Vandenberghe, Luc. "Keeping the focus on clinically relevant behavior: supervision for functional analytic psychotherapy." The International Journal of Behavioral Consultation and Therapy 5.2 (2009): 209+. Academic OneFile. Web. 29 Mar. 2010.

Introduction
Functional Analytic Psychotherapy (FAP) is an experiential interpersonal treatment belonging to the third wave in behavior therapy. It is compatible with state-of-the-art cognitive behavioral approaches. It is different, however, in that cognition is understood in terms of verbal behavior. In addition, the influence of beliefs on other behavior is analyzed in terms of rule -following and as depending on the effect of real-life contingencies (Kohlenberg & Tsai, 1994). The translation of cognitive therapy strategies into a behavioral framework is not typical of FAP. It has, been a theme in a broader section of the third wave movement from its beginnings (e.g. Zettle & Hayes, 1982).

The hallmark of FAP, however, is its emphasis on clients' direct learning through experiencing their problems in-session, as opposed to learning about their problems. In order to appreciate this point, we may remember that cognitive behavior therapy concentrates on discussing actions, feelings and thoughts that arise in the client's daily-life settings. The main target for change is what the client believes and thinks while he or she is experiencing problems outside the therapy session. In contrast, the FAP therapist works directly on client behavior while it is affecting the therapist-client relationship. And only when therapeutic change is noticeable within the boundaries of the relationship will the therapist monitor and (when necessary) promote generalization of in-session improvement to daily-life settings.

The whole process rests upon the idea that the therapist-client relationship offers the therapist an opportunity to observe the client's clinically relevant patterns firsthand and to respond to them in ways that promote change. For instance, a client whose romantic life has been on hold for years because she only feels attracted to inaccessible men may develop a crush on her therapist. In order to be able to work on this, the therapist must quickly become aware that what is happening in the relationship is a sample of the client's daily life problem. The therapist will also need to identify exactly what the client does that contributes to the problem pattern, both in her daily life environment and in-session. In this way, those client behaviors through which the client unwittingly brings her problem directly into the relationship with the therapist are identified. It is also possible to determine what clinical improvement would look like if it were to happen in-session. The therapist needs to have a clear view on what such improvement may look like, because he or she will need to respond to that improvement when it happens. For instance, if the client were to label her positive feelings towards the therapist in a better way and relate to the therapist in ways that helped the therapist help her professionally, such in-session improvements would need to be reinforced.

Once therapist and client have agreed on what the target behaviors will be, the therapist will allow his or her reactions, which are the natural consequences of the client's actions, to affect the target behavior in-session. Sometimes the problem pattern will need to be evoked intentionally in order to give the client the opportunity to deal with it. The therapist may, for instance, appropriately express his or her positive non-romantic feelings toward the client so that she can react to them emotionally. The central process in FAP is to gradually shape improvement by patiently reinforcing progressive changes in the right direction.

Therefore, the biggest challenge for the therapist is to identify initial shifts toward improvement in client behavior. By missing slight in-vivo improvements or mislabeling them as problem behavior, the therapist may be responsible for stalling therapeutic change. In our example, a distracted therapist may react aloofly to an appropriate approach behavior by the client. The therapist may thus miss the opportunity to reinforce the client's first move toward relating in more productive ways to him or her. A complementary error may be committed. The therapist who does not identify the client's languishing approach behavior as related to her daily life problem may unwittingly reinforce it.

For thorough explanations of how to identify and classify in-session client behavior as being clinically relevant, the reader should consult Kohlenberg and Tsai (1991) and Kanter et al. (2008). For the purposes of the present article, however, it is sufficient to distinguish two kinds of Clinically Relevant Behavior (CRB): in-vivo occurrences of client behavior that is part of the client's problem and in-vivo improvements. As the examples of possible therapist errors given above make clear, it is crucial to immediately identify both types of CRB in order to make contingent responding possible. And the task of improving the therapist's focus on CRBs makes supervision of FAP therapists different in some fundamental ways. We could say that Rose's (1977) definition of supervision as assisting professionals in improving their therapeutic skills and helping them resolve problems they may be experiencing with their clients still applies. But the concepts of problems with clients and therapeutic skills take on new meanings.

In traditional cognitive behavior therapy, problems with clients are most often seen as a hindrance to treatment progress. They are to be avoided or otherwise dealt with quickly so they do not take away time from work on daily life issues. For this purpose, the therapist needs to learn to get problems out of the way as smoothly as possible. On special occasions, problems with clients are focused on differently, namely as special therapeutic opportunities. This is more likely to happen when working with personality disorders (e.g. Beck, Freeman, Davis & Associates 1990) or when a rupture of the alliance occurs (e.g. Safran & Muran, 2000). In FAP, however, work on problems between client and therapist is at all times the very fabric of the treatment process. It is therefore a fundamental rule of FAP to seek out, and when useful, intentionally evoke problems in the relationship that may be worked through for the client's benefit. In our example above, the therapist did not maintain a safe emotional distance from the client in order to keep her difficulties in dealing with romantically inaccessible persons from threatening the collaborative relationship. Instead, the therapist made the relationship closer, expressing positive feelings towards the client and thus evoking the client's difficulties.

The concept of therapeutic skill is also approached differently in FAP. Like the mainstream behavior therapist, the FAP therapist still needs conceptual skills to define classes of responses involved in the client's problems and to specify target behaviors and related contingencies. But when it comes to the treatment process, other skills are involved. As may be surmised from the example above, the therapist needs to respond continuously to the effects client problems and target behaviors have on him or her as a person. These skills include being watchful for, expressing and evoking emotions (V. Follette & Batten, 2000), dealing with emotions, interpersonal closeness and conflict, bi-directional communication (giving and receiving feedback), and discriminating and expressing what the therapist needs from the client in the relationship (Callaghan, 2006,a). Although intended for in-session CRB work, all of these skills are also related to core interpersonal abilities. Improving those skills may change the therapist's interpersonal style in fundamental ways.

All this brings out a parallel between personal growth and professional progress as a therapist. Consider a remark that may sound familiar to many supervisors: "When I compare myself today to that shallow, quiet girl I was only one year ago, I seem to be a completely different person now. And I attribute the change to this supervision experience." Another supervisee confessed: "I only understood what made me hide from my clients, or exactly why I did that, when I was able to share what I felt in this supervision group. It hurt, it hurt badly, but I became a better therapist because of it." Now compare these two statements to what clients often say at the end of therapy and you may find a clear resemblance.

School-based approaches to supervision serve as illustrations of how close supervision can come to treatment. Back when behavior therapy was still young, Rose (1977) described the use of behavioral group therapy as a supervision method. Supervision was also been described as a kind of treatment for the therapist in Rational Emotive Behavior Therapy (Woods & Ellis, 1996) and Dialectical Behavior Therapy (Fruzzetti, Waltz & Linehan, 1997). This is of course only metaphorically true, as the treatment does not concern the therapist's personal problems per se, but only his or her functioning as a professional. However, our point is that different schools of therapy use their treatment principles in supervision. This is not surprising at all, since these principles embody each school's understanding of the mechanisms of personal improvement.

When a treatment model specifies contingent reinforcement as the critical process for behavioral change, this is also reflected in the supervision strategies developed from this model. W. Follette and Callaghan (1995) described a procedure in which the behavior of the FAP therapist is shaped in-vivo during sessions with the client through direct contingent feedback provided by the observing supervisor. However, the parallel between FAP and this strategy of supervision is only partial. The FAP therapist does not shape the client's behavior in daily life settings, but rather responds to it when it occurs within the boundaries of the session. Still, a supervisor can use contingent reinforcement to influence the supervisee's behavior during supervision sessions in ways that will improve the latter's performance as a therapist. This insight has allowed a school-based conception of FAP supervision to evolve (Callaghan, 2006a; Tsai et al., 2008). The relevant literature will be discussed below under the heading Comparing the model to the state of the art.

The present article attempts to expand on existing school-based supervision practice within FAP. Criteria will be proposed for deciding when FAP-style contingent responding is a desirable supervision strategy and when other principles of change may be preferable. The proposed model distinguishes three different functions of supervision (Vandenberghe, 1997). Each of these functions should prompt a different choice of supervision strategies. The model evolved during an effort to introduce FAP in an undergraduate training program. Admittedly, it may be more advisable to teach FAP to seasoned therapists who have had extensive exposure to the therapist-client relationship and its vicissitudes. In the latter case, one can take advantage of sophisticated interpersonal repertoires and clinical wisdom shaped by years of in-session experience. These can provide skills and sensibilities that may need to be rearranged but can still serve as building blocks for learning FAP. As a result, training seasoned therapists may not give a clear picture of how much is involved in learning to identify, evoke and respond to in-session client behavior. In contrast, working with fledgling therapists made it clear how complex a task this can be. It is not the intention of the present paper to report on the training program or its outcome, but only to describe the model for supervising FAP that was developed in the course of it.

The first context: introducing skills.
Supervisors of inexperienced therapists need to make sure that the therapists know what to do. At this stage, strong instructional control over therapeutic activity is needed. As long as the student does not know what to do, the instructor retains therapeutic responsibility and does much of the thinking. He or she explains or shows how to proceed in developing and using the client case conceptualization and in making interventions. This protects the client from the obvious risks involved in a beginner's lack of experience. At the same time, it allows the novice to practice real life therapy without having the needed experience. By applying the conceptual and technical skills as instructed, he or she will have the opportunity to learn from experience. This condition mimics mediation therapy, in which the behavior analyst technically prepares a parent or other caretaker who will implement treatment. This is an approach with a long tradition in applied behavior analysis (e.g. Moreland, Schwebel, Beck & Wells, 1982).

Often the needed instructional control is provided through discussion of session recordings and reports. Scrutinizing these recordings and reports, the supervisor can monitor how well supervisees followed through with previous instructions and then lay out what to do in the next sessions. In reviewing what happened in-session, the supervisor may detect, for instance, that the therapist failed to attend to an opportunity to reinforce an in-vivo improvement of the client. When this failure is due to a lack of conceptual or technical skills, the first context is invoked: the therapist does not know what to do. In this case, the supervisor needs to give clear instructions, and after the following session, needs to check to see if they were followed. As an example, he or she may need to explain to the therapist how to trace parallels between what happens in-session and the client's daily life issues or how to respond therapeutically to the client's difficulties and improvements.

One potential risk at this level is that the mediator may do literally what he or she is instructed to and does not learn from practice. An extreme case would be when the therapist follows instructions regardless of unforeseen developments during the session. Excessive dependence on rules is particularly deleterious for therapist development. In the first place, it hinders learning from practice. Rigid instructional control can overrule the effects of experience that should change and enrich skills. In the second place, excessive rule -following makes interpersonally directed psychotherapies like FAP virtually impossible. It entails decreased responsiveness to changes in the interpersonal contingencies during the session and thus to CRBs. When this happens, the skill the supervisee learned is following instructions. This skill will certainly not turn him or her into a skillful therapist.

The supervisor-supervisee relationship may inadvertently provide contingencies that promote rigid rule -following and thus keep the therapist from acquiring conceptual and technical skills. As an example, therapist rule -following may easily come under the control of how deviating from instructions might influence the evaluation their supervisor may give them. As will be discussed further on, the supervisor can help avoid this by selecting instructions that promote contact with the in-session contingencies and weaken excessive control through approval and disapproval during the supervision encounter.

The second context: overcoming therapist difficulties.
Even therapists who reached intermediate and advanced skill levels may show dysfunctional avoidance patterns or repeatedly misunderstand a certain type of event in-session. Myths and taboos about therapy which therapists bring from their academic training (Pope, Sonne & Greene, 2006) or overly generalized assumptions and rigid viewpoints stemming from their personal background (Ellis, 1984) may put therapists on the wrong track. They may then fail to apply the skills they have already mastered, or apply them inadequately.

Therapists may report rules such as "I must hide that I'm upset about what my client said, or I will lose credibility" or "I am younger than my client. I have no right to challenge her" to justify not tackling a CRB. Supervision must address the therapist's fears and misconceptions lest they obstruct his or her development. Sometimes dysfunctional rules may be discussed and clarified in a single meeting and the effect of the discussion on the therapist's in-session performance can be monitored subsequently. But many difficulties will require more time and effort to analyze and remediate. In doing so, supervision may mimic traditional cognitive behavior therapy.

Again, this comparison is only partially valid. Although cognitive therapy techniques of challenging automatic thoughts, labeling distorted interpretations of reality and restructuring irrational beliefs are particularly recommended in this context, they need to be used specifically to improve the therapist's work in-session and not to heal his or her emotional problems. The goal of supervision on this level is to eliminate dysfunctional verbal control and avoidance patterns that compete with an adequate focus in-session. This entails progress which needs to be visible in therapy reports and recordings. Factual in-session improvement in noticing, evoking and responding to CRBs will signal that the work in supervision was effective.

A danger at this level is that by targeting supervisees' fears and unreasonable assumptions, supervision may slide into unsolicited cognitive therapy for the supervisee. The supervisor can avoid this danger by proposing clear targets for change. Besides this, a restriction of the goals of supervision to the professional realm must be made explicit in the supervision contract.

The third context: the shaping of a therapist.

Finally, a skill may not blend in well with the therapist's interpersonal repertoire. This may be the case when a therapist can use the skill while following a rule, but aspects of his or her style hinder spontaneously putting these skills to work. The therapist may know how to focus on CRBs that occur in interpersonal situations like conflict or intense emotional closeness. However, when actual therapist-client conflict occurs in session, behaviors other than focusing on CRBs may come much more naturally to the therapist. In these cases, the therapist has learned the skills on an intellectual level, meaning he or she can execute them under verbal control, but has not really made them his or her own. An example would be the therapist whose failure to work on an alliance rupture is related to his or her detached interpersonal style.

The novice starts out following rules. But in building up clinical experience, his or her behavior undergoes contingency shaping. Thus, clinical skills initially used under instructional control gradually come under the control of the contingencies. Extensive exposure to in-session contingencies gradually provides the needed integration of skills in the therapist's repertoire. However, this process may be long and unreliable. Aversive initial experiences with clients may, for instance, adventitiously shape patterns of avoidance behavior in therapists. Therefore additional experiential learning may be desirable.

The supervisor-supervisee relationship itself offers opportunities for such experiential learning. It provides interactions that will be functionally similar to what happens when the supervisee is with his or her client. These can contribute to shaping the subtle repertoires that make a skillful therapist. Dealing with conflict, closeness and disclosure, for example, are as important in supervision as they are in therapy. To provide experiential learning opportunities, the supervisor starts out by identifying therapist problems and targeting behaviors in recordings and reports of therapy sessions. Skills that can be targeted in this context are diverse. They include such subtle abilities as seeing positive aspects in the other's behavior or being responsive to changes in the relationship. The supervisor then watches for functionally similar behavior in supervision encounters. After identifying such parallels, the supervisor can respond in-vivo to supervisee behavior to promote interpersonal repertoires that will be more effective for therapy. As will be discussed more extensively further on, supervision in this context mimics FAP itself.

This context is not, however, risk-free. It is the context in which the boundaries of supervision are hardest to maintain. A clear supervisee case conceptualization that makes sense of the therapist's problems and learning goals (Callaghan, 2006,a) is essential. Also, the supervisee's responsibility to give feedback to the supervisor when he or she feels boundaries are being crossed must be openly discussed and set out in the supervision contract. The supervisee must be aware of his or her right to set things straight when the supervisor unwittingly invades his or her private life or targets issues beyond the scope of supervision. Asserting limits in a close relationship is an example of a skill that is also essential in managing the intimacy of the therapist-client relationship. At the same time, it is a skill the supervisee can work on in-vivo in his or her interaction with the supervisor.

Applying the model to supervising FAP.

Comparing the model to the state of the art.
Are the three contexts specified above relevant to state-of-the-art FAP supervision? A quick overview will allow us to answer this question. Callaghan (2006,a) emphasizes what he calls fappervision. This term refers to in-vivo strategies, as in the third context of our model. But his discussion of supervision also includes training in conceptual and treatment principles, comparable to our second level and "instruction in attempting strategies that may be more effective in the next session" (Callaghan, 2006,a, p. 422), which corresponds to our first context. More recently, Tsai, Callaghan, Kohlenberg, Follette and Darrow (2008) distinguish between "knowing that" (intellectual knowing) and "knowing how" (emotional knowing). The former is accomplished via instructions, reading assignments and feedback on performance, among other strategies. The best fit in our model is with the first context. But if we assume that "knowing that" also involves changing mistaken assumptions, part of the work may also shift to our second context.

The emotional knowledge discussed by Tai and cols. is acquired through shaping and modeling in the context of the supervisory relationship. Knowing how to respond to a CRB is related to subtle issues. These include being in touch with one's feelings and sensitive to one's impact on the client. Instead of learning about such issues, one can experience them directly in the supervisory relationship. The authors describe two strategies. One is contextual modeling by the supervisor, with the supervisee as a participant in the interaction. This form of modeling is called contextual "because it is based on what is happening in the moment in the relationship" (Tsai et al., 2008, p. 173). The other is described as evoking and reinforcing in-vivo improvements in the supervisee's target behavior in the relationship with the supervisor. The latter strategy best matches Callaghan's fappervision and our third context.

The same distinction between the experiential and intellectual ingredients of supervision is made by V. Follette and Batten (2000), who also illustrate the dynamic shifting between the two. They suggest, for instance, that after exploring an issue experientially, the supervisor may shift to a didactic stance. This shifting can also be translated into our model. As an example, an issue that was worked on experientially in-vivo in the third context would be brought down to the first context, where the supervisor would teach how to take advantage of the new learning in future sessions. Alternatively, work could be carried on in the second context if the supervisee demonstrated erroneous conceptions regarding the application in-session of what was learned in third-context supervision.

This dynamic shifting is compatible with the needs criterion our model is built on. According to this principle, the three contexts are not developmental stages the supervisee grows through. Rather, they occur in response to the demands of a particular situation and may intertwine in any single supervision session. Admittedly, the first context is more applicable to fledgling therapists, but more seasoned practitioners can also benefit from learning what to do in new situations. Similar comments can be made concerning the other two contexts.

Summing up, the present model would appear to be compatible with the FAP literature on supervision. It also promises to make a useful contribution of its own, as will be discussed in the sections below. The explicit separation of the three contexts makes selecting the best supervision strategy easier. Being aware of the most relevant context at any given moment will also help the supervisor prepare for the specific challenges that are to be expected on that level. It will make it easier for the supervisor to decide how best to meet the needs of the therapist.

Keeping the focus on CRBs: First context.
As discussed before, following instructions is an excellent way to lose sight of CRBs. W. Follette and Callaghan (1995) have pointed out that general rules for therapists do not cover unique circumstances, and they do not work for all therapists. As we have seen, the authors resolve the issue by using direct shaping in-session, so the reliance on verbal control is greatly reduced. However, instructions are also useful as they may aid the therapist in knowing what to do (Tsai et al., 2008). Rules help the therapist remember to watch for, evoke and respond to CRBs. Our concern in this context is how to avoid rigid rule -following in supervisees who depend on instructions. Vandenberghe (1997) approached this question in terms of the distinction between plyance, which is maintained by approval in the supervision session, and tracking, in which natural consequences of following these instructions shape the therapist's behavior.

Zettle and Hayes (1982) defined plyance as rule -following maintained by social reinforcement, and tracking as rule-following reinforced by contact with the natural consequences of following the instruction. The natural consequences of the therapist's behavior will be changes in the case conceptualization or the client's behavior. For instance, a natural reinforcer contacted by watching for CRBs would be the actual identification of a CRB. As another example, evoking CRBs should be reinforced by the swift occurrence of a workable CRB. To promote tracking, instructions should be given that prompt more sensitivity to such natural consequences.

The supervisor can give instructions that must be completed with in-session information before the therapist can decide how to proceed. An example of an instruction that promotes tracking could be: "Disclose to your client an effect she had on you and watch her react. Then compare that reaction to the information in the case conceptualization before deciding what to do next." As an alternative technique, the supervisor can ask the questions a more experienced therapist would ask herself: "Would this behavior contribute to solving this particular daily-life problem? And how would you respond to it?"

As another strategy, the supervisor can instruct the therapist to observe and label the effects clients have on him or her. An example could be: "Label all feelings you become aware of and ask yourself what the client just did when you felt them." This instruction will help the therapist focus on functional classes of client behavior. A functional understanding of client behavior will be easier because the effect of the behavior defines its function. In-session, this effect will necessarily be on the therapist. Hence instructing the therapist to explore his or her feelings in the exchanges with the client increases his or her chances of coming across CRBs and grasping the contingencies they are related to.
Keeping the focus on CRBs: Second context.

Imagine a client who bends over backward to please others to avoid being abandoned. In therapy, she agrees with all the therapist's statements and suggestions and unquestioningly accepts all assignments and all proposed activities. Now imagine that the therapist is following a rule like this: "For therapy to go ahead, I must at all times support collaborative behavior." Under the influence of such a rule, the therapist may not perceive the client's submissive behavior as a CRB. The rule specifies a desired process and the client's pleasing behavior will fit the process too well for the therapist to perceive it as an in-vivo problem behavior.

Rules that hinder attunement to CRBs do not need to be irrational or outlandish. Often, rigid rule following even looks like prudent practice. This makes it harder for the supervisor to identify it as related to the therapist's difficulties. An example could be a belief like "Feeling sexually attracted to a client would be horrible. I would be a piece of filth if that happened." Even theoretically and empirically well supported rules become counter-productive in specific situations when taken literally and absolutely. This point deserves special attention, because the supervisor must be aware that quite reasonable principles may become a hindrance when they are used unbendingly.

Examples from other theoretical approaches in the literature show that experienced therapists do intentionally depart from well established rules. Goldfried (2000) recalls a demonstration for students in which sticking to an assertiveness protocol became a hindrance. It kept him from what he felt he should be doing, namely addressing the lack of assertiveness with which his client was treating him. Zurr (2007) reports the case of a male therapist who was asked by a female client to hold her hands while she was processing her molestation experience. Theoretical assumptions did not help him figure out what to do until supervision alerted him that he could discuss the question with his client. By opening up with the client, he not only got his answer, but also directly addressed her control over emotional, physical and sexual boundaries in relation to himself. From a FAP perspective, we would contend that the therapists in these two anecdotes broke rules so they could focus on what was happening in the relationship.

The examples above illustrate that assumptions on how to conduct therapy should be flexible (Ellis, 1984). This also goes for the basic assumptions of FAP. Paradoxical as this may seem, even these assumptions can take the focus away from CRBs. Rigidly and literally following a FAP rule like "I must share my interpretations with the client" may keep the therapist from detecting a CRB. That may be the case when a client's daily life is excessively controlled by cues given by others. When this client is overly receptive to and eager for interpretations by the therapist, that may be an in-vivo problem behavior our therapist may not detect.

The following vignette illustrates how a therapist's difficulty in focusing on CRBs was tackled by helping the therapist identify his bias and correct it. Pedro, a 24-year-old undergraduate therapist, was treating Maria , an older child-abuse survivor who had difficulties in trusting people and felt excluded in most social situations. When she confessed she had lied to him about being single, and that she was in a relationship with another woman, he told her that her homosexual option was avoidance behavior related to her sexual abuse history. Maria had hidden personal information for which she might be judged. Taking the risk of trusting a person she had been close to for some months (Pedro) with such information looked to the supervisor like an in-vivo improvement because such opening up could promote social inclusion.

The supervisor disputed the psychopathological assumption before Pedro was willing to consider that he had missed the opportunity to reinforce an in-vivo improvement. A balance sheet with arguments in favor and against labeling Maria's disclosure as a CRB and the downward arrow technique were used to address the theoretical justifications Pedro used. During this work, it became clear that Pedro used pathologizing interpretations of client disclosures to reestablish the distance between himself and the client after a client had shared personal information with him. Subsequent occurrences of this escape strategy were easily tackled. They gave way to experiential work in which Pedro learned to deal with intimacy in relationships.

Keeping the focus on CRBs: Third context.
As discussed above, FAP supervision comes into its own when the supervisor uses the relationship with the client as a space for in-vivo learning. Third context remediation work is illustrated in the following paragraphs. As part of her graduate work, Mandy, 31, was supervising the first attempts at therapy by Ana, a 20-year-old undergraduate, at a free community clinic.

Mandy observed that Ana did not identify CRBs that involved her clients' expressing needs or dealing with conflict. Ana punished what Mandy thought were reasonable attempts by clients to give her feedback about her cold, uncaring attitudes. Case conceptualizations in hand, Mandy argued that giving feedback to the therapist was an obvious in-vivo improvement for several of these clients.
Ana, however, was able to defend her attitudes and Mandy, as a first-time supervisor, was careful to give the therapist as much credit as possible and not to impose her own perception. However, during later supervision encounters, the pattern was repeated. When Mandy discussed Ana's intolerance of poorly educated clients' non-standard language, Ana once again rejected the criticism out of hand. And the same happened when Mandy addressed her intolerance of no-shows by clients who could barely afford the often hours-long bus rides from the slum to the clinic. When Mandy finally shared the feeling of helplessness this caused her, Ana became surprisingly upset and considered giving up training altogether. While discussing what was happening, they found out that Ana reacted excessively to negative evaluations in all areas of her life. To function well in interpersonal settings, she depended on signs of deference from others, such as coming on time (which was difficult for clients who depended on long irregular bus trips) and addressing her respectfully (including the use of polished language), which signaled that no evaluation was coming up.

In this case-example, longstanding avoidance strategies which had been successful in previous situations were a threat to the therapist's efficiency. As the supervisor spoke in an educated manner and was punctual, the required safety signals had always been in place in the supervision relationship. Mandy had also reinforced Ana's escape behavior by endorsing the reasons Ana gave for her rigid attitudes and allowed her immunity from evaluation in supervision. But finally, Mandy's sharing the impact Ana's behavior had on her provided an in-vivo learning opportunity in which Ana needed to deal with negative feedback. This observation finally made it possible to include this learning goal in Ana's supervisee case conceptualization.

Direct shaping of the relevant therapist repertoires was chosen as the way to remediate the problem. It was agreed that Mandy would respond contingently to improvements in Ana's receiving of feedback and in her way of dealing with a lack of deference during future supervision encounters. With contextual modeling chosen as the strategy of change, Mandy would need to respond firmly and compassionately when Ana criticized her or at moments when Ana lacked deference to her supervisor. These responses from Mandy would then serve as a model for Ana's behavior toward her clients.

The third context is not only relevant for remediating pre-existing problems with the therapist's style. It can also serve to shape new behavior in interpersonal situations the therapist has never been exposed to before. Behavioral deficits may occur for the first time when the therapist starts the practical part of his or her training or shifts to a new client population. Another special situation is when work with a specific client group shapes dysfunctional therapist behavior. The latter risk has perhaps been best described in the literature on treating borderline personality disorder (Masterson, 1976; Linehan, 1993; Fruzzetti, Waltz & Linehan, 1997).

In a series of case studies, Sousa and Vandenberghe (2007) identifie d two categories of inadequate behaviors in inexperienced therapists who treated borderline clients. These therapist behaviors were (1) avoiding unpleasant interpersonal experiences and giving elaborate reasons for doing so and (2) making dramatic demands for progress while exaggerating difficulties and rejecting reasonable options. These two categories of behaviors hindered both therapy and supervision. Paradoxically, but in line with FAP principles, this provided in-vivo learning opportunities for the supervisee during supervision sessions. The supervisor was able to identify supervisee behaviors toward her that belonged to the same classes as the therapists' undesirable behaviors towards their clients. As an example, some therapists would demand unreasonably rapid progress from the client and miss small but important in-vivo client improvements. As supervisees, these therapists would call the supervisor at inappropriate times for immediate solutions. Contingently responding to this behavior in-vivo in the supervisor-supervisee relationship would than be the strategy of choice. Generalization of the changes in the therapist's behavior to the client can than be monitored by the supervisor.

The Sousa and Vandenberghe (2007) study favored the perspective that therapists acted the same way toward the supervisor and toward their borderline client. They showed similar counter-productive behavior (namely experiential avoidance accompanied by dysfunctional reason-giving and unreasonable demandingness) with both their client and their supervisor. This comparison allows the supervisor to shape better ways of responding in-vivo in the supervisory relationship. The supervisee can then put these better ways of responding to work in his or her relationship with the client.

However, a different comparison is also possible. We could observe that the supervisee's behavior toward the supervisor is similar to the borderline client's behavior toward the therapist. In the latter case, the process in therapy (dysfunctional avoidance and unreasonable demands by the client towards the therapist) parallels the process in supervision (avoidance and demands by the supervisee toward the supervisor). In this case, contextual modeling can be used (Tsai et al., 2008). Both options allow for an experiential approach in which resolving the problem with the supervisee may lead to a resolution in the therapist-client relationship.

Conclusion.
We can now abstract a set of suggestions from this model. The aim is to make it easier for the supervisor to decide, at any particular juncture with any particular supervisee, what he or she should do to keep therapists focused on CRBs.

One suggestion for the first context is the use of specific types of instructions. Instructions that select actions leading to natural reinforcement for the therapist in-session are to be preferred. When a therapist obtains sufficiently reinforcing effects in-session while following through with supervisor instructions, this will lead to increased control by in-session contingencies over the therapist's behavior. In other words, he or she will continue using the instructed skills, not because of the supervisor's control over his or her rule -following, but because these skills work for him or her in-session. How can the supervisor predict if selected instructions will help the therapist to contact natural reinforcement? Information about both the supervisee and the client is needed for such a decision. For instance, the supervisor can provide the therapist with instructions for actions that are well within his or her technical reach, and which will yield immediate results with the client, given what the supervisor knows about the client. An example might be asking an affectionate therapist to increase closeness in response to in-vivo improvements of a client who finds closeness desirable.

Another specific class of instructions the supervisor may use in the first context is a category we call incomplete instructions. The supervisor does not state the entire action to be undertaken by the therapist. The latter will need to complete the instructions with information to be identified in-session. As an example, the supervisor can instruct the therapist to give the client feedback about behaviors that may help solve the problem for which she sought treatment. To make this instruction workable, the therapist will need to think the client case conceptualization over and engage in keen observation of what is happening in the relationship.

Alternatively, the supervisor can avoid using instructions and instead ask direct questions that prompt the therapist to stay closely attuned to the client's behavior. An example would be to ask the therapist what in-session behaviors he or she imagines the client may also emit in the relationships

where the client's daily life problems are most salient. Therapists then learn to make theirinterventions on the basis of the answers they find to these questions.

A related technique that can be used in this context is having the therapist label the effects the client has on him or her. The supervisor may, for instance, ask frequent questions about what the therapist is feeling in-session with the client. Such questions will make the therapist focus on those client behaviors that have interpersonal impact. And these are most frequently the client problem or target behaviors. In answering such questions, the therapist learns to monitor his or her feelings towards the client. As a result, the therapist's sensitivity to promising in-vivo learning opportunities as well as his or her awareness of subtle client improvements will increase.

When the supervisor faces problems in the second context, the use of traditional cognitive behavioral techniques is recommended. To challenge therapist assumptions and prejudices that compete with CRBs for control over the therapist's behavior, the downward arrow technique, Socratic questioning and behavioral experiments may be used. Dysfunctional rule -following should be discussed explicitly and replaced with more appropriate practices and more flexible rules.

In the third context, one typical technique is in-vivo shaping of supervisee behavior. Supervisee target behavior is evoked and shaped in the relationship with the supervisor. Typical examples of target skills include responding contingently to supervisor behavior and attending to the supervisor's needs in the interaction.

Contextual modeling was discussed as a second technique for the third context. As an example, the supervisor can respond compassionately when the supervisee admits a failure committed in-session. In this way, the therapist can learn to react compassionately to his or her client when the latter opens up and admits a shameful mistake in another relationship.

In all cases, supervisee improvement may be gauged from tapes of therapy sessions and supervisee reports that contain increasing evidence of interventions that target in-session therapist-client interactions identified as clinically relevant according to the client case conceptualization. No improvement signals that the markers of supervisee needs and difficulties may have been missed or misunderstood. The problems presented by the therapist must then be reviewed and his or her learning aims reconsidered so that work can be shifted to a more relevant context.

When the context evoked in supervision does not fit supervisee needs, little progress may be obtained. For instance, a male supervisor may work in vain on his or her relationship with a male supervisee if the problem at hand concerns specific prejudices the supervisee holds about the gender-appropriate behavior of a female client. The appropriate context would have been the second instead of the third. As another example, interventions that target a supervisee's irrational assumptions may have little effect on the performance of a therapist who misses in-session opportunities because he or she has never acquired the needed skills. The supervisor should consider the first context.

As the choice of supervision techniques depends on the context, the key to the model lies in identifying the context. To do this, both the supervisee's level of mastery and the specific difficulties at hand must be evaluated. Both will provide the markers for selecting the most appropriate context at any given moment. When the relevant variable is a lack of repertoire, the supervisor will need to seek to provide the kind of instructions and questions that best prompt the skills the therapist needs to learn (first context). If dysfunctional verbal control hinders focusing on CRBs, cognitive restructuring and other ways of weakening rigid rule -following will be needed (second context).

And finally, when broader problems involving the therapist's functioning are the issue, direct shaping of the relevant repertoires will be the strategy of choice (third context). The integrity of supervision, in this model, thus depends on accurately identifying the needs of the therapist.

Abramson, M. Reflections on knowing oneself ethically: Toward a working framework for social work practice. Families in Society, 1996a, 77(4), 195-202.

American Association for Marriage and Family Therapy, Code of Ethics http://www.aamft.org/iMIS15/AAMFT/Content/Legal_Ethics/Code_of_Ethics.aspx

American Counseling Association’s Code of Ethics http://www.counseling.org/docs/ethics/2014-aca-code-of-ethics.pdf?sfvrsn=4

American Psychological Association's (APA) Ethical Principles of Psychologists and Code of Conduct http://www.apa.org/ethics/code/

Barker, Robert L. Milestones in the Development of Social Work and Social Welfare, Washington, DC NASW Press, 1998. Rankin, Mary, JD employed by the Healthcare Training Institute to research and write the article Evolution of Social Work Ethics. Edited by Tracy Appleton, LCSW. The reference Barker, Robert L. Milestones in the Development of Social Work and Social Welfare. Washington, DC NASW Press, 1998

Bemecker, Samantha L. Helping Clients Help Themselves: Managing Ethical Concerns When Offering Guided Self-Help Interventions in Psychotherapy Practice. Professional Psychology: Research & Practice. Apr2014, Vol. 45 Issue 2, p111-119.

Brems, Christiane PhD, Dealing with Challenges in Psychotherapy and Counseling, Brooks and Cloe, New York, 2000.

Brill, Naomi PhD, Working with People, Longman, New York, 1997.

Butheil, TC PhD, "The Concept of Boundaries in Clinical Practice; theoretical and risk-management dimensions." American Journal of Psychiatry 1999, 188-96.

Robison, Wade PhD, Allyn and Bacon, Ethical Decision Making. Boston, 2000.

Gartrell, N. PhD, Boundaries in Lesbian Therapy Relationships. Women and Therapy 2002; 29-50.

Henderson, Kathryn L. Mandated Reporting of Child Abuse: Considerations and Guidelines for Mental Health Counselors. Journal of Mental Health Counseling. Oct2013, Vol. 35 Issue 4, p296-309.

Kangas, Julie L.; Calvert, James D. Ethical Issues in Mental Health Background Checks for Firearm Ownership. Professional Psychology: Research & Practice. Feb2014, Vol. 45 Issue 1, p76-83.

Kottler, Jeffrey A. MA, Compassionate Therapy. Jossey-Bass Publishers. San Francisco 1999

Loewenberg, F. M., & Dolgoff, R. & Harrington, D. Ethical decisions for social work practice. Itasca, Illinois: F.E. Peacock., 2000.

Moe, Cathrine; Kvig, Erling I; Brinchmann, Beate; Brinchmann, Berit S. Working behind the scenes’ An ethical view of mental health nursing and first-episode psychosis. Nursing Ethics. Aug2013, Vol. 20 Issue 5, p517-527.

National Association of Social Workers, Code of Ethics http://www.naswdc.org/pubs/code/code.asp

National Board for Certified Counselors (NBCC) Code of Ethics http://www.nbcc.org/assets/ethics/nbcc-codeofethics.pdf

Penford, Susan PhD, Sexual Abuse by Health Professionals, University of Toronto, Buffalo, 1998.

Reamer, F. G. The social work ethics audit: A risk-management strategy. Social Work, 2000, 45(4), 355-366.

Reamer, Frederic G., Social Work Ethics Casebook: Cases and Commentary. Washington, DC: NASW Press, 2009

Reamer, Frederic G., Social Work Malpractice and Liability: Strategies for Prevention, 2nd ed. (New York: Columbia University Press, 2003)

Reamer, Frederic G., Social Work Values and Ethics, 3rd ed. (New York: Columbia University Press, 2006)

Reamer, Frederic G., Ethical Standards in Social Work: A Review of the NASW Code of Ethics, 2nd ed. (Washington, DC: NASW Press, 2006)

Reamer, Frederic G., The Social Work Ethics Audit: A Risk-management Tool (Washington, DC: NASW Press, 2001)

Reamer, Frederic G., Tangled Relationships: Managing Boundaries in the Human Services (New York: Columbia University Press, 2001)

Reamer, Frederic G., Ethics Education in Social Work (Alexandria, VA: Council on Social Work Education, 2001)

Rothstein, Mark A., Tarasoff Duties after Newtown, Journal of Law, Medicine & Ethics. Spring2014, Vol. 42 Issue 1, p104-109.

Schoener, GR PhD, Boundaries in Group Therapy: Ethical and Practice Issues. Women and Group Psychotherapy, New York, Builford Press, 1997.

Stone, M. PhD,Boundary Violations between Therapist and Patient. Psychiatric Annals, 1999, 670-7.

Vandenberghe, Luc. "Keeping the focus on clinically relevant behavior: supervision for functional analytic psychotherapy." The International Journal of Behavioral Consultation and Therapy 5.2 (2009): 209+. Academic OneFile. Web. 29 Mar. 2010.

Coordinating Author/Instructor: Tracy Appleton, LCSW, MEd