COMMON FACTORS IN PSYCHOTHERAPY

Jacqueline A. Sparks

Barry L. Duncan

Scott D. Miller

The great tragedy of science—the slaying of a beautiful hypothesis by an ugly fact.

—Thomas Henry Huxley,

Presidential Address to the British

Association for the Advancement of Science

This chapter addresses the efforts of researchers and clinicians alike to answer this question. Covering historical and empirical foundations, we explore the evolution of common factors as a metatheoretical framework for research and the emergence of transformative clinical practice. By definition, a common factors framework is not a model, psychotherapy, or specific set of techniques. As such, it cannot be “manualized” but informs the immediate therapeutic encounter one client at a time. We incorporate this distinction throughout while exploring past and present interpretations and clinical applications. Specifically, we discuss how this insight has inspired client-directed, outcome-informed practice (CDOI), where client views, not models or theories, guide the therapy process. We describe CDOI at the end of key sections and in “Description of a Specific Approach,” tracing the progression of this particular viewpoint from a common factors heritage. In so doing, we make the case that CDOI practice is a logical heir to the conceptual and empirical common factors tradition.

HISTORY AND VARIATIONS

Beginnings

Saul Rosenzweig’s classic 1936 article, “Some Implicit Common Factors in Diverse Methods of Psychotherapy,” is likely the beginning of common factors as it is known today. Rosenzweig, a 1932 Harvard PhD and schoolmate of B. F. Skinner and Jerome Frank, suggested that the effectiveness of different therapy approaches had more to do with their common elements than with the theoretical tenets on which they were based. He summarized these common factors:

. . . the operation of implicit, unverbalized factors, such as catharsis, and the yet undefined effect of the personality of the good therapist; the formal consistency of the therapeutic ideology as a basis for reintegration; the alternative formulation of psychological events and the interdependence of personality organization. (p. 415)

Luborsky (1995), whose comprehensive review of comparative trials later confirmed Rosenzweig’s insight (see Luborsky, Singer, & Luborsky, 1975), said that Rosenzweig’s paper “deserves a laurel in recognition of its being the first systematic presentation of the idea that common factors across diverse forms of psychotherapy are so omnipresent that comparative treatment studies should show nonsignificant differences in outcomes” (p. 106).

If we were to apply to our colleagues the distinction, so important with patients, between what they tell us and what they do, we might find that agreement is greater in practice than in theory. . . . We have agreed further. . . that our techniques cannot be uniform and rigid, but vary with the age, problems and potentialities of the individual client and with the unique personality of the therapist. . . . A therapist has nothing to offer but himself. (p. 29)

On this panel, Rosenzweig outlined his implicit common factors with some further elaboration, and Carl Rogers presented about areas of agreement in working with children. Rogers highlights this panel as recommended reading in his first book, Counseling and Psychotherapy (1942), and also references Rosenzweig’s 1936 paper. Sollod (1981) notes that the 1940 panel, especially the ideas offered by Watson, significantly influenced Rogers.

Following this auspicious beginning, little was published about the common factors until an interesting study by Heine (1953) foreshadowed later comparative investigations. Heine credits the questions raised by Rosenzweig as providing the impetus to conduct a study that compared several prevailing methods of the day. Given comparable results, Heine supported Rosenzweig’s analysis by concluding that a common factor(s) was operating in the different forms of psychotherapy investigated. Heine suggested that theory and technique are less important than the characteristics of the individual applying them—a conclusion that reiterates the 1940 panel’s assertions and has since gained much empirical support. He recommended that the field devote itself to developing a psychotherapy rather than a variety of psychotherapies. Heine’s influential study was often referenced by later scholars. Heine was also acknowledged in Fiedler’s (1950) classic investigation of the ideal therapeutic relationship.

Nineteen years after Rosenzweig’s original article, Paul Hoch echoed Rosenzweig’s words in a 1955 article:

If we have the opportunity to watch many patients treated by many different therapists using different techniques, we are struck by the divergencies in theory and in practical application and similarity in therapeutic results. . . . There are only two logical conclusions. . . first that the different methods regardless of their theoretical background are equally effective, and that theoretical formulations are not as important as some unclear common factors present in all such therapies. (p. 323)

Rosenzweig said:

What. . . accounts for the result that apparently diverse forms of psychotherapy prove successful in similar cases? Or if they are only apparently diverse, what do these therapies actually have in common that makes them equally successful?. . . it is justifiable to wonder. . . whether the factors that actually are in operation in several different therapies may not have much more in common than have the factors alleged to be operating. (pp. 412–413)

The same year as Garfield’s (1957) exploration of common factors, Carl Rogers published the profoundly influential paper, “The Necessary and Sufficient Conditions of Therapeutic Personality Change,” in the Journal of Consulting Psychology. Rogers proposed that, in effective psychotherapy, therapists create core relational conditions of empathy, respect, and genuineness. Although the recognition of the importance of the therapeutic relationship was widespread as early as 1940 (see Watson, 1940), Rogers raised the stakes by suggesting that therapist-provided variables were “sufficient” for therapeutic change. Remarkably, Rosenzweig (1936), 21 years earlier, comments:

Observers seem intuitively to sense the characteristics of the good therapist time and again. . . sometimes being so impressed as almost to believe that the personality of the therapist would be sufficient (emphasis added) in itself, apart from everything else, to account for the cure of many a patient by a sort of catalytic effect. (p. 413)

This may be the first report of the sufficient nature of therapist-provided variables as popularized by Rogers’s groundbreaking 1957 article.

Building on Rogers’s understanding of therapist-provided variables, Truax and Carkhuff (1967) define empathy as the therapist’s ability to be “accurately empathic, be with the client, be understanding, or grasp the client’s meaning” (p. 25). Genuineness, or congruence, speaks to the therapist’s ability to relate transparently and honestly with the client, casting aside the façade of the professional role. Respect, according to Rogers (1957), means the ability to prize or value the client as a person with worth and dignity; it refers to the unconditional acceptance of the client, including a positive, nonjudgmental caring and a willingness to abandon suspicions regarding the authenticity of the client’s account. While these definitions describe therapist-provided variables, they do not describe the idiosyncratic interpretations of therapist behavior by the client. Duncan, Solovey, and Rusk (1992) argue that “the therapist’s reliance on standby responses to convey empathy [genuineness, or respect] will not be equally productive . . .” (p. 34), and make the point that the implementation of Rogers’s core conditions must rely instead on a fit with the client. Bachelor’s (1988) study of client perceptions of empathy concluded that this factor had different meanings for different clients and should not be viewed or practiced as a universal construct. Nevertheless, Rogers’s work spawned great practical and theoretical interest, influencing clinical training, practice, and a wave of research focused on the role of the relationship as a core variable across therapy models.

Key Figures and Variations

If Rosenzweig wrote the first notes of the call to the common factors, Johns Hopkins University’s Jerome Frank composed an entire symphony. Frank’s (1961) book, Persuasion and Healing, was the first entirely devoted to the commonalities cutting across approaches. In it, he incorporated much of Rosenzweig’s brief proposal, but articulated a much expanded theoretical and empirical context, especially regarding the profound effects of hope and expectancy in healing endeavors. In this and later editions (1973, 1991), Frank placed therapy within the larger family of projects designed to bring about healing. He (joined by his daughter, Julia, in the last edition) looked for the threads linking such different activities as traditional psychotherapy, group and family therapies, inpatient treatment, drug therapy, medicine, religiomagical healing in nonindustrialized societies, cults, and revivals.

Two apparently very different psychotherapies, such as psychoanalysis and systematic desensitization, might be analogous to penicillin and digitalis—totally different pharmacological agents suitable for totally different conditions. On the other hand, the active ingredient of both may be the same, analogous to two compounds marketed under different names, both of which contain aspirin. I believe the second alternative is closer to the mark. (pp. 313–314)

Frank also identified four features shared by all effective therapies: (1) an emotionally charged, confiding relationship with a helping person; (2) a healing setting; (3) a rationale, conceptual scheme, or myth that plausibly explains the patient’s symptoms and prescribes a ritual or procedure for resolving them; and (4) a ritual or procedure that requires the active participation of both patient and therapist and that is believed by both to be the means of restoring the patient’s health.

Frank’s common factors bear a resemblance to Rosenzweig’s original formulations, especially the notions of a conceptual scheme and alternative explanation, and the therapeutic relationship. In addition, Frank’s “single entity” concept resembles Heine’s idea of developing “a psychotherapy” (see Heine, 1953).

During the 1970s, theorists picked up on Frank’s far-reaching discussion of hope and expectancy (referred to in the literature as placebo effects), conceptualizing the common factors in these terms (e.g., A. Shapiro, 1971; A. Shapiro & Morris, 1978). The 1970s also ushered a more refined definition of the basic ingredients of psychotherapy (e.g., Garfield, 1973; Strupp, 1973), an increased empirical argument for the common factors (e.g., Strupp & Hadley, 1979), and the empirical confirmation of Rosenzweig’s original insight that diverse psychotherapies have equivalent outcomes. This finding has since been summarized by quoting the dodo bird from Alice’s Adventures in Wonderland, who said, “Everybody has won and all must have prizes” (Carroll, 1962/1865). It was Saul Rosenzweig (1936), a devotee of Lewis Carroll, who first invoked the words of the now infamous dodo bird to illustrate his prophetic observation of this phenomenon (Duncan, 2002a). Almost 40 years later, Luborsky et al. (1975) empirically validated Rosenzweig’s conclusion in their now classic review of comparative clinical trials. They dubbed their findings of no differences among models the “dodo bird verdict.”

From 1980 forward, the debate concerning what works in psychotherapy was fueled by an increasing interest in common factors (Weinberger, 1995). Grencavage and Norcross (1990) collected articles addressing common factors and noted that a positive relationship exists between year of publication and the number of common factors proposals offered. Perhaps in response to the comparative studies and reviews of the 1970s and 1980s (e.g., Luborsky et al., 1975; D. Shapiro & Shapiro, 1982; Smith et al., 1980; Stiles, Shapiro, & Elliot, 1986) reflecting the equivalence of outcome, these decades gave rise to greater prominence of common factors ideas, particularly in the eclecticism/integration movement (see Lazarus; Stricker & Gold, this volume). Many noteworthy common factors proposals appeared (e.g., Garfield, 1982; Goldfried, 1982; S. Miller, Duncan, & Hubble, 1997; Patterson, 1989; Weinberger, 1993).

Moving beyond theory-based treatment toward client-informed practice avoids the common factors paradox—how to use what is known about common processes of change without losing a shared, or common, orientation. Instead of adding one more model to the plethora already in existence, CDOI necessarily requires the tailoring of treatment to each unique situation based on client feedback. The systematic collection and incorporation of client feedback throughout therapy operationalizes this insight (see “Description of a Specific Approach” later in this chapter). A client-directed, outcome-informed, approach represents a logical evolution of the ideas first expounded by the earliest common factors theorists and offers a progressive perspective on psychotherapy theory, research, and practice in the twenty-first century.

GENERAL THEORY OF PERSONALITY AND PSYCHOPATHOLOGY

The common factors literature has largely focused on therapist-generated events (Grencavage & Norcross, 1990). Tallman and Bohart astutely note that even the language of psychotherapy (e.g., therapist mtervention and client response^) maintains a therapist-centric perspective and denotes clients as passive recipients. Theories of personality and psychopathology traditionally have viewed clients as deficient—possessing more or less stable core traits that when identified require remediation. Indeed, the field has a long history of disparaging clients, perhaps reflecting the view of people in general held by psychotherapy’s “founding father,” Sigmund Freud (1909/1953), who once said, “I have found little that is good about human beings. In my experience, most of them are trash” (p. 56). Duncan and Miller (2000b) write:

Whether portrayed as the “unactualized” message bearers of family dysfunction, manufacturers of resistance, or targets for the presumably all-important technical intervention, clients are rarely cast in the role as chief agents of change or worthy of mention in advertisements announcing the newest line of fashions in the therapy boutique of techniques. (p. 57)

The Diagnostic and Statistical Manual of Mental Disorders (DSM), “the professional digest of human disasters” (Duncan, Miller, & Sparks, 2004, p. 23), enjoys virtually unquestioned acceptance and widespread use in everyday practice, carrying on the field’s preoccupation with client dysfunction. This is the case, even though the DSM fails basic parameters of validity and reliability, and psychiatric diagnoses do not correlate with treatment outcome (Duncan et al., 2004; Kutchins & Kirk, 1997; Sparks, Duncan, & Miller, 2006).

Among the client variables frequently mentioned as salient to outcome are severity of disturbance, motivation, capacity to relate, ego strength, psychological mindedness, and the ability to identify a focal problem (Asay & Lambert, 1999). However, in the absence of compelling evidence for any of the specific client variables to predict outcome or account for the unexplained variance, this most potent source of variance remains largely uncharted. This suggests that the largest source of variance cannot be generalized because the factors affecting the variance differ with each client. Studies indicating that people overcome significant difficulties even without formal intervention support the evidence of client resourcefulness and resiliency in psychotherapy outcome (Tallman & Bohart, 1999). Prochaska and his colleagues have asserted, “in fact, it can be argued that all change is self-change, and that therapy is simply professionally coached self-change” (Prochaska, Norcross, & DiClemente, 1994, p. 17). The picture emerging from the literature is of client strength rather than pathology. In fact, although clients may bring different vulnerabilities to the therapy endeavor, client engagement far outweighs specific diagnoses in predicting outcome (Duncan et al., 2004). In sum, the common factors literature puts forward no specific frameworks of client personality or psychopathology as empirically correlated with outcome, but affirms the preeminent role of nonspecified client factors across therapies and self-generated change. Moreover, the burden of evidence points toward the resourceful engagement of clients as pivotal regardless of predetermined assessments of dysfunction.

GENERAL THEORY OF PSYCHOTHERAPY

By definition, common factors is a pantheoretical framework defined by factors shared by all treatment approaches. Common factors are nonmodel-specific and considered efficacious above specific treatment effects (Wampold, 2001). Theorists have attempted to organize levels of common factors within various frameworks. To bridge diverse groupings, Goldfried (1982) suggested that therapies contained, if not shared techniques, shared general strategies, such as providing corrective experiences and offering direct feedback. Patterson (1989) made a convincing argument that specific factors common to all theories could provide a foundation for a systematic eclecticism, particularly therapist acceptance, permissiveness, warmth, respect, nonjudgmentalism, honesty, genuineness, and empathic understanding. Castonguay (1993) proposed three categories of meaning: (1) global aspects of all therapies; (2) social and interpersonal variables, including the therapeutic alliance; and (3) nontherapy variables such as client expectancy and involvement. Grencavage and Norcross (1990) compiled five overarching clusters encompassing client characteristics, therapist characteristics, change processes, treatment structures, and relationship elements. Frank and Frank’s (1991) four components of effective therapy (see “Key Figures and Variations” earlier in this chapter) highlights the role of the therapist-client relationship in client remoralization and the fit of client beliefs with the therapist’s explanation for the problem and rationale for a ritualistic method for resolving it.

My position is not that technique is irrelevant to outcome. Rather, I maintain that, as developed in the text, the success of all techniques depends on the patient’s sense of alliance with an actual or symbolic healer. This implies that ideally therapists should select for each patient the therapy that accords, or can be brought to accord, with the patient’s personal characteristics and view of the problem. Also implied is that therapists should seek to learn as many approaches as they find congenial and convincing. (p. xv)

Following this insight, the actualization of effective strategies takes place within each singular therapy experience regardless of therapy modality (Duncan & Moynihan, 1994).

Without specific ingredients as a focal point, describing a common factors perspective represents a challenge. First, it is impossible to divorce any discussion of goals, interventions, and a therapy process in general without framing these within the context of potent common factors such as the therapy alliance, client and therapist variables, and the role of hope and expectancy. Second, to expound on treatment specifics disembodied from the largest source of change, the client—his or her engagement in therapy and extratherapeutic world—misses the mark of 50 years of empirical evidence. The following sections conceptualize psychotherapy practice within a metaframework (common factors) that provides practical, empirically grounded information for clinicians without enshrining techniques that must be practiced universally regardless of their fit with the client in each unique therapy relationship.

Goals

Bachelor and Horvath (1999) convincingly argue that next to what the client brings to therapy, the therapeutic relationship is responsible for most of the gains resulting from therapy. Referred to typically as the alliance, this common factor is the most mentioned in the therapy literature (Grencavage & Norcross, 1990) and has been called “the quintessential integrative variable” (Wolfe & Goldfried, 1988, p. 449). The fact that the role of the alliance has captivated psychotherapy researchers these past 50 years can well be traced to the pioneering work of Carl Rogers (1951). Roger’s “core” or “necessary and sufficient [conditions] to effect change in clients” (empathy, respect, genuineness [Meador & Rogers, 1979, p. 151]) have not only galvanized pivotal research, but have long provided a key emphasis in training programs and clinical practice. Patterson (1984) concluded:

There are few things in the field of psychology for which the evidence is so strong as that supporting the necessity, if not sufficiency, of the therapist conditions of accurate empathy, respect or warmth, and therapeutic genuineness. (p. 437)

The alliance is one of the most researched variables in all psychotherapy outcome literature reflecting over 1,000 findings and counting (Orlinsky, Ronnestad, & Willutzki, 2004). Researchers repeatedly find that a positive alliance is one of the best predictors of outcome (Horvath & Symonds, 1991; Martin, Garske, & Davis, 2000). Data from the Treatment of Depression Collaborative Research Project (TDCRP; Elkin et al., 1989), the landmark National Institute of Mental Health (NIMH) project considered one of the most sophisticated comparative trials ever done, found that the alliance was predictive of success for all conditions while the treatment model and the severity of the presenting problem were not (Blatt, Zuroff, Quinlan, & Pilkonis, 1996; Krupnick et al., 1996). In another large study of diverse therapies for alcoholism, the alliance was also significantly predictive of success (sobriety), even at 1-year follow-up (Connors, DiClemente, Carroll, Longabaugh, & Donovan, 1997). Moreover, the data suggest that the alliance quality is an active factor (Gaston, Marmar, Thompson, & Gallagher, 1991). Thus, the relationship produces change and is not only a reflection of beneficial results (Lambert & Bergin, 1994). Finally, based on the Horvath and Symonds (1991) meta-analysis, Wampold (2001) portions 7% of the overall variance of outcome to the alliance. Putting this into perspective, the amount of change attributable to the alliance is about seven times that of a specific model or technique. As another point of comparison, in the TDCRP, mean alliance scores accounted for up to 21% of the variance, while treatment differences accounted for at most 2% of outcome variance (Wampold, 2001), over a 10-fold difference.

Bordin’s alliance elements have been combined under the concept of the “client’s theory of change” (Duncan & Miller, 2000a; Duncan & Moynihan, 1994; Duncan et al., 1992; Hubble et al., 1999a; S. Miller et al., 1997). This concept suggests that each client has an idiosyncratic set of ideas about the nature of the problem as well as preferred ways to resolve it. To the degree that the therapist matches the client’s theory of change—provides a therapy that fits the client’s view of the desired type of therapist/client connection, goals, and therapy activities (e.g., steps to reach goals, homework assignments, in-session interventions)—the chance of a positive outcome increases. Studies, in fact, support the notion of matching clients’ theories of change. Hester, Miller, Delaney, and Meyers (1990) compared the efficacy of traditional alcohol treatment with learning-based counseling approach. While no differences were found at the conclusion of the study between the two groups, at 6-month follow-up, differences emerged stemming from beliefs clients held about the nature of alcohol problems prior to the initiation of treatment. Similarly, a post hoc analysis of the TDCRP data found that congruence between a person’s beliefs about the causes of his or her problems and the treatment approach offered resulted in stronger therapeutic alliances, increased duration, and improved treatment outcomes (Elkin et al., 1999). Essentially, the client’s theory of change is the seat of a three-legged stool that connects the legs of the empirical evidence for what constitutes a positive therapy alliance; it ties together client preferences into a stable platform on which the entire therapy rests (Duncan et al., 2004).

Assessment

Assessment, whether based on psychiatric diagnosis or other problem frameworks, consists of specific problem identification strategies based on the theoretical, or model, assumptions. Assessment procedures can be consistent with Wampold’s definition of a contextual model where “specific ingredients are necessary to construct a coherent treatment that therapists have faith in and that provides a convincing rationale to clients” (Wampold, 2001, p. 25). More often, however, psychological assessments conform to a “medical model” in which the methods of problem definition are largely theory-derived, consist of more or less invariant prescribed steps, and stand apart from the client’s frame of reference and worldview. In these instances, assessment procedures are specific, not common, factors. Outcome data spanning nearly 50 years has consistently failed to support the assumption that specific therapist technical operations are largely responsible for client improvement (Duncan & Miller, 2006; Duncan et al., 2004; Luborsky et al., 1975; Wampold, 2001). After his meticulous review of the literature, Wampold concluded that the evidence that specific ingredients account for treatment effectiveness remains weak to nonexistent. Wampold emphatically asserts, “Decades of psychotherapy research have failed to find a scintilla of evidence that any specific ingredient is necessary for therapeutic change” (p. 204). Diagnosis, a distilled assessment, is assumed to provide a blueprint for correct procedure and is, therefore, frequently required before intervention. The literature, however, indicates that diagnosis is not correlated with outcome or length of stay, and cannot tell clinicians or clients the best approach to resolving a problem (Brown, Dreis, & Nace, 1999; Wampold, 2001). Similarly, diagnosis tells clinicians little that is relevant to why people enter therapy or how they change (Duncan et al., 2004). Nevertheless, diagnoses proliferate each year, making Jerome Frank’s ironic observation (1973)—that psychotherapy might be the only treatment that creates the illness it treats—particularly salient.

Similarly, Prochaska and colleagues (Prochaska, 1999; Prochaska & Norcross, 2002; Prochaska et al., 1994) focus on the client’s readiness to change, or stage of change, as critical information required not only prior to treatment but as treatment progresses to ensure engagement in the change process. The word stage implies the client’s specific state of motivational readiness that the therapist must accommodate to make progress. Clients in the precontemplation stage have not, as yet, made a connection between a problem in their lives and their contribution to its formation or continuation. Consequently, precontemplative clients usually do not establish an alliance with a helping professional (Prochaska, 1995). Clients in the contemplation stage recognize that a change is needed, but may be unsure whether the change is worth the cost in time and energy and are ambivalent about the losses attendant to any change they might make (S. Miller et al., 1997). Clients in the preparation stage perceive a problem as well as their role in it and actively seek help in formulating solutions. In the final action stage, clients are firmly committed to and actively pursue a plan for change. Failure of the therapist to assess the client’s stage of change and match treatment strategies accordingly is likely to result in an unsatisfactory outcome, particularly in settings that serve mandated clients (e.g., court-ordered addictions counseling) due to client disinterest in or disengagement from the process. Prochaska’s stages-of-change framework has focused the field away from a preoccupation with theoretical content toward an assessment of client motivation and client engagement as central, common factors across models.

Both Prochaska’s readiness for change assessment and a client-directed, outcome-informed approach make the case for using client feedback throughout therapy to determine the strength of the alliance and whether measurable progress is being made. This type of real-time assessment is based not only on the alliance literature, but on an entire tradition of using outcome to inform process that enjoys a substantial empirical base. Outcome research indicates that the general trajectory of change in successful therapy is highly predictable, with most change occurring earlier rather than later in the treatment process (J. Brown et al., 1999; Haas, Hill, Lambert, & Morrell, 2002; Hansen & Lambert, 2003; Howard, Moras, Brill, Martinovich, & Lutz, 1996; Smith et al., 1980; Steenbarger, 1992; Whipple et al., 2003). More recently, researchers have been using early improvement—specifically, the client’s subjective experience of meaningful change in the first few visits—to predict whether a given pairing of client and therapist or treatment system will result in a successful outcome (Garfield, 1994; Haas et al., 2002; Lambert et al., 2001). To take advantage of what is empirically known about the fit of clients’ views of the alliance and their perceptions of meaningful change in the early stages of therapy, CDOI uses brief alliance and progress measures at each session or point of service (see “Description of a Specific Approach” later in this chapter). Continual feedback allows therapists to adjust their approach to better fit the client’s stage and preferences, enhancing the possibility for success. From this point of view, assessment is a living, ongoing process that engages clients, heightens hope for improvement, and is a core feature of therapeutic change.

The Process of Psychotherapy

Role of the Therapist

There is substantial evidence of differences in effectiveness between clinicians and treatment settings (Lambert et al., 2003; Luborsky et al., 1986; S. Miller et al., in press; Wampold, 2001). Conservative estimates indicate that between 6% (Crits-Christoph et al., 1991) and 9% (Project MATCH Research Group, 1998) of the variance in outcomes is attributable to therapist effects. These percentages are particularly noteworthy when compared with the variability among treatments (1%). Some therapists are simply better than others, regardless of adherence to a given treatment protocol, a point Wampold states once again supports a contextual/common factors over a medical paradigm. The TDCRP offers a case in point. Blatt, Zuroff, Quinlan, and Pilkonis (1996) reanalyzed the data to determine the characteristics of effective therapists. This is a telling investigation because the TDCRP was well-controlled, used manuals, and employed a nested design in which the therapists were committed to and skilled in the treatments they delivered. A significant variation among the therapists emerged in this study, related not to the type of treatment provided or the therapist’s level of experience, but rather to his or her orientation toward a psychological versus biological perspective, and longer term treatment.

Length of Therapy

Outcome research has much to tell us about when change happens in therapy and, once again supports the role that common factors, specifically client variables, play in the change process. Researchers Howard, Kopta, Krause, and Orlinsky (1986), estimated that approximately 15% of clients show measurable improvement prior to the first session of treatment, pointing to extratherapeutic factors as key in helping clients resolve difficulties. In therapy beyond one session, the research shows that improvement between treatment sessions is the rule rather than the exception. In one pioneering study, Reuterlov, Lofgren, Nordstrom, Ternstrom, and Miller (2000) followed 175 cases over the course of treatment and found that at the beginning of any given session 70% of clients reported complaint-related improvement. Even more encouraging, however, was the finding that half of the 30% of clients who initially reported no between-session improvement did identify specific, complaint-related improvement by the conclusion of any given session.

Change happens often and early in the therapy process. Howard et al. (1986), in their now classic meta-analytic study of nearly 2,500 clients, found that as many as 65% were measurably improved by the seventh session and 75% within 6 months. These same findings further showed “a course of diminishing returns with more and more effort required to achieve just noticeable differences in patient improvement” as time in treatment lengthens (p. 361). To illustrate, Howard, Lueger, Maling, and Martinovich (1993) not only confirmed that most change in treatment took place earlier than later, but also found that an absence of early improvement in the client’s subjective sense of well-being significantly decreased the chances of achieving symptomatic relief and healthier life functioning by the end of treatment. Similarly, in a study of more than 2,000 therapists and thousands of clients, researchers J. Brown et al. (1999) found that therapeutic relationships in which no improvement occurred by the third visit did not on average result in improvement over the entire course of treatment. This study also showed that clients who worsened by the third visit were twice as likely to drop out as those reporting progress. Variables such as diagnosis, severity, family support, and type of therapy were “not. . . as important [in predicting eventual outcome] as knowing whether or not the treatment being provided [was] actually working” (J. Brown et al., 1999, p. 404).

The empirical data about when clients change provide a golden opportunity for therapy practice. The research is clear that, rather than therapy being an arduous task based on models of stability and deficit with contingent notions of resistance, chronicity, and lengthy treatment, change happens frequently and early in the therapy process. As mentioned, most standardized measures of pathology cannot predict whether a client will change or when. Instead, studies increasingly find that clients’ views of change in the first few sessions provide more accurate predictions of eventual treatment success. Client-directed, outcome-informed practitioners capitalize on this empirical window of opportunity by creating formal feedback loops with clients from the first encounter and at each subsequent session to proactively attune therapy to client preferences (S. Miller et al., 2004, in press; see “Description of a Specific Approach” later in this chapter).

The importance of the therapeutic alliance as a crucial common factor has been unequivocally confirmed in outcome literature (Horvath & Symonds, 1991; Orlinsky et al., 2004). Recall that Wampold (2001) portions 7% of the overall variance of outcome to the alliance, or about seven times the amount of change than that attributable to a specific model or technique. Horvath (2001) concludes that as much as 50% of the variance of treatment effects is due to the alliance. Recognition of the disparity between alliance and technique effects has led to the creation of a counterbalancing movement by the APA Division of Psychotherapy to identify elements of effective therapy relationships (Norcross, 2001).

Data on the importance of client factors and the alliance, when combined with “the observed superior value, across numerous studies, of clients” assessment of the relationship in predicting the outcome (Bachelor & Horvath, 1999, p. 140), makes a strong empirical case for putting the client in the “driver’s seat” of therapy. At the conclusion of each session or point of service, CDOI practitioners use client-report alliance rating scales to obtain invaluable clues about the fit of therapy with client expectations, including the method, congruence on goals, and felt sense of connection with the therapist. In turn, this information serves as pivotal guideposts for the re-alignment of therapy to the client’s preferences in the interest of enhancing outcome (Duncan et al., 2004; Duncan, Miller, & Sparks, 2007; S. Miller et al., 2004).

Strategies and Interventions

Recall that techniques account for as little as 1% of the overall outcome in psychotherapy (Wampold, 2001). Nevertheless, techniques can provide a cogent structure and rationale for therapy, engender hope, and foster strong therapy alliances. The following discussion of strategies and interventions spans these two levels of abstraction—the role of technique as a potential catalyst for common factors and as a minor outcome variable relative to extratherapeutic, client, and alliance factors.

Major Strategies and Techniques

In a narrow sense, model/technique factors may be regarded as beliefs and procedures unique to specific treatments. The miracle question in solution-focused therapy, the use of thought restructuring in cognitive-behavioral therapy, hypnosis, systematic desensitization, biofeedback, transference interpretations, and the respective theoretical premises attending these practices are exemplary. In concert with Frank and Rosenzweig, model/technique factors can be interpreted more broadly as therapeutic or healing rituals. When viewed as a healing ritual, even the latest therapies (e.g., EMDR) offer nothing new. Healing rituals have been a part of psychotherapy dating back to the modern origins of the field (Wolberg, 1977). Whether instructing clients to lie on a couch, talk to an empty chair, or chart negative self-talk, mental health professionals are engaging in healing rituals—technically inert, but nonetheless powerful, organized methods for enhancing the effects of placebo factors. These methods include providing a rationale, offering a novel explanation for the client’s difficulties, and establishing strategies or procedures to follow for resolving them. Depending on the clinician’s theoretical orientation, different content is emphasized. Rosenzweig proposed that whether the therapist talks in terms of psychoanalysis or Christian Science is unimportant. Rather it is the formal consistency in adhering to the selected doctrine that offers a systematic basis for change and an alternative formulation to the client.

Typical Sequences in Intervention

Techniques are often sequentially arranged in evidence-based practice (EBP). Evidence-based practice assumes that specific ingredients of a given approach account for change and that adherence to these strategies will result in better outcomes. Hence, the proliferation of manuals detailing the precise model-specific steps, including sequences of intervention required to bring about change. When manualized psychotherapy is portrayed in the literature, it is easy to form the impression of technological precision.

The illusion is that the manual is like a silver bullet, potent and transferable from research setting to clinical practice. Any therapist need only to load the silver bullet into any psychotherapy revolver, and shoot the psychic werewolf terrorizing the client. (Duncan & Miller, 2006, p. 143)

However, well-controlled studies argue the opposite point. While research shows that manuals can effectively train therapists in a given psychotherapy approach, the same research shows no resulting improvement in outcome and the strong possibility of untoward negative consequences (Beutler et al., 2004; Lambert & Ogles, 2004). High levels of adherence to specific technical procedures may actually interfere with the development of a good relationship (Henry, Strupp, et al., 1993), and with positive outcomes (Castonguay, Goldfried, Wiser, Raue, & Hayes, 1996). In a study of 30 depressed clients, Castonguay et al. (1996) compared the impact of a technique specific to cognitive therapy—the focus on correcting distorted cognitions—with two other nonspecific factors: the alliance and the client’s emotional involvement with the therapist. Results revealed that while the two common factors were highly related to progress, the technique unique to cognitive-behavioral therapy—eliminating negative emotions by changing distorted cognitions—was negatively related to successful outcome. Duncan and Miller (2006) observe, “In effect, therapists who do therapy by the book develop better relationships with their manuals than with clients and seem to lose the ability to respond creatively” (p. 145). Little evidence, therefore, exists to support manualized treatments with sequenced, stepwise interventions, offering additional confirmation of the minor role played by technique compared with more robust common factors in therapy outcome.

Typical Clinical Decision Process

This trend reflects the belief that therapists’ technical operations are responsible for client improvement and that taking the guesswork out of treatment through manuals will produce consistently better outcomes. As we have seen, this is not the case. In light of the failure of manualized formulas to reliably improve outcomes, some researchers and clinicians have turned to an alternative empirical basis for guiding clinical decisions—in particular, research regarding the trajectory of therapeutic change and the role of client perceptions of progress and the alliance in predicting of outcome. In the mid-1990s, some researchers began using data generated during treatment to improve the quality and outcome of care. In 1996, Howard et al. (1996) demonstrated how measures of client progress could be used to “determine the appropriateness of the current treatment. . . the need for further treatment. . . [and] prompt a clinical consultation for patients who [were] not progressing at expected rates” (p. 1063). That same year, Lambert and Brown (1996) made a similar argument using a shorter, and hence more feasible, outcome tool. Finally, Johnson and Shaha (1996) were the first to document the impact of outcome and process tools on the quality and outcome of psychotherapy as well as show how such data could foster a cooperative, accountable relationship with payers.

More recently, CDOI practitioners have used brief alliance and outcome measures to guide clinical process (S. Miller et al., 2004, in press). Gathering and responding to clients’ views of change and the alliance as therapy progresses brings the largest portion of known variance in psychotherapy outcome—the client—center stage in clinical decision making. Moreover, continuous incorporation of client feedback gives clinicians a chance to intervene in accordance with client views prior to client dropout or negative outcome. To date, research on this approach is promising, indicating improved efficiency in overall service utilization and enhanced outcome across client populations and presenting problems (S. Miller et al., in press).

Homework

Between-session tasks or activities designed to further therapy goals are key components of many therapies, particularly behavioral, cognitive, and systemic (Kazantzis & Ronan, 2006). Badgio, Halperin, and Barber (1999) suggest that homework, defined as the acquisition of skills acquired through work done between sessions, is shared by both behavioral and dynamic therapies. Whether therapist or client generated, homework may represent a common process variable across therapy models. A common factors rationale exists for homework assignment, given the significant role of extratherapeutic factors in psychotherapy outcome. Assigning tasks for clients to perform at home presumably can increase client engagement in the therapy process, activate extratherapeutic resources, and provide a structure for therapy that enhances client expectation for success (Duncan et al., 1992). Kazantzis & Ronan recommend research to determine more precisely the mechanisms by which homework produces effects beyond classic and operant conditioning specific to behavioral and cognitive therapies and in line with common processes across diverse approaches.

Consistent with client-directed, outcome-informed work is the notion that each client will experience therapist intervention, including homework assignment, idiosyncratically. The refusal of clients to accomplish homework provides information to the therapist that the client may not perceive the task as being relevant to his or her situation or view of problem resolution. Similarly, clients who refashion assignments to fit their own tastes indicate that homework has a meaningful role despite its lack of congruence with the specifics of the therapist’s original task. From a CDOI perspective, client utilization of therapist tasks is indicative of engagement and the fit of the method with client expectations, ultimately boding well for a positive outcome. At the same time, client nonparticipation in tasks does not speak to client deficiency (e.g., resistance or “not a good candidate for psychotherapy”) but simply to a lack of fit for the client and the particular intervention. In this case, therapists can use the information either to continue on a certain track (assigning tasks, etc.) or to try something different.

Adaptation of Strategies to Specific Presenting Problems

View of Medication

When medication is associated with positive outcomes, we ask, “What are the factors that produce an effect?” We propose that how one answers this question makes a difference in the choices clients are offered, particularly in the required use of medications in some settings (see, e.g., Whitaker, 2002) and whether the adverse events (side effects) associated with psychotropic medications, especially in children, can be avoided. Our analysis suggests that medication may work, not by virtue of its impact on neurotransmitters and the like, but as the result of common factors. In particular, the research indicates that the belief by clients that they are getting a powerful healing agent and the hope for improvement this engenders, play powerful roles in outcome. Similarly, who administers the medication (therapist effects), and, presumably, the relationship he or she establishes with the client, significantly determine whether the treatment is effective (Wampold, 2001, 2006). In the TDCRP, differences in therapist effects were significant and) independent of whether the therapist was in the medication or therapy group (Blatt et al., 1996). Nonetheless, outcomes were better for therapists who held a psychological rather than a medical orientation, suggesting that clients resonated more in this treatment context with this mindset and were less inclined to view their problems or the solutions to their problems as medical (Blatt et al., 1996).

Therapist allegiance may also play a role when medication has a desired effect. Wampold (2001) notes that not only the client’s but the therapist’s belief in the efficacy of an approach greatly enhances treatment effects (see also, Greenberg, 1999). When therapists have allegiance to a medical approach, they are likely to reinforce expectancy for improvement. Similarly, the ritual of medicine—the diagnostic interview, the formal explanation (diagnosis), and the prescriptive treatment (medication)—holds all the allure of healing rituals that are part of the cultural scripts characteristic of human societies (Frank’s framework of common factors particularly suits this understanding of medication efficacy). Greenberg (1999) summarizes these common elements in psychiatric drug therapy:

Medication response can be readily altered by who delivers the drug, how its properties are described, the degree of familiarity with the setting in which it is presented, and the ethnic identity or socioeconomic status of the person ingesting it. (p. 301)

The argument we make is that, based on an analysis of the empirical evidence, the specific ingredients of medication and their alleged biochemical impact are secondary to common factors effects in producing desired outcomes.

Curative Factors

Even if model ingredients appear to be the magic bullet, their effects are not due to their unique properties but to the client’s use of them within a strong therapy relationship. An empirically appropriate frame for curative factors in psychotherapy may be called client utilization. Here, focus is on how clients take what therapy offers to fashion unique and perfectly fitting solutions for even the most daunting dilemmas (Duncan & Moynihan, 1994; Sparks, 2000; Tallman & Bohart, 1999). From this point of view, therapists may well think of themselves as simply not getting in the way of their clients. One way that therapists accomplish this is to work with clients to create a context, with the appropriate exercise of intervention palatable to the client’s worldview, wherein clients essentially cure themselves.

Another important common factor interweaves with clients’ own curative capacities—known variously as placebo, hope, or expectancy. Asay and Lambert (1999) put the contribution of these variables to psychotherapy outcome at 15%. In part, this class of therapeutic factors refers to the portion of improvement deriving from client’s knowledge of being treated and assessment of the credibility of the therapy’s rationale and related techniques. Expectancy parallels Frank’s idea that in successful therapies both client and therapist believe in the restorative power of the treatment’s procedures or rituals. These curative effects, therefore, are not thought to derive specifically from a given treatment procedure; they come from the positive and hopeful expectations that accompany the use and implementation of the method. Frank’s (1973) classic discussion of remoralization as the final common pathway of all therapeutic intervention speaks to the power of hope to counter the most demoralized client.

That the procedures are not in and of themselves the causal agents of change matters little (Kottler, 1991). What matters is that therapy participants have a structured, concrete method for mobilizing placebo factors. From this perspective, any technique from any model may be viewed as a healing ritual, rich in the possibility that hope and expectancy can inspire (see Hubble et al., 1999b). Similarly, therapist allegiance to a model is a prominent constituent of positive change (see Wampold, 2001), supporting Frank’s (1973) idea of the importance of therapist and client belief in a given therapy procedure. As researchers Benson and Epstein have noted, treatment professionals “who have faith in the efficacy of their treatments. . . are the most successful in producing positive placebo effects” (O’Regan, 1985; p. 17). Broadly speaking, the simple belief in the possibility, indeed inevitability, of change in general is sufficient to inspire clients and promote problem resolution.

Special Issues

There is nothing wrong with wanting to know which approaches are effective. However, one should always ask, “Whose evidence is it and what kind of evidence is it really?” Only then can it be determined whether this evidence warrants privilege of this approach or any mandate of its practices. (p. 6)

The notion that specific technical operations are largely responsible for client improvement is empirically bankrupt. Similarly, the preponderance of the data refutes any claim of superiority when two or more bona fide treatments fully intended to be therapeutic are compared (see “Empirical Support” later in this chapter). If there are no specific technical operations that can be reliably shown to produce a specific effect, then mandating EBP seems to make little sense. More damning to EBP, perhaps, is that the repeated demonstration that superiority over placebo or treatment as usual is not really saying that much; psychotherapy has demonstrated its superiority over placebo for nearly 50 years! Therapy is about twice as efficacious as placebo and about four times better than no treatment at all. This research tells us nothing that we already do not know—therapy works. Demonstrating efficacy over placebo is not the same as demonstrating efficacy over other approaches.

When differential efficacy is claimed, a critical analysis is warranted. First, the studies that find differences between bona fide approaches are no more than one would expect from chance. Further, closer inspection of studies that claim superiority reveals two major issues that must be considered: allegiance effects (whose evidence?) and indirect comparisons (what kind of evidence? Wampold, 2001). Allegiance effects are those that are attributable to the therapist or researcher’s affinity toward the treatment at hand; Wampold (2001) suggests that allegiance accounts for up to 40% of any treatment effects. Though some reviews have found a very small advantage for cognitive-behavioral approaches, later studies found that the differences disappeared completely when the allegiance of the experimenters to the methods being investigated was taken into account (Lambert & Ogles, 2004; Miller, Wampold, & Varhely, in press). Another important issue in evaluating claims of differential efficacy is whether the study really presents a fair contest—is the comparison actually a contrast between two approaches fully intended to be therapeutic? Or is it, in fact, the pet approach of the experimenters pitted against a treatment as usual or less than ideal opponent? Wampold (2001) calls such unfair matches indirect comparisons. An inspection of one such comparison involving serious juvenile offenders (Borduin et al., 1995) reveals that multisystemic therapy (an intensive family and community-based treatment for juvenile offenders, e.g., Henggeler, Melton, & Smith, 1992), conducted in the home, involving parents and other interacting systems, by therapists regularly supervised by founders of the approach is compared with therapy of the adolescent only, with little to no outside input of parents or others, conducted in an outpatient clinic by therapists with no special supervision or allegiance. This type of comparison represents a treatment as usual contrast rather than a bona fide treatment comparison.

The evidence-based practice debate may create an unfortunate dichotomy that detracts from efforts to capitalize on the known evidence of what makes therapy effective. Addressing this dilemma, the APA Presidential Task Force on Evidence-Based Practice (APAEBP) defined evidence-based practice in psychology as “the integration of the best available research with clinical expertise in the context of patient characteristics, culture, and preferences” (2006, p. 273). The task force recommends:

• Clinical decisions should be made in collaboration with the patient, based on the best clinically relevant evidence, and with consideration for the probable costs, benefits, and available resources and options.
  
• Psychological services are most effective when responsive to the patient’s specific problems, strengths, personality, sociocultural context, and preferences.
  
• The application of research evidence to a given patient always involves probabilistic inferences. Therefore, ongoing monitoring of patient progress and adjustment of treatment as needed are essential (APAEBP, 2006, pp. 284–285, italics added).

Evidence here, then, means the findings of nearly 5 decades of research on the common factors of change and known predictors of success and evidence of the progress and fit of services collected collaboratively with clients that significantly improves effectiveness and efficiency in real clinical settings—or what we call practice-based evidence. The APA’s definition and its implications reject mandating approaches without client input and, instead, embrace a collaborative, contextualized, culturally responsive and client-informed practice.

Culture and Gender

Given the long reach of psychotherapy’s arm into the everyday lives of so many people, it is crucial to examine its differential impact across dimensions of diversity: race, gender, ethnicity, social class, age, sexual orientation, religion, immigrant, refugee, and colonization heritage, and disability status. Recent government reports have concluded that nondominant groups face treatments that fail to consider their unique contexts and are, therefore, ineffective (Sue & Zane, 2006). Emphasis on cultural competency in training and professional discourse, however, may inadvertently reinforce psychotherapy’s blindness to the inherent power and privilege disparity between typically white, heterosexual, middle-class mental health professionals and their diverse clientele (Levant & Silverstein, 2006). Simply learning about difference does not address contexts of oppression, prejudice, and discrimination that nondominant groups face. The ghettoization of “diverse populations” continues the marginalization of the very groups it seeks to mainstream (L. Brown, 2006). Similarly, focusing on individual pathology detracts from an analysis of systems of power and privilege that underpin oppressive relationships (Duncan et al., 2004; Levant & Silverstein, 2006; Waldegrave, 1990). Finally, simple knowledge of diversity fails to deconstruct the underlying premises of psychotherapy theory in which unequal relations of power are embedded.

A common factors perspective, particularly the distinction between practice-based evidence and evidence-based practice, has relevance for providing culturally aware, respectful, and effective services to diverse client groups. Despite the evidence that clients—their resources, networks, and life circumstances—make up by far the largest portion of variance in psychotherapy outcome, evidence-based practice focuses on the matching of specific treatment approaches to particular identified problems. This script forms the bedrock of much of psychotherapy practice. It also draws clients as “cardboard cutouts”; more a diagnosis or problem to be corrected by the EBP. As a result, a culture of client disability and passivity can take shape, devoid of context and stripped of the richness of client histories and culture. Despite well-intentioned efforts to invite clients’ voices, the machinery of psychotherapy (paperwork, policies, procedures, and professional language) codifies noncontextualized descriptions of client dysfunction and effectively silences client views, goals, and preferences.

The blind adherence to “best practice” without a cultural critique is not without consequence. The history of the treatment of those labeled mentally ill in the United States is replete with the best practices of the day, including forced drugging, restraint, social isolation, brain surgery, and shock (see, e.g., Whitaker, 2002). While current evidence-based treatments are heralded as advances over earlier methods, those most affected still may have little voice to either reject them or to recommend something different. The call to expand evidence-based research to diverse groups (see Sue & Zane, 2006) does not go far enough. A move to honor the rightful roles of clients in psychotherapy, based not simply on a desire to sidestep the mistakes of the past but to engage clients as the most potent common factor, requires a culture of feedback. This milieu is grounded in knowledgeable and affirming practice (L. Brown, 2006) and an appreciation of context. It also entails asking for, listening to, and valuing each client’s meanings, hopes, and preferred forms of help at each therapy encounter (Duncan et al., 1992). Culturally competent practitioners enhance outcomes by not imposing goals derived from unquestioned theory or personal bias but by tailoring the intervention process to each person being served. A psychotherapy where evidence flows up from clients rather than down from theory can provide an antidote to the sometimes dehumanizing aspects of prescriptive care; it ensures that clients’ unique world-views, preferences, and values are not only respected but central to the therapy process.

Adaptation to Specific Problem Areas

The previous review of the literature by now should be sufficient to support the claim that nonspecific, common factors play the major role in psychotherapy outcome across problem areas and specified disorders; Adaptation, or application, to a specific problem area again calls forth the square peg/round hole conundrum; since common factors is not a model, it cannot be adapted or applied. Nevertheless, it is worthwhile to consider whether research on common factors, based primarily on adults, bears up when considering psychotherapy with children. It is also relevant to examine recent efforts to establish evidence-based practice for children and adolescents, and the role common factors can play in designing more effective services for youths and their families.

The news about what works for our youngest clients is mixed. Kazdin (2004), citing numerous research reviews, concludes that youth psychotherapy is effective when compared with no treatment. The effect size in child efficacy studies is relatively large (.70), rivaling similar estimates in the adult literature. However, effectiveness studies carried out in real-world settings tell a different story. According to Bickman (2002), the literature regarding effectiveness of treatment as usual for children and adolescents in the community is, “depressingly consistent in its poor outcomes” (p. 195). Dropout rates for young people being treated in community settings are as much as 40% to 60% (Kazdin, Holland, & Crowley, 1997). One study reports that most children who start psychotherapy never make it past the 2nd session (Armbruster & Fallon, 1994). Despite the disconnect between upbeat outcomes based on clinical trials and the disappointing results of community-based studies, recommendations based on trials forge full speed ahead with expanding lists of evidence-based practices and treatment guidelines (see, e.g., Kazdin, 2002).

The science [of diagnosis] is challenging because of the ongoing process of development. The normally developing child hardly stays the same long enough to make stable measurements. Adult criteria for illness can be difficult to apply to children and adolescents, when the signs and symptoms of mental disorders are often also the characteristics of normal development. (U.S. Department of Health and Human Services, 1999)

Similarly, the World Health Organization states, “Childhood and adolescence being developmental phases, it is difficult to draw clear boundaries between phenomena that are part of normal development and others that are abnormal” (World Health Organization, 2001).

Bickman (2002) pointedly questions the utility of child and adolescent diagnosis:

Suffice it to say that I do not believe that there is adequate scientific evidence to support the diagnostic approach in developing services. Our own research. . . has suggested that diagnostic categories have a great deal of overlap (i.e., show low discriminate validity), are arbitrary in setting standards for caseness, and are most often used for economic and social reasons. (p. 196)

Bickman (2002) further notes that as much as 50% of children seen in clinic practice have more than one diagnosis. Consistent with a medical model, diagnosis should provide a path toward effective treatment. However, as with adults, virtually no evidence exists to support the notion that a diagnosis can lead to an effective treatment-matching system (Bickman):

It is my observation that most of the interventions are developed by individuals who have a particular interest in a diagnostic category and design their intervention based on the theory of that disorder. . . it creates silos of intervention that have little relationship to each other. (p. 196)

Apart from questionable validity and reliability, giving a psychiatric diagnosis to a child can overlook cultural and societal factors, missing key opportunities to provide effective social intervention beyond changing internal chemistry or cognition. Diagnosis also carries a significant stigma, labeling the child as impaired, potentially creating a lifelong deficit identity that influences the child’s ability to succeed.

Just as treatment characteristics did not predict outcome or moderate associations between the alliance and outcome in Shirk and Karver’s (2003) study, other youth studies have failed to demonstrate differential efficacy between various approaches. The CYT (Dennis et al., 2004) compared five different intervention models at different doses; overall, the different types and doses of intervention worked with about the same effectiveness. A recent meta-analytic study of the child/adolescent literature, after controlling for allegiance effects, found a treatment effect size of .22, comparable to the effect size (ES) in Wampold et al.’s 1997 adult meta-analysis, indicating a miniscule portion of the variance in youth treatment attributable to specific factors; no differences were found between bona fide child and adolescent approaches for key problem domains (Miller, Wampold, et al., in press). Another meta-analysis failed to find superiority for behavioral interventions over other bona fide child treatments when allegiance confounds were considered (Spielmans, Pasek, & McFall, in press).

All signs point to significant similarities in the psychotherapy literature between child/adolescent and adult approaches related to evidence for common factors and against specific ingredients. Notably, studies also support similar change trajectories. In the CYT, change occurred within the first 3 months. Despite a growing cadre of studies that proclaim superiority of one approach over another, critical analysis reveals the same flaws that taint adult claims of differential efficacy—allegiance effects and indirect comparisons (Duncan & Miller, 2006). The CYT found no clear superiority of best practice or researched based intervention (Godley, Jones, Funk, Ives, & Passetti, 2004). Miller, Wampold, et al. (in press) conclude:

None of the effects of any of the approaches deemed “evidence-based” by the American Academy of Child and Adolescent Psychiatry and the Task Force on the Promotion and Dissemination of Psychological Procedures has been shown to be demonstrably superior to other treatments intended to be therapeutic for the disorder treated. (p. 3)

What is slowly gaining focus in the child/adolescent research is the critical role young people, within a strong therapy alliance, play in their own change (Karver et al., 2005). Client variables frequently mentioned for youth treatment include youth age-developmental status, youth-parent interpersonal functioning, parental status, family environment, and youth-parent expectancies of efficacy (Karver et al., 2005). From a common factors perspective, these are potent elements to engage in the interest of better outcomes. As with adults, a growing body of common factors research recommends demoting EBPs and, instead, promoting children, adolescents, and their significant networks to the forefront of therapy.

Empirical Support

The dodo bird hypothesis has since garnered increasingly unequivocal support. Ushering in the age of the randomized clinical trial, the TDCRP (Elkin et al., 1989), a study that represented the state-of-the-art in outcome research, found that the four investigative treatments—including placebo—achieved about the same results. Further confirmation of the dodo bird’s assertion of uniform efficacy across treatment models was found in the Wampold et al. (1997) study addressing methodological problems of earlier meta-analyses. This meta-analysis included some 277 studies conducted from 1970 to 1995 and verified that no approach has reliably demonstrated superiority over any other. At most, the effect size (ES) of treatment differences was a weak .2. “Why,” Wampold et al. ask, “[do] researchers persist in attempts to find treatment differences, when they know that these effects are small?” (p. 211). Finally, an enormous real-world study conducted by Human Affairs International of over 2,000 therapists and 20,000 clients revealed no differences in outcome among 13 approaches, including medication as well as family therapy (J. Brown et al., 1999).

The fact that the dodo bird verdict has emerged by accident—while researchers were trying to prove the superiority of their own models—makes it even more compelling. It is a finding free of researcher bias. As Rosenzweig amazingly said some 71 years ago, because all approaches appear equal in effectiveness, there must be common factors that overshadow any perceived or presumed differences among approaches. Intervention works, but our understanding of how it works cannot be found in the insular explanations of theoretical orientations.

DESCRIPTION OF A SPECIFIC APPROACH

For a field as intent on identifying and codifying the methods of treatment as therapy is, abandoning process in favor of outcome may seem radical indeed. Research provides a rich source of data concerning how change happens, providing therapists with readily usable tools to make optimal clinical decisions. Specifically, this research indicates that (a) change in successful therapy is highly predictable, with most occurring early in the treatment process; (b) the client’s experience of change early in the treatment is predictive of outcome; and (c) the client’s early ratings of the therapeutic alliance are highly correlated with outcome (see “Typical Clinical Decision Process” earlier in this chapter). Recognition and deliberate utilization of extant knowledge of change and the importance of client feedback in psychotherapy led to the development of an “outcome-informed approach” (Duncan et al., 2004). The diverse number of approaches encompassed in the change data hinted that the particular brand of therapy employed was of less importance than whether the current relationship was a good fit. Obtaining clients’ views of fit using client-rated outcome tools was already underway, though most of these efforts occurred in laboratory settings using lengthy measures not suitable for everyday clinical practice (e.g., see Howard et al., 1996; Johnson & Shaha, 1996, 1997; Lambert & Brown, 1996).