Reflecting on Current Challenges and Future Directions in Psychotherapy: What Can be Learned from Dialogues between Clinicians, Researchers, and Policy Makers? Ä

Michelle G. Newman and Louis G. Castonguay The Pennsylvania State University

The challenges that confront contemporary psychotherapy and that have been highlighted in the previous series of articles rest on many competing interests. Unless genuine dialogues and serious efforts of conciliation are launched to address these challenges, irreconcilable factions may result between clinicians, researchers, and policy makers. A number of therapists certainly have deep concerns about the influence of empirically supported therapy research. On the other hand, many applied researchers are worried about private practitioners’ limited use of empirical findings. At the same time, those who make decisions about federal funding for psychotherapy research are concerned about monetary allocation. Nonetheless, the previous four panel discussions demonstrates that although there are some differences of opinion as to the optimal directions that should guide the future of psychotherapy, there are also many points of agreement. Further, researchers, clinicians, and federal policy makers in these panels have converging ideas as to how to address some of the challenges and unresolved issues that are salient to contemporary psychotherapy. When attempting to identify and address current challenges and issues, a central focus of each of the panel discussions deals with the empirical standards of randomized controlled trials. Influenced (if not regimented) by federal funding policies, these standards focus on maximizing internal validity at the expense of external clinical validity. As a result, most clinical trials have abided by a number of rules that have stirred serious controversies in the field. These rules include the reliance on DSM-IV diagnoses and the adherence to treatment manuals. For many, controlled trials have also failed to be informed sufficiently by theory. Moreover, debates over the merits of clinical trials have been fueled by the expectation that research findings soon may play a pivotal role in the reimbursement policies of HMOs and other third party payers. Preparation of this article was supported in part by National Institute of Mental Health Research Grant MH-58593. Correspondence concerning this article should be addressed to Michelle G. Newman, Ph.D., Department of Psychology, 310 Moore Building, Penn State University, University Park, PA 16802; e-mail: [email protected]

JCLP/In Session: Psychotherapy in Practice, Vol. 55(11), 1407–1413 (1999) © 1999 John Wiley & Sons, Inc.

CCC 0021-9762/99/111407-07

1408

JCLP/In Session, November 1999

Controversies related to the strengths and limitations of the controlled clinical trials are at the core of a more global source of concern about the current state and near future of psychotherapy: the link, or lack thereof, between research (basic and applied) and practice. Not surprisingly, the tenuous bridge between the work of researchers and clinicians was the most predominant theme within and across the four previous panels. The goal of this paper was to summarize the highlights of the panel discussions. These discussions addressed specific issues related to the viability and usefulness of controlled clinical trials, as well as the challenges that are associated, more globally, with the lack of connection between research and practice. The Influence of Federal Funding Policies Federal funding policies significantly influence the direction of psychotherapy research. As noted by Pilkonis, the goal of these policies is to ensure the safety of treatments and to provide evidence for their efficacy. To achieve this goal, such agencies tend to support small-scale, cross-sectional, time-limited, symptom-oriented, randomized controlled trials conducted by academic researchers. According to Pilkonis, the view that currently guides policy makers is that a small number of such trials are sufficient to provide evidence of the efficacy of a particular treatment. Many panelists expressed the belief that federal funding goals do not always further the development of the most efficacious therapies. For example, one problematic federal policy is to fund only research studies that utilize short-term therapy. Whereas such a limitation assumes that most clients can achieve change in response to short-term therapy, clinical experience suggests that some disorders (e.g., personality disorders), tend to require longer-term treatment to achieve meaningful change (L. Benjamin). Further, whereas many clinical trials exclude comorbidity, such exclusions do not reflect the clientele seen by private practitioners (B. Taylor), whose complications make it more likely that they will need longer-term treatment. Another focus of federal funding policy that came under fire was the sufficiency of symptom-change demonstration to gauge the efficacy of psychotherapies. Many panelists suggested that fundamental change would not be reflected by symptomatic improvement. Instead, a better means of assessing basic change would be in measures of interpersonal relatedness, functioning, and quality of life (S. Blatt, L. Benjamin, E. Jones). Moreover, panelists indicated that factors reflecting acute outcome (such as symptom relief ) might be distinct from factors that reflect long-term outcome (B. Arnow, P. Pilkonis, Z. Segal). An additional problematic issue raised by panel members was that federal funding sources do not appear to be interested in research geared toward understanding the mechanisms of change. The focus tends to be on outcome rather than on the delineation of the therapeutic components that lead to change (P. Pilkonis). Several panelists felt that this approach was shortsighted and that an understanding of the mechanisms of change would eventually better serve the federal goal of determining optimal outcome (S. Blatt, E. Jones). Because many scientists are aware that research on change mechanisms is less valued by federal agencies, some have attempted to meet outcome and change-mechanism goals in the same research grant (D. Hope). However, there are many factors that are likely to act against this goal. First, as noted by Kozak, if a grant writer attempts to add basic research to an outcome-oriented project, it substantially increases the cost. If such a grant is then submitted to a review board whose priority is outcome, the basic science component is likely to be removed to reduce cost. A second factor working against the attempt

Reflecting on Challenges

1409

to meet both goals in the same project is that it greatly increases the complexity of the project and makes it more vulnerable to criticism and more likely to be rejected (M. Kozak). Moreover, because the National Institute of Mental Health has been incorporated into the National Institute of Health, it may become more and more difficult to obtain funding for a grant that combines basic mechanism and applied research (T. Borkovec). Although federal agencies appear to be reluctant to support the combination of basic mechanism and applied research, some of the more recent funding directions are geared toward addressing other concerns raised by panelists. First, as indicated by Borkovec and Pilkonis, there is a movement to provide funding for actual clinical effectiveness research that is less controlled methodologically and more reflective of real-world therapy. Further, as noted by Pilkonis, the emphases will be on larger-scale community-based longitudinal studies that incorporate acute, continuation, and maintenance phases. In addition, there is a movement to include a broader array of outcome measures that go beyond symptoms (P. Pilkonis). Therefore, there is the hope that funding sources will promote research that is more likely to reflect what clinicians do. The Current Diagnostic Categorization Directly related to the empirical standards promulgated by current public funding policies, panelists discussed their reaction to the ways in which randomized controlled trials have maximized experimental control. One way that studies have attempted to do this has been to adhere strictly to DSM diagnostic classifications. However, many panelists recognized the weaknesses in the DSM-IV categorization system. One noted weakness is that there is no theoretical basis to DSM-IV diagnosis (L. Benjamin). In addition, it is often difficult to differentiate distinct diagnostic categories such as anxiety and depression (E. Jones). Moreover, DSM-IV diagnostic categorization is viewed as inadequate in aiding the determination of treatment for many disorders (J. Clarkin, E. Jones, J. Magnavita). As a result, many psychotherapists do not find DSM-IV diagnoses useful (J. Clarkin, E. Jones, J. Magnavita, G. Parry). As a solution to the problems with DSM diagnosis, many panelists felt that categorization of psychopathology might do better to follow a dimensional model rather than a categorical model (E. Jones). Benjamin suggested that we also need a theory-based diagnostic system that reflects normal and pathological development, as well as attachment. The general consensus was that diagnosis should do a better job at reflecting the client characteristics that clinicians often assess and that would guide the selection of treatment. Pros and Cons of Treatment Manuals Also related to the issue of maximizing experimental control, many of the panelists commented on the requirement to use treatment manuals in randomized controlled trials. There were a number of concerns related to this requirement. First, psychotherapy for some disorders (e.g., personality disorders) and reflecting some theoretical orientations (e.g., psychodynamic) is hard to manualize (L. Benjamin). Second, manuals create the impression that a particular therapy can be mastered easily and conducted in a cookbook fashion (B. Arnow, J. Clarkin, E. Jones, J. Magnavita). Third, because there is no infrastructure to gain artful training in an empirically supported therapy approach, clinicians are likely either to apply a manual in a rigid manner or not apply the manualized approach at all (B. Arnow). In addition, therapy manuals tend to specify techniques (S. Blatt) and to focus on specific symptoms (B. Arnow) and may oversimplify the complexity of treatment for a particular disorder (S. Blatt, E. Jones).

1410

JCLP/In Session, November 1999

Although there were many concerns about therapy manuals, there was an acknowledgement that some of the newer manuals have begun to address the weaknesses of the earlier manuals. In particular, the newer manuals tend to be more process and strategies oriented and are more likely to incorporate therapist-client relationship factors (P. Pilkonis). Many panel members felt that this trend should continue. Although panelists were happy with some of the recent innovations in manuals, there were a number of additional changes recommended by panel members. First, authors of manuals should make it clear that the therapy cannot be mastered simply by reading their manual and/or by attending a time-limited workshop (B. Arnow). Second, manuals should include sections on clients who are resistant or more difficult to treat (P. Pilkonis). Third, manuals should be constructed to address a variety of clinical issues that may arise, as well as the various stages of therapy (J. Magnavita). Finally, once a therapy approach has been developed and appears promising, training opportunities need to be created that allow therapists to master the therapy in an artful manner (B. Arnow, J. Clarkin). Such opportunities are likely to increase the dissemination, acceptance, and effectiveness of the manualized therapies (B. Arnow). The Use of Theory to Guide Intervention Research Another issue that was addressed by panelists was the role of theory in the development of treatment strategies used in randomized controlled trials. In general, panelists representing anxiety disorders felt that for the most part, theory had guided treatment research (T. Borkovec, R. McNally, B. Taylor). However, some of the anxiety panelists also felt that there were some ways in which theory had not guided empirically tested therapies. For example, Kozak noted that cognitive therapy tended to be based more on folk theory than on cognitive-psychology theory. He also noted that it was often the case that therapies found to be effective for one problem were applied to another problem with no theoretical basis. Moreover, McNally made the point that, at times, a theory had led to an effective therapy technique; however, the theory later turned out to be unsubstantiated. Nonetheless, the therapy technique continued to be used because it worked. Those on the depression panel generally felt that theory had not played a significant enough role toward research progress. Jones made the point that depression theories were inconsistent and contradictory and that such theories tended not to be empirically linked to treatment approaches. Similarly, Pilkonis indicated that there were no current theories to guide treatment for refractory depression. Blatt also felt that theories that were related to symptoms were less helpful and relevant than were theories related to actual life experiences, past and current, that may be at the root of vulnerabilities to psychopathology. Despite the mixed views on the role of theory, many panelists felt that there were some promising theories that had and would continue to lead to fruitful research endeavors. For example, McNally indicated that a “levels-of-analysis” theoretical approach (i.e., the view that explanation in psychology entails three levels of analysis: categorizing what the system accomplishes, determining the information-processing mechanisms that carry out the task, and analyzing how the system is wired in the brain) could be particularly helpful in guiding the conceptualization and treatment of psychopathology. In addition, Borkovec suggested that mind-body theories might help explain why psychotherapy and pharmacologic agents lead to similar outcomes. Both McNally and Blatt also referred to the strong influence of Freudian theory. In particular, Blatt indicated that current research on issues of self-definition and interpersonal relatedness is important to our understanding of the treatment of psychopathology. Panelists generally felt that the most useful

Reflecting on Challenges

1411

theories were those that reflected the nature and course of psychopathology, as well as the most effective treatment approach. Health-Maintenance Organizations Despite the concern that findings from randomized controlled trials have limited generalizability to clinician reality, HMOs are using such data in a fashion that does not acknowledge that psychologists are still in an early stage of psychotherapy research. As a result, there are a number of problematic trends within the HMOs. One such trend is to reimburse only for empirically supported therapies (L. Benjamin). Another potential problem is the blanket requirement that therapists employ brief therapies to all clients. Such a requirement does not account for more complicated or resistant clients. Panelists also expressed concern about the trend to view successful optimal outcome in terms of a reduction of acute symptoms rather than the larger goal of life enhancement (L. Benjamin, J. Clarkin, E. Jones). Further, many HMOs perceive personality disorders as nonexistent (L. Benjamin) or untreatable (J. Magnavita). Finally, HMOs tend to treat therapy as a portable skill that can be acquired by anyone—a “provider”—rather than as an art. In addition to being concerned about some of the trends, panelists were unhappy with the overarching goal of HMOs to spend as little money as possible in order to optimize short-term profits. Several panel members indicated that if HMOs spent a little bit more money up front, they could save more money in the end. For example, ignoring the treatment of personality disorders may end up costing the HMOs money in medical visits (J. Clarkin). Further, the panel indicated that it would be preferable if HMOs modeled health care after the British system, which has the broader goal of optimizing the allocation of resources to maximize the quality of care (L. Johnson, G. Parry). The Link between Research and Practice The challenging issues and controversies generated by controlled clinical trials are at the center of what was the predominant theme of discussion, both within and across panels: the connection of basic and applied research to practice, or more precisely, the general lack of clinical usefulness of empirical investigations in psychotherapy. In terms of basic mechanism research, panelists discussed the ways in which such research tends to influence the development of new interventions. Several important observations were made. First, basic research does not systematically influence the development of new interventions (M. Kozak, B. Taylor). This is mostly because basic scientists are interested in developing new theories and therefore direct applications of their work are not always obvious (M. Kozak). Another reason why basic research does not systematically influence applied-therapy research is that to make use of basic science, one needs to have the specialized knowledge to read and critically understand the published works of basic scientists (M. Kozak). As psychology is becoming increasingly specialized, more and more research is being published. Such specialization and increased publication makes it increasingly difficult for researchers to keep up with new innovations outside of their specialized area (R. McNally). Nonetheless, panelists felt that basic research was an important way to determine mechanisms of change in psychotherapy, and they had a few suggestions as to how scientists could make its applied value more apparent. One suggestion was for basic and applied researchers to collaborate (T. Borkovec, M. Kozak, R. McNally) and to try to conduct research in a way that would systematically address unanswered questions (B. Taylor). Another suggestion was for basic researchers to attempt to identify psychologi-

1412

JCLP/In Session, November 1999

cal risk factors for a particular disorder. Once the risk factor is identified, applied research can be conducted to target that risk factor (B. Taylor). If the treatment successfully alleviates the risk factor, the client will be less likely to relapse (Z. Segal). In addition to addressing the link between basic research and practice, panelists commented on the link between applied research and practice. One approach to this link was to ask researchers whether their own clinical practice influenced their research. Whereas several panelists indicated that many of the questions they tested in their research arose from clinical experience (T. Borkovec, M. Kozak, R. McNally), Taylor felt that his clinical practice had not influenced his research. The other way in which this question was addressed was the degree to which practitioners made use of empirical research. In response to this issue, Benjamin observed that the HMO requirement for empirically supported therapy techniques had led to a tighter link between research and practice than was previously the case. Nonetheless, the consensus was that the utilization of research by practicing therapists was generally low (L. Benjamin, J. Clarkin, L. Johnson, E. Jones, J. Magnavita, G. Parry, B. Taylor). Some of the factors that may explain the relative lack of utilization of research findings by practitioners have been linked to the empirical standards guiding clinical trials that were previously described. First, the populations addressed by randomized controlled trials are viewed as being dissimilar to “real world clients” due to the focus on DSM-IV diagnoses and restrictions on comorbidity. Second, therapy in research studies tends to be manualized. Manualized therapies allow the therapist less flexibility to deviate from addressing particular issues, from using specific techniques, from a specified timeline, or from following more than one theoretical approach (B. Arnow, E. Jones, G. Parry). This differs from the therapy conducted by practitioners who adjust their approach and length of treatment dependent on how the client responds (E. Jones, B. Taylor). Third, randomized controlled trials examine symptomatic improvement, and many practicing therapists tend to address issues such as interpersonal relatedness and general functioning (S. Blatt, E. Jones). Moreover, whereas real-world clients choose their therapist and type of therapy, there is no such choice in controlled trials. Other obstacles to an optimal link between research and practice include the fact that research publications are not informative about how the reader might conduct a particular therapy (J. Magnavita). In addition, there is no infrastructure for interested parties to learn empirically supported therapies (B. Arnow, J. Clarkin, L. Johnson, M. Kozak). Furthermore, many practitioners do not have access to academic libraries (L. Johnson) and therefore, little access to published research or the clinical measures used in them. Although there is a low utilization of research by practitioners, panelists had many ideas about how to make psychotherapy research more appealing to clinicians. For example, panelists recommended that researchers do a better job of advertising their findings (B. Taylor) and that they take the time to write separate articles geared toward practicing therapists (L. Johnson). Panel members also suggested that there be numerous inexpensive opportunities for empirically based therapy training using videotapes or the Internet (L. Johnson). In addition to suggestions about making current research more appealing, it was suggested that researchers employ greater variation in their methodology to include a whole range of approaches. For example, in addition to randomized controlled trials, such a range could include process research (L. Benjamin, J. Clarkin, D. Hope, E. Jones), basic research (T. Borkovec), case analyses, qualitative research (J. Magnavita, G. Parry), and naturalistic research (B. Arnow, T Borkovec, E. Jones, B. Taylor). Each approach would add a dimension to our understanding of a particular therapy, such as mechanisms of change, timing, context, efficacy, and effectiveness (L. Benjamin, Z. Segal). Another

Reflecting on Challenges

1413

suggestion was that researchers broaden their measurement focus to include measures of well being (D. Hope, E. Jones, P. Pilkonis), inner processes, developmental disruptions (S. Blatt), and long-term follow up (J. Clarkin). In addition, several panelists felt it would be important for researchers to gain ongoing therapist input about what they view as important (T. Borkovec, J. Clarkin). Another area of discussion pertained to methods that might motivate therapists to participate in ongoing effectiveness research. Such participation would require them to conduct continuing assessments of their clientele during therapy and at follow-up points. One suggestion was to create financial incentives. Another idea was to provide free continuing-education training in empirically supported therapy techniques to those willing to conduct research (T. Borkovec). Panel members generally believed that such clinician involvement was an important key to the goal of naturalistic and applied research. Conclusion Recent pressures for accountability and effectiveness coming from within and outside of the field of psychotherapy have forced clinicians to consider seriously the role that empirically supported treatments play in their practice. These pressures are not likely to disappear in the near future. It therefore seems essential that extensive efforts be made to help converge the work and interests of all parties involved in psychotherapy and, hopefully, to bridge the gap between therapists, basic and applied researchers, and policy makers. The most encouraging aspect of the prior panel discussions comprising this issue of In Session was that researchers and funding agencies have started to address some of the issues that have created the separation, and that there are some ideas put forth about how to continue to bridge the gap in the future. Panel members also had many suggestions about numerous additional changes that will need to be made before the gap can truly be bridged. Such changes will require a massive collaborative effort and include such things as: a) a better link between theory, research, and practice; b) ongoing and direct collaboration between clinicians and researchers; c) a greater priority to fund and apply basic and process research; d) a more-continuous, less-rigid view of psychopathology diagnosis and outcome assessment; e) a more-explicit recognition of interpersonal variables and personality disorders; and f ) more comprehensive-training opportunities in empirically supported therapies. It is hoped that these suggestions will lead to a refinement of current research strategies and, ultimately, to an improvement of our therapeutic effectiveness.