Interpreting A Quantitative Journal Article 1&2. What I Need Specifically Is Someone To Read The 6-Page Article And Then Read My Paper Draft To Advise What I Have Missed Or Misinterpreted

Quantitative Journal Article

Running head: ASSIGNMENT 1: QUANTITATIVE JAR 1

ASSIGNMENT 1: QUANTITATIVE JAR 2

Assignment 1: Quantitative Journal Article Review and Topic Declaration

Abstract

This assignment is comprised of a Journal Article Review (JAR) related to the topic chosen for the Final Paper: Factors contributing to the success of Cognitive-Behavioral Therapy (CBT) in treating Panic Disorder. The article under review, The Effects of Treatment Compliance on Outcome in Cognitive-Behavioral Therapy for Panic Disorder: Quality Versus Quantity, employs a quasi-experimental design and is relevant to the topic for the Final Paper because it demonstrates the effectiveness of group CBT in the treatment of panic disorder, particularly when there is quality compliance with group exercises related to target outcomes, and that quality of work when compared to quantity of work, better predicts therapy outcome. The authors argue these findings are important for therapists to consider when designing good quality work and training patients to complete high quality assignments. Such considerations can contribute to the success of CBT in treating panic disorder.

Keywords: cognitive-behavioral therapy, CBT, panic disorder, anxiety

Table of Contents

Abstract ……………………………………………………………………………………………………….2

JAR: Quantitative Review…………………………………………………………………………………….4

References…………………………………………………………………………………………………….?

Assignment 1: Quantitative Journal Article Review and Topic Declaration

Introduction

Statement of Problem

This research aims to fill a gap in the literature on addressing quality versus quantity of assignments completed by patients receiving cognitive-behavioral therapy (CBT) for treatment of panic disorder (PD). The researchers evaluate the relationship between homework quality and outcome as they argue, based on clinical experience, that quality of homework completed is more important to recovery.

Literature Review

The authors review the literature evaluating the effectiveness of homework compliance on outcome of CBT. They argue a collaborative-empiricist approach to treatment requires treatment compliance for therapeutic change. The authors cite literature which supports clinical improvement for homework compliance (Edelman & Chambless, 1993; Barlow, O’Brien, & Last, 1984; Michelson, Mavissakalian, Marchione, Dancu, & Greenwald, 1986) and research by Mavissakalian and Michelson (1983) which does not support this relationship. Primakoff, Epstein, and Covi (1986) noted the gap in the literature on assessing quality versus quantity of homework, which has influenced the present study.

Hypotheses To Be Tested

The hypotheses to be tested are: 1. The overall quantity of the work completed by patients will be positively associated with treatment outcome; and 2. The relative quality of homework will better predict outcome.

Method

Smith and Woolaway-Bickel (2000) employ a quasi-experimental one-group pretest-posttest design. To control for within-group differences and increase internal validity, researchers placed stability requirements on medication use and accepted participants who were not suicidal or abusing substances, and had no history of schizophrenia, bipolar disorder, or organic mental disorder.

Participants

The sample consisted of 48 patients, mostly White (86%), female (66%), married (63%), and employed (52%) who met the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV) criteria for PD with or without agoraphobia. Most patients (52%) were medicated. The participants were recruited by convenience sampling.

Materials

Participants were assessed pre- and post-treatment for clinical dimensions of PD. If panic frequency, anxiety, and phobic avoidance scores fell within the normal range post-treatment, participants were classified as recovered. Panic attack symptoms ratings by patients had adequate test-retest reliability throughout treatment and were significantly associated with corresponding self-report measures and clinic ratings (Schmidt & Woolaway-Bickel, 2000, p. 14). Videotaped interviews were randomly selected for viewing to establish interrater agreement on DSM-IV Axis I diagnoses. Interval scale measures such as The Multicenter Panic Anxiety Scale (MC-PAS), Sheehan Patient-Rated Anxiety Scale (SPRAS), Mobility Inventory for Agoraphobia (MI-Alone; MI-Accompanied), Sheehan Disability Scale, and Beck Depression Inventory (BDI) were administered pre- and post-treatment to measure self-rated disability and clinician-rated impairment. Researchers monitored participants’ panic symptoms and compliance through the Uniformed Services University of the Health Sciences (USUHS) Weekly Monitoring System.

Procedure

Patients received 12 sessions over 12 weeks of group CBT for panic disorder. Homework was assigned based on session exercises and individually-tailored to type of panic experienced by participants. Treatments were structured, manualized, well-defined, and administered by an experienced psychologist. Participants were assessed weekly on their subjective sense of time spent on homework and compliance. A quality ideal was operationally defined for therapists and participants. Participant quality compliance was measured with a 6-point Likert scale (0 = poor to 5 = excellent). To control for therapist bias on compliance ratings, an independent rater completed weekly quality and quantity ratings, thus increasing inter-rater reliability and internal validity. Inter-rater reliability for quantity and quality were high.

Statistics

Descriptive statistics, including measures of central tendency, frequency, and variation were used to describe data. Additionally, inferential statistics such as Pearson’s r and paired t-test values were calculated, and a stepwise multiple regression analysis was performed.

Results

Overall, patients completed 60% of home assignments, as rated by the therapist, with good quality (mean 2.8; SD – 0.9; range = 0.3 – 4.8). Quantity remained consistent while quality increased over time. Patient and therapist rating correlations were significant at r = .28, ps < .05. Researchers found that older and unemployed patients submitted higher quality work; thus, “age and employment status were covaried for subsequent analysis with the quality ratings” (p. 15). Paired t tests showed significant improvement to panic disorder based on treatment protocol, with high recovery from panic attacks (94%), anxiety (83%), and phobic avoidance (71%). Importantly, the authors demonstrated that, consistent with pervious research, group-administered CBT is “highly effective for panic disorder” (p. 15). Therapist ratings indicated that quality ratings are “somewhat better predictors of outcome” (p. 16); whereas, researchers found they significantly predicted outcome in .

Stepwise multiple regression analyses indicated that clients ratings of compliance were not correlated with outcome, but therapist ratings were and that quality ratings are “somewhat better predictors of outcome” ; independent evaluator ratings and outcomes mimicked those of the therapist, though were “less strongly predictive.” In six of the sessions, compliance was significantly associated with “some outcome variable” and “compliance with particular skills is associated with changes in a related symptom variable” Found that “compliance exerts incremental effects on symptomatology that is best evidence over the entire course of treatment” (p. 17). Researchers found “compliance with particular skills is associated with canges in related symptom variable” Researchers observed that impacts of compliance were not immediate but were noted over time.

Implications

Implications for Counselors, Clients, and Counselling

As noted by the authors, quality of work and training of patients in quality of assignment completion are important factors in CBT success. Also, group CBT therapy is an effective way to employ CBT for panic disorder. Counsellors should understand that immediate effects of CBT may not be immediately evident and may incrementally improve over time. It is important that counsellor and client agree on meaning of quality work. This study is a great example of applied research as findings monitoring of quality and quantity or work between counsellor and client, developing a common understanding “The goal of CBT is to get patients actively involved in their treatment plan” (https://www.psychologytoday.com/ca/therapy-types/cognitive-behavioral-therapy-0 ). Having clients assess their own compliance and understand and be trained in quality completion can enhance treatment success. Support your assertions with citations, using the text and additional references where possible.]

Discussion

Summary

Group CBT is highly effective in treating panic disorder. Homework assignments administered to patients based on clinical status can build skills necessary to reduce panic symptoms. Note that these findings are important in that more attention could be given to quality of individualized homework over quantity. Strengths of this research include control of: ( …..) plus inter-rate reliability which lends internal validity. However, there are limitations to this study to consider: Weaknesses = no random assignment, hard to control for extraneous variables, but medication and mental health controlled for; variation in treatment makes it difficult to directly compare, small sample size; [Summarize the main results with reference to the original hypotheses. To what extent were the hypotheses supported by the findings?] Threats to internal validity: history (repeat applicants); maturation (better as study progresses); test and homework fatigue, statistical conclusion validity? Power? Variable types of homework administered to patients. Control for researcher and participant bias (psychologist) with independent rater.

Interpretation

[Are the conclusions made by the author(s) warranted? What are the possible alternative explanations of the results? What are the strengths of the study? What was done well? What are the weaknesses/limitations of the study? Identify the specific design and describe the salient features (it must be a quantitative design for JAR 1; if you have chosen an experiment for JAR 1, explain whether it is a true experiment or a quasi-experimental design and why). Incorporate your knowledge of course-related concepts in your discussion of these questions (e.g., ethical considerations, reliability/validity of measures, ability to draw causal conclusions, threats to the internal/external validity of the study, generalizability of the study, confounding variables and sources of bias). Support your claims with material from the text and other scholarly references, as appropriate.] power was noted to be compromised due to independent rater examining a subset of participants (p. 16) Researchers found no significant different in “demographics, symptom severity at pretreatment, or average level of compliance during completed portions of the treatment (ps > .05)” between attritors (n = 10) and completers (n – 38)

For Further Study

It would be interesting to perform this study with a control group receiving individual CBT to those receiving group CBT to monitor how individuals may or may not be motivated to complete homework based on group dynamic. Motivation to complete these assignments could have come a Hawthorne effect, participant bias, being observed by other individuals.

References

Surname, Initial., Surname, Initial., & Surname, Initial. (Year). Title of journal article with no caps. Title of Journal, Volume number (issue number), page numbers.

Surname, Initial. (Year). Book title: No caps except after colon or if the capitalization is for proper nouns like names of assessment tools. City, State or Province written as 2 letter abbreviation, such as ON: Publisher Name.

Assignment 1 and Assignment 2 Templates

General Description of and Template for Assignments 1 and 2: Journal Article Reviews

(Assignment 1 must include Final Paper Topic declaration)

For Units 3 and 4 of PSYC 6213, complete Assignments 1 & 2 using the template that follows. Refer to this description and the instructions in the syllabus. consult the rubric to see how points are allotted for this assignment. For Assignment 1, you will be declaring the topic for the Final Paper (in the abstract) and reviewing a quantitative journal article relating to the topic you have chosen for the Final Paper. Your topic must be relevant to the field of counselling psychology. For Assignment 2, you will be reviewing a qualitative journal article relating to the topic you have chosen for the final paper.

Please find full text articles in the Yorkville University online library. You may also use scholarly peer-reviewed journals with full text articles available online. Ideally, you will use this assignment to contribute to your topic selection and develop your understanding of the subject for your final paper.

The purpose of an article review is to communicate the main points of a study and to discuss both the implications and strengths and weaknesses of the design. The review is not meant to re-iterate the details of the study and paraphrase each and every point; rather, it should summarize and discuss the study. These assignments should be seen as an opportunity for you to demonstrate your understanding of the course concepts, particularly in the Discussion section. Quotes should be used sparingly or not at all. Your JAR assignments should be about two – three pages in length, not counting the title and reference pages. There are no penalties for going slightly over this limit, but the final length should be close. As a rule of thumb, aim to write about 1/3 page for each of the five main section headings in the template (Introduction, Method, Results, Implications, Discussion).

Submissions should be double-spaced and follow APA format. Please note that APA includes attention to word choice, grammar, punctuation, syntax, page and heading formatting, as well as in-text citation and referencing style. Do not use a numeral to start a sentence (use word form or re-word the sentence to avoid beginning with a number), do not use bullet points, and if you are presenting information that you have not read first-hand (i.e., a secondary source), indicate this using the (as cited in….) convention, as per APA. Note that the secondary source should not be included on your reference page. See Appendix A in the Goodwin text, pp. 379-398, for important guidelines to follow in your writing.

For JAR 1, be sure that the research design in the article selected is quantitative. You MUST distinguish the design type in the discussion. For JAR 2, the design must be qualitative and utilize a formal qualitative methodology (rather than simply asking open-ended questions and reporting the results).

Note: Single-subject case studies are not permitted in either review. For each article, the methodology (statistics for the quantitative studies) and results should be clearly discussed. Meta-analysis is a way to quantitatively synthesize research results; it is not a formal quantitative research design.

Running head: JAR 1 1

Assignment 1: Quantitative Journal Article Review and Topic Declaration

Your Name Here

JAR 1 2

Abstract

(In the abstract, present the study you are reviewing, name the design, and briefly explain how it relates to the topic chosen for the Final. The topic for the final will be presented as one single sentence that will serve as your working title as you undertake your research. See example abstract below, which you may use as a guide.)

Keywords:

This assignment is comprised of a Journal Article Review (JAR) related to the topic chosen for the Final Paper: [Insert one sentence describing the topic you have chosen for the Final Paper (e.g., The use of Animal-Assisted Therapy for treating Selective Mutism in preschool children)]. This article, Insert Title Here, employs a [insert name of the quantitative design (e.g. factorial, correlational, quasi-experimental)] design and is relevant to the topic of the Final because […].

JAR 1 3

Table of Contents

Abstract ……………………………………………………………………………………………………….ii

JAR: Quantitative Review [2 – 3 pages, double-spaced] ……………………………………..1

References…………………………………………………………………………………………………….5

JAR 1 4

Title of Paper

Introduction

Statement of Problem

[What is the research problem being addressed and the specific research question(s)? What are the researchers trying to do? Why?]

Literature Review

[Briefly describe important research that guided or precipitated this study. Include citations. Note: this should not be an exhaustive restatement of the article’s introduction; just summarize the key theory or empirical findings that provide the background and rationale for the study being reviewed.]

Hypotheses To Be Tested

[What are the hypotheses of this study? What did the researchers predict they would find?]

Method

[Note: the content under each of these headings shouldn’t take more than one or two lines.]

Participants

[Who were the participants (demographics) and how were they recruited/selected?]

JAR 15

Materials

[Briefly make note of the materials/measures that were used to assess the variables and assess reliability and validity.]

Procedure

[Briefly describe the procedure – what was done in the study?]

Statistics

[Name the primary statistical design(s) employed. How were the data analysed? What inferential analyses were performed? Which statistics and/or tests of significance?]

Results

[Briefly describe the main findings (include numerical values and alpha levels where appropriate).]

Implications

Implications for Counselors, Clients, and Counselling

[What are the implications of the findings? Support your assertions with citations, using the text and additional references where possible.]

Discussion

Summary

[Summarize the main results with reference to the original hypotheses. To what extent were the hypotheses supported by the findings?]

JAR 1 6

Interpretation

[Are the conclusions made by the author(s) warranted? What are the possible alternative explanations of the results? What are the strengths of the study? What was done well? What are the weaknesses/limitations of the study? Identify the specific design and describe the salient features (it must be a quantitative design for JAR 1; if you have chosen an experiment for JAR 1, explain whether it is a true experiment or a quasi-experimental design and why). Incorporate your knowledge of course-related concepts in your discussion of these questions (e.g., ethical considerations, reliability/validity of measures, ability to draw causal conclusions, threats to the internal/external validity of the study, generalizability of the study, confounding variables and sources of bias). Support your claims with material from the text and other scholarly references, as appropriate.]

For Further Study

[Recommend what should be done next in this area. How can this work be followed up or extended?]

JAR 1 7

References [alphabetical order]

Surname, Initial., Surname, Initial., & Surname, Initial. (Year). Title of journal article with no caps. Title of Journal, Volume number (issue number), page numbers.

Surname, Initial. (Year). Book title: No caps except after colon or if the capitalization is for proper nouns like names of assessment tools. City, State or Province written as 2 letter abbreviation, such as ON: Publisher Name.

Running head: JAR 2 1

Assignment 2: Qualitative Journal Article Review

Your Name Here

JAR 2 2

Abstract

(In the abstract, present the study you are reviewing, name the design, and briefly explain how it relates to the topic chosen for the Final. The topic for the final will be presented as one single sentence that will serve as your working title as you undertake your research. See example abstract below, which you may use as a guide.)

Keywords:

This assignment is comprised of a Journal Article Review (JAR) related to the topic chosen for the Final Paper: [Insert one sentence describing the topic you have chosen for the Final Paper (e.g., The use of Animal-Assisted Therapy for treating Selective Mutism in preschool children)]. This article, Insert Title Here, uses a [insert name of formal qualitative methodology (e.g., grounded theory, phenomenological, ethnographic)] design and is relevant to the topic of the Final because […]. )

JAR 2 3

Table of Contents

Abstract ……………………………………………………………………………………………………….ii

JAR: Qualitative Review [2 – 3 pages, double-spaced] ………………………………………..1

References…………………………………………………………………………………………………….5

JAR 2 4

Title of Paper

Introduction

Statement of Problem

[What is the research problem being addressed and the specific research question(s)? What are the researchers trying to do? Why?]

Literature Review

[Briefly describe important research that guided or precipitated this study. Include citations. Note: this should not be an exhaustive restatement of the article’s introduction; just summarize the key theory or empirical findings that provide the background and rationale for the study being reviewed.]

Research Questions Being Investigated

[What are the primary research questions of this study? What were the researchers interested in exploring?]

Method

[Note: the content under each of these headings should not take more than one or two lines.]

Participants

[Who were the participants (demographics) and how were they recruited/selected?]

JAR 2 5

Materials

[Briefly describe the materials that were used to collect the data.]

Procedure

[Briefly describe the procedure – what was done in the study?]

Data Analysis

[How were the data collected? How were the data analysed?

Results

[Briefly describe the main findings. Discuss the clinical relevance of the findings.]

Implications

Implications for Counselors, Clients, and Counselling

[What are the implications of the findings? Support your assertions with citations, using the text and additional references where possible.]

Discussion

Summary

[Summarize the main results with reference to the original research question. To what extent did the findings support the research question?]

Interpretation

JAR 2 6

[Are the conclusions made by the author(s) warranted? What are the possible alternative explanations of the results? What are the strengths of the study? What was done well? What are the weaknesses/limitations of the study? Identify the specific design and describe the salient features. Incorporate your knowledge of course-related concepts in your discussion of these questions (e.g., ethical considerations). Support your claims with material from the text and other scholarly references, as appropriate.]

For Further Study

[Recommend what should be done next in this area. How can this work be followed up or extended?]

JAR 2 7

References [alphabetical order]

Surname, Initial., Surname, Initial., & Surname, Initial. (Year). Title of journal article with no caps. Title of Journal, Volume number (issue number), page numbers.

Surname, Initial. (Year). Book title: No caps except after colon or if the capitalization is for proper nouns like names of assessment tools. City, State or Province written as 2 letter abbreviation, such as ON: Publisher Name.

PSYC 6213 – Journal Article Reviews Template 16

Journal of Consulting and Clinical Psychology 2000, Vol. 68, No. 1, 13-18

Copyright 2000 by the American Psychological Association, Inc. 0022-006X/00/$5.00 DOI: 10.1037//0022-006X.68.1.13

The Effects of Treatment Compliance on Outcome in Cognitive-Behavioral Therapy for Panic Disorder: Quality Versus Quantity

Norman B. Schmidt and Kelly Woolaway-Bickel Ohio State University

Cognitive-behavioral therapy (CBT) is skill based and assumes active patient participation in regard to treatment-related assignments. The effects of patient compliance in CBT outcome studies are equivocal, however, and 1 gap in the literature concerns the need to account for the quality versus the quantity of assigned work. In this study, both quality and quantity of home-based practice were assessed to better evaluate the effects of treatment compliance in patients with panic disorder (N = 48) who participated in a 12-session CBT protocol. Patient estimates of compliance were not significantly associated with most outcome measures. On the other hand, therapist ratings of compliance significantly predicted positive changes on most outcome measures. Moreover, therapist and independent rater estimates of the quality of the participant’s work, relative to the quantity of the work, were relatively better predictors of outcome.

Acquisition of new skills and knowledge is a distinctive feature of cognitive-behavioral therapy (CBT). Skill acquisition assumes active patient participation in regard to relevant treatment-related assignments. Such assignments, often conducted out of session as “homework,” are one of the hallmarks of CBT. In fact, a collaborative-empiricist approach to treatment presumes that compliance with treatment, within and outside of sessions, is a central and necessary mechanism for therapeutic change.

Surprisingly, the literature evaluating the effects of homework compliance on outcome has been inconclusive. In their review, Edelman and Chambless (1993) found equivocal evidence for a relationship between completion of between-sessions assignments and clinical improvement. Many CBT outcome studies of depres- sion and anxiety indicate that compliance is significantly related to treatment outcome, although a similar number do not show the predicted relationship.

CBT for panic disorder is a case in point. Newer CBT protocols for panic disorder emphasize a number of skills, including cogni- tive restructuring, interoceptive exposure, breathing control proce- dures, education, and in vivo therapy techniques (Wolfe & Maser, 1994). Despite the consistent support for the efficacy of CBT for panic disorder, evaluation of homework compliance has not been entirely consistent with some studies that suggest a significant relationship (Barlow, O’Brien, & Last, 1984; Edelman & Chamb- less, 1993; Michelson, Mavissakalian, Marchione, Dancu, & Greenwald, 1986) and other studies that do not (Mavissakalian & Michelson, 1983).

Norman B. Schmidt and Kelly Woolaway-Bickel, Department of Psy- chology, Ohio State University.

This research was supported by Grant R072CF from the Uniformed Services University of the Health Sciences.

Correspondence concerning this article should be addressed to Norman B. Schmidt, Department of Psychology, Ohio State University, 245 Town- shend Hall, 1885 Neil Avenue Mall, Columbus, Ohio 43210-1222. Elec- tronic mail may be sent to schmidt.283@osu.edu.

Previous reviews of the compliance literature have suggested a number of limitations, but one major design gap in previous work has been a failure to assess the quality (vs. the quantity) of homework (Primakoff, Epstein, & Covi, 1986). Clinical experi- ence suggests that the absolute level of homework completed might be less important to recovery than the quality of the work done. For example, some patients may do quite a bit of work but do it improperly or incompletely, whereas others carefully adhere to the prescribed exercises. In this study, we were particularly interested in evaluating the relationship between homework qual- ity and outcome. Consistent with prior work (Edelman & Chamb- less, 1993), we hypothesized that the overall quantity of the work completed by patients would be positively associated with out- come. However, we hypothesized that the relative quality of home- work completed would be a better predictor of outcome.

Method

Participants

The sample consisted of 48 patients meeting the following criteria: (a) principal Axis I diagnosis of panic disorder with or without agoraphobia, as described in the Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association, 1994); (b) no change in medication type or dose during the 12 weeks prior to treatment; (c) no evidence of serious suicidal intent; (d) no evidence of current substance abuse; and (e) no evidence of current or past schizophrenia, bipolar disorder, or organic mental disorder. Sixty-six percent of the patients were women with an average age of 35 years (SD = 11.8). A majority of the patients were White (86%), married (63%), and employed (86%). Fifty-two percent of the patients were taking psychotropic medications (37% benzo- diazepines, 25% antidepressants), and 18% of the patients were taking both benzodiazepines and antidepressants. Patients taking medications met sta- bility requirements such that initial administration of the medication and dose had been maintained for at least 3 months prior to treatment and maintained until posttreatment.

Procedure

Patients were consecutive applicants who presented for evaluation at an academic research center specializing in the assessment and treatment of

13

14 SCHMIDT AND WOOLAWAY-BICKEL

anxiety disorders and met the study criteria. Diagnostic assessment was based on an initial phone screening interview followed by a face-to-face structured clinical interview with the Structured Clinical Interview for the DSM-IV—Patient Edition (SCID-NP; First, Spitzer, Gibbon, & Williams, 1994). Randomly selected videotaped interviews from our laboratory have demonstrated acceptable kappa coefficients for interrater agreement for all DSM-TV Axis I diagnoses (see Schmidt, Trakowski, & Staab, 1997).

Patients received a group-administered cognitive-behavioral treatment for panic disorder that consisted of 12 sessions over a 12-week period. The treatment protocol included four major components: (a) education and corrective informatioE regarding the etiology and maintenance of panic disorder, (b) cognitive restructuring, (c) interoceptive exposure, and (d) in vivo exposure (cf. Schmidt, Staab, Trakowski, & Sammons, 1997).

Treatment integrity was maintained with a structured and manualized treatment protocol (Schmidt, 1994) that describes specific interventions for each session. At the end of each session, patients were assigned one or more homework tasks on the basis of the types of skills covered during the session (e.g., sensation induction procedures followed the introduction of interoceptive exposure). These tasks were individualized on the basis of each patient’s clinical profile (e.g., patients with hypersensitivity to dys- pnea cues might be assigned to repeatedly hyperventilate, whereas those with cardiac fears might be assigned to complete aerobic exercises).

The treatment was administered by a licensed psychologist with over 10 years of experience with CBT for anxiety disorders. An independent rater, Kelly Woolaway-Bickel, evaluated videotape from 48 hr of randomly selected treatment sessions to assess adherence to the treatment protocol. Adherence was assessed in terms of the content of materials presented during the prescribed treatment session relative to its description in the treatment manual (e.g., therapist reviews homework for each patient and therapist describes cognitive model of panic). This evaluation yielded 100% adherence according to an adherence rating scale (Young, Beck, & Budenz, 1983, as cited in Primakoff et al., 1986).

Approximately 80% (38/48) of the patients who began treatment were assessed at posttreatment. Reasons for dropout included moving (n = 3), transportation difficulties (n = 2), scheduling conflicts (n = 2), and unknown (n = 3). Comparisons between completers and noncompleters indicated no significant differences in demographics, symptom severity at pretreatment, or average level of compliance during the completed portions of the treatment (ps > .05).

Measures

An assessment battery tapping the major clinical dimensions of panic disorder was administered to all the patients at baseline and posttreatment. This battery consisted of clinician-rated measures of panic frequency, intensity, anticipatory anxiety, and impairment (Multicenter Panic Anxiety Scale [MC-PASJ; Shear et al., 1997), as well as self-rated measures of anxiety (Sheehan Patient-Rated Anxiety Scale [SPRAS]; Sheehan, 1983), phobic avoidance when alone and accompanied (Mobility Inventory for Agoraphobia (Mi-Alone, Mi-Accompanied]; Chambless, Caputo, Jasin, Gracely, & Williams, 1985), disability (Sheehan Disability Scale [SDS]; Ballenger et al., 1988), and depression (Beck Depression Inventory [BDI]; Beck & Steer, 1993).

In addition to self-rated disability and clinician-rated impairment, a composite measure of clinically significant change was computed for evaluation of end-state functioning at posttreatment. A patient was classi- fied as recovered when scores on each of three central symptom dimen- sions of panic disorder fell within the normal range of functioning (i.e., panic frequency = 0, anxiety [SPRAS] < 30, and phobic avoidance [Mi-Alone] < 1.5).

Patients also completed a weekly monitoring form (Uniformed Services University of the Health Sciences [USUHS] Weekly Monitoring Form) at the beginning of each treatment session. The USUHS Monitoring Form provides an average estimate of panic disorder—related symptoms, as well

as compliance estimates. At the beginning of each session, patients record the number of panic attacks and overall symptom ratings for the previous week. Ratings of anxiety, panic-related worry, fear of bodily sensations, and depression are made on an 11-point scale ranging from 0 (none) to 10 (extreme). In this sample, each of these estimates was found to possess adequate test-retest reliability across sessions (r = .65-.81), and each was significantly associated with corresponding self-report measures (r = .48-69) and clinician ratings (r = ,43-.79).

At each session, every patient was assigned a number of home-based assignments that were based on interventions covered in the group and individually tailored to the patient’s presenting concerns. After the first 2 weeks, during which assignments were more limited, most patients were routinely assigned four or more tasks to complete prior to the next session. Assignments were accompanied by practice forms consistent with the assigned tasks. For example, patients would be given three cognitive restructuring practice forms and told to complete all three during the next week.

The patients’ subjective sense of the amount of time practicing was assessed on a weekly basis. At the beginning of each session, patients made two compliance estimates, including the number of days and hours spent conducting treatment-related assignments on the weekly monitoring form.

The therapist made ratings of quantity and quality on a weekly basis. After each session, the therapist, Norman B. Schmidt, collected the home- work practice forms and compared those forms with the prescribed home- work for each patient. The quantity compliance rating was the percentage of assigned homework that was completed (0%-100%). Patients were also credited for appropriately substituting assignments for therapist-prescribed work.

The quality compliance rating was based on the overall quality of the work conducted during the previous week (0 = poor, 1 = marginal, 2 = fair, 3 = good, 4 = very good, 5 = excellent). Quality estimates captured the type of work conducted relative to an ideal. We established the ideal criteria and explained them to patients in the session where we introduced the technique. For example, in completing in vivo or interoceptive expo- sure exercises, we considered five main criteria in evaluating homework quality: (a) whether a specific task was identified; (b) whether the task generated moderate levels of fear; (c) whether the task was repeated until fear was extinguished; (d) whether the patient identified any “safety aids” (i.e., coping strategies used to assist them in dealing with their fear during exposure); and (e) whether safety aids were successfully faded during practice and, in the case of interoceptive exposure, whether the exercise produced a sufficiently high level of sensation. We considered four criteria in rating cognitive restructuring exercises: (a) whether an anxiogenic cognition was appropriately identified, (b) whether the patient evaluated the evidence in support and against this thought, (c) whether an appropriate behavioral experiment was developed in response to identification of an unrealistic thought, and (d) whether the patient completed the behavioral experiment. We completed these quality ratings for each assignment and averaged them for an overall weekly rating.

Therapist ratings of compliance could be biased by clinical status (i.e., ratings based on symptom change rather than work conducted). An inde- pendent rater, Kelly Woolaway-Bickel, who did not conduct therapy ses- sions and was unaware of patient outcome and therapist ratings also made quantity and quality ratings on the basis of 48 hr of randomly selected videotape from therapy sessions, along with homework monitoring forms. Reliability estimates between these ratings and those of the therapist were high for both quantity (r = .80) and quality ratings (r = .84).

Results

Compliance Ratings

Evaluation of compliance ratings indicated that patients reported working an average of 3 hr/week (SD = 1.6, range = 0.4-7.4) and

COMPLIANCE AND PANIC 15

Table 1 Means and Standard Deviations for Major Outcome Measures From Pretreatment to Posttreatment

Pretreatment Posttreatment

Measure M SD M SD r(37)

Clinician ratings Panic frequency (0-4) Panic intensity (0-4) Anticipatory anxiety (0-4) Overall severity (1-7)

Self-report SPRAS (0-140) Mi-Accompanied (1-5) Mi-Alone (1-5) Disability (1-5) BDI (0-63)

1.9 2.6 2.2 4.6

54.1 1.7 2.2 3.5

14.4

1.1 1.1 1.2 1.7

27.2 0.6 0.9 1.0 8.8

0.7 1.3 0.7 1.6

16.0 1.2 1.4 2.1 6.9

0.8 1.0 0.8 1.0

14.8 0.3 0.6 1.1 7.3

6.83** 5.71** 9.42** 8.98**

8.12** 5.71** 5.56** 5.59** 6.40**

Note. Panic attack frequency, panic intensity, anticipatory anxiety, and overall severity are derived from the Multicenter Panic Anxiety Scale. SPRAS = Sheehan Patient-Rated Anxiety Scale; Mi-Accompanied = When Accompanied subscale of the Mobility Inventory for Agoraphobia; Mi-Alone = When Alone subscale of the Mobility Inventory for Agora- phobia; Disability = Work and Social Disability subscale of the Sheehan Disability Scale; BDI = Beck Depression Inventory. **p < .0001.

an average of 3 days/week (SD = 1.4, range = 0.5-6.4) on treatment-related assignments. Overall, the level of work reported remained relatively stable across the first 10 weeks (3-4 hr/week) and increased to 5-6 hr/week during Sessions 11 and 12. Therapist compliance ratings indicated that patients completed 60% (SD = 17.9, range = 14.1-93.6) of assignments for the entire treatment. Quality ratings were largely rated in the good range with an overall mean rating of 2.8 (SD = 0.9, range = 0.3-4.8). Quantity ratings remained in the 55% to 65% range across all sessions. Quality ratings tended to increase over time, with ratings averaging below 3.0 through Session 6 and ratings averaging above 3.0 during Sessions 9-12, suggesting that patients were learning to conduct higher quality assignments. Patient compliance ratings (hours vs. days) were significantly and moderately corre- lated (r = .43, p < .01), and therapist ratings (quantity vs. quality) were highly correlated (r = .93, p < .01). Correlations between patient and therapist ratings were somewhat lower, but all were significant (r = .28-.37.ps < .05).

Relationship Between Demographic Variables and Compliance

Preliminary analyses for the association between demographic variables, medication status, and compliance revealed significant positive relationships between the quality compliance rating and age (r = .27, p < .05) and employment status, F(2, 55) = 5.50, p < .01, with older and unemployed individuals showing higher quality work. Medication status was not significantly predictive of compliance (ps > .05). Age and employment status were covaried for subsequent analyses with the quality ratings.

Response to Treatment

At posttreatment, a consistent pattern of findings emerged (see Table 1). Paired t tests indicated significant improvement on all

measures (ps < .0001). We examined the clinical significance of the treatment findings by evaluating the composite recovery index that represents a conservative estimate of recovery from panic disorder. We calculated the percentage of patients falling within the recovered range on each of the three main clinical dimensions of panic disorder. The recovery rate at postreatment was 94% for panic attacks, 83% for anxiety, and 71% for phobic avoidance. When we used the more stringent criterion, 63% of the treated group evidenced recovery on all three measures. These findings are consistent with controlled treatment outcome studies, suggest- ing that group-administered CBT is highly effective for panic disorder (Schmidt, Staab, et al., 1997; Telch et al., 1993).

Relationships Between Compliance and Outcome

We used stepwise multiple regression analyses to evaluate the relationship between compliance and treatment outcome. These analyses are summarized in Table 2. First, we entered each pre- treatment outcome measure (outcome variables were analyzed separately), thereby creating a residualized change score. In the case of analyses using the quality ratings, age and employment status were also entered. Finally, each of the four compliance variables was entered separately.

As can be seen in Table 2, a fairly consistent pattern of findings emerged, with the patient ratings showing no significant relation- ship to outcome. The only significant finding for the patient ratings suggested that the overall number of days spent conducting home- work was related to increased levels of self-rated disability (j3 = 0.34, p < .05). However, therapist ratings of compliance were consistently and significantly related to clinical improve-

Table 2 Standardized Beta Weights for Patient and Therapist Ratings of Compliance With Reactivity From Pre- to Posttreatment on Major Outcome Measures

Patient ratings

Posttreatment measure

Clinician ratings Panic frequency Panic intensity Anticipatory anxiety Impairment

Self-report SPRAS Mi-Accompanied Mi-Alone Disability BDI

Days

-0.07 0.09

-0.09 -0.22

0.00 -0.07 -0.18

0.34* 0.13

Hours

0.00 -0.18 -0.14 -0.16

-0.06 0.02

-0.03 0.19 0.01

Therapist ratings

Quantity

-0.17 -0.27 -0.30* -0.50*

-0.32* -0.20 -0.36* -0.23 -0.11

Quality

-0.22 -0.42a* -0.32* -0.51*

-0.45″* -0.31″* -0.32* -0.34″* -0.21

Note. All analyses controlled for pretreatment severity, and analyses of quality ratings additionally controlled for age and employment status. Panic frequency = logarithmically transformed frequency of panic attacks; SPRAS = Sheehan Patient-Rated Anxiety Scale; Mi-Accompanied = When Accompanied subscale of the Mobility Inventory for Agoraphobia; Mi-Alone = When Alone subscale of the Mobility Inventory for Agora- phobia; Disability = Work and Social Disability subscale of the Sheehan Disability Scale; BDI = Beck Depression Inventory. ” Tests of correlated correlations indicated significant differences between quantity and quality ratings. * p < .05.

16 SCHMIDT AND WOOLAWAY-BICKEL

ment. Although therapist ratings of quantity and quality are highly related, the pattern of findings indicates that the quality ratings are somewhat better predictors of outcome.

Next, we directly compared the therapist quality and quantity ratings for each outcome variable by using a method for comparing correlated correlations (see Meng, Rosenthal, & Rubin, 1992). In this method, the difference in correlation between each rating and the same outcome can be transformed to a z score and compared. Our comparisons of these correlations indicated that the quality ratings were significantly better predictors than the quantity ratings in four of the seven cases where the quality rating was a significant predictor.1

Independent Evaluator Ratings and Outcome

A similar regression-based approach was used to evaluate the relationship between quality and quantity ratings and outcome for the independent rater. Because the evaluator examined only a subset of patients, power is somewhat compromised. The correla- tion between quality and quantity ratings was high and significant (r = .69). As can be seen in Table 3, this pattern of findings is generally consistent with those obtained from the therapist ratings, although the independent evaluator ratings were generally not as strongly predictive. Once again, the quality ratings are somewhat better predictors of outcome relative to the quantity ratings. Spe- cifically, quality ratings significantly predicted outcomes in four measures compared with only one measure for quantity ratings. In two cases, quality ratings were significantly better predictors rel- ative to quantity ratings (Meng et al., 1992).

Table 3 Standardized Beta Weights for Independent Evaluator Ratings of Compliance With Reactivity From Pre- to Posttreatment on Major Outcome Measures

Evaluator ratings

Posttreatment measure Quantity Quality

Clinician ratings Panic frequency Panic intensity Anticipatory anxiety Impairment

Self-report SPRAS Mi-Accompanied Mi-Alone Disability BDI

0.12 -0.15 -0.13 -0.15

-0.27 -0.11 -0.26* -0.28 -0.02

-0.09″ -0.14a

-0.29′ -0.48ab*

-0.37C* -0.16d

-0.29d* -0.52ab* -0.01a

Note. All analyses controlled for pretreatment severity, and analyses of quality ratings additionally controlled for age and employment status. Panic frequency = logarithmically transformed frequency of panic attacks; SPRAS = Sheehan Patient-Rated Anxiety Scale; Mi-Accompanied = When Accompanied subscale of the Mobility Inventory for Agoraphobia; MI Alone = When Alone subscale of the Mobility Inventory for Agora- phobia; Disability = Work and Social Disability subscale of the Sheehan Disability Scale; BDI = Beck Depression Inventory. a n = 25. b Tests of correlated correlations indicated significant differ- ences between quantity and quality ratings. c n = 22. d n = 24. * p < .05.

Effects of Individual Session Quality of Compliance on Outcome

A more microanalytic evaluation of the effects of compliance quality2 for particular sessions on outcomes was conducted. First, the relationship between compliance during each session and over- all change on outcome variables was evaluated. We used regres- sion analyses, controlling for pretreatment severity, age, and edu- cation, to- separately assess the relationship between level of compliance quality for each session (starting at the 2nd week of treatment) and residualized change in symptoms. Compliance dur- ing six sessions (Sessions 2, 5, 6, 8, 9, and 10) was significantly associated with some outcome variable.

Interestingly, many of our significant findings were consistent with the idea that compliance with particular skills is associated with changes in a related symptom variable. For example, com- pliance during Week 8, which largely focused on completing in vivo exposure to phobic situations, was a significant predictor of change in phobic avoidance (Mi-Alone: /3 = -0.51, p < .05; Mi-Accompanied: ft = —0.37, p < .05). Level of compliance for this week was also significantly associated with reduction in depression (BDI: /3 = -0.42, p < .05) and disability (SDS: |3 = -0.52, p < .05). Compliance during Week 9, also focusing on in vivo exposure, was also predictive of decreased phobic avoidance (Mi-Accompanied: ]8 = —0.50, p < .05) and decreased disability (SDS: /3 = —0.53, p < .05). Similarly, compliance during Week 10, a session that focused on interoceptive exposure, was significantly associated with reductions in fear of bodily sensations (/3 = -0.43, p < .05).

Compliance for Session 6, focusing on cognitive restructuring, was the only session associated with significant decreases in panic frequency (/3 = —0.48, p < .05) and decreases in anticipatory anxiety (/3 = — 0.44, p < .05). Session 5, also focused on cognitive restructuring, was associated with reductions in panic intensity (/3 = -0.54, p < .05) and reductions in anxiety (SPRAS: |3 = —0.41, p < .05). Two other sessions were also associated with reductions in anxiety, including Session 2, which focused on educational materials (SPRAS: j3 = -0.37, p < .05), and Ses- sion 8 (SPRAS: 0 = -0.60, p < .05).

We conducted additional analyses to determine the immediate impact of compliance on symptoms (anxiety, depression, worry, and fear of bodily sensations). Using the weekly monitoring forms, we calculated partial correlations between compliance during the index week with symptoms for that same week, controlling for prior symptom level (i.e., index week — 1), age, and employment. In addition, we calculated impact of compliance for each session on changes in symptoms for the 2-week period following the index week.3 These regression analyses also controlled for compliance at the later weeks. For example, analyses predicting change in symp- toms (from the index week) at 1-week postindex (week +1) also included week + 1 compliance as a covariate. Predicting change in

1 The quality rating was a better predictor even after controlling for attenuation owing to somewhat lower reliability of the quantity rating (see Nunnally & Bernstein, 1994, p. 257).

2 To control for Type I error to some degree, we used only the quality measure in these analyses because it proved to be the best predictor.

3 We extended these analyses out only 2 weeks from the index week because of missing data with the use of higher numbers of covariates.

COMPLIANCE AND PANIC 17

symptoms (from the index week) at 2 weeks postindex (week + 2) included week + 1 and week + 2 compliance ratings as covariates to isolate the effects of the initial week’s compliance ratings.

These analyses generally suggested that compliance during any given week has relatively little impact on immediate symptom change. Exceptions included findings indicating that Session 9 significantly affected symptoms during that week, leading to re- ductions in worry (j3 = —0.35, p < .05) and depression (/3 = —0.46, p < .05). Compliance during Session 10 seemed to sub- stantially reduce fear of bodily sensations 1 week later (j3 = -0.70, p < .05) and 2 weeks later (/3 = -0.47, p < .05). In general, the pattern of findings is consistent with the idea that compliance exerts incremental effects on symptomatology that is best evidenced over the entire course of treatment.

Moderator Analysis

In an attempt to further elucidate the relationship between compliance and outcome, exploratory analyses were conducted to evaluate whether demographic variables (e.g., age, gender, ethnic- ity, marital status, and employment status) and medication status (i.e., medicated vs. not medicated) moderated the relationship between therapist-rated compliance and outcome. Moderator anal- yses were conducted as described by Baron and Kenny (1986). Regression analysis was used to evaluate main effects and the interaction between compliance and each potential moderator (e.g., age). There was no evidence for moderator effects as indicated by no significant interactions between any of the hypothesized mod- erators and the compliance ratings (ps > .05).

Discussion

Compliance with treatment-related prescriptions is generally believed to be a mediator of treatment outcome. Although formal mediation analyses could not be conducted in this study, findings from this study are consistent with the idea that homework com- pliance may be an important factor in determining response to treatment. Moreover, there are data suggesting that the quality of compliance that is associated with particular exercises, such as in vivo and interoceptive exposure, is significantly associated with related outcome variables (i.e., phobic avoidance and fear of bodily sensations, respectively).

The most important finding generated by the present report is the suggestion that quality of work conducted by patients (relative to the absolute amount of work) is a better predictor of outcome. This finding has clear clinical implications as it suggests that therapists may want to pay particular attention to the details that determine good quality work and that therapists spend sufficient time training patients in being able to complete the highest quality assignments possible. In this study, patients appeared to gain understanding of the elements necessary for higher quality work as the ratings of homework quality increased across sessions. Ther- apists may want to worry less about motivating patients to do as much work as possible and instead may want to devote attention to ensuring that the work that gets done is very good in quality.

What mechanism underlies the relationship between homework quality and outcome? The emotional processing and related en- gagement literature provides an explanatory framework for under- standing the effects of quality work and end-state functioning.

Emotional processing is believed to occur when new information is incorporated into existing memory structures (Foa & Kozak, 1986; Rachman, 1980). Recovery is facilitated by treatments that encourage engagement (i.e., the activation of relevant pathological structures) in the context of providing information at odds with existing beliefs. In the present study, indices of quality homework are entirely consistent with the parameters deemed to be critical for emotional processing, including high levels of engagement and significant extinction of fear responses during repeated exposure to the feared situation or internal bodily cue. Conversely, patients who exhibited lower quality homework were more likely to show low levels of engagement and poor habituation. Conceivably, the assessment of homework quality may be one method of indexing the level of emotional processing taking place during therapy- related homework.

As far as we can determine, this is the first study to assess the relationship between compliance quality and quantity. However, this positive finding must be viewed cautiously. It is important to qualify the relative importance of the quality versus quantity ratings by reiterating the high level of correlation between these ratings. It is not surprising that these ratings are significantly associated, as both ratings are likely to reflect common underlying factors, such as patient motivation (Primakoff et al., 1986). Patient motivation and other similar factors are likely to contribute to the level of energy devoted to the treatment, which in turn would be reflected in amount and quality of homework. The extremely high level of correlation among these measures suggests that our indices are assessing virtually the same thing, despite the fact that the quality ratings proved to be better predictors in some cases. Outlier analysis suggests little evidence for outliers accounting for the differential predictive power of these indices. Examination of the level of covariation among quality and quantity ratings at each session suggests a somewhat lower range of correlations (r = .78-.90), indicating relatively more divergence among these mea- sures at this level of analysis. It is also notable that the regression analyses included demographic covariates in the case of the quality ratings. Exclusion of these covariates leads to some convergence between quality and quantity ratings, suggesting that use of co- variates may also help to account for the differential predictions.

There are a number of study limitations. First, further work is needed to assess the generalizability of the present findings. These findings may be specific to panic disorder patients treated in a group format. Additional work with other disorders using other treatment modalities is required to document the relationship be- tween the quality of compliance and outcome. Another issue that affects generalizability is the fact that only one therapist delivered the treatment. Use of one therapist is both a potential strength and a potential limitation. A strength arising from the use of one therapist is the somewhat greater assurance that the treatment was delivered consistently and that the compliance ratings were made consistently. The use of only one therapist should not affect generalizability unless the treatment was delivered in such a unique fashion that others would not be capable of replicating it. Fortunately, this was unlikely to be the case in this study because we used a highly structured treatment protocol that is very similar to CBT protocols used at other research centers specializing in the treatment of panic disorder (Telch et al., 1993).

The subjective nature of the compliance measures used in this study is another limitation. Patient reports, as well as the thera-

18 SCHMIDT AND WOOLAWAY-BICKEL

pist’s assessments of these reports, introduce a variety of possible biases. Although our use of structured homework forms and well- defined criteria for judging the work is likely to limit some confounding influences, future studies might consider the use of videotaped practice, direct therapist observation of homework exercises, and brief momentary assessments with handheld com- puters to provide more objective assessments of compliance.

Consistent with previous literature, patients’ self-reported rat- ings of compliance, relative to clinician ratings, appear to be less predictive of outcome. It is unclear, however, whether this lower predictive power is the result of some type of reporting bias, the result of a memory bias, or the result of assessing different things (i.e., time spent practicing vs. percentage of practice completed). It is worth noting that patients were not given the opportunity to offer estimations of their perceptions regarding quality of homework. In future studies on differences between therapist and patient ratings, both parties should make identical ratings.

References

American Psychiatric Association. (1994). Diagnostic and statistical man- ual of mental disorders (4th ed.). Washington, DC: Author.

Ballenger, J. C, Burrows, G. D., DuPont, R. L., Jr., Lesser, I. M, Noyes, R., Jr., Pecknold, J. C., Rifkin, A., & Swinson, R. P. (1988). Alprazolam in panic disorder and agoraphobia: Results from a multicenter trial. I: Efficacy in short-term treatment. Archives of General Psychiatry, 45, 413-422.

Barlow, D. H., O’Brien, G. T., & Last, C. G. (1984). Couples treatment of agoraphobia. Behavior Therapy, 15, 41-58.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychol- ogy, 51, 1173-1182.

Beck, A. T., & Steer, R. A. (1993). Beck Depression Inventory. San Antonio, TX: Psychological Corporation.

Chambless, D. L., Caputo, G. C., Jasin, S. E., Gracely, E. J., & Williams, C. (1985). The Mobility Inventory for Agoraphobia. Behaviour Re- search and Therapy, 23, 35—44.

Edelman, R. E., & Chambless, D. L. (1993). Compliance during sessions and homework in exposure-based treatment of agoraphobia. Behaviour Research and Therapy, 31, 767-773.

First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. (1994). Structured Clinical Interview for DSM-IV—Patient Edition (SCID-I/P, Version 2.0). New York: Biometrics Research Department.

Foa, E. B., & Kozak, M. J. (1986). Emotional processing of fear: Exposure to corrective information. Psychological Bulletin, 99, 20-35.

Mavissakalian, M., & Michelson, L. (1983). Self-directed in vivo exposure practice in behavioral and pharmacological treatments of agoraphobia. Behavior Therapy, 14, 506-519.

Meng, X., Rosenthal, R., & Rubin, D. B. (1992). Comparing correlated correlation coefficients. Psychological Bulletin, 111, 172-175.

Michelson, L., Mavissakalian, M., Marchione, K., Dancu, C., & Green- wald, M. (1986). The role of self-directed in vivo exposure in cognitive, behavioral, and psychophysiological treatments of agoraphobia. Behav- ior Therapy, 17, 91-108.

Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory. New York: McGraw-Hill.

Primakoff, L., Epstein, N., & Covi, L. (1986). Homework compliance: An uncontrolled variable in cognitive therapy outcome research. Behavior Therapy, 17, 433-446.

Rachman, S. (1980). Emotional processing. Behaviour Research and Ther- apy, 18, 51-60.

Schmidt, N. B. (1994). Safety maneuver elimination for panic disorder. Unpublished treatment manual, Uniformed Services University of the Health Sciences, Bethesda, MD.

Schmidt, N. B., Staab, J. P., Trakowski, J. H., & Sammons, M. (1997). Efficacy of a brief psychosocial treatment for panic disorder in an active duty sample: Implications for military readiness. Military Medicine, 162, 123-129.

Schmidt, N. B., Trakowski, J. H., & Staab, J. P. (1997). Extinction of panicogenic effects of a 35% CO2 challenge in patients with panic disorder. Journal of Abnormal Psychology, 106, 630-638.

Shear, M. K., Brown, T. A., Barlow, D. H., Money, R., Sholomskas, D. E., Woods, S. W., Gorman, J. M., & Papp, L. A. (1997). Multicenter Collaborative Panic Disorder Severity Scale. American Journal of Psy- chiatry, 154, 1571-1575.

Sheehan, D. V. (1983). The anxiety disease. New York: Scribner. Telch, M. J., Lucas, J. A., Schmidt, N. B., Hanna, H. H., Jaimez, T. L., &

Lucas, R. (1993). Group cognitive-behavioral treatment of panic disor- der. Behaviour Research and Therapy, 31, 279-287.

Wolfe, B. E., & Maser, J. D. (1994). Treatment of panic disorder: A consensus development conference. Washington, DC: American Psychi- atric Press.

Young, J., Beck, A. T., & Budenz, D. (1983). Assessment of compliance in cognitive therapy. Unpublished manuscript, University of Pennsylvania, Philadelphia.

Received December 7, 1998

Revision received May 4, 1999

Accepted May 10, 1999