Description of the STS System

The STS is one of three general approaches that have sought to identify the scientific foundations of mental health treatments and to establish treatment applications upon a scientific framework (Beutler, 2002; Beutler, Clinton, Gierlach, & Klajic, in press). The most usual method of identifying what treatments have been demonstrated to be effective, scientifically, entails comparing specific treatment models to a no-treatment or placebo control condition, among a diagnostically homogeneous group of patients (Nathan & Gorman, 2003). This approach relies on a random clinical trial methodology and seeks establish efficacy rather than effectiveness (e.g., impact of an ideal treatment in a non-representative setting). This approach, by its nature, gives scant attention to factors that are not directly related to the treatment model but that, in the mental health arena, are critically important contributors to establishing the effectiveness of treatments in actual practice. Namely, this approach ignores the important role played by characteristics of the therapeutic relationships and of the particular participants who are involved in the treatment (i.e., patients, therapists, family members). Because of these omissions, some scholars have urged the use of alternative procedures for assessing the scientific validity of an intervention in mental health. One of these alternative methods identifies the role of various patient characteristics and participant qualities, resulting in a list of qualities that can be assigned or varied in treatment in order to enhance the effectiveness of interventions (Norcross, 2002). Unlike the first method, which addresses the question of efficacy, effectiveness research such as embodied in the second approach, considers the range of actual factors and settings in which treatment is actually practiced.

All three of these approaches assume that scientific research is the most useful and valid foundation on which to establish the efficacy and effectiveness of an intervention. Thereby, they serve to complement one another, rather than being competitive systems. They do, however, stand in contrast to the many approaches that eschew scientific evidence in favor of personal experience and strong beliefs (c.f., Beutler, 2000). Specifically, these general, research-informed systems view sound scientific research as absolutely critical to establishing a credible treatment and to preventing fraud and misuse of treatment procedures. Indeed, the STS system provides a broad framework that allows integration and balance among the systems’ relative emphasis on ideal treatments and the practical significance of participant, contextual, and relationship domains that are addressed in effectiveness research.

The Development of the STS

There has always been an implicit belief among mental health practitioners, and especially among psychotherapy researchers, that there is some optimal or “best fit” between a given collection of interventions and the characteristics of a particular patient. The most usual tactic for exploring this illusive “fit” has been based on a model derived from pharmacology and medicine, the randomized clinical trial (RCT). While RCT research designs are good for assessing the effects of pure treatment packages, isolated techniques, and specified groups of interventions, these designs are not adequate for identifying and blending the many extra-therapy and non-diagnostic contributors to outcome that comprise the more part of treatment. A “treatment”, in the area of mental health, must be considered to extend beyond the particular methods applied. It includes these, and also important qualities of the person who applies these techniques, the receptivity and preferences of the person who is subjected to these procedures, and the quality of the interactions that comprise the medium through which the procedures are applied.

The STS system comes from a different perspective than that from which efficacy studies derive. Instead of simply constructing and testing another rationally-based treatment to add to the 400+ already in existence, Beutler and colleagues, over the years, have inspected patients, clinicians, and treatments at more micro levels than that subsumed either by the theoretical model used or by the diagnosis of the patient on which it is applied. Instead, we have been drawn to inspect research through a lens than identifies dimensions of patient, therapist, treatment, and relationships, that optimally fit together and that induce positive effects, regardless of the model of treatment from which these qualities were conceived. Thus, a growing number of scholars have joined the effort to match specific therapeutic technologies with equally specific non-diagnostic patient dimensions (Shoham-Salomon, 1991; Karno & Longabaugh, 2005; Norcross, 2002; Beutler, 1983; Beutler & Clarkin, 1990: Beutler, Clarkin & Bongar, 2000; Prochaska, 1984; Harwood & Williams, 2003).

The STS--Clinician Rating form (STS-CRF) was first described in 1995 (Beutler & Williams, 1995) and first was published in 1999 (Fisher, Beutler, & Williams, 1999). It represented an extrapolation of findings from a research program that was initiated in 1979 with a re-analysis of the dodo bird verdict proclaimed by Luborsky, Singer, and Luborsky in 1975. That analysis (Beutler, 1979) extracted all research comparisons of two or more psychotherapies from the original Luborsky et al review, and undertook a systematic inspection of the patient characteristics that were present when one of the treatments disaffirmed the dodo bird verdict by proving to be more effective than another treatment. This analysis, contrary to Luborsky et al’s original conclusions, suggested that patients who presented with certain kinds of characteristics were more and less responsive to different classes of therapeutic procedures. That is, procedures that share a particular form or structure, may be equally well received by a given patient, even when their objectives and founding model are different.

Two additional studies, one a post-hoc, naturalistic analysis of patients and therapists in outpatient settings (Beutler & Mitchell, 1981), and the other a quasi-experimental comparison of three types of therapy with various Abest fit@ matches with procedures used in outpatient psychotherapy (Beutler & Thornby, 1982) provided early confirmatory evidence to support the 1979 findings. These studies each constituted comparisons of three different psychotherapy approaches, as represented in therapist behaviors (Behavoral/Cognitive, Insight, Experiential). The results demonstrated the value of fitting therapist directiveness with patient resistance and of fitting the insight or symptom focus of therapy to select patient coping styles.

Complemented by a comprehensive review of accumulated research from numerous research programs, up to that time, the first book based on this research presented a multi-dimensional approach to matching patients with treatments. Systematic Eclectic Psychotherapy (Beutler) was published in 1983. The identified patient and therapist matching variables that had been found, with modest consistency, to be predictive of positive outcomes among various kinds of patients. These dimensions for tailoring the therapy to the patient established the framework for a decision model for planning an integrated approach to treatment. Research on this model then began in earnest with a randomized, prospective study of group therapy among psychiatric inpatients (Beutler, Frank, Scheiber, Calvert, & Gaines, 1984) and a companion study of treatment among psychiatric outpatients (Calvert, Beutler, & Crago, 1988).

As research accumulated, dimensions for optimizing the effects of psychotherapy were extended to include therapist, as well as therapy characteristics and prognostic as well as matching factors. A wide range of research methods were used to identify and refine the matching and prognostic dimensions, including interviews of psychotherapists (e.g., Glueck & Beutler, 1987), prospective studies using randomized clinical trials (RCT) methodologies applied to both the processes of psychotherapy (e.g.; Beutler, Daldrup, Engle, Oro'-Beutler, Meredith, & Boyer, 1987; Beutler, Daldrup, Engle, Guest, Corbishley, & Meridith, 1988; Corbishley, Hendrickson, Beutler, & Engle, 1990; Hill, Beutler, & Daldrup, 1989) and the outcome (e.g., Beutler, Engle, Mohr, Daldrup, Bergan, Meredith, & Merry, 1991; Beutler, Machado, & Engle, 1993; Scogin, Bowman, Jamison, Beutler, & Machado, 1994 ), and post-hoc or quasi-experimental designs (e.g; Beutler, 1991a; Beutler, Mohr, Grawe, Engle, & MacDonald, 1991; Calvert, Beutler, & Crago, 1988; Lafferty, Beutler, & Crago, 1989; Mohr, Beutler, Engle, Shoham-Salomon, Bergan, Kaszniak, & Yost, 1990).

The results of these studies began to converge around a number of demonstrably important principles that described the conditions under which patient, treatment, therapist, and process dimensions could promote change and expanded on the sketchy model that had been described in 1983 (e.g., Beutler, 1991a, 1991b; Beutler & Consoli, 1992; Beutler, Patterson, Jacob, Shoham, Yost, & Rohrbaugh, 1994; Beutler, Sandowicz, Fisher, & Albanese, 1996;Daldrup, Engle, Holiman, & Beutler, 1994; Lazarus & Beutler, 1992).

From early on, however, it became clear that the process of planning a broad ranging treatment for a particular patient, and carrying out that treatment plan through psychotherapy involved somewhat different processes. The new approach to identifying and using the patient and treatment matching dimensions to construct a multi-faceted treatment plan was officially christened, “Systematic Eclectic Psychotherapy” (Beutler & Consoli, 1992) but later was changed to “Systematic Treatment Selection” (Beutler & Clarkin, 1990; Beutler, & Rosner, 1997; Beutler, Zetzer, & Williams, 1996; Norcross, Beutler, & Clarkin, 1998). The particular psychotherapeutic intervention based on these plans was dubbed, “Prescriptive Psychotherapy” (Beutler & Hodgson, 1993; Beutler & Harwood, 2000). Eventually, the term Systematic Treatment Selection, or STS was used to describe the assessment process and Prescriptive Psychotherapy was used to describe the collection of interventions that comprised the particular shape of the psychotherapy that derived from this planning process (Beutler, Zetzer, & Williams, 1996; Beutler & Rosner, 1997; Norcross & Beutler, in press; Norcross, Beutler, & Clarkin, 1998; Beutler, Clarkin, & Bongar, 2000; Beutler, Alomohamed, Moleiro, & Romanelli, 2002).

Refinement and Extension of the STS.

Rather than representing a new therapy model, Systematic Treatment Selection was conceived and has remained a method of circumventing the narrow theoretical models and proscriptive lists of procedures and techniques that traditionally have characterized rationally-derived approaches that comprise the fields of psychotherapy and the clinical experiments that have comprised much of psychotherapy research. In contrast to these rational, theory-based approaches, the empirical approach of the STS has ensured that it is an open system in which all research-grounded dimensions of patients, therapists, processes, and treatments could be incorporated as available research was able to articulate the principles that underlay their application. As a result, the principles that comprise the application engine of the STS can be modified and extended almost limitlessly, with the major consideration being only that the characteristics and matching dimensions employed: (1) are well founded in sound research methodology, (2) can be reliably measured, (3) can be articulated as either common or specific principles, and (4) successfully predict a patient=s or a group of like-patients= course of treatment, prognosis, and outcome.

As the STS system identifies the classes or types of procedures and principles that should guide treatment, the Prescriptive Therapy that derives from this planning is varied in order to both fit the proclivities of the clinician and to conform to the particular needs of each patient. STS, as a planning system, directs the clinician to think about and measure a given patient=s status on a dynamic blend of dimensions that are all research-informed and grounded. Prescriptive Therapy, as an application of these research-informed principles, can focus on different types of problems and different settings, challenging the clinician to develop ways of reliably tailoring the application of the research-informed principles of STS for the unique needs of particular patients (e.g., Beutler & Harwood, 2000; Housley & Beutler, in press; Norcross & Beutler, in press).

However, the specific number and identity of the STS planning dimensions are subject to change over time as their measurement and predictive power become more refined and better established. Thus, in the original Beutler (1983) book on Systematic Treatment Selection, a list of 8 particular dimensions were identified; by 1990, Beutler and Clarkin identified nearly 40 dimensions of prediction. From a number of perspectives, it is desirable that the number of dimensions used be kept small and manageable. Thus, the 40 dimensions outlined and advocated by Beutler and Clarkin (1990) were not conducive to efficient clinical decision making. Clinicians can only hold in mind and mentally manipulate a small number of dimensions, on their own. Thus, by 2000, these dimensions had been consolidated and combined through the development and testing of measurement instruments back into 12 dimensions (4 patient factors, 4 treatment factors, and four “fit” or “matching” factors) that can be applied to the prediction of outcomes in individual therapy, and four additional patient dimensions that are related to the broad planning of treatment modality (psychosocial and pharmacological) and format (group vs family vs individual). For special purposes or because of special interests of the user, additional dimensions can be used as well (see Norcross & Beutler, in press).

Even 12 factors are difficult for a clinician to manage and keep in mind while initiating a complex treatment. The availability of computer technology has enabled a clinician to expand the number of dimensions considered and to integrate these dimensions into workable treatment plans. Thus, in the late 1990s, a process of computerizing the decision process was initiated and this process has continued to evolve to the present. The contemporary use of the STS and particularly the STS-CRF has been described in a major graduate textbook on psychological assessment (Beutler & Groth-Marnat, 2003; Groth-Marnat, 2002; Harwood, & Williams, 2003). The self-report version also has been adapted to this format, but does not yet have the history of clinical application associated with it optimally to establish the limits of its potential.

Validation Research on the STS Dimensions

Since those early years in which the concepts and ideas of STS were formulated, hundreds of related research and theoretical papers and several books have been published (See Appendix A). Several major research grants, including large grants from the National Institute of Mental Health, the National Institute for Alcohol, Alcohol Abuse, and Alcoholism, and the National Institute for Drug Abuse, as well as other benevolent groups, supported validation and refinement studies (See Appendix B). This research, complimented by that of many students and colleagues who gradually gravitated to the ideas presented in the STS system, has made it one of the most widely researched models of integrative treatment available today (Norcross & Beutler, in press). Mention will be made of only a few of the contemporary studies that have supported the STS principles and the Prescriptive Therapy that is derived from applying these principles.

A seminal study for the development of the contemporary STS system (Beutler, Engle, & Mohr, et al., 1991) utilized three manualized treatments for Major Depressive Disorder. These treatments were selected to ensure that they systematically varied along two dimensions: The first dimension of therapy addressed, was the mechanism of change. Two of the three specific therapies (Focused Expressive Psychotherapy [FEP] and a new supportive, self directed procedures or S/SD) were designed to facilitate insight and emotional awareness as agents of change, while the second (CT) focused directly on retraining symptoms. The second dimension was one of therapist role, ranging from being very directive (FEP and CT) to being quite non-directive (S/SD).

The average differences in outcomes among the three therapy conditions (CT, FEP, S/SD) were relatively small. This is exactly what was expected----mean outcome differences are minimal between broad therapeutic models. More specific and refined analysis of variables denoting the fit of the specific interventions used to patient characteristics were more important predictors of outcome than the treatment models alone. The persistently small or negligible differences in average outcomes among treatment models (the dodo bird verdict), contrasted with the strong relationships attributable to specific patient qualities and treatment strategies, raises serious questions about the value of inspecting treatment outcomes solely through the models of treatment studied without addressing the moderating effects of patient factors.

Patient factors, particularly patient coping style and level of resistance have been found in STS research to be particularly strong moderators of treatment effects. For example, Beutler, Engler, and colleagues found that broad and robust differences occurred when outcome was looked at as a function of patient coping style. Among patients assigned to CT, treatment outcomes were greater among patients whose personality profiles revealed the reliance on acting out and other externalized coping strategies, as compared to those whose profiles revealed few of these qualities. Conversely, among patients seen in the two insight-oriented treatments (FEP, S/SD), the reverse was true. While mean outcomes for this latter group were nearly identical to those in CT, it was those patients whose personalities were characterized by internalized coping behaviors (i.e., self-reflection, emotional withdrawal, social introversion) who experienced the greatest amounts of symptom reduction when insight facilitation strategies were employed.

The results also indicated that patients who were initially assessed to be very sensitive and avoidant of being controlled by others who did best when assigned to the non-directive treatments (S/SD) than when assigned to either of the directive treatments. Conversely, those with low levels of initial defensive sensitivity, as assessed at intake, performed comparably and well when assigned to the two treatments, which employed a relatively frequent amount of therapist directives (Beutler, Engle, Mohr, et al., 1991).

The patterns of results for both patient variables and both types of treatment were subsequently cross-validated in a cross-cultural study of a Swiss population. In this study, a sample of anxious and depressed patients from the Bern (Switzerland) Psychotherapy Research program (Beutler, Mohr, Grawe, Engle, & MacDonald, 1991) was studied utilizing a randomized clinical trial design. Coping style significantly predicted the differential value of symptom focused (behavior therapy) and insight focused (client-centered therapy) interventions. Likewise, resistant sensitivity was differentially predictive of the value of these directive and non-directive procedures.

Karno, Beutler & Harwood (2002) replicated the above studies once again, this time in a randomized clinical trial of alcoholic men, wherein they pitted cognitive therapy (CT) and family systems therapy (FT) against each other, once again looking at patient qualities that differentially affect the outcomes. STS dimensions of the clients were assessed, again with the expectation that patient coping style and sensitivity to avoiding control (i.e., resistance traits) would interact with the level of directiveness and the level of emotional focus to enhance outcomes in well matched treatments.

This study again validated the previous findings and took them to an additional level. As opposed to looking at techniques broadly classified as “directive” and “nondirective” Karno, Beutler and Harwood applied a fine-grained analysis of process variables across both therapies as they were being conducted. The TPRS analyzes the therapy process according to relevant STS dimensions, including 1) the extent to which the therapist is attempting to increase or decrease emotional arousal, 2) the extent to which the therapist is focusing on insight- or behavior-oriented change, and 3) the extent to which therapy is directive or nondirective (Beutler, Sandowicz, Fisher, & Albanese, 1997). Each of these dimensions was found to interact with corresponding patient dimensions such as problem severity, subjective distress, and client coping style.

The interactions between process variables and therapy techniques was striking: Patients assessed as being high in resistant traits improved most when therapy was non-directive, and those low in resistance traits responded the best with non-directive therapy techniques. Similar results were found regarding emotional distress levels within session: patients with high emotional distress appeared to respond best to emotion-focused techniques, while patients low in emotional distress displayed the opposite pattern.

Collective analysis of the contribution of the matching dimensions alone accounted for 76% of abstinence outcome. A combination of therapist directiveness and low levels of patient impairment was also a positive predictor of patient change, independent of the fit of these treatment qualities with patient characteristics. However, when these two independent variables were added to the analysis, they accounted for an impressive 82% of unique variance contributing to drinking outcomes.

Of note, Karno and Longabaugh (2003, 2004, 2005) further replicated several of these findings in a re-analysis of Project Match data. Project match is the largest and most carefully monitored study on patient-treatment matching undertaken to date. It is telling, therefore, that the STS dimenions performed so well when many other matching dimensions had been rejected as significant predictors of treatment effects.

Beutler, Moleiro, Malik et al. (2003) constructed an even more fine-grained study of therapy-client matching within the context of a randomized clinical trial of chemically abusing and depressed patients. Forty stimulant-dependent, comorbidly depressed patients were assigned either to Cognitive Therapy (CT), Narrative Therapy (NT), or Prescriptive Therapy (PT). CT and NT were applied to maximize their contrasting qualities (such as focus, directiveness, and level of arousal). The results indicated that while all psychotherapies appeared to exert similarly powerful effects (the by-now redundant dodo bird verdict), STS-based PT produced significantly superior outcomes, and the STS matching dimensions were particularly strong predictors of both change in chemical abuse and depression. In this study, over 90% of the long term maintenance of decrease in depression (6 month follow up) was attributed to the STS factors. The maintenance of improvements in drug abuse was somewhat less closely associated with the STS dimensions, but still resulted in over a 60% predictive efficiency, even including the attenuating effects of drop out rates and non-compliance.

Development of Treatment Guidelines and Principles

Beutler, Clarkin & Bongar (2000) describe guidelines as following a series of steps, which generally are in the order of 1) identification and measurement of prognostic indicators, 2) assigning the context of treatment , and 3) managing risk. None of these steps can possibly be followed properly by a therapist without thorough, ongoing assessment of the client. The various dimensions and recommendations that evolve from the STS treatment planning system are bound together, not by reliance on an archaic theory of psychopathology or by a reliance on a stagnant set of techniques and procedures, but by the development of research-grounded principles of therapeutic influence. As research accumulates sufficiently that relationships among participant factors (patient, therapist, and sometimes family), relationship factors, and treatment factors, emerge as general or specific principles that can guide clinical work, they are easily incorporated into the STS system.

Thus, Beutler, Clarkin, and Bongar distilled 15 principles from a comprehensive review of over 2000 research studies on depression and drug abuse and then tested these principles using a cross-validation procedure. Their sample consisted of 284 patients, males and females ranging from age 17 to 79 who presented with a variety of problems and a range of functional impairment levels. These patients were given one of nine different structured treatments that varied in context, intensity, format, and modality. Treatments included insight and symptom approaches, pharmacologic and non-pharmacological approaches, couples and individual approaches, and short and long term approaches.

Prior to treatments, all patients were assessed on the STS dimensions by experienced clinicians using the Clinician Rating Form (Fisher, Beutler, & Williams, 1999) and through the use of standardized tests that captured these dimensions. All treatments were video recorded and reviewed by independent judges to determine if they were following the principles extracted from the research findings. The results confirmed the predictive value of 13 of the15 principles that had been extracted from several comprehensive reviews of psychotherapy literature on depression and chemical abuse. Five additional principles were extracted from a consensus panel of expert clinicians who addressed the issue of low frequency behaviors in which patients were dangerous to themselves or others. These final list of 18 principles identified patients who were most likely to benefit from treatment, the acts of therapists that would be most protective of dangerous patients, and 10 principles that offered guidance in how to fit the treatment to those patients who were judged to have a poor prognosis.

The movement to identify research-based principles, rather than research supported theories and models, has gained ground in the past several years. For example, in 2003 a joint task force (comprised of 45 scholars) of the American Psychological Association=s Division 12 and the North American Society for Psychotherapy Research was formed to identify what cross-cutting principles of treatment could be extracted from extant research literature to guide treatment among depressed, anxious, chemically abusing patients and those with personality disorders. The results were published in 2006 (Castonguay & Beutler). The joint task force found sufficient research evidence to expand and refine Beutler, Clarkin, and Bongar’s list of 18 principles to a list of over 60 research-informed principles that were extractable to clinical practice. Whereas the earlier list (Beutler, Clarkin, & Bongar, 2000) did not differentiate among the treatments for different disorders, nearly half of the principles identified by the joint task force were specific to the treatment of one of four problem areas discussed (depression, anxiety, personality disorders, substance use disorders).

Return to Home