Our Best Selves

The Herculean Task

Most of us psychologists laboring in the vineyards of therapy practice and education strive to establish mere competence among our students and colleagues. Even the assessment of competence is relatively new and challenging for professional psychologists (Kaslow, 2004). The competence of our graduates and the adequacy of our education have been typically assumed rather than verified. Competency requirements, now standard fare in other health care professions, will soon be on us and represent a transformational moment (Kaslow, Falender, & Grus, 2012).

Hill et al. (2017) and Tracey et al. (Tracey, Wampold, Goodyear, & Lichtenberg, 2015; Tracey, Wampold, Lichtenberg, & Goodyear, 2014) aspire to greater heights. Their pursuit of expertise strikes us as Hercules assuming another labor, like slaying the nine-headed Hydra or cleaning the Augean stables in a single day. Perhaps it is more akin to capturing Cerberus from the underworld and bringing him back to the light. Identifying and developing expertise requires more strength and courage than most mortals can muster. We admire and congratulate all of those taking on the Herculean labor.

Hill et al. (2017) have seized and synthesized disparate elements from a vast literature related to, but alas infrequently on, expertise. We suspect that our quibbles with their contribution are largely rooted in their ambition of trying to address expertise too broadly, and without particular context. It is not always clear how or whether the material fits. For example, what is “ability” in the conduct of psychotherapy, and which part of the reviewed literature is relevant to it? Why are years of experience and reputation included, and how do they relate to the authors’ definition? How well do the sections on therapist training fit with the stated goal of identifying expertise? It is not always evident to what purpose information about expertise would be applied and, at times, that makes it difficult to know what to measure. Furthermore, expertise is unnecessarily treated as an acontextual therapist factor, when performance in the conduct of psychotherapy is not only an individual trait but also an emergent feature of the psychotherapy interaction.

Points of Convergence

The psychotherapy research clearly favors the therapy relationship, responsively adapted to the diagnostic, and especially the transdiagnostic, features of the client (Norcross, 2011). We were delighted to see relationship skills and responsiveness at the top of Table 1 on the criteria and related measures for assessing expertise. A steady focus on the therapeutic relationship is consistent with Hill et al.’s (2017) proposed definition of expertise and with the process-outcome literature. Although the relationship develops differently in diverse treatment models and modalities, focusing at the relationship level spares us from getting bogged down in theory-specific details when addressing expertise.

In terms of the relationship, we would advocate, for training and assessment purposes, for more operationalization of those particular therapist behaviors associated with, and predictive of, client outcome. In place of the global, multidimensional working alliance and real relationship, greater specificity might help. On the short list would be empathy, congruence, collaboration, support/affirmation, requesting client feedback, repairing alliance ruptures, and the like (Norcross, 2011).

Likewise, in terms of responsiveness, educators and supervisors will naturally require more specific guidance than enjoining their students to “go out there and be responsive.” Expert therapists probably possess the ability to recognize certain client markers and in part consciously tailor therapy accordingly. These “when . . . thens”—when the client is highly reactant, then ease off the directiveness and emphasize his or her control—lie at the heart of a responsive repertoire. Expertise may exist in the responsive use of the multiple skills and methods supported by the research evidence (Hatcher, 2015).

It certainly could prove useful to know with confidence to whom to send a challenging client, someone for whom the “good enough” therapist might not be quite good enough. This is another possible application of therapist expertise, and one that further highlights the fact that expertise does not reside outside the therapy interaction. Ability, skill, professional competence, and effectiveness in the conduct of psychotherapy are concepts that require consideration of the context and the client. In session, these factors are heavily influenced by the client. For example, the ability to adapt to the needs of an angry or aggressive client, or a client with a personality disorder, is probably an important component of any measure of expertise.

We also concur with and delight in the emphasis on deliberate practice and feedback, both in the Hill et al. (2017) and in the Tracey et al. (2014, 2015) contributions. Deliberate practice may well make the master. Consistent with the research on expert performance in other domains, the amount of time spent in improving therapeutic skills significantly predicts client outcomes (e.g., Chow et al., 2015; Persons, Hong, Eidelman, & Owen, 2016). That dovetails with collecting—and then using—client feedback (Lambert, 2010). We echo the lesson that it is not mindlessly gathering client impressions and preferences that matters, but passionately working with those data to improve one’s performance and client outcomes. Multiple methods of feedback and deliberate practice probably lead to the most performance improvement.

Points of Contention

Hill et al. (2017) begin by proposing a reasonable definition of therapist expertise, “restricted to performance in the conduct of psychotherapy” (p. 7), covering “ability, skill, professional competence, and effectiveness” (p. 9), and conceptualized as a dimensional characteristic on which therapists vary, not a qualitative categorical difference. Throughout the rest of the article the authors work to organize an unruly literature to fit the components of this definition.

After reading Hill et al.’s (2017) operational and concise definition of expertise, we excitedly expected to find a review of the literature that addressed their definition, that is, expertise. Instead, to our surprise and disappointment, we discovered a literature review on clinical experience. Although Hill et al. acknowledge that experience is not equivalent to, nor highly correlated with, expertise, they review the literature that assesses experience (e.g., years of clinical experience, professional level) rather than expertise anyway. We sympathize with and understand their plight—the paucity of literature on psychotherapy expertise—but are puzzled by their use of therapist experience as a proxy for expertise.

Expertise and experience are barely correlated in the research literature (although as we age and acquire decades of experience, we are rooting for a strong positive correlation). Indeed, a recent large-scale study found that the barely existent correlation between therapist experience and client outcome had a negative sign (Goldberg et al., 2016). We are convinced by the bulk of the research findings and the conclusions of Tracey et al. (2014) that there is no robust evidence that experienced therapists routinely achieve better client outcomes than do inexperienced therapists.

We are not convinced of Hill et al.’s (2017) justification or reasoning for using experience instead. Yes, it would prove a substantially briefer review of the literature had they remained on the topic of expertise per se. But for authors cogently arguing for improved operationalization of expertise, to use an orthogonal criterion as a proxy seemed inconsistent and self-defeating to us.

While taking collegial exception, let us add a few other caveats and alternatives. In our interpretation of the research literature, therapist self-assessment and clinical experience should not be presented in Table 1 as criteria and related measures for assessing expertise. They bear no discernable relationship to performance! As one of us illustrates the point to students, “My grandmother drove an auto for 65 years and prided herself on her driving ability, and she was absolutely the worst driver I have ever witnessed.”

We respectfully disagree with the proposition that the therapist’s cognitive functioning, as an indicator of expertise, should be rated higher in “relative importance” (Hill et al., 2017, p.10) than client outcomes in Table 1. Scant evidence attests to its reliable contribution to actual performance or client outcomes in psychotherapy (Tracey et al., 2015). We share Hill et al.’s (2017) conclusion that experienced therapists exhibit “more complex, deeper, and more integrated knowledge structures and case formulation” (p. 19) than inexperienced therapists, but without a palpable link between therapists’ cognitive functioning and subsequent clients’ outcome, it seems as much as an academic exercise in longevity as a clinical demonstration of expertise. Nor are we convinced of reputation as a criterion or indicator of expertise. One of us (J. C. N.), who occasionally travels the international speaking circuit, chuckled upon reading that “invitations to demonstrate methods in workshops, videos, or books” (p. 29) served as criteria for expertise in the performance of psychotherapy in Table 1 (but again, we passionately root for that to be the case).

We draw a firm distinction between performing as an expert therapist and being an expert about therapy. This demarcation, paralleling that of operational/procedural knowledge versus declarative knowledge, may clear up some of the confusion in Table 1. An expert at therapy primarily proves more effective than colleagues with comparable patients; that is, the criteria are performance and enhanced client outcomes. These experts are laboring away in the field, possessing varied years of experience and probably manifesting few “reputation” indicators.

An expert on psychotherapy, by contrast, would need not even to practice therapy. When we survey “psychotherapy experts” on, say, the future of psychotherapy (Norcross, Pfund, & Prochaska, 2013), we have no reasonable expectation or evidence of their clinical performance or their clients’ outcomes. Rather, we are concerned centrally with their knowledge base, cognitive functioning, years of experience, credentials, and reputation (as indicated in Table 1). This is a distinction with a profound difference.

A gedankenexperiment (thought experiment) might clarify our distinction, which is lost on occasion in the Hill et al. (2017) presentation. Can a psychotherapist be properly characterized as an expert at psychotherapy without consistently achieving advanced outcomes with similar patients? We think not. But an expert on psychotherapy would not need to meet that performance criterion; reputation, titles, and academics would prevail. Several world-renown psychotherapy experts have confided to us over the years that, when participating in clinical trials or other outcome studies, they discovered that they were among the worst performing therapists in the study! They were experts about therapy, not experts in performing it.

Thus, we would respectfully recommend narrowing the list of indicators in the article and in Table 1 to those aspects of performance that take place and can be measured “in the conduct of psychotherapy,” (Hill et al., 2017, p. 7) an interpersonal pursuit with real clients. The sections of the article on relational expertise, multicultural competence, responsiveness, and empathy measured in the context of the specific client all hold sway here. Alternatively, two lists of criteria and indicators could be constructed: one for procedural expertise (expert at therapy) and one for declarative expertise (expert about therapy).

Moreover, we conjecture that the relation between expertise on psychotherapy and expertise at conducting therapy is close to zero, like that of the relation between research and teaching in universities (Hattie & Marsh, 1996) and between research and practice (Goldfried & Wolfe, 1996). That is not a negative association, just a null one. As adherents of the scientist–practitioner tradition, we are rooting and hoping for at least a modest positive correlation, but as scientist–practitioners, that is not what the science will probably tell us.

In any definition of expert on or about therapy, we reject therapist self-assessment of outcome as a legitimate or empirical criteria. Study after study attests to therapists’ inflated estimates of their clinical and diagnostic abilities (Lambert, 2010; Tracey et al., 2014). We are still searching for a single psychotherapist who admits that his or her clinical performance or relational skills are well below the norm. The Lake Wobegon effect (Kruger, 1999) apparently afflicts our discipline as well—all are above average in skill and performance.

On the other hand, perhaps accurate therapist self-assessment and subsequent efforts to improve may well be an indicator of expertise. Here we agree with Hill et al. (2017) that honest reflection on performance and remediation of deficiencies may in fact be associated with expertise and with Tracey et al. (2015) that a hypothesis-driven approach might be the way to do this with any accuracy.

Neither the Hill et al. (2017) nor the Tracey et al. (2014) articles make clear why therapist expertise should be pursued. Perhaps it is too obvious to be stated? But as we read we found ourselves frequently wondering, to what end?

Many other fields have clear indicators of expertise, but not much is said about the purpose served or whether it is relevant to psychological practice. As a reader of the Hill et al. (2017) article might conclude, is measurement of expertise primarily to inform therapist training and development? That would explain the inclusion of measures that are only distally related to expert behavior in the therapy session (e.g., therapist self-actualization). If, on the other hand, the purpose is to help identify people to whom we will refer clients, little is needed beyond indicators of effectiveness. In either case, it probably is not realistic or necessary to require expertise—performance in the top 10%. Outside of the community of Lake Wobegon, few of our trainees will achieve performance in the top 10%. Most clients will experience satisfactory outcomes with “good enough” therapists.

In the particular case of public health, we would suggest better use of the hypothetical expertise distribution by reliably ascertaining the bottom quarter of therapists in performance and outcomes. That is, ascertain those who exhibit relatively poor relational skills and have more bad outcomes than expected, and launch remediation efforts. Of course, identifying and helping the poor performers and simultaneously defining and promoting expertise will prove complimentary strategies, but we suspect the former would assist more suffering humans.

Our Best Selves

The person of the therapist will probably prove the elemental quality, the sine qua non, of the psychotherapy expert: that is for expertise in doing psychotherapy, not as an expert about psychotherapy. We probably grow the best therapists by selecting and then growing the best people. This unoriginal position has managed to survive decades of disembodied research (Orlinsky & Ronnestad, 2005) and rivalrous schools of psychotherapy (Norcross & Goldfried, 2005).

Mind you, that is not the perfect person or the one with the ideal life, on one hand, or the seriously wounded healers, on the other. Much like Maslow’s self-actualizers, our best selves have weathered adversities, confronted life, and struggled with its vicissitudes. They have likely benefitted from several courses of their own personal treatment and personal development activities. Our best selves emerge, in the words of George Eliot/Mary Ann Evans in The Mill on the Floss (1860), “from a life vivid and intense enough to have created a wide fellow-feeling with all that is human” (p. 527).

These and related personal qualities of the therapist are manifested in session and experienced repeatedly by patients. Elevating the consideration of the client in the definition of expertise, perhaps “expert” is a term that should be reserved for therapists who consistently repair ruptured alliances, successfully treat difficult clients, effectively manage their countertransference, and reliably rally when presented with data that treatment is not (yet) succeeding. Those therapists walk with the gods.

Our best selves evince an abiding, passionate devotion to the craft. Our reading of the research literature tells us, as it did Hill et al. (2017), that expertise is less about mastering the therapy method and more about the relationship, responsiveness, and commitment to improvement. In Thoreau’s (1854) words,

I learned this, at least, by my experiment [living at Walden pond]: that if one advances confidently in the direction of his dreams, and endeavors to live the life which he has imagined, he will meet with a success unexpected in common hours. (p. 427)

Here we experience deep respect for, and empathy with, our intrepid colleagues Hill, Spiegel, Hoffman, Kivlighan, and Gelso, and all those who struggle to study expertise, due to our limited ability to operationalize, let alone measure, these components of therapeutic expertise. Despite our reservations about a couple of Hill et al.’s (2017) conclusions and criteria, we confess that we would not have even attempted to accomplish what they have done.

Hill et al. (2017) wrote that “the overarching purpose” of their article was “to advance the dialogue about the assessment and development of expertise” (p. 8). They have done so admirably. We trust that our collegial points of convergence and contention may contribute in some small measure to that urgent dialogue and promote additional research on how best to grow expert therapists.