Clinicians’ Experiences in using an Empirically Support Treatment (EST) for Panic Disorder: Results of a Survey

American Psychological Association (APA) Division 12 Committee on Building a Two-Way Bridge Between Research and Practice

             As part of its effort to build a two-way bridge between research and practice, the Society of Clinical Psychology recently surveyed therapists about the variables they found to limit the effectiveness of cognitive behavior therapy (CBT) in clinical practice for treating panic disorder—the only current EST. The goal of this initiative was to close the gap between research and practice by providing clinicians with a voice in the research process. In essence, it could allow them to make use of their clinical experience–the context of discovery–to highlight researchable questions and hypotheses that could help improve the effectiveness of our interventions. In having a two-way, rather than a one-way bridge, it is also hoped that it will also encourage practitioners to make use of research findings to guide their clinical work. With growing demands for accountability, it is important for both researcher and therapist to have a collaborative voice in deciding which treatments work.

            We began by surveying experiences in the treatment of panic disorder because it is a clinical problem that therapists are likely to encounter in their practice, and one where there exists research findings indicating that it is efficacious—although not 100% effective. The items in the survey, which involved patient, therapist, treatment and contextual variables, were generated from open-ended interviews with a group of clinicians who were experienced in using CBT for the treatment of panic. For their invaluable help in constructing the survey, we thankDianne Chambless, Steven Fishman, Joann Galst, Alan Goldstein, Steven Gordon, Steven Holland, Philip Levendusky, Barry Lubetkin, Charles Mansuto, Cory Newman, Bethany Teachman, Dina Vivian, and Barry Wolfe.
            The next two surveys will focus on the use of ESTs in the treatment of general anxiety disorder, and of social phobia (social anxiety disorder).
            Invitations to participate in the current survey were announced on Web sites, listservs, and in newsletters of numerous professional organizations in the US, Canada, the UK, Europe, and Australia. The survey included the following instructions:
Once a drug has been approved by the Food and Drug Administration (FDA) as a result of clinical trials, practitioners have the opportunity to offer feedback to the FDA on any shortcomings in the use of the drug in clinical practice. The Society of Clinical Psychology, Division 12 of the American Psychological Association, has established a mechanism whereby practicing psychotherapists can report their clinical experiences using empirically supported treatments (ESTs). This is not only an opportunity for clinicians to share their experiences with other therapists, but also to offer information that can encourage researchers to investigate ways of overcoming these limitations.
This questionnaire provides the opportunity for therapists using cognitive-behavior therapy (CBT) in treating panic disorder to share their clinical experiences about those variables they have found to limit the successful reduction of symptomotology. By identifying the obstacles to successful treatment, we can then take steps to overcome these shortcomings.
Your responses, which will be anonymous, will be tallied with those of other therapists and posted on the Division 12 Web site at a later time, with links made to it from other relevant Web sites. The results of the feedback we receive from clinicians will be provided to researchers, in the hope they can investigate ways of overcoming these obstacles.
                Participants were asked to indicate which of the items listed under the following categories limited successful symptom reductionPatient’s symptoms related to panic; Other patient problems or characteristics; Patient expectations; Patient beliefs about panic; Patient motivation; Social system (home, work, other); Problems/limitations associated with the CBT intervention method; and Therapy relationship issues. The results of their responses to these questions appear in the Tables appearingbelow.
            We received a total of 326 usable surveys, containing responses to the questions and the necessary demographic information. The median age of respondents was 45 years (25 to 81 years), 52% of whom were female and 86% Caucasian. Fifty-six percent had a PhD in clinical psychology, and the median year of the highest degree was 2000. Experience level varied widely, with 35% having over 20 years of experience, 29% between 10 and 20 years, and the remaining 36 % less than 10 years of experience. Closely paralleling this were the number of panic patients treated: 36% having seen over 51 patients, 28% between 21 and 50, and 36% less than 20 patients. The patients had been seen in varying settings, including outpatient clinic (61%), private practice (54%), counseling centers (10%), and inpatient (3%). The length of treatment varied, from less than 3 months (21%), between 3 and 6 months (49%), 6 months to a year (25%), and over a year (5%). In identifying the degree to which different theoretical orientations guided their work, cognitive (41%) and behavioral (38%) ranked the highest, with psychodynamic (15%), experiential/humanistic (12%), family/systems (11%) and other (12%) playing some role.
            The Division 12 committee that has been overseeing this initiative and discussing the findings consists of Louis G. Castonguay: President, Society for Psychotherapy Research; Marvin R. Goldfried (Chair): Past-President, Society for Psychotherapy Research, President, APA Society of Clinical Psychology; Jeffrey J. Magnavita: President, APA Division of Psychotherapy; Michelle G. Newman: Psychotherapy researcher in anxiety disorders, Associate Editor, Behavior Therapy; Linda Sobell: Past-President, of Association for Behavioral and Cognitive Therapies, Past-President, APA Society of Clinical Psychology; and Abraham W. Wolf: Past-President, APA Division of Psychotherapy. The findings of this study, many of which we believe are interesting and important, are presented in the Tablesbelow. Some of results have been discussed by the committee, and a transcription of this discussion appears below:
MARVIN GOLDFRIED: There are a number of interesting points that resulted from our survey of clinicians about their experiences in using cognitive-behavior therapy (CBT)–which is considered an empirically supported treatment for panic disorders. Perhaps we can start by looking at the characteristics of the therapists who participated. As seen by what they indicated to be the orientations that guided their work, approximately 79% of them reported that they use cognitive and behavioral interventions, but also suggest that they are not doing pure CBT
ABE WOLF: What you mean by “pure CBT?”
GOLDFRIED: Therapy that would be essentially following the CBT manual for the treatment of panic, which is used in clinical trials.
NEWMAN: I think it is likely that therapists interpreted the question more broadly to indicate whether they used only techniques that are based on cognitive and behavioral principles of change (manualized or not) in their treatment of panic disorder.
GOLDFRIED:  This is often discussed in the literature, namely that people go beyond what is described in the manual. When we come to the discussion of the experience level of the therapists, we’ll see that there are different cohorts of therapists who do CBT. What do people think about their report that they have an estimated 80% success rate?
WOLF:  (laughing). Well, we all like to think that we are very effective in what we do. 
JEFFREY MAGNAVITA:  And it depends on what you mean by “effective.”
GOLDFRIED:  Either total elimination of panic or significant reduction of panic. 
MAGNAVITA: In my experience, I often don’t see people who are totally free of panic, but rather are not as adversely affected by it.
WOLF:  Yes. Perhaps the panic attacks become less intense and less debilitating.
WOLF: So it’s really hard to know what people mean when they talk about being successful with the treatment.
NEWMAN: An 80% success rate, however it is interpreted, is consistent with randomized controlled trials on the efficacy of panic control therapy. So if clinicians’ report of success is accurate, it does say something about the generalizability of this treatment to the real world. In my experience with using panic control therapy with a variety of clients, it is quite helpful in reducing their avoidance and associated symptoms. The goal of the treatment is not to eliminate panic attacks but to eliminate fear and avoidance of panic attacks and associated situations. By doing this, reduction in panic follows.
MAGNAVITA: The question deals specifically with reducing the symptoms of panic disorder, not necessarily the other problems that might be associated with panic. 
GOLDFRIED: The question deals with the reduction of panic symptoms, not necessarily “cure.”
WOLF: The other part of what I find interesting is that the respondents reported that 50% of their patients are taking medication. This speaks to the difficulty in doing psychotherapy research these days, as most of the patients seen in actual clinical practice are on medication—especially if they have been referred by a pharmacologist or primary care provider. It’s very rare to see patients that are not currently taking medication for panic disorder.
MAGNAVITA: And we don’t know what other modalities are being used to treat these patients. 
GOLDFRIED: This is, no doubt, why it’s difficult to do good effectiveness research, as there are all these other variables that may be operating. 
NEWMAN: When patients are taking medications, it makes it very difficult for therapists and CBT techniques to teach the client that they can cope with their panic attacks. These clients are also more likely to attribute any gains to the medication, rather than to something they learned or did.
LINDA SOBELL:  Do we know what primary diagnoses were given with these patients?
GOLDFRIED: No, we don’t have information on that.
SOBELL: In subsequent surveys, it would be a good idea to ask whether or not the clinical problem that they are reporting on was the primary diagnosis or not.
MAGNAVITA: Or what percentage of the patient’s with this diagnoses have other diagnoses.  
NEWMAN:  I think it would be helpful to know how closely therapists follow DSM criteria when making diagnoses and whether they do formal full structured interviews or unstructured interviews. It would also be helpful to know whether therapists routinely determine primacy of diagnoses and how they make this determination.
WOLF: We do ask about substance abuse and premorbid functioning, and how it affects treatment. This is presented in Table 1.
Table 1
Other Patient Problems or Characteristics
Response Rate
Inability to work independently between sessions
Unwillingness to give up safety behaviors
Chaotic life style
Personality disorder(s)
Reliance on psychotropic medication
Substance abuse
Fear of exposure and associated emotional reactions
Premorbid functioning is limited
Resistance to directiveness of treatment
Intellectual/cognitive/introspective limitations
Depressed mood/mood disorder
Perfectionistic/obsessive style
Low self-esteem/self-efficacy
Negative emotions not recognized
Poor interpersonal skills
Physical problems
Low socioeconomic status
Diversity issues
GOLDFRIED: In Table 2 there’s an indication about the problem of CBT not dealing with comorbid problems, which were reported by 37.5%of the respondents.
Table 2
Problems/Limitations Associated with the CBT Intervention Method
Response Rate
Patient’s reluctance to eliminate safety behaviors
Exposure in vivo has logistical problems
Doesn’t deal with comorbid problems/symptoms
Simulating panic in session is difficult
Triggers to panic not evident
Strict adherence to CBT protocol
Relaxation doesn’t work or causes anxiety
Absence of guidelines for dealing with resistance/noncompliance
Doesn’t deal with patient’s anger
Doesn’t deal with fear of interpersonal loss
Triggers for panic are not linked to client’s past history
Doesn’t deal with comprehensive or lasting change
Current coping skills are not linked to past
WOLF:  One of the things that I think is important is that, in essence, there are different kinds of panic disorders. Some people have panic in the middle of the night, others out of the blue during the day, and so on.
MAGNAVITA:  Again, we’re also dealing with the very important issue of comorbidity, which makes it more difficult to use CBT for simple symptom reduction. 

NEWMAN:  Tsao and colleagues (Tsao, Lewin, & Craske, 1998; Tsao, Mystkowski, Zucker, & Craske, 2005) have looked at this within the context of RCTs and they have found a significant impact of CBT for panic disorder on comorbid anxiety disorders. However, we do not really know anything about whether other types of comorbidity interfere with the efficacy of CBT. In addition, we don’t know anything about how this plays out in the real world. Most RCTs do exclude substance abuse and dependence.
GOLDFRIED: Table 1, which deals with findings about other patient problems or characteristics that can undermine treatment, is relevant here. For example, 55.9% report the patient’s chaotic life style creates treatment problems. In writing about agoraphobia some years back, Chambless and Goldstein (1982) made a distinction between “simple” and “complex” agoraphobics, depending upon the precipitant of the problem (e.g., a specific, isolated trigger or more pervasive life problems). This was before the notion of comorbidity became popular. So, complexity/comorbidity represents a major part prognostic factor.
 MAGNAVITA: Absolutely!  And Mark Lenzenweger and his associates (Lenzenweger et al., 2007) have written about how ubiquitous comorbidity is in patients with personality disorders. So it is possible that many patients who are unresponsive to first line treatment protocols are much more complicated and challenging to psychotherapists.
CASTONGUAY: These and other issues endorsed by our participants point the importance of asking clinicians about their day-to-day practice. One thing that is quite clear to me is that some of our findings are consistent with what the empirical literature is telling us. As most of you know, Michelle published two important reviews (chapters in a book that Larry Beutler and I edited) of the literature on principles of change in the treatment of anxiety disorders. One of the reviews focuses on pre-treatment client characteristics and the other provides an integration of client characteristics, technique, and relationship variables. A number of our findings are consistent with her reviews, such as the negative relationship between outcome and symptoms/impairment, as well as personality disorders. Such convergence of information collected via different methods of knowledge acquisition is very reassuring, as it should increases our confidence in the reliability of our current knowledge.
Other findings add to what we know from the empirical literature by providing very detailed information that can help us understand the link (or lack of thereof) between outcome and other client characteristics, such as expectations and beliefs about panic, motivation, and issues related to therapeutic relationship.
Our findings go one step further by providing information that, at least to my knowledge, is not reported in empirical journal articles (while being very complementary to it). Some of this information is specific to the treatment of panic disorder (e.g., unwillingness to give up safety behavior, reliance on medication, fear of exposure), while other is relevant to all forms of clinical problems (e.g., inability to work independently between sessions, chaotic life style, bad experiences with previous therapy, issues related to social system).
Related to the previous point, our findings regarding problems and limitations of CBT are not likely to be highlighted in peer reviews of RCT trials (and, I would venture to guess in treatment manuals associated with such RCT). Yet they provide important information about how to improve our current gold standard treatment for panic disorders.
GOLDFRIED: In some ways, that may tie into the question of how many sessions people need when being treated for panic disorder. Close to a third of our respondents say they see patients six months or more, which might have to do with the need to deal with these interfering problems. If we look at the findings for the question that asks whether or not more than symptom reduction is needed in working with panic patients, an overwhelming 73percent indicated “yes. “
SOBELL: In addition to the fact that patients are seen longer in practice than in clinical trials, there are a couple of things in Table 3 dealing with therapy relationship issues that are really striking. That deals with therapist frustration and negative reactions to the patient. There clearly is something happening in this relationship that often does not get discussed in clinical trials.
Table 3
Therapy Relationship Issues
Response Rate
Therapy alliance not strong enough
Therapist’s frustration with progress
Therapist’s negative feelings toward patient
Distress not sufficiently understood/validated
CASTONGUAY: This is very much in line with what I mentioned before. Alliance has been linked with outcome in treatment of anxiety disorder, but the quantitative results don’t tell us much about the toxic or difficult issues that are involved when the relationship between client and therapist is not good. Our findings address these issues in a way that are very much in line with intensive qualitative analyses that have been conducted.
MAGNAVITA: Maybe that’s what we’ve been talking about, namely that therapists are trying this first-line treatment for Axis I disorders, but they are actually working with something much more complex. Under such circumstances, therapists may very well have a sense of frustration with therapeutic progress—and, at times, with the patient.
GOLDFRIED:  It also may have to do with the fact that the literature gives one the impression that the intervention is straightforward, and that there will not be any problems–which then takes the therapist by surprise. That can be very frustrating, causing therapists to become impatient because the clients are not doing what they are “supposed to be doing. . .”
MAGNAVITA:  . . . they’re not doing the work. . .
WOLF:   . . . right. . .
GOLDFRIED:  . . . as a opposed to the intervention not working as well in certain instances.
SOBELL: It’s very important to get this information out there to clinicians and researchers. It’s not saying that CBT is not working, but rather that there are certain moderators that cause it to be less effective. And this is precisely the kind of thing that really needs to be studied in clinical trials. 
CASTONGUAY: I could not agree more! In addition, this has clear implications for training. I remember Bruce Arnow making the point that many people believe that you can be trained in CBT for a specific disorder by attending a workshop at ABCT – and that in a few hours you have attained a minimal level of competence. As Bruce also mentioned, nobody would dare have the same thought about psychodynamic treatments! My view has always been that CBT is complex and requires quite extensive training and supervision in the application of its manuals and beyond; our findings are consistent with this.
GOLDFRIED: Some of these moderating variables seem to be related to general principles associated with success in therapy, such as, patient expectations, motivation, and the nature of the therapy alliance (see Tables 3, 4, and 5). And these are the kinds of things that really need to be spelled out very clearly in treatment manuals. We can’t simply go along with the assumption that
Table 4
Patient Expectations
Response Rate
They will be free of all anxiety
Therapist will do all the work to make things better
They need medication to reduce panic
Successful exposure means not having panic/anxiety
Pessimism due to disappointment with past therapy
Treatment will be brief and easy
Symptom reduction is not enough
Table 5
Patient Motivation
Response Rate
Minimal motivation at outset
Premature termination
Motivation decreased as some improvement occurs
Motivation decreased when patient learns reasons for having panic
because somebody is entering therapy—either in a clinical trial in actual clinical practice–that he or she is motivated, has positive expectations, and is able to form a good alliance with the therapist. Without adequate motivation, even the best of treatments won’t work; if therapists are not attuned to and aligned with the patient’s expectations, there are likely to be problems in the treatment; and if the therapist’s directiveness and behavioral assignments are not tempered with empathy and compassion, a rupture in the alliance can occur.  
NEWMAN: I think that many of the newer manuals and books on CBT try to provide information about working on the therapeutic alliance, as well as with working with unmotivated or uncooperative clients. However, it isn’t easy and there is no simple cookbook for it.
CASTONGUAY: Again our finding can hopefully help newer manuals to do a better job
MAGNAVITA: One of the criteria for entering patients in clinical trials it is that they are indeed motivated—-at least at the outset. But we know from clinical experience that this can change over time. However, this is not usually monitored over the course of treatment in clinical trials—which is also true of other important moderating variables.
GOLDFRIED: And those individuals that refuse to be entered into a protocol are sometimes the kinds of individuals that we see in clinical practice. The findings of clinical trials need to be interpreted with this in mind; these are the kind of patients on which we don’t have the research data that could help the clinician.
MAGNAVITA:  And these may be the more comorbid people.
NEWMAN:  I think it would be really interesting to study if those who were screened out of RCTs and/or refused randomization do go on to pursue psychotherapy at all. We assume that they do, but we don’t really know. Some of these people may just seek medication because it doesn’t require as much effort on their part, and some may decide that they can do the work on their own. 
GOLDFRIED: There are also those individuals who drop out of clinical trials—which could be considered treatment failures. These are also the kind of individuals we see in clinical practice. In many ways, the practicing clinician has to deal with more problematic patients than does the therapy researcher. 

NEWMAN: Again, I wonder if this is true or if these are the same clients who will drop out of any therapy (RCT or private practice), given that the modal numbers of private practice sessions is one. Also, we assume that people drop out because they don’t like the treatment or because the treatment wasn’t working, but when asked why they drop most clients report that they just don’t have time to schedule regular sessions and/or do homework.

CASTONGUAY: We have to be careful about this. Researchers at Penn (Barber, DeRubeis) have made very persuasive arguments that some of the patients seen in clinical trials (in downtown Philadelphia, for examples) are very difficult and are those that private practitioners do not see!
WOLF:  A lot of this may speak to the issue of adherence–the therapist’s adherence to a treatment manual—in contrast to the decisions that clinicians have to make. The practicing therapist may err on the side of not following the treatment manual closely enough, and a fair amount of clinical judgment is required in doing that. That also translates into the question of whether or not therapists are really doing CBT or some variation of it.
GOLDFRIED: This relates to the question: what is CBT? Are there different cohorts of therapists, depending on when they were trained or is it their amount of clinical experience? The survey points to some very interesting findings in this regard. As can be seen in Table 6, the extent to which therapists more closely adhere to a CBT protocol seems to be more characteristic of less experienced therapists. Depending upon how you consider it, more experienced therapists can be seen as
Table 6
Please Indicate All Those Aspects of CBT That You Usually Use in Treating Panic:
Response Rate
Years of Experience
< 21 years (N = 211)
> 21 years (N = 115)
Psychoeducation about nature of panic
Cognitive restructuring of general beliefs associated with panic
Cognitive restructuring of feared outcomes associated with panic attacks
Cognitive relabeling of sensations triggering panic
Identification of emotional reactions to situations associated with panic
In vivo exposure to travel, open spaces and other agoraphobic situations
Breathing retraining
Simulation of panic sensations within the session
Resolution of stressful conflicts leading to panic (e.g., relationships, work)
Relaxation training
Helping patient understand developmental roots of fears
Motivational enhancement
Assertiveness training
Communication training
Independence training
either diluting a CBT intervention with other procedures, or enhancing its effectiveness by being integrative in nature. And this difference between the more and less experienced individuals may also be a function of which vintage of CBT they learned, and also when they learned CBT in their career—either as their first orientation, or after practicing from within another orientation.  
NEWMAN:  I think it also speaks to the fact that the accessibility of treatment manuals for a variety of disorders is fairly new. The newer therapists may be those who are more likely to have been introduced to the manuals in their training. If therapists were not introduced to a manual in their training, perhaps they are less likely to use one.
CASTONGUAY: Consistent with the points made by Marv and Michelle, I was struck by how more experienced clinicians are using more traditional behavioral interventions (e.g., assertiveness training) and less CBT-specific interventions to PD.
WOLF: Someone who has been trained in rational emotive therapy and someone who has been trained in CBT can both refer to themselves as “cognitive-behavior therapists,” but can be very different in what they do.  
GOLDFRIED: Exactly. As we can see in Table 6 dealing with the breakdown according to whether the therapist is more or less experienced, there are a number of interesting differences in which aspects of CBT are used.
WOLF: It certainly looks like experience is a very relevant moderator.
GOLDFRIED:  Absolutely. Although we know that level of experience is playing a role, it is not clear as to why, or as to whether it helps or hinders. But it looks like the less experienced therapist use cognitive restructuring more often—-95% as opposed to 87%, which reflects the growing ascendancy of cognitive therapy within CBT. Another practice difference that is a function of experience is the extent to which therapists simulate the sensations of panic within the session– which has been an addition to CBT practice over the years. In speaking to a number of my CBT colleagues who are practitioners, there seems to be a fair amount of variation in their attempt to use this simulation, and also their success in doing so. And while the findings of the survey indicate that younger CBT therapists make greater use of in-session simulation, it is not quite clear as to how much success they have with it.
NEWMAN: It would be important for future surveys if we discriminated between therapists’ years of practice of psychotherapy and their experience with using a particular treatment like panic control therapy. I don’t believe these are necessarily the same thing, and I think they are important distinctions to understand. Also, as noted by Marv and Linda, interoceptive exposure is a fairly new technique in the scheme of panic treatment, so this may reflect different training. In my own training of therapists, I find that therapists who haven’t used interoceptive exposure are reluctant to ask clients to do something that exacerbates their anxiety, and I believe this is probably true of experienced therapists who have not used this technique in the past. However, therapists are more likely to use these techniques and to instill confidence in their patients around the techniques when they believe that this technique is a key to helping clients and when they have seen the long-term positive impact of these techniques.
WOLF: Another difference is that more experienced therapists tend to use relaxation training, more than those with less experience. 
NEWMAN: This is also consistent with the possibility that these therapists prefer to use techniques that decrease anxiety as opposed to exposure techniques, which increase anxiety.
GOLDFRIED: And more experienced therapists are more likely to explore those situational conflicts that might be driving the panic
MAGNAVITA: The triggers of panic.
GOLDFRIED: There is a dramatic difference here, in that 74 per cent of more experienced focus on those stressful factors contributing to the panic, as compared to only 47 per cent of less experienced clinicians.
SOBELL: Why do you think that would be?  Is it because there was a greater emphasis in the past on looking at the antecedents of panic than there is now?
WOLF: The answer to that question is very important. It addresses the question of the developmental trajectory of the therapist over time. As we considered earlier, it is a question of whether or not this is the way we as therapists have been trained, and/or how the way our practice has evolved over time. We really can’t answer that question from the results of this survey, and it is very definitely an important question that needs to be researched. 
SOBELL: In many ways, it is not surprising to see that finding, as a focus on antecedents of clinical problems was the way I was originally trained in CBT some years ago.
MAGNAVITA: If we think of many of the people in this survey as being integrative in their practice—especially the more experienced ones—there is another difference in experience level, one that was not included in the survey. Based on my experience in training therapists, it is very difficult for beginning therapists to focus on what is happening within the session that might be a sample of the patient’s problem, or at least something that is relevant to it—immediacy. This is also more characteristic of a psychodynamic and experiential approach to therapy. 
GOLDFRIED:   Interesting enough, the results of a multi-site clinical trial in the use of CBT for treating panic found that more experienced therapists tended to be more successful, even though they did not differ from less experienced therapists on the basis of adherence or competence in administering the intervention (Huppert, Bufka, Barlow, Gorman, & Woods, 2001). So, clinically important things may have been happening during the sessions that go beyond what exists in the manual, which can contribute to success. But we don’t know exactly what that is.
NEWMAN: This was the finding of an RCT, so perhaps we should also examine the reported success rate of our therapists by level of experience to see whether this effect is replicated in our sample. If we replicate this effect in our sample, this may be due to the ability of more experienced therapists who know well the specific techniques in the manual and their underlying rationales and are more able to apply the treatment more flexibly and more tailored to the individual. Also, a more experienced therapist may be more confident in the efficacy of the techniques and may more readily instill that confidence in their patients. I have seen new therapists and/or those who are experienced with other approaches but just learning CBT introduce a technique reluctantly and apologetically and this almost invariably instills doubt in clients about whether they really want to engage in it.
CASTONGUAY: Another thing that strikes me about possible differences between therapists, which may have an impact on the question that Michelle raised earlier and the points discussed by many of you so far, is that experienced therapists appear to see more complex cases. I am inferring this, perhaps incorrectly, by their emphasis on issues such as substance abuse, limited premorbid functioning, perfectionistic style, and loss of family member.
MAGNAVITA: It is possible that it is related to more experienced therapists being better able to tolerate the patient’s distress level—but we don’t know for sure. In many ways, this survey highlights those gaps in knowledge about what we do that is effective. Although we have learned a great deal in recent years, these knowledge gaps continue to exist. As a practicing clinician, I see it as being very important to have research inform us more about what is actually going on that is helpful.
SOBELL: I totally agree. Indeed, I think this is one of the most exciting things that I have seen in my career.  It opens up so many questions that we really need to address empirically. 
MAGNAVITA: Speaking as a clinician who sits in a room all day working with patients, I find these questions both fascinating and important in helping me know more about what I’m doing. And clearly, these are questions that are relevant to both clinicians and researchers.
GOLDFRIED: That’s an interesting point. When we first conceived of having this survey, the thought was that it would be useful in providing clinically-driven hypotheses that would be useful for researchers to investigate. However, it also appears that the findings can be of considerable interest to practitioners, in that they can compare their own experiences to what their colleagues have been doing and experiencing.
MAGNAVITA: A lot of the findings are of really interest for the clinician.
GOLDFRIED: When we talk about differences in clinical experience regarding the limitations of CBT (Table 7), there were six interesting differences. The more experienced therapists don’t seem to have as much of a problem in getting the patient to eliminate safety behaviors. So there may be something about experience, or having a greater array of methods, that makes this relatively less of a problem. 
Table 7
Problems/Limitations Associated with the CBT Intervention Method:
Response Rate
Years of Experience
< 21 years (N = 211)
> 21 years (N = 115)
Patient’s reluctance to eliminate safety behaviors
Exposure in vivo has logistical problems
Doesn’t deal with comorbid problems/symptoms
Simulating panic in session is difficult
Triggers to panic not evident
Strict adherence to CBT protocol
Relaxation doesn’t work or causes anxiety
Absence of guidelines for dealing with resistance/noncompliance
Doesn’t deal with patient’s anger
Doesn’t deal with fear of interpersonal loss
Triggers for panic are not linked to client’s past history
Doesn’t deal with comprehensive or lasting change
Current coping skills are not linked to past
WOLF: Perhaps they have a greater tolerance for the problem.
MAGNAVITA: Yes, maybe they get less frustrated. 
NEWMAN:  Maybe more experienced therapists are more creative in searching for and defining safety behaviors. I have seen some very idiosyncratic safety behaviors in some clients and I have learned over the years to ask the question in many different ways to get at this issue. Most safety behaviors are subtle and not things you will always find on a typical list. In my mind they include things such as where in the room somebody will sit, internal focus, what they are doing during exposure, etc.
GOLDFRIED: And more experienced therapists seem to have less of a problem in dealing with comorbid issues. Perhaps this is also related to experienced therapists saying they see strict adherence to a CBT protocol as being a limitation in clinical practice.
WOLF: There is also an interesting difference in the more experienced therapist seeing the need to deal with the patient’s anger. The less experienced therapists might not identify that as a problem, perhaps because of their adherence to the CBT protocol. More experienced therapists may have a broader conceptualization of aroused states, whether we’re dealing with anxiety or anger.
NEWMAN:  I do think it may be due to a different definition of aroused states. In my own work with anxiety disorders I would agree that anger is another common expression of arousal.
GOLDFRIED: What is interesting is that there is a two-site clinical trial for treating panic disorder in progress—Penn Medical Center and Cornell Medical Center—where CBT is being compared to a psychodynamic intervention. What is relevant to what we’re discussing is that the psychodynamic approach does focus on the possible role of anger in panic disorder—particularly triggered by interpersonal conflict.
WOLF:  From a clinical point of view, I see anger as sometimes playing an important role in panic.
GOLDFRIED:  If nothing else, it can result in increased panic symptomotology—such as increased heart rate, tension, and hyperventilation.

WOLF: And also anxiety.
MAGNAVITA: It definitely results in an aroused emotional state.
NEWMAN:  Yet anger may require some different treatment techniques than the treatment of anxiety. Exposure to anger may only exacerbate the anger.
GOLDFRIED:  Another difference that can be seen in Table 7 is that more experienced therapists are more likely to see a limitation of CBT as not linking patients’ current coping skills to their past—such as becoming aware of the development roots to current coping.
NEWMAN: Certainly, the manuals do not really focus on the developmental roots to current coping beyond suggesting that clients’ first panic attack, their concurrent stress level, how they coped with their first panic attack, and whether or not they developed full-blown panic disorder. The manuals do not focus on clients learning history that may have created their predisposition to panic.
WOLF:  In many respects, while CBT views problems as having been learned, it doesn’t seem to make as much use of a developmental approach to psychopathology in its interventions. It’s very much present-focused.

NEWMAN:  Yes, an interesting contradiction. The way I have heard this explained is that what led to the development of a disorder is not necessarily what is currently maintaining the disorder, and it is more important to focus on current maintaining variables. However, this is a theoretical proposition that has not been empirically tested.
GOLDFRIED:   Some years ago, Sarason(1979) once pointed out that a limitation of CBT is the failure to recognize that cognitions have histories, and that sometimes knowing about this history can be helpful in an intervention.
WOLFE: I think that’s a very important point.
MAGNAVITA: Jeff Young’s work talks about that with regard to the historical roots of the maladaptive schemas (1994). And the important things about that is the period in one’s life when the schemas was developed, with the earlier development as creating more problems for intervention.
WOLF: We should also just mention that there was some other interesting finding about those patient symptoms, their beliefs about panic, and very importantly their social system, all of which must be taken into consideration in working with panic patients, as they can undermine treatment effectiveness. These finding are summarized in Tables 8, 9, and 10. 
Table 8
Patient’s Symptoms Related to Panic
Response Rate
Tendency to dissociate
Functional impairment
Fainting history
Table 9
Patient’s Beliefs about Panic
Response Rate
Belief that their fears are realistic
Their problems are due to external factors
Being anxious is abnormal/dangerous
Panic is biologically based
Symptom reduction will negatively impact relationships
Table 10
Social System (Home, Work, Other)
Response Rate
Symptoms/dependency is reinforced/supported
Trapped in a dysfunctional home, work, or social situation
Stress very high at home, work, or socially
Family does not support treatment
Social isolation of patient
Family is controlling and critical
Family members are very anxious
Loss of family member, partner, employment
GOLDFRIED:   We can certainly go on at greater length about these findings, and will in future conference presentations and in articles. In many respects, the survey has raised as many questions as it has answered. But then again, the purpose of surveying clinicians about their experiences in treating panic was to generate clinically relevant hypotheses for further research.


Chambless, D. L. & Goldstein, A. J. (Eds.) (1982). Agoraphobia: Multiple perspectives on theory and         treatment. NY: Wiley.
Huppert, J. D., Bufka, L. F., Barlow, D. H., Gorman, J. M., Shear, M. K., & Woods, S. W. (2001).  Therapists, therapist variables, and cognitive-behavioral therapy outcome on a multicenter trial  for panic disorder. Journal of Consulting and Clinical Psychology, 69, 747-755.
Lenzenweger, M. F., Lane, M. C., Loranger, A. W., & Kessler, R. C. (2007). DSM-III personality     disorders in the national comorbidity study. Biological Psychiatry, 62, 553-564.
Sarason, I. G. (1979). Three lacunae of cognitive therapy.  Cognitive Therapy and Research, 3, 223-235.
Tsao, J. C., Lewin, M. R., & Craske, M. G. (1998). The effects of cognitive-behavior therapy for panic disorder on comorbid conditions. Journal of Anxiety Disorders, 12, 357-371.
Tsao, J. C., Mystkowski, J. L., Zucker, B. G., & Craske, M. G. (2005). Impact of cognitive-behavioral therapy for panic disorder on comorbidity: A controlled investigation. Behaviour Research and Therapy, 43, 959-970.   
Young, J. E. (1994). Cognitive therapy for personality disorders: A schema-focused approach (Rev. ed.), Sarasota, FL: Professional Resource Exchange.