In the past half a century psychotherapy research has blossomed, with thousands of studies confirming its positive effects for a wide array of clinical problems, including depression, anxiety, eating disorders and sexual dysfunction. Yet in recent years, intense controversy over whether and how to put these findings into practice has erupted, further widening the "scientist-practitioner gap," the deep gulf that has separated many researchers and psychotherapists for decades.
The current debate centers on the growing use of empirically supported therapies, or ESTs, which are specific therapies for specific problems--for example, depression and bulimia--that meet certain criteria (such as a given number of well-designed studies showing positive effects) for treatment efficacy. Proponents have welcomed ESTs for their clear guidelines on what works for patients and their explicit manuals prescribing administration of treatment. Critics have sharply questioned ESTs on a number of grounds, namely, whether their research base is adequate, whether their one-size-fits-all approach can address the needs of individual patients, and whether their focus should be primarily alleviation of symptomatic distress or changes in underlying dispositions and vulnerabilities.
The debate's resolution bears important implications for treatments that psychotherapy patients seek and receive. A survey of nearly 10,000 adults published in 2005 showed that one out of four Americans meets the criteria for a diagnosis of a psychological disorder in any given year and that slightly less than half of all people in the U.S. will suffer from a psychological disorder over the course of their lifetimes [see "Half Are Mentally Ill," by Jamie Talan, Head Lines; Scientific American Mind, Vol. 16, No. 3; 2005].
Before we wrote this article, one of us (Arkowitz) had been highly critical of ESTs (though not of placing psychotherapy on a more scientific basis). The other one of us (Lilienfeld) had been a strong advocate of ESTs. Ultimately we found considerable common ground on many points regarding the proper role of research in informing clinical practice. In this feature, we hope to offer a modest step toward reconciling opposing views on ESTs.
Laying the Groundwork
Fifty years ago the foundations of modern psychotherapy research were just being laid. One participant at a 1950 conference was being only partially facetious when he commented: "Psychotherapy is an undefined technique applied to unspecified problems with unpredictable outcomes. For this technique we recommend rigorous training."
Just two years later an eminent British psychologist named Hans Eysenck questioned the scientific basis of talk therapy in a landmark paper--asserting that it was no more effective than the absence of treatment. Researchers soon rose to Eysenck's challenge, and thousands of studies over the ensuing decades demonstrated conclusively that psychotherapy does help many patients. But which are the most effective therapies and for which problems? Further studies sought answers.
In 1995 a task force of a division of the American Psychological Association (APA), chaired by Boston University psychologist David H. Barlow, issued the first of several reports that set forth initial criteria for ESTs, along with lists of therapies that met those criteria. The current task force list is widely used today, especially in university settings in which future clinical psychologists are educated [see box on page 47].
We should note that the list tells only whether a treatment has been found to work in controlled studies but not necessarily in clinical practice outside the laboratory. Most experiments have examined cognitive-behavioral therapy; psychoanalytic, humanistic and integrative methods have received less research attention [see box on page 48]. If a treatment is absent from the list, it means one of two things: either studies have shown that the treatment does not work, or it has not been tested and, therefore, we do not know whether or not it works. Most of the more than 500 "brands" of psychotherapy are not on the EST list, because they fall in the second category.
The Case for ESTs
Advocates have advanced three major arguments in favor of a list of efficacious therapies for specific disorders: it protects patients against fringe psychotherapies, it empowers mental health consumers to make appropriate choices for their care, and it aids in training future therapists.
First, in recent years consumers have been beset by a seemingly endless parade of fad therapies of various stripes [see box on page 49]. Despite scant scientific support--or sometimes outright debunking--some fringe treatments continue to be used widely. For example, surveys of doctoral-level therapists in the 1990s indicated that about one quarter regularly employed two or more recovered-memory techniques. Facilitated communication, discredited by scientific research in the 1990s, is still popular in some communities. Counselors who administer crisis debriefing number in the thousands; in the aftermath of the September 11 terrorist attacks, one crisis-debriefing outfit in Atlanta alone dispatched therapists to 200 companies. All these treatments have been found to be ineffective or even harmful. Some studies have discovered that crisis debriefing, for example, increased the risk of post-traumatic stress disorder in trauma-exposed individuals. The EST list makes it harder for practitioners who administer these and other questionable techniques to claim that they are operating scientifically.
Second, the EST list benefits patients because by providing them with information regarding which treatments have been proven to work, it puts them in a better position to make good choices for their care. Like the Food and Drug Administration's list of approved medications, the EST list performs a quality-control function. It serves a similar purpose for managed care organizations and health care agencies, which want to make scientifically informed decisions about which treatments should--and should not--be reimbursed. By placing the burden of proof on a treatment's proponents to show that it is efficacious, the EST list helps to ensure that therapies promoted to the general public have met basic standards.
Third, the EST list can improve the education and training of graduate students in clinical psychology, social work and other mental health fields. The sprawling psychotherapy research literature is often confusing and contradictory; without such a list, novice clinicians have no clear research guidance concerning which treatments to administer and which to avoid.
The Case against ESTs
Critics have responded with four concerns: EST research findings may not apply to psychotherapy as practiced in the "real world;" the list may be biased toward cognitive-behavioral therapies; the EST view of psychotherapy is narrow; and techniques emphasized by such lists may not be the key ingredients of therapeutic change.
First, critics have attacked ESTs for both the science underlying their "empirical support" and their applicability to clinical practice. "The move to worship at the altar of these scientific treatments has been destructive to clients in practice, because the methods tell you very little about how to read the real and complex people who actually come in for therapy," said psychiatrist Glen O. Gabbard of the Baylor College of Medicine in a 2004 New York Times article.
To satisfy requirements for good research, which seeks to eliminate any variables that could confound the results, investigators must sacrifice a great deal of what practicing psychotherapists believe is important. EST manuals often sharply constrain therapists' flexibility to tailor the treatments to clients' needs, resulting in a one-size-fits-all approach. Researchers reject up to 90 percent of subjects who are initially recruited, in the name of ensuring a "pure" group with the diagnosis of interest. As a result, participants in these studies typically represent only a small percentage of those who might be seen in actual practice.
The all-or-none nature of the EST list also has been criticized. By categorizing treatments as either empirically supported or not, the list omits potentially useful information, such as the degree of efficacy of different EST therapies. Further, many of the ESTs have modest or even relatively weak effects. That is, they leave many clients slightly improved or not helped at all, with a high likelihood of relapse. Is it reasonable to call such therapies "empirically supported?"
In 2001 psychotherapy researchers Drew Westen, now at Emory University, and Catherine M. Novotny, now at the Department of Veterans Affairs Medical Center in San Francisco, published an analysis of a large number of efficacy studies for depression and some anxiety disorders. Most of the therapies they examined were variants of cognitive-behavioral therapy. Their findings revealed a glass that is both half-full and half-empty. On the positive side, they learned that 51 percent of depressed clients and 63 percent of those with panic disorder were significantly better or no longer had symptoms. But the glass seems emptier if we recognize that many patients who had improved still exhibited symptoms at the end of treatment and that others were not helped at all. If we include people who dropped out of therapy, the success percentages plunge considerably. In addition, follow-up studies reveal high rates of relapse. For example, only 37 percent of those depressed clients who completed treatment remained improved one to two years later.
Second, some critics have argued that EST therapies are biased in favor of cognitive-behavioral techniques. Reviews of research on psychoanalytic and humanistic therapies suggest positive effects broadly comparable to those of cognitive-behavioral therapies. Although less research has been conducted on these therapies than on cognitive-behavioral therapy, their underrepresentation on EST lists raises questions of bias.
Third, ESTs focus almost exclusively on symptoms and distress to the exclusion of other important factors that lead people to seek therapy. These considerations include predispositions, vulnerabilities and personality characteristics that often persist after the symptoms are gone. Many psychotherapists believe that it is important to focus on these types of problems in therapy, in order to enhance the quality of the client's life and help reduce the chances of a relapse. The emphasis of ESTs on standardized techniques similarly ignores not only the uniqueness of individuals but also the salutary power of the therapist-client relationship.
Fourth, the techniques emphasized by the EST list may not be what produces change in many cases. Most studies comparing the efficacy of two or more therapies find that they all do about equally well. This surprising result is termed the "Dodo Bird verdict," after the Dodo Bird in Alice's Adventures in Wonderland, who declares (following a race) that "everybody has won and all must have prizes." Psychotherapy researchers intensely debate the meaning of the Dodo Bird verdict. Some argue that actual important differences exist among therapies but that problems with study design have masked them. Such problems include small samples and the limited range of therapies that have been compared. It is also possible that although average outcomes of various therapies may not differ, some clients may do better with one therapy, whereas other clients may do better with another.
Still other researchers have accepted the Dodo Bird verdict and attempted to account for it. One explanation suggests that therapeutic change is caused more by "common factors" that therapies share rather than by specific techniques. Such factors include instilling hope and providing a believable theoretical rationale with associated therapeutic "rituals," which can make clients feel that they are taking positive action to solve their problems. This perspective also emphasizes the healing power of the therapist-patient relationship.
The EST movement has succeeded in placing the importance of evidence-based practice squarely on the agenda of clinical psychology. Because EST lists have many inherent problems, however, they may prove more useful as a catalyst for helping the field move toward scientifically informed practice than they will be as the final word.
Several promising proposals recently have attempted to refine or replace ESTs in ways that retain their emphasis on science-based practice. One comes from the work of University of New Mexico psychologist William R. Miller. Miller constructed a list of all researched therapies for alcoholism, ranking them by the quality of the research and magnitude of the effects. His method provides access to all relevant information about all therapies studied, not just those that meet the all-or-none criteria for inclusion on the EST list.
Others have suggested that we seek empirically based "principles of change" rather than empirically supported therapies. For example, repeated exposure to feared objects and events is a central principle underlying most effective treatments for anxiety disorders. Therapists can derive many ways of flexibly implementing a principle of change to fit clients without being constrained by a specific technique or manual. In a similar vein, others have recently suggested that we focus on "empirically supported relationship factors," such as therapist empathy and warmth. But there is not yet sufficient agreement concerning which change or relationship principles should qualify as empirically supported.
Another alternative to ESTs was proposed by a committee appointed by past APA president Ronald F. Levant. The concept, which is called evidence-based practice, has been widely embraced in many areas of medicine. In its 2005 policy statement, the APA committee defined evidence-based practice as "the integration of the best available research with clinical expertise in the context of patient characteristics, culture, and preferences."
The term "best available research" is much broader than evidence based on psychotherapy studies alone. It encompasses research across the entire field of psychology, including personality, psychopathology and social psychology. "Clinical expertise" relates to therapist competencies that are not tied directly to research but that are believed to promote positive therapeutic outcomes. These capabilities inform the ability to form therapeutic relationships with clients and to devise and implement treatment plans. Finally, inclusion of client characteristics, culture and preferences points to the importance of tailoring treatments to individuals.
Although this APA report is a noble effort to grapple with some of the controversies, its long-term impact remains unclear. Many EST proponents have been dissatisfied with the recommendation to employ "the best available research" as being so vague, at least compared with the specificity of ESTs, as to be of little value. Many EST advocates have also objected to the inclusion of clinical expertise in a definition of evidence-based practice.
Given the shortcomings of ESTs and the existing alternatives to them, it is clear that the field is just beginning to incorporate science-based practice. Nevertheless, we can begin to see the broad outlines of promising positions that are less dogmatic than earlier ones. Such trends may help assuage the legitimate concerns of both researchers and practitioners. Ultimately we believe that the field must move beyond a narrow definition of ESTs toward views that bridge the gap between researchers and practitioners. After all, whatever their differences may be, aren't all clinical psychologists seeking better ways to help troubled people feel happier and live enriching lives?
HAL ARKOWITZ and SCOTT O. LILIENFELD hope to bring some insights to the contentious discussion surrounding empirically supported therapies. Arkowitz, associate professor of psychology at the University of Arizona, has served as editor of the Journal of Psychotherapy Integration. He has received two awards from the Arizona State Psychological Association for distinguished contributions to the practice of psychology and distinguished contributions to the science of psychology. Lilienfeld, associate professor in the department of psychology at Emory University, is former president of the Society for a Science of Clinical Psychology and editor of Scientific Review of Mental Health Practice.
- The Great Psychotherapy Debate: Models, Methods, and Findings. Bruce E. Wampold. Lawrence Erlbaum Associates, 2001.
- Evidence-Based Practices in Mental Health: Debate and Dialogue on the Fundamental Questions. Edited by John C. Norcross, L. E. Beutler and R. F. Levant. American Psychological Association Press, 2005.