FOCUS ON RESEARCH Which Therapies Work Best for Which Problems

519
Evaluating Psychotherapy
■ What conclusions are most reasonable?
Statistical analyses show that various treatment approaches appear about equally effective overall. But this does not mean that every psychotherapy experience will be equally
helpful. Potential clients must realize that the success of their treatment can still be
affected by how severe their problems are, by the quality of the relationship they form
with a therapist, by their motivation to change, and by the appropriateness of the therapy methods chosen for their problems (Goldfried & Davila, 2005).
Like those seeking treatment, many clinical psychologists, too, are eager for more
specific scientific evidence about the effectiveness of particular therapies for particular
kinds of clients and disorders. These empirically oriented clinicians are concerned that,
all too often, therapists’ choice of therapy methods depends more heavily on personal
preferences or current trends than on scientific evidence of effectiveness (Lynn,
Lilienfeld, & Lohr, 2003; Nathan, Stuart, & Dolan, 2000; Norcross, Beutler, & Levant,
2005; Tavris, 2003). They believe that advocates of any treatment—whether it is object
relations therapy or systematic desensitization—must demonstrate that its benefits are
the result of the treatment itself and not just of the passage of time, the effects of
repeated assessment, the client’s motivation and personal characteristics, or other confounding factors (Chambless & Hollon, 1998). A movement toward the same kind of
evidence-based practice has also appeared in the medical and dental professions (Borry,
Schotsmans, & Dierickx, 2006; Niederman & Richards, 2005).
Empirically oriented clinicians also want to see evidence that the benefits of treatment are clinically significant. To be clinically significant, therapeutic changes must be
great enough to make treated clients’ feelings and actions similar to those of people
who have not experienced these clients’ disorders (Kendall, 1999; Kendall & Sheldrick,
2000). The need to demonstrate clinical significance has become more important than
ever as increasingly cost-conscious clients—and their health insurance companies—
decide whether, and how much, to pay for various psychotherapy services (Levant,
2005; Makeover, 2004; Nelson & Steele, 2006). The most scientific way to evaluate
treatment effects is through experiments in which clients are randomly assigned to various treatments or control conditions and their progress is objectively measured.
T
o help clinicians select treatment methods on the basis of this kind of empirical evidence, the American Psychological
Association’s Division of Clinical Psychology
created a task force on effective psychotherapies (Task Force on Promotion and Dissemination of Psychological Procedures, 1995).
FOCUS ON RESEARCH
Which Therapies Work Best
for Which Problems?
■ What was the researchers’ question?
The question addressed by this task force was, “What therapies have proven themselves
most effective in treating various kinds of psychological disorders?”
■ How did the researchers answer the question?
Working with other empirically oriented clinical psychologists, members of this task
force examined the outcomes of thousands of well-controlled experiments that evaluated psychotherapy methods used to treat mental disorder, marital distress, and healthrelated behavior problems in children, adolescents, and adults (Baucom et al., 1998;
Chambless & Ollendick, 2001; Compas et al., 1998; DeRubeis & Crits-Christoph, 1998;
Foley, 2004; Kazdin & Weisz, 1998; Kendall & Chambless, 1998).
520
TA B L E
Chapter 13 Treatment of Psychological Disorders
13.4
Examples of Empirically Supported Therapies for Selected Disorders
Treatments listed as “efficacious and specific” (pronounced “effeh-KAY-shus”) were shown to be superior to no treatment or to some alternative
treatment in at least two experiments in which clients were randomly assigned to various treatment conditions. These experiments are called
randomized clinical trials, or RCTs. Also included in this category are treatments supported by scientific outcome measures from a large number
of carefully conducted case studies. Treatments listed as “probably efficacious” are supported by at least one RCT or by a smaller number of rigorously evaluated case studies. Those listed as “possibly efficacious” are supported by a mixture of data, generally from single-case studies or
other nonexperimental studies (Chambless & Ollendick, 2001).
Problem
Efficacious and Specific
Probably Efficacious
Major depressive disorder
Behavior therapy
Cognitive-behavior therapy
Interpersonal therapy
Brief dynamic therapy
Social problem solving
Self-control therapy
Specific phobia
Exposure therapy
Systematic desensitization
Agoraphobia/panic disorder
Cognitive-behavior
therapy
Couples training exposure therapy
Generalized anxiety disorder
Cognitive-behavior therapy
Applied relaxation therapy
Obsessive-compulsive
disorder
Exposure therapy response
prevention
Cognitive therapy
Family-assisted exposure
therapy response
prevention relaxation
Rational emotive
behavior therapy exposure therapy
Exposure
Stress inoculation
Cognitive therapy stress
inoculation exposure
Structured
psychodynamic
treatment
Cognitive therapy
(for delusions)
Posttraumatic stress
disorder
Schizophrenia
Behavioral family therapy
Family systems therapy
Social skills training
Supportive group therapy
Alcohol abuse and
dependence
Community reinforcement
Cue exposure therapy
Behavioral marital therapy anti-alcohol drug, disulfiram
Social skills training (with inpatients)
Opiate abuse and
dependence
Marital discord
Possibly Efficacious
Behavior therapy
Brief dynamic therapy
Cognitive therapy
Behavioral marital therapy
Insight-oriented marital therapy
Source: Chambless & Ollendick (2001).
■ What did the researchers find?
empirically supported therapies (ESTs)
Treatments for psychological disorders
whose effectiveness has been validated
by controlled experimental research.
The task force found that a number of treatments—known as empirically supported
therapies or ESTs—have been validated by controlled experimental research (Chambless
& Ollendick, 2001; DeRubeis & Crits-Christoph, 1998; Kendall & Chambless, 1998). Table
13.4 contains some examples of these therapies. Notice that the treatments identified as
effective for particular problems in adult clients are mainly behavioral, cognitive, and cognitive-behavioral methods, but that certain psychodynamic therapies (e.g., interpersonal
521
Evaluating Psychotherapy
therapy and brief dynamic therapy) also made the list (Chambless & Ollendick, 2001;
Svartberg, Stiles, & Seltzer, 2004).
■ What do the results mean?
The authors of the report on empirically supported therapies and those who support
their efforts claim that by relying on analysis of experimental research they have scientifically evaluated various treatments and generated a list of methods from which clinicians and consumers can choose with confidence when facing specific disorders (e.g.,
Hunsley & Rumstein-McKean, 1999; Kendall & Chambless, 1998). Therapists are even
being urged to follow treatment manuals stemming from this research to help them
deliver empirically supported therapies exactly as they were intended (Addis, 1997;
Wade, Treat, & Stuart, 1998). The state of Kansas has been the first in the United States
to formally encourage the use of empirically supported treatments for children
(Roberts, 2002).
Not everyone agrees with the conclusions and recommendations of the APA Task
Force. Critics note, first, that treatments missing from the latest list of ESTs haven’t
necessarily been discredited. Some of those treatments might not yet have been studied or validated according to the efficacy criteria selected by the task force. These critics also have doubts about the value of some of those criteria. They point to research
showing that had the task force used different outcome criteria, it might have
reached different—and perhaps less optimistic—conclusions about the value of some
empirically supported treatments (Bradley et al., 2005; Thompson-Brenner, Glass, &
Westen, 2003). There is concern, too, about the wisdom of categorizing treatments as
either “supported” or “unsupported.” These simple either-or judgments seem reassuring, but may fail to give a complete picture of the impact of various treatments
on various clients with various problems (Westen & Bradley, 2005). Critics argue further that the list of empirically supported therapies is based on research that may not
be relevant to clinicians working in the real world of clinical practice. They note that
experimental studies of psychotherapy have focused mainly on relatively brief treatments for highly specific disorders, even though most clients’ problems tend to be far
more complex (Westen & Bradley, 2005). These studies focus, too, on the therapeutic procedures used rather than on the characteristics and interactions of therapists
and clients (Cornelius-White, 2002; Garfield, 1998; Hilliard, Henry, & Strupp, 2000;
Westen et al., 2004). This emphasis on procedure is a problem, critics say, because
the outcome of therapy in these experiments might have been strongly affected by
client-therapist factors, such as whether the random pairing of clients and therapists
resulted in a match or a mismatch on certain personal characteristics. In real clinical situations, clients and therapists are not paired up at random (Goldfried & Davila,
2005; Hill, 2005; Hohman & Shear, 2002). Finally, because therapists participating in
experimental research were required to follow standard treatment manuals, they were
not free to adapt treatment methods, as they normally would, to the needs of particular clients (Garfield, 1998). Perhaps, say critics, when there is less experimental
control over the treatment situation, all therapies really are about equally effective,
as suggested by the statistical analyses of outcome research we mentioned earlier
(Shadish et al., 2000; Smith et al., 1980).
In short, critics reject the empirically supported therapies list as a useful guide. In
fact, some see it as an incomplete and ultimately misleading document (Westen &
Bradley, 2005). They also worry that widespread use of treatment manuals would make
psychotherapy mechanical and less effective and would discourage therapists’ from
developing new treatment methods (Addis & Krasnow, 2000; Beutler, 2000; Garfield,
1998).
■ What do we still need to know?
The effort to identify empirically supported treatments and to develop evidencebased practice in clinical psychology (Levant, 2005) represent important steps in
responding to Paul’s (1969) “ultimate question” about psychotherapy: “What