|Home | About | Journals | Submit | Contact Us | Français|
To improve cognitive and behavioral therapies (CBT) for depression, several approaches recommend an increased focus on the occurrence of problems as they occur in the therapeutic relationship or in relation to the live therapy process, referred to as present-focused. A lingering question has been the degree to which CBT therapists already engage in present-focused work. This study utilized sessions from recent trials of CBT for depression and, in Phase I, raters identified present-focused interventions on a turn-by-turn basis. Phase II raters used a qualitative analysis to determine categories of present-focused interventions. Results indicated that therapists rarely focused on the therapeutic relationship; when they did it was often transient and lacking in the elaborations suggested by newer approaches. Therapists more often performed therapy process and emotion focused interventions, but these also tended to lack elaboration.
Both cognitive therapy (CT; Beck, Rusch, Shaw, & Emery, 1979) and behavioral activation (BA; Martell, Addis, & Jacobson, 2001) for depression are considered to be empirically supported (DeRubeis & Crits-Christoph, 1998; Dobson, 1989; Ekers, Richards, & Gilbody, 2008). However, both approaches leave ample room for improvement, in terms of acute response, remission, and prevention of relapse and recurrence (e.g., Gortner, Gollan, Dobson, & Jacobson, 1998; Jacobson et al., 1996). For example, in a recent meta-analysis of 12 trials comparing behavior therapy and cognitive therapy, it was found that at the end of treatment only 55% of clients were classified as recovered (Ekers et al., 2008). In response to such data, multiple efforts to modify, present alternatives to, and otherwise improve cognitive and behavioral therapies (CBT) are occurring.
One avenue for improvement has been an increased focus on the occurrence of problems as they occur in the here-and-now therapy moment (e.g., in relation to the therapeutic relationship or therapy process), which we refer to as present-focused. This is not to say the present moment is ignored in CBT. In CT for depression, for example, the therapeutic relationship has been seen as an important common factor, necessary for the establishment of collaborative empiricism but not the core mechanism of change (Goldfried & Davila, 2005; Waddington, 2002). Traditional CT manuals (Beck et al., 1979; J. Beck, 1995) note the importance of establishing a strong collaborative relationship, contain instructions for how to resolve difficulties and impasses in the relationship, and include several examples of relationship-focused interactions. The relationship in BA is similar—the therapist is seen as a coach or collaborator with the client, but the relationship is not conceptualized as an active mechanism of change (Martell et al., 2001). In other words, in both CT and BA, the relationship is an implicit rather than explicit priority.
Likewise, CT and BA's focus on the therapy process is clearly articulated but not seen as essential to the change process. At the beginning and end of sessions, therapists are encouraged to discuss the therapy process and elicit feedback from the client about how he or she is reacting to the process. This feedback serves to identify whether the client understands what the therapist is saying and to elicit potentially counter-productive client responses to the treatment (e.g., Beck et al., 1979, pp. 100–101). In addition, CT therapists are taught to see “hot cognitions” and current emotions as opportunities for key cognitive changes to occur. The primary focus in traditional CBT approaches, however, is on changing thoughts and behavior related to the world outside of therapy, not in relation to the here-and-now therapy moment, the therapy relationship or the therapy process.
Several recent CBT variants shift these priorities and emphasize the present moment in specific ways linked directly to core change processes. These treatments include J. Beck's (2005) approach to CT for challenging problems, Leahy's (2001) guidelines for overcoming resistance in CT, the Cognitive-Behavioral Analysis System of Psychotherapy (CBASP; McCullough, 2000), Cognitive Therapy for Personality Disorders (CT-PD; Beck, Freeman, Davis, & Associates, 1990), Schema Focused Therapy (SFT; Young, 1999), Short Term Cognitive Therapy (STCT; Newman, 1998), and Integrated CT (Castonguay et al., 2004). In addition, Functional Analytic Psychotherapy (FAP; Kohlenberg and Tsai, 1991; Tsai et al., 2008) has been suggested as a framework with which to enhance the therapeutic relationship, and therefore outcomes in CT (resulting in FAP-Enhanced CT, or FECT; Kohlenberg, Kanter, Bolling, Parker, & Tsai, 2002; Kohlenberg & Tsai, 1991) and BA (FAP-Enhanced BA, or FEBA; Kanter, Manos, Busch, & Rusch, 2008). While these treatments vary from each other in terms of the theoretical models that underlie the specific present-focused techniques incorporated, they share a common emphasis on work within the present moment as a proposed active mechanism of change.
Much, but not all, of the emphasis in these newer treatments is on the therapy relationship or therapeutic alliance. For example, Integrated CT (Castonguay et al., 2004) provides techniques for therapists to identify, explore and address alliance ruptures as they arise in therapy, with the goal being a client's direct exploration of their feelings toward the therapist, followed by therapist empathic validation. CBASP (McCullough, 2000), STCT (Newman, 1998), CT-PD (Beck et al., 1990), FECT (Kohlenberg et al., 2002), and FEBA (Kanter et al., 2008), focusing more broadly on the therapy relationship than just the alliance per se (Gelso & Samstag, 2008; Lejuez, Hopko, Levine, Gholkar, & Collins, 2005), also emphasize identifying and resolving problems that may occur between the client and therapist as key facets of treatment. For example, CBASP (McCullough, 2000) contains detailed instructions for dealing with a host of issues that may disrupt the therapeutic alliance, including therapist feelings of frustration and incompetence, client resistance to the therapeutic work, lack of client improvement, hostile and angry clients, and passive-dependent clients. Similarly, STCT (Newman, 1998) specifically highlights potential alliance ruptures due to clients feeling a loss of autonomy and control to the therapist, feeling rushed into tasks before they are ready, and anticipating that abandonment will occur when the therapy relationship ends, among other issues. CT-PD (Beck et al., 1990) likewise focuses on problems with non-collaboration in the therapeutic relationship, offering a number of strategies to target and explore possible causes of non-collaboration, including, but not limited to, lack of client skill, the role of culture, dysfunctional beliefs, poor timing of interventions, problems with goals of therapy, and client or therapist frustration with lack of progress.
Several approaches, including SFT (Young, 1999), CT-PD (Beck et al., 1990), STCT (Newman, 1998), and FECT (Kohlenberg et al., 2002) stress the importance of present-focused cognitive restructuring, wherein therapists pay particular attention to clients' immediate emotional reactions to the therapist, therapy process, or other aspects of the present moment and apply cognitive restructuring strategies, including the use of dysfunctional thought records, as these reactions occur. These strategies are largely consistent with traditional CT techniques; the primary difference is that the beliefs targeted for change are in-session present-focused beliefs rather than beliefs about daily life events and relationships.
Another common theme is strategic therapist self-disclosure, advocated by BRT (Safran & Segal, 1990), SFT (Young, 1999), STCT (Newman, 1998), FECT (Kohlenberg et al., 2002), FEBA (Kanter et al., 2008), and CBASP (McCullough, 2000). The rationale for disclosure varies; some employ it to challenge maladaptive and present-focused cognitions (e.g., SFT, STCT, and FECT), while others employ therapist self-disclosure of reactions to client behavior as punishment or reinforcement (e.g., CBASP, FECT, and FEBA).
Finally, several approaches, including CBASP (McCullough, 2000), FECT (Kohlenberg et al., 2002), and FEBA (Kanter et al., 2008) highlight the importance of comparing and contrasting present-focused interactions with daily life situations. This is seen as important both in generating hypotheses about potential transference reactions or other present-focused situations requiring therapeutic attention, as well facilitating generalization of gains made during present-focused interactions to daily life.
Research on the use of present-focused interactions in these approaches is limited but encouraging, and largely consists of research on the therapy relationship. Of all of the treatments highlighted above, CBASP is the most empirically supported. In a large trial, CBASP performed as well as nefazadone (Keller et al., 2000) and was effective at preventing relapse (Klein et al., 2004). Furthermore, the best predictor of positive outcomes for CBASP was the overall degree of emphasis therapists placed on discussing the therapeutic relationship (Vocisano et al., 2004). In addition, FECT has been evaluated in one non-randomized trial (Kohlenberg et al., 2002) compared to CT. In this trial, both CT and FECT performed equally, however FECT clients demonstrated large improvements on several measures of interpersonal functioning compared to CT clients. Additional analyses of the FECT study suggested that present-focused interventions, broadly defined, predicted client-reported progress during treatment (Kanter, Schildcrout, & Kohlenberg, 2005), and present-focused cognitive restructuring specifically predicted improved outcomes at the end of treatment (Kohlenberg et al., 2002). ICT has been evaluated in two trials. The first (Castonguay et al., 2004) compared ICT to a wait-list control, with encouraging results, and the second (Constantino et al., 2008) compared ICT to traditional CT. In this trial, ICT performed well, evidencing improved depression rates, fewer dropouts, better perceived alliances, and higher perceived therapist empathy compared to traditional CT.
Taken together, these findings suggest that not only are these treatments promising in terms of empirical outcomes, but there is accumulating support for the importance of the specific present-focused interventions employed in these treatments. A lingering question has been the degree to which traditional CBT therapists already engage in present-focused work, and if they do, the degree to which this work parallels what is recommended by these newer approaches. For example, the original CT manual (Beck et al., 1979) notes the importance of addressing transference reactions and provides several examples of techniques specifically focused on the therapy or the therapeutic relationship. J. Beck (1995) also presents guidelines for using similar techniques when working with challenging clients. The claim that CBT therapists already engage in this work also has been made more directly. Lazarus (2003) argued, “…would any competent therapist misunderstand and not address, in session, hostile behavior from a somewhat belligerent youth who has problems with authority? Would any well schooled-CBT practitioner find it farfetched to point out to a client that she appears to react to her therapist in the same way the she views her abusive husband, and then use in-session cues to deal with the problem?”
Kanter and colleagues (2005) and several studies by Goldfried and colleagues (Castonguay, Hayes, Goldfried, & DeRubeis, 1995; Goldfried, Castonguay, Hayes, Drozd, & Shapiro, 1997; Goldfried, Raue, & Castonguay, 1998), all using turn-by-turn coding systems to identify present-focused work during psychotherapy sessions, collectively suggest that research CT therapists typically do not engage in it. However, Kanter et al. (2005) also suggested that even small incremental improvements in the amount of present-focused work may predict progress, so it may be the case that, even though these therapists did not often employ present-focused interventions, the occasional well-timed intervention was significant. In addition, previous research has not indicated the nature or type of present-focused interventions employed, such as the degree to which observed interventions are consistent with procedures recommended by recent variants in psychotherapy. Finally, as the recent variants gain empirical support, leakage of these techniques into traditional CBT may occur; thus, it is important to assess the work of recent CT therapists. Assessing the degree to which this occurs in BA is also important, since this has not yet been investigated.
The present study examines the degree to which CT and BA therapists engage in present-focused work, when they do, and what exactly they do. We selected sessions from the CT condition of DeRubeis et al. (2005) and sessions from the CT and BA conditions of Dimidjian et al. (2006). To ensure that any and all present-focused interventions in these sessions would be identified, we engaged in a two-phase procedure. In Phase I, a very wide and inclusive definition of present focus was developed. Based on this definition, trained raters identified interventions on a turn-by-turn basis in selected sessions. In Phase II, the identified present-focused interventions were submitted to a qualitative analysis to determine categories of present-focused interventions. The identified interventions were then categorized to provide a more detailed understanding of the nature of present-focused work in these treatments.
The therapy sessions from two clinical trials, DeRubeis et al. (2005) and Dimidjian et al. (2006), were coded for the current project (see the original reports for more detailed descriptions of conditions, inclusion criteria, client demographic characteristics, and therapeutic procedures). In brief, DeRubeis et al. randomly assigned 240 participants to 16 weeks of CT, 16 weeks of medication, or 8 weeks of pill placebo. Dimidjian et al. randomly assigned 241 participants to 16 weeks of CT, 16 weeks of BA, 16 weeks of medication, or 8 weeks of pill placebo. In addition, both clinical trials required clients to be diagnosed with Major Depressive Disorder according to the DSM-IV (American Psychiatric Association, 2000) and to exceed a cut-off score on the 17-item version of the Hamilton Rating Scale for Depression (HRSD; Hamilton, 1967; Williams, 1960; 20 or greater for DeRubeis et al., and 14 or greater for Dimidjian et al.) indicating moderate to severe depression symptom severity.
In DeRubeis et al. (2005), 60 clients were randomly assigned to receive CT of which 51 completed treatment. Sessions from 19 of the treatment completers were randomly selected and coded for the current study, equally distributed across the two sites (9 from one site, 10 from the other). Treatment lasted 16 weeks, however the number of sessions varied as treatment was held twice weekly during the first four weeks, once or twice weekly during the middle eight weeks, and once weekly during the final four weeks. Three sessions were selected and coded for each client representing a session from the early (sessions 4–9), middle (sessions 8–17), and late (sessions 12–25) phases of therapy. In total, 55 sessions were coded (two sessions were not coded due to administrative error). The mean age of the sample used in the current study from this condition was 42.4 (SD = 9.5; range 25–56) years and 58.0% of the sample was female. Regarding ethnicity, 88.0% of the sample was Caucasian.
In Dimidjian et al. (2006), 45 clients were randomly assigned to receive CT of which 39 completed treatment. Sessions from 38 of the treatment completers were coded for the current study (one did not have tapes available to code). Treatment lasted 16 weeks with a maximum of 24 sessions. Sessions were held twice weekly for the first 8 weeks and once weekly for the next 8 weeks. Three sessions were selected for each client representing a session from the early (sessions 4–6), middle (sessions 7–12), and late (sessions 11–19) phases of therapy. Treatment length was not equal in length for each client and as a result early, middle, and late phases of therapy were based on total number of sessions for each client. In total, 112 sessions were coded (two sessions were not coded due to administrative error). The mean age of the sample used in the current study from this condition was 39.5 (SD = 11.1; range 19–60) years and 73.7% of the sample was female. Regarding ethnicity, 86.9% of the sample was Caucasian.
In Dimidjian et al. (2006), 43 clients were randomly assigned to receive BA of which 36 completed treatment. Sessions from 33 of the treatment completers were coded for the current study (3 were unavailable for coding). Similar to the Dimidjian et al. CT condition, three sessions were selected for each client representing a session from the early (sessions 4–6), middle (sessions 9–12), and late (sessions 13–18) phases of therapy. In total, 92 sessions were coded (seven sessions were not coded due to administrative error). The mean age of the sample used in the current study from this condition was 42.4 (SD = 12.2; range 18–60) years and 60.6% of the sample was female. Regarding ethnicity, 78.8% of the sample was Caucasian.
The CT therapists in DeRubeis et al. (2005) were five licensed clinical psychologists and one psychiatric nurse practitioner. The CT therapists in Dimidjian et al. (2006) were two licensed clinical psychologists and one licensed clinical social worker. CT therapists from both conditions followed standard CT protocols (Beck et al., 1990; Beck et al., 1979; J. Beck, 1995). The BA therapists in Dimidjian et al. consisted of three licensed clinical psychologists. BA therapists followed the BA protocol by Martell and colleagues (2001).
Raters used a revised coding scheme designed for a previous similar study (Kanter et al., 2005). In this scheme, each therapist turn of speech is categorized into one of two mutually exclusive categories: “Present-Focused” (PF) and “Other”. PF was defined as talk aimed at working on client problems that occurred in therapy in relation to the therapy process, the therapy relationship, or anything else having to do with therapy. “Other” was simply defined as everything other than PF. In order to maximize identification of PF turns, raters were instructed to rate a turn as PF if they identified both PF and Other material in a single turn. Categorizations for each turn were based both on the content of speech, as well as the context. For example, a turn would be coded as PF if it did not specifically refer to PF material, but was part of a larger discussion that was focused on PF material. These instructions ensured that all potential PF turns would be coded; at this stage the aim was to minimize false negatives by allowing for more false positives. The following are examples of PF therapist turns:
“How did you feel about today's session?”
“Well…I think we need to stop. But I'm really tickled with your progress, and I think you're doing a great job.”
“Remember last time as we would talk about things sometimes you would kind of, you'd already come up with how a solution might not work before we even would finish talking about it, and I was pointing out that you were doing that. Is that something you think happens at home as well?
“Are you worried about doing that with me? Because you preface this by saying `sometimes I can't even tell you what's going on.'” (The client was talking about draining other people around her with her problems.)
“Okay. Uh…can you put the paper down? Thanks. Because it's a…it's a little distracting for me. And it makes me think that what I'm saying may not really be picked up by you because you're not that tuned in to what I'm saying. Not…I don't think that everything I have to say is so important, but some things are important and I…as long as you're here I'd like you to kind of take them in. I appreciate that. I know it's not easy for you.”
“We're really talking about the hardest thing right now. You know, so it really makes sense to me that you'd feel lousy talking about this.”
“Well along those lines, I actually want to check out…if this ties into that. So I'll do it real quick right here. I want to keep checking in with you about how you feel about this treatment, does it feel like it's what you need…and about our relationship, the whole thing. How are you feeling about you and me?”
Pairs of raters coded DVDs of whole therapy sessions. Sessions were assigned to rater pairs at random from the dataset. Rater pair assignments rotated weekly so that each rater pair occurred equally. Kanter et al. (2005) provide evidence for the validity of the coding system. Using the same coding system and coding process, in that study, as predicted, traditional CT therapists engaged in very low rates of present-focused interventions as identified by the coding system (turns per session M = 3.4, SD = 4.9), while FECT therapists, trained in present-focus, demonstrated approximately triple the rate of present-focused turns per session (M = 9.2, SD = 9.9).
Coding occurred at two sites. Raters at Site 1 consisted of 7 undergraduates and 2 graduate students and raters at Site 2 consisted of 10 undergraduates. Training consisted of approximately 15 hours of didactic instruction, practice, and feedback at both sites. Raters were unaware of the overall purpose of the study and blind to treatment condition. To assess interrater reliability, a random subset of 6 raters at site 1 coded 3 sessions and 10 raters at site 2 coded 4 sessions randomly pulled from the larger dataset, and overall agreement between raters was calculated based on the percent of each session identified as PF by each rater. The formula of intraclass correlation (ICC) recommended by Shrout and Fleiss (1979) when each session is coded by a different set of raters randomly selected from a larger pool of raters was used. Site 1 demonstrated acceptable reliability (ICC = .75) and Site 2 demonstrated excellent reliability (ICC = .88). During the course of the rating project, weekly meetings were held to prevent rater drift and provide corrective feedback.
In Phase II, all PF turns of speech identified in Phase I were categorized into 9 categories. As there were no previous coding systems designed for this purpose, categories were developed based on a qualitative review of the treatments described above and a content analysis of the Phase I codes, and were intended to capture the broad themes of present-focused work that cut across several treatment approaches. Four categories derived from the literature, described below, are Ruptures and Resistance, Present-Focused Cognitive Restructuring, Comparisons, and Disclosure. In addition, the PF turns of speech identified in Phase I were transcribed according to standard psychotherapy transcription guidelines (Mergenthaler & Stinson, 1992). These PF turns were reviewed by the current authors to determine additional categories necessary to capture the full range of PF content. These additional categories, described below, are General Process, Supportive, Direct Instruction, Emotion Focus, and Relationship Focus. A final category was included that captured turns that did not fit into the above categories and should not have been coded as present-focused in Phase I (false positives). This category was labeled Not Present-Focused. A coding manual (available from the first author) containing descriptions and hypothetical examples of these nine categories was created, and a rater pair categorized turns into these categories using the manual and transcripts of PF turns in each session. Raters were instructed to each code the transcribed turn. Upon coding, they were instructed to compare codes and if they disagreed they were to discuss their codes and come to an agreement.
Ruptures and Resistance turns were defined to be consistent with the therapeutic procedures outlined by Safran and Muran (2000) and the other approaches described above. For example, this category included therapists pointing out ruptures or resistance in the therapy, inviting clients to explore these issues in session, and strategic attempts to resolve or overcome the rupture or resistance. An example of a ruptures and resistance turn is “You know, you tell me how difficult life is for you, but when I offer alternatives, you aren't willing to try any of them.”
Present-Focused Cognitive restructuring turns included therapist attempts to explore and challenge the client's thoughts, beliefs, assumptions, and schemas that were occurring live in the therapy session and were related to the therapeutic relationship or therapy process. An example of a present-focused cognitive restructuring turn is “You could test out that idea or belief that everyone is judging you right now. You could ask me if I'm judging you right now.”
Comparisons turns included therapist attempts to compare or contrast situations between the client and the therapist to either current or historical situations in the client's daily life. An example of a comparisons turn is “I think that you were really taking responsibility today, when you told me what you needed from me in terms of support. Is that the same kind of thing you need to do with your wife?”
Disclosure turns included therapist self-disclosures to clients about their feelings or thoughts about the client or the client's behavior. An example of a disclosure turn is “When you get angry and raise your voice, it makes me feel uncomfortable.”
General Process turns were defined as turns dealing directly with the standard therapy process, including agenda setting, feedback about what was useful or helpful about the session, asking for or providing a summary of the session, goal setting, and getting depression ratings. These therapist behaviors are explicitly recommended by Beck et al. (1979) and J. Beck (1995) as important strategies to be used in order to establish and maintain structure and continuity between sessions. They were conceptualized as present-focused strategies because they focus on what is happening in the moment in the session. For example, “Can you give me some feedback on our session today?” requires that the client consider his or her immediate reactions to the session and possibly to the therapist, and the therapist may use this question to explore a PF issue in further detail.
Supportive turns were defined as turns in which the therapist offers support or otherwise lets the client know he or she is supportive in a generally empathic way. Supportive turns may be seen as non-specific present-focused turns and can be conceptualized as a Rogerian response to clients. An example of a supportive turn is “Good job last session. I really have a lot of empathy for how difficult that was for you.”
Direct Instruction turns were defined as turns in which the therapist provides direct instructions to the client during the session or participates in role plays. This occurs in sessions when the therapist is teaching techniques such as progressive muscle relaxation (PMR) or assertiveness skills. An example of a direct instruction turn occurring during PMR is: “Leave your feet on the floor and sit back and relax… I'll count and we'll do the breathing together.” These turns were conceptualized as present-focused because they focus the client on what he or she is feeling or doing as it is happening in the moment.
Emotion Focus turns were defined as turns during which the client's immediate emotions were discussed, without connection to the therapeutic relationship and without specific follow-up. An example of an emotion focus turn is: “What are you feeling right now as we talk about this?” It is important to note that if the therapist asked about the client's emotions in the context of a discussion of ruptures and resistance, the turn would receive a rupture and resistance code instead. Similarly, asking about the relationship between present-focused cognition and affect, which is an important component of present-focused cognitive restructuring (e.g., “When you think that I do not like you how do you feel?”) would have received an emotion focus code if there was no additional restructuring attempted and would have received an present-focused cognitive restructuring code if there was additional restructuring attempted.
Relationship Focus turns were defined as therapist attempts to discuss the therapy relationship or feelings toward the therapist that were not specific enough or lacked subsequent follow-through to be coded as any of the above codes. These turns are expected to occur quite often in CBASP, FECT, BRT, and the other present-focused approaches as the background on which more specific present-focused techniques, such as a focus on ruptures or resistance and present-focused cognitive restructuring occur. An example of a relationship focus turn is: “How do you feel about our relationship?” Like emotion focus codes, these codes would not have been used if another specific code was appropriate.
Not Present-Focused turns were defined as turns that were not correctly coded during Phase I and were not considered to be present-focused turns by the Phase II raters.
Two advanced undergraduate research assistants served as raters and received approximately 6 hours of training over the course of 2 weeks. Training included didactic instruction by the first author, practice on distinguishing between types of PF turns, and discussion of what occurs during a therapy session. Raters were unaware of the overall purpose of the study and were unaware that different treatment conditions existed. To assess reliability, 15 turns were randomly selected and coded by criterion raters (the first and second authors), resulting in 93% agreement between criterion and Rater 1, 80% agreement between criterion and Rater 2, and 87% agreement when the raters met and agreed on the specific turn.
Table 1 summarizes the percent of sessions by condition that were coded as present-focused as well as the percent of total sessions for each condition in which specific ranges of PF turns were found. PF turns per session occurred at similar and very low rates for all three conditions, with the percent of PF turns falling between 1% and 3%. None of the conditions demonstrated a session in which there were more than 30% PF turns and only the Dimidjian et al. CT condition had sessions in which there were 20–30% PF turns (2.7% of total sessions). Dimidjian et al. CT also had more sessions with 10–20% PF turns (7.1%), while Dimidjian et al. BA (3.3%) and DeRubeis et al. CT (1.9%) demonstrated fewer sessions in this range. The vast majority of sessions in all conditions contained 0–10% PF turns.
Even though PF turns occurred at low rates in all three conditions, it was still possible that statistically significant differences between conditions in the percentage of PF turns would be found. Another interest in the current study was the possibility that the percentage of PF turns would vary depending on when in the course of therapy the sessions coded took place and on the remission status of the clients (non-remitters = HRSD of 13 or high at post-treatment, partial remitters = HRSD of 8 to 12 at post-treatment, and full remitters = HRSD of 0 to 7 at post-treatment, as defined by DeRubeis et al., 2005). To explore these possibilities, a multilevel growth curve analysis was conducted using HLM 6.02 (Raudenbush, Bryk, Cheong, Congdon, & du Toit, 2004) to determine the effects of treatment condition and remission status on percent PF turns and to model change over the course of therapy (i.e., when in the course of therapy the sessions rated took place). This procedure is optimal because it allows the number and spacing of observations to vary between participants and can include participants who do not have data for all time points (Raudenbush & Bryk, 2002). The final model included a random intercept, a random linear time effect (session position), and fixed between person effects (treatment condition, remission status). Results from the growth curve analysis are presented in Table 2. Results indicated that, despite overall low rates, treatment condition was significantly related to PF turns. Specifically, the Dimidjian et al. CT condition demonstrated a significantly higher percentage of present-focused turns than both the Dimidjian et al. BA and DeRubeis et al. CT conditions. On average, therapists in the Dimidjian et al. CT condition had 1.4% more PF turns than therapists in the Dimidjian et al. BA condition and 2.3% more PF turns than therapists in the DeRubeis et al. CT condition. This difference can be seen graphically in Figure 1. There was no significant time effect. That is, the percentage of PF turns did not differ depending on whether the session coded was an early, middle, or late session. Finally, remission status also was not significantly related to percent PF turns.
In Phase II, all PF turns identified in Phase I were categorized into the nine categories described above, as well as the not present-focused code (to identify any miscodes from Phase I). Because of the exploratory nature of this analysis, these results are presented and described descriptively. The percent and number of PF turns coded in each category are summarized in Table 3. The general process and relationship focus codes were the most frequent codes in all three conditions, each accounting for roughly one third of the overall codes. As expected, the present-focused cognitive restructuring code was also frequently used in both CT conditions and not in the BA condition. The least frequent PF techniques used included focusing on ruptures and resistance, therapist disclosure, and direct instruction. In terms of miscodes from Phase I, a marginal number of turns were identified as not present-focused.
The broad conclusion from this investigation is that the research therapists in these trials of CT and BA for depression rarely focused on the present moment, and when they did the focus was often transient and lacking in the elaborations suggested by newer variants of CBT. Overall, the percentage of present-focused turns per session was very low across all three treatment conditions. Occasionally, a session occurred with a significantly higher present focus, particularly in the Dimidjian et al. CT condition, but no session in any condition evidenced more than 30% of a focus on the therapy process or therapy relationship. Rates of present-focus found in the current study were consistent with those found for traditional CT therapists in Kanter et al. (2005) and much lower than rates for therapists trained in present-focus, suggesting that the system is valid and does pick up present-focus when it occurs.
In addition, the results of the Phase II analysis suggest that, in the relatively rare instance when present-focused techniques were applied, the techniques more often concerned general therapy processes, such as agenda setting or asking for feedback about the session, rather than the strategic applications suggested by the newer approaches reviewed in this article. In particular, although the cognitive therapists in Dimidjian et al. (2006) demonstrated significantly higher rates of present-focused techniques than did the other therapists, the Dimidjian et al. CT therapists did not engage in the highlighted techniques—ruptures and resistance, present-focused cognitive restructuring, disclosure, and comparisons— to any significant degree or any more than did the other therapists.
Instead, our descriptive analysis suggested that the Dimidjian et al. (2006) CT therapists appeared to engage in more emotion focus than the other therapists and in more relationship focus than the therapists in DeRubeis et al. (2005). These two sets of codes represent what can be seen as opportunities lost. In other words, these codes were applied when the therapists noticed and commented on an present emotion (emotion focus) or a live client reaction to the therapist (relationship focus) but did not explore the issue in depth or otherwise follow up on the opportunity with any of the specific strategies recommended by the newer approaches. The DeRubeis et al. (2005) CT therapists displayed a slightly different pattern. While they engaged in the least amount of present-focused techniques overall, the work they did engage in tended to include the highlighted strategies more often than that of the other therapists. They demonstrated fewer opportunities lost compared to the Dimidjian et al. therapists, and higher rates of cognitive restructuring and ruptures and resistance. This is most likely due to the inclusion in the DeRubeis et al. study of Beck and colleagues (1990) treatment manual on CT for personality disorders. This manual highlights the importance of live cognitive restructuring and resolving difficulties with the therapeutic alliance (e.g., ruptures and resistance) for clients with personality disorders. In fact, all 13 of the present-focused cognitive restructuring turns and all 2 of the ruptures and resistance turns identified in that study occurred with clients with co-morbid personality disorders (PD). The present-focused cognitive restructuring turns occurred in a single session with one client while the ruptures and resistance turns occurred across four sessions with four clients. Although this is somewhat encouraging, given that a high proportion of clients in that study received co-morbid PD diagnoses (10 of the 19 clients sampled for the current study received such diagnoses), one may wonder why higher rates of present-focused work did not occur.
Although the current research suggests that many present-focused opportunities were lost in these studies, more research is needed to demonstrate that a more intense focus on the present moment would improve outcomes in traditional forms of CBT. To date, two studies have now compared versions of CT augmented with a present focus to traditional CT, with promising results (Constantino et al., 2008; Kohlenberg et al., 2002), and two studies have linked present-focused processes (largely relationship focus) to outcomes (Kanter et al., 2005; Vocisano et al., 2004), but this research is far from definitive and more is needed in this area. Furthermore, the current study found no relationship between the use of present-focused interventions and outcome, suggesting that they are not necessary in standard CBT with well-trained therapists. However, this analysis was limited by the low rates of present-focused interventions found, the potential high rate of falsely identified present-focused interventions due to the study's priority to minimize false negatives, and the possibility that study therapists engaged in present-focused interventions with harder clients. Assessing the degree to which present-focus predicts outcomes will require samples in which present-focused interventions were actually performed at significant rates and hopefully future research will address this issue.
If future research does suggest that these interventions are important, it is important to note that it is difficult to follow through on present-focused issues. Live cognitive restructuring, for example, requires that the therapist, during an emotionally-charged interaction with the client, become aware of his or her own emotional reactions as well as those of the client, apply a traditional restructuring intervention, and potentially self-disclose his or her own emotional reactions strategically, genuinely, empathically, and at the proper time. The complexity of such an intervention may result in therapists' employing it only rarely, even in approaches that emphasize such moves. For example, Kohlenberg et al. (2002), in an analysis of FECT, found that even after specific training in FAP's present-focused interventions, experienced CT therapists engaged in less than ideal amounts of present-focused work in the FECT phase of the study. Future research focusing not just on if present-focused work is important but on how to facilitate the integration of present-focused work into a therapist's repertoire will be helpful.
It is important to discuss the skill level and generality of the specific therapists in the current study compared to the larger community of practicing cognitive and behavioral therapists who treat depression. First, Dimidjian et al. (2006), but not DeRubeis et al. (2005), provided data demonstrating therapists' adherence to their respective treatment manuals; therefore the Dimidjian et al. therapists clearly conducted CT and BA as indicated by the manuals. Although DeRubeis et al. did not report such data, it is important to note that the therapists at one of this study's sites (Penn) were considerably more experienced and expert in CT than were therapists at the other site (Vanderbilt). However, post-hoc analyses of rates of present-focused work for each therapist in DeRubeis et al. found no indication of therapist effects in our data - the therapists individually and collectively did not engage in this work to any meaningful degree. Therefore, this data suggests well-trained and adherent CBT research therapists do not focus on the present moment. These therapists had strong backgrounds in CBT and were well-trained in the research protocols, so generalization to the larger community remains an open question, as practicing CBT clinicians may be more eclectic and may be more likely to include a psychodynamic, interpersonal, humanistic, intersubjective or other formulation in their clinical work that addresses the momentary interpersonal process to a substantial degree. Furthermore, the current research therapists may have prioritized adherence to their treatment manuals over explorations of the present moment not explicitly described in the manual, as is necessary for such trials to be internally valid.
An additional limitation of the current study is that only three sessions per client were coded; thus, important but sporadic present-focused work easily could have been overlooked. In other words, a focus on the therapy relationship could occur in a small number of sessions, and have a major impact, rather than continuously across many sessions. Furthermore, the current study did not code sessions from clients who prematurely terminated treatment; these clients may have evoked present-focused work at higher frequencies. Research by Kanter et al. (2005), however, exhaustively coded every session of every client in the comparison of CT to FECT (Kohlenberg et al., 2002) and found rates of present-focused work in the CT condition to be very similar to those presented here. A further limitation of the study is that a highly inclusive definition of present-focus was used, including basic therapeutic techniques, such as agenda setting, thus diluting the potential impact of more powerful present-focused techniques that were also coded. This highly inclusive definition may have played a role in the lack of findings of a significant relationship between present-focused responses and remission.
The current study adds additional support to previous research (Castonguay et al., 1995; Goldfried et al., 1997; Goldfried et al., 1998; Kanter et al., 2005) that suggests that research CBT therapists only rarely engage in present-focused work. While this conclusion is becoming increasingly clear, it is by no means a criticism of research therapists who are trying to adhere to treatment manuals that do not emphasize present-focused work. However, given the increasing number of variants of CBT that emphasize the therapy relationship as an active mechanism, and compelling albeit still preliminary data that such a focus may be beneficial (Kanter et al., 2005; Muran, Safran, Samstag, & Winston, 2005; Vocisano et al., 2004), it does suggest a window of opportunity for potentially increasing the effectiveness of CBT, especially for clients for whom traditional CBT has failed.