Psychology

https://doi.org/10.1177/2167702618805513

Clinical Psychological Science 1 –5 © The Author(s) 2018 Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/2167702618805513 www.psychologicalscience.org/CPS

ASSOCIATION FOR PSYCHOLOGICAL SCIENCEShort Communication/Commentary

We are pleased indeed to read the opinions of such an excellent group of clinical scholars. Each commentary on our article (Hofmann & Hayes, 2019; this issue) provided a thoughtful and suitable analysis that is likely to be a reflection of the field at large. There were some notable similarities but also interesting differences between these views.

Each commentary agreed with our basic message: Intervention science needs to move from the protocols- for-syndrome approach (governed by the fifth edition of the Diagnostic and Statistical Manual of Mental Dis- orders [DSM–5]; American Psychological Association, 2013) to process-based treatments. These processes need to be grounded in testable and heuristically useful theories. Interestingly, however, there were some dif- ferences in the implications and perhaps also the sig- nificance of our call toward process-based therapies.

Perhaps the most cautious view was voiced by Gerald Davison (2019). In essence, he saw our article “more as a restatement of previous scholarship than as novel and innovative,” suggesting that “it should be more clearly placed in historical and scholarly context” (p. XXX). We appreciate that context, and both of us have written exten- sively about the history of our field in other venues (e.g., Barlow, Hayes, & Nelson, 1984; Hayes & Hofmann, 2018c; Hofmann, 2011). We are pleased that students or others who do not know the history will have Davison’s response available as a well-crafted and succinct reference.

Like a walk up a spiral staircase, as the field returns to old themes in intervention science, it does so from a newly advantaged perspective. We are better able to see in hindsight what worked and what did not. In Davison’s (2019) reply, four fifths of the references are more than 2 decades old and nearly 60% are over 3 decades old. The practical and scholarly context we are responding to in our target articles includes what has happen in those decades.

Functional analysis began based on Skinner’s approach to the analysis of action in its historical and situational context. As we noted in the target article, it has been a guiding principle since the early days of behavior therapy and has been embraced by many notable scholars, including Davison. But something happened along the way. A search of the term “func- tional analysis” in Web of Science shows that if you limit the search to the fields of psychology and psy- chiatry, the number of articles that used the term last year is virtually that same as 20 years ago. It continues inside applied behavior analysis in a robust but limited form, but in the context of the enormous growth of

805513 CPXXXX10.1177/2167702618805513Hofmann, HayesLong Live Functional Analysis research-article2018

Corresponding Author: Stefan G. Hofmann, Department of Psychological and Brain Sciences, Boston University, 900 Commonwealth Ave., 2nd Floor, Boston, MA 02215 E-mail: [email protected]

Functional Analysis Is Dead: Long Live Functional Analysis

Stefan G. Hofmann 1 and Steven C. Hayes2 1Department of Psychological and Brain Sciences, Boston University, and 2Department of Psychology, University of Nevada, Reno

Abstract In this rejoinder, we discuss the commonalities and differences of the commentaries to our target article. Each commentary agreed with our basic message that intervention science needs to move from the DSM-governed protocols- for-syndrome approach to process-based treatments. Functional analysis has been a guiding principle since the early days of behavior therapy but lost its dominance with the rise of the latent disease model of psychiatry. This model gave rise to disorder-specific treatments with limited benefit to patients and science. We now have the tools and expertise to study human complexity grounded in an understanding of processes of change drawn from and fully applicable to the psychological level of analysis.

Keywords psychopathology, randomized controlled trials, psychotherapy

2 Hofmann, Hayes

intervention science generally, the interest in functional analysis is currently feeble. We join with our friend and colleague in an embrace of functional thinking, but we need also to deal seriously with why that solid start petered out and what will be different this time around.

In our opinion, above all, the toxic effects of a latent- disease model narrowed our vision and our science, strangling functional analysis in its intellectual crib. When modern psychiatry adopted structuralism for its nosology, psychological issues (such as emotional dis- tress or behavioral problems) became expressions of a latent disease. It was believed by many that biological psychiatry would eventually develop drugs to effec- tively treat these latent diseases. Billions of taxpayers’ dollars went into randomized controlled trials to test the efficacy of specific compounds for DSM-defined disorders. Creative psychological scientists, often under the broad term cognitive behavioral therapy (CBT), developed psychological models of the DSM-defined disorders and developed treatment approaches based on them, but the net effect was to foster syndromal thinking and its latent-disease model. Each year the percentage of clients receiving evidence-based psycho- social interventions decreased.

How and why did this happen? For one thing, many psychological scientists went along for the ride, driven by funding agencies and policies. Major psychological scientists served on the DSM panels. It was common for psychosocial treatments for DSM disorders to serve as separate arms of the many randomized controlled trials designed to test the efficacy of specific drugs.

This had notable scientific benefits: CBT became the most-researched psychological intervention. It became clear that the efficacy of these drug treatments showed low treatment specificity and produced effects that were generally disappointing, while the comparator condi- tion, often CBT, was at least as good as and often better than the drugs that were tested, had lower side-effect profiles, was produced less expensively, and often had better long-term adjustment.

Those data are hugely important and provide us all with a solid foundation for moving forward, but it needs to be noted that, as a public-health strategy, it has so far failed. Instead of fostering a new wave of dissemina- tion and use of evidence-based psychosocial methods, the exact opposite has occurred. It appears that once psychological scientists fully take on the assumptions of the latent-disease model, it is no longer possible to be part of a serious public-health discussion about human misery. In turns out that almost any outcome can be used successfully by the marketing arm of a half-trillion-dollar industry. We played a rigged game structured by forces and interests foreign to our field.

Although psychosocial interventions (especially CBT writ large) are now undoubtedly efficacious, this period

had clear negative impacts on our underlying science. Protocols and manuals trumped processes and mecha- nisms, and entire generations of psychological scientists were socialized into the assumptions of a latent-disease model. The field essentially lost its functional and con- textual behavioral roots. We moved away from identify- ing the crucial and controllable causal functional relationships for an individual client in the effort to succeed inside a protocol-for-disorder strategy.

But again we ask: How and why did this happen? It is here that we must part company with our colleague. We believe that the positive functional start of the evidence-based-treatment movement collapsed because these early models of functional analysis failed the field scientifically and practically. Much as an elderly patient with a weakened immune system will succumb to any one of a number of diseases, early behavior therapy succumbed to the siren call of protocols for syndromes because it had no robust and viable alternative to offer, given the limits of the day.

We did not review much of that later history in our article, nor did our esteemed colleague. The source of failure included all of the following and more: the lim- ited range of direct contingency principles; the lack of reliability in functional analysis; limited data on treat- ment components and kernels; the failure of classical statistical methods to deal with the individual; the absence of extensive and high-density longitudinal data sets; the absence of ready technology to record client processes in situ regularly over time; the absence of well-specified, robust, and empirically viable theories and models; weaknesses in the underlying basic sci- ences of genetics, epigenetics, neurobiology, emotion, culture, and cognition, among other areas; the lack of methods available to properly test moderation, media- tion, and processes of change; and bulky assessment instruments not designed for repeated use.

To put it simply, the field was not ready. Now, we believe, it is. That part is new, even if the core ideas we are arguing for certainly are not. In the words of Teeters and Dimidijan (2019), “not being new does not mean not being important” (p. XXX). Other than honor- ing our past, the primary reason to take the history seriously is to learn from the mistakes of the intellectual cul-de-sacs we entered.

In hindsight, the medicalization of psychological suf- fering needs to be seen as the dead end it was. But that will happen only if we now rise to the challenge we had not met as a field as the DSM–III (American Psy- chiatric Association, 1980) arose. What biopsychosocial processes are best targeted, and how, with this person, given this goal? In effect, we need a viable alternative to the DSM. That alternative will not be a better disease model, nor better microtheories of DSM disorders. What we need are broad and effective models of human

Long Live Functional Analysis 3

suffering and prosperity that specify the processes that need to be changed and tell how best to do so. That is what can rise to the public-health challenge pointed to by Kazdin and Blase (2011). If successful, it will lead to the functional clustering of people and issues but not as an armchair task. It needs to be data-driven, multidisciplinary, and grounded in nomothetic princi- ples driven from large sets of idiographic analyses.

The fact that the National Institute of Mental Health promotes alternatives such as RDoC (Insel et al., 2010) shows that our future is not going to be a simple con- tinuation of the past. RDoC is heavily guided by neu- roscientists, but its larger message is that a more process-based approach is back on the agenda. If we can take the next steps—finally stripping out the latent- disease model, confronting afresh how to alleviate human problems, and promoting human prosperity inside a functional, contextual, process-oriented model—we have tools at hand that simply did not exist 3 or 4 decades ago.

Complex network approaches (Hofmann, Curtiss, & McNally, 2016), for example, give us modern method- ological tools that are entirely consistent with functional behavioral analysis. Statistical methods for scaling indi- vidual data into nomothetic generalizations have been developed (Beltz, Wright, Sprague, & Molenaar, 2016; Fisher, Medaglia, & Jeronimus, 2018). Technological advances allow us to gather larger data sets on a single individual (e.g., though ecological momentary assess- ments). We know much more about functional princi- ples of behavior, cognition, emotion, motivation, culture, genetics, epigenetics, and neuroscience. Test- able and highly specified models of change exist, with large data sets relevant to understanding moderation and principles of change.

These more recent historical and contextual events have the clear potential to transform the field of inter- vention science in a way that was not possible only a few years ago. Thus, there might be a solution to an old problem right around the corner. The future of interven- tion science is bright, with exciting new possibilities built on the solid foundations of a more distant past but using the advancements in knowledge and methods that have occurred since. In our view, the very best way to honor our traditions and to stand on the shoulders of giants is to reach again for what was once out of grasp.

Teachman (2019) acknowledged the problem we out- lined and basically agreed with our solution. However, she expressed some initial trepidation to “jump into the water” (p. XXX) and fully embrace a process-based approach. As Teachman noted, one of the obvious prob- lems is which processes should be considered. We have begun to describe some of them in our recent book (Hayes & Hofmann, 2018a), based on the report of the

interorganizational task force on cognitive and behav- ioral psychology doctoral education (Klepac et  al., 2012). After some thoughtful deliberation, Teachman concluded that the time is ripe to jump. We agree. Her cautionary note to initially identify and train students in empirically supported strategies to target specific pro- cesses is well taken and is an example of how we can use what we have learned more recently to pursue older aims.

The last time clinical scientists took it upon them- selves to recommend strategies for specific psychologi- cal problems was at the creation of the infamous list of empirically supported treatments by the American Psychiatric Association’s Division 12 (Chambless & Ollendick, 2001). This has become perhaps the most contentious list in modern clinical science. The many reasons why this list has become so contentious should be well known to the majority of readers. The primary controversies centered around the fact that most treat- ments from “the list” were CBT-oriented, leaving other orientations on the sideline and marginalized. The list was dominated by DSM categories, and there was noth- ing in the requirements to prevent “purple-shirted desensitization” from making it onto the list as a result of a failure to insist on process evidence. A process- based approach avoids all of these problems. It is unlikely to be associated narrowly with any specific treatment orientation because the only entry pass needed is to provide empirical support for the treat- ment process and the underlying theory. It does not require an assumption of latent disease; in fact, it encourages the abandonment of that assumption, and it inherently organizes the field according to function- ally important differences.

Theory needs to be the foundation upon which we can build our intervention science, as acknowledged by Stephen Hollon (2019). He eloquently elaborated on the theoretical significance of these processes by associating them with Robert Sapolsky’s (2017) views that these processes may be distinguishable by the immediate, intermediate, and distal causes. We fully agree with Hollon that modern neuroscience is likely to inform the psychological processes that are targeted in therapy. Linking these processes to evolutionary sci- ence is fully consistent with our own ideas (Hayes & Sanford, 2015), and for the first time in modern CBT, so far as we are aware, our text on process-based CBT includes a basic chapter on principles of evolution sci- ence (Hayes, Monestès, & Wilson, 2017). Hollon’s (2019) hypothesis that cognitive restructuring is linked to higher cortical processes, whereas several “third-wave” (p. XXX) behavioral processes are more associated with evolutionary, conserved limbic structures, seems entirely plausible and is a good example of how tests of

4 Hofmann, Hayes

processes of change might bring together competing wings and traditions. Work by LeDoux (2000), in particu- lar, provides, in fact, some support for this notion.

Finally, we appreciate Teeters and Dimidijan’s (2019) call to consider the larger social and cultural context of an individual when examining the processes. The medical-illness model has isolated the individual from its context, thereby creating artificial groups while ignoring essential commonalities and differences. Human suffering and well-being can only be under- stood in the larger context in which the individual is embedded. Evolution occurs at multiple levels, in mul- tiple dimensions, and at different time scales, all nested and intertwined in a dynamic system (Wilson & Hayes, 2018). Our understanding of any one dimension, level, or time frame (such as how psychological processes foster or inhibit change within the lifetime of the indi- vidual) is dependent on our understanding of processes at other levels, dimensions, and time frames (such as how cultural contexts structure the assumptions, beliefs, and practices of the individual).

In fact, one can interpret our target article as a con- frontation of exactly that problem. In a recent article on how to make intervention science more relevant (Hayes & Hofmann, 2018b), we noted that dissemina- tion of intervention science is inhibited by a model that “oozes privilege” (Introduction, para. 6). Noting that “[c]lients were not demanding that they be labeled with four out of seven signs, or five out of nine symptoms” (Consider the Needs and Wants of Others section, para. 1), we asked, “Where are the clinicians and their goals in this picture? Where are the individual clients and their needs? Where are the human beings with lives unfolding as they are actually lived?” (Introduction, para. 8). The kind of research that emanated from the latent-disease model rode roughshod over personal and cultural beliefs, goals, and practices, and increasingly could be mounted only by academic medical centers in a handful of Western, White countries. A process- based approach opens the door to countries, research- ers, cultures, and ideas that have been silenced for too long by the hegemony of a latent-disease assumption and the need for large and supposedly “homogenous” (read “decontextualized”) groups that research on these imaginary entities demanded.

It is possible that treatment will become more com- plex in order to match the complexity of human suf- fering. The path ahead is unlikely to be smooth. But a process-based vision seems more likely to be progres- sive. We now have the tools and expertise to study human complexity grounded in an understanding of processes of change drawn from and fully applicable to the psychological level of analysis.

The spiral staircase of knowledge itself has circled intervention science back to its future. We now have the methods at hand to simplify complexity based on principles that illuminate the functional processes that lead to the forms of human suffering and prosperity we see. We believe that the time is ripe to use them.

Action Editor

Scott O. Lilienfeld served as action editor for this article.

Author Contributions

S. G. Hofmann and S. C. Hayes jointly wrote and approved the final manuscript for submission.

ORCID iD

Stefan G. Hofmann https://orcid.org/0000-0002-3548-9681

Declaration of Conflicting Interests

The author(s) declared that there were no conflicts of interest with respect to the authorship or the publication of this article.

References

American Psychiatric Association. (1980). Diagnostic and sta- tistical manual of mental disorders (3rd ed.). Washington, DC: Author.

American Psychiatric Association. (2013). Diagnostic and sta- tistical manual of mental disorders (5th ed.). Washington, DC: Author.

Barlow, D. H., Hayes, S. C., & Nelson, R. O. (1984). The scien- tist-practitioner: Research and accountability in clinical and educational settings. New York, NY: Pergamon.

Beltz, A. M., Wright, A. G. C., Sprague, B. N., & Molenaar, P. C. M. (2016). Bridging the nomothetic and idiographic approaches to the analysis of clinical data. Assessment, 23, 447-458. doi:10.1177/1073191116648209

Chambless, D. L., & Ollendick, T. H. (2001). Empirically sup- ported psychological interventions: Controversies and evidence. Annual Review of Psychology, 52, 685–716. doi:10.1146/annurev.psych.52.1.685

Davison, G. (2019). A return to functional analysis, the search for mechanisms of change, and the nomothetic- idiographic issue in psychosocial interventions. Clinical Psychological Science. Advance online publication. doi:10.1177/2167702618794924.

Fisher, A. J., Medaglia, J. D., & Jeronimus, B. F. (2018). Lack of group-to-individual generalizability is a threat to human subject research. Proceedings of the National Academy of Science. Advance online publication. doi:10.1073/ panas.1711978115

Hayes, S. C., & Hofmann, S. G. (2018a). (Eds.) Process-based CBT: The science and core clinical competencies of cog- nitive behavioral therapy. Oakland, CA: New Harbinger.

Hayes, S. C., & Hofmann, S. G. (2018b). A psychological model of the use of psychological intervention science:

Long Live Functional Analysis 5

Seven rules for making a difference. Clinical Psychology: Research and Practice. Advance online publication. doi:10.1111/cpsp.12259

Hayes, S. C., & Hofmann, S. G. (2018c). Survival circuits and therapy: From automaticity to the conscious experience of fear and anxiety. Current Opinion in Behavioral Sciences, 24, 21–25. doi:10.1016/j.cobeha.2018.02.006

Hayes, S. C., Monestès, J.-L., & Wilson, D. S. (2017). Evolutionary principles for applied psychology. In S. C. Hayes & S. G. Hofmann (Eds.), Process-based CBT: The science and core clinical competencies of cognitive behavioral therapy (pp. 155–171). Oakland, CA: Context Press/New Harbinger.

Hayes, S. C., & Sanford, B. (2015). Modern psychotherapy as a multidimensional multilevel evolutionary process. Current Opinion in Psychology, 2, 16–20. doi:10.1016/j .copsyc.2015.01.009

Hofmann, S. G. (2011). An introduction to modern CBT: Psychological solutions to mental health problems. New York, NY: Wiley-Blackwell.

Hofmann, S. G., Curtiss, J., & McNally, M. J. (2016). A complex network perspective on clinical science. Perspectives on Psychological Science, 11, 597–605. doi:10.1177/1745691616639283

Hofmann, S. G., & Hayes, S. C. (2019). The future of intervention science: Process-based therapy. Clinical Psychological Science. Advance online publication. doi:10.1177/2167702618772296

Hollon, S. D. (2019). Process trumps protocol: What I liked so much about Hofmann and Hayes. Clinical Psychological Science. Advance online publication. doi:10.1177/2167702618797111

Insel, T., Cuthbert, B., Carvey, M., Heinssen, R., Pine, D. S., Quinn, K., . . . Wang, P. (2010). Research domain criteria (RDoC): Toward a new classification framework for research on mental disorders. American Journal of Psychiatry, 167, 748–751. doi:10.1176/appi.ajp.2010.09091379

Kazdin, A. E., & Blase, S. L. (2011). Rebooting psychotherapy research and practice to reduce the burden of mental illness. Perspectives on Psychological Science, 6, 21–37.

Klepac, R. K., Ronan, G. F., Andrasik, F., Arnold, K. D., Belar, C. D., Berry, S. L., . . . Strauman, T. J. (2012). Guidelines for cognitive behavioral training within doctoral psychol- ogy programs in the United States: Report of the inter- organizational task force on cognitive and behavioral psychology doctoral education. Behavior Therapy, 43, 687–697. doi:10.1016/j.beth.2012.05.002

LeDoux, J. E. (2000). Emotion circuits in the brain. Annual Review of Neuroscience, 23(1), 155–184. doi:10.1146/ annurev.neuro.23.1.155

Sapolsky, R. (2017). Behave: The biology of humans at our best and worst. New York, NY: Penguin.

Teachman, B. A. (2019). Commentary on Hofmann and Hayes: The water looks inviting, but how and when do we jump in? Clinical Psychological Science. Advance online publi- cation. doi:10.1177/2167702618797104

Teeters, L., & Dimidjian, S. (2019). The promise of a par- ticipatory approach in clinical psychology. Clinical Psychological Science. Advance online publication. doi:10.1177/2167702618805082

Wilson, D. S., & Hayes, S. C. (Eds.). (2018). Evolution and contextual behavioral science: An integrated framework for understanding, predicting, and influencing human behav- ior. Oakland, CA: Context Press/New Harbinger.