, 2004)
Meta analysis of Correlational Studies
• 21 studies with 51 correlations investigated the relationship between
the AAQ and quality of life (QOL) outcomes (e.g., depression, anxiety,
PTSD, trichotillomania, stress, pain, job performance, and negative
affectivity).
• The Q statistic indicated that the magnitude of these 51 associations
varied significantly. Subsequent analyses indicated that these
correlations could be separated into two groups, in each of which the
magnitude of the correlations was significantly similar or homogenous.
• Group 1: 26 correlations, with a total sample size of 6,024:
Medium size effect: aggregated correlation 0.28 (95% confidence
interval: 0.26 – 0.31).
• Group 2: 25 correlations, with a total sample size of 4,817:
Large size effect: aggregated correlation of 0.54 (95% confidence
interval: 0.52 – 0.56)