Panic disorder is fairly common, with a 12-month prevalence rate of 2.7% and a lifetime prevalence rate of 4.7% (1,2). The course of panic disorder tends to be chronic, with high rates of recurrence after remission, particularly for panic disorder with agoraphobia (3,4,5). Furthermore, individuals with panic disorder experience considerable impairment and disability, including occupational difficulties (6,7,8,9), impaired well-being (10,11,12), and reduced quality of life (9,10,11,12,13,14). They also have higher rates of health care use, with a greater number of outpatient visits, emergency room visits, and hospitalizations compared with those without the disorder (8,10,15).

Individuals with panic disorder typically present to the primary care setting, with estimates suggesting that as many as 80% of cases first present to primary care (16). Thus the rate of the disorder is higher in primary care settings, with a reported median prevalence of 4% to 6% (8). Furthermore, the majority of individuals with panic disorder obtain their mental health treatment in the primary care setting (17,18). Despite these findings, research suggests that panic disorder often goes unrecognized (19,20) and is inadequately treated in both primary care (8,21,22,23) and psychiatric settings (24,25,26).

A number of effective pharmacologic treatments for panic disorder exist, including tricyclic antidepressants (TCAs), selective serotonin reuptake inhibitors (SSRIs), serotonin and norepinephrine reuptake inhibitors (SNRIs), and benzodiazepines (27,28,29,30). Likewise, psychosocial treatments, namely cognitive-behavioral therapy (30,31) and possibly a specific form of psychoanalytic treatment (32), have been found to be effective. Despite this, estimates suggest that over 40% of individuals with panic disorder go untreated (33). Certain demographic characteristics (for example, gender, education, and race) and clinical variables (for example, comorbid diagnoses) appear to be related to mental health service use in general (34,35,36). Additionally, there may be other factors that have an impact on service use, such as not perceiving oneself in need of treatment (37). For individuals with panic disorder who do receive treatment, little is known about the treatment typically received, and no studies have examined whether there are differences in treatment between persons with panic disorder with agoraphobia and those with panic disorder without agoraphobia.

This study was part of and used data from the Primary Care Anxiety Project (PCAP), which is a naturalistic, longitudinal study of anxiety disorders among primary care patients. The purpose of the study presented here was to explore characteristics of psychiatric treatment received by those with panic disorder with agoraphobia versus those with panic disorder without agoraphobia, including the rates of receiving different psychotropic medications and specific psychotherapeutic techniques. This study also sought to examine demographic and clinical predictors of receiving psychotropic medications, psychotherapy, and concurrent medication and therapy. Finally, this study provides descriptive information concerning reasons for not receiving treatment.

Participants

Participants were recruited from 15 primary care, internal medicine, and family medicine clinics in New England, which included ten sites in urban or suburban areas and five sites in rural areas; four of the sites were in small, private practices, four in university-affiliated clinics, and seven in large teaching hospitals. Participant recruitment occurred between July 1997 and May 2001.

The entire PCAP sample consisted of 539 participants with one or more of the following index anxiety disorders: generalized anxiety disorder, panic disorder with or without agoraphobia, agoraphobia without a history of panic disorder, posttraumatic stress disorder, social phobia, mixed anxiety-depressive disorder, or generalized anxiety disorder features occurring within a mood disorder. Inclusion criteria required participants to be English speaking, at least 18 years of age, and scheduled for a medical appointment on a day of study recruitment. Participants were excluded if they had active psychosis, were currently pregnant, or had no current address or telephone number. This study reported data from 235 individuals with panic disorder at intake: 150 had agoraphobia and 85 did not have agoraphobia.

Procedure

Research assistants approached all patients in primary care waiting rooms during times of recruitment and asked whether they were interested in participating in a study on stress or nervousness. The number of days per week, particular days of the week, and times of the day for study recruitment varied based on the availability of research assistants. A complete description of the study was provided to potential participants, and written informed consent was obtained. PCAP was approved by the institutional review board of Brown University and all other sites. Of the 14,320 patients who were approached, 31% (N=4,383) were interested in participating and completed the anxiety screening questionnaire. Of those who completed the questionnaire, 63% (N=2,755) screened positive for anxiety symptoms. These individuals were then invited to participate in an intake assessment, which included the Structured Clinical Interview for DSM-IV (SCID-IV) (38). Of those invited to complete the intake assessment, 456 declined participation and 665 repeatedly cancelled or did not show up to the appointment. Thus the SCID-IV was administered to 1,634 individuals, and those who had one or more index disorders (N=539) were invited to participate in PCAP (39,40). After the intake assessment, participants were contacted to complete follow-up assessments at six months, 12 months, and annually thereafter.

Measures

Anxiety screening questionnaire. The anxiety screening questionnaire, which was developed specifically for PCAP, is a self-report instrument containing 32 items assessing essential features of anxiety disorders. It was designed to be highly sensitive to the presence of any anxiety disorder symptoms in order to reduce the number of false negatives. In a separate validation study of this measure, it was found to have a sensitivity of 1.0 and a specificity of .67, with no individuals who screened negative on the questionnaire being found to have an anxiety disorder on the SCID-IV (39).

Diagnostic clinical interview. All psychiatric diagnoses were established by means of an in-person diagnostic interview using the SCID-IV (38). The SCID-IV is a well-validated assessment instrument that has been found to have excellent interrater reliability and diagnostic accuracy (41). For PCAP, participants first completed the psychotic screener followed by the anxiety disorders module. Participants who received an anxiety diagnosis then completed the mood, alcohol and substance use, and eating disorders modules. As part of the interview, Global Assessment of Functioning (GAF) scores were assigned. All interviewers had at least a bachelor's degree in psychology and underwent a rigorous, multistage training program employed in other large-scale studies conducted through the Brown University Department of Psychiatry and Human Behavior.

Mental health treatment. Information regarding psychotropic medication was obtained with the psychotropic/auxiliary drug treatment schedule, which is an interviewer-administered form about current medications. It is part of the Longitudinal Interval Follow-up Evaluation (LIFE) (42), which has been found to have excellent interrater reliability and long-term test-retest reliability (43). Participants were asked about psychotherapy currently being received by using the types of mental health treatment form, which is an interviewer-administered measure developed for PCAP that assesses involvement in various types of psychosocial treatments. Information regarding specific therapeutic techniques used was obtained with the Psychosocial Treatment Interview-Revised (PTI-R) (44). The PTI-R, an interviewer-administered measure, assesses whether 39 techniques have been used in a patient's therapy. It includes a number of therapy techniques, including supportive, dynamic (for example, discuss childhood experiences and the patient-therapist relationship), family systems (for example, examine problems from a systemic approach and change family interactions), relaxation, exposure (for example, imaginal, interoceptive, and in vivo), and cognitive (for example, identify dysfunctional thinking and substitute more rational thoughts) therapy techniques. The PTI-R has been found to have good interrater reliability and validity (44).

Treatment not received. For those not receiving treatment at intake, interviewers administered the treatment not received form, which was designed specifically for PCAP. This measure was added to the intake assessment partway through the recruitment phase; therefore, only a subset of participants was administered this measure. Furthermore, this measure was administered only to patients, not to their providers. For patients who were not receiving psychotropic medication, it inquires about reasons for not seeking medication, and for those not receiving psychotherapy, it inquires about reasons for not seeking therapy. Reasons were coded by the interviewer into the following categories: doctor did not recommend treatment, financial reasons, lack of insurance coverage for treatment, treatment was ineffective in the past, patient does not believe he or she has a problem, patient does not believe in using this treatment for emotional problems, too busy to seek treatment or treatment is inconvenient, worried about treatment record, concerns about stigma or embarrassment, not knowing how to obtain treatment, and concerns about medication side effects.

Statistical analyses

Analyses were conducted using SAS, version 9.1.3 (45). Descriptive statistics were used to examine characteristics of psychiatric treatment received by those with panic disorder with and without agoraphobia, in addition to reasons for not receiving treatment. Comparisons between the two groups were made with chi square statistics and t tests. Logistic regression analyses were used to examine predictors of receiving psychotropic medication, psychosocial treatment, and concurrent medication and therapy.

Sample characteristics

At intake 150 participants were diagnosed as having panic disorder with agoraphobia and 85 were diagnosed as having panic disorder without agoraphobia. Table 1 displays the demographic and clinical characteristics of the sample at intake. No significant differences were found between the groups on any demographic variables. However, in terms of clinical variables, the group with panic disorder with agoraphobia had an earlier age of panic disorder onset (t=1.98, df=233, p<.05) and lower GAF scores (t=2.96, df=233, p<.01), which is indicative of worse functioning or more severe symptoms. The group with panic disorder with agoraphobia also had higher rates of social phobia (χ²=5.19, df=1, p<.05) and a greater number of comorbid anxiety disorder diagnoses (t=-2.15, df=233, p<.05).

Treatment received

Over half of those with panic disorder (N=145, 62%) were receiving psychiatric treatment at intake, with those with agoraphobia being more likely to be in treatment than those without agoraphobia (χ²=21.1, df=1, p<.001). Table 2 displays the frequencies and characteristics of treatment received at intake. Significantly more patients with panic disorder with agoraphobia were receiving psychotropic medication at intake (χ²=20.05, df=1, p<.001). In particular, more patients with panic disorder with agoraphobia were receiving benzodiazepines (χ²=20.95, df=1, p<.001). SSRIs or SNRIs were the most commonly received psychotropic medication for both groups. More specifically, the most common SSRIs or SNRIs were sertraline (N=27, 11%; mean dose=84.37±58.78 mg), paroxetine (N=21, 9%; mean=23.95±19.82 mg), fluoxetine (N=19, 8%; mean=31.05±16.63 mg), and venlafaxine (N=4, 2%; mean=243.75±166.30 mg), all of which have approval by the U.S. Food and Drug Administration for treatment of panic disorder (46). A subset of patients receiving medication (N=87, 66%) were asked about medication adherence, with 15% (N=13) reporting that they had taken their medication less frequently or at a lower dosage than prescribed.

In terms of psychotherapy, patients with panic disorder with agoraphobia were more likely that those with panic disorder without agoraphobia to be receiving psychotherapy (χ²=5.48, df=1, p=.02). More specifically, patients with panic disorder with agoraphobia were more likely to be receiving therapy using supportive (χ²=5.19, df=1, p=.02), psychodynamic (χ²=3.86, df=1, p<.05), exposure (χ²=7.9, df=1, p<.01), and cognitive (χ²=4.79, df=1, p<.05) therapy techniques. For both groups, therapy using supportive techniques was the most commonly received, whereas exposure therapy was one of the least frequently received interventions. In terms of concurrent medication and psychotherapy, those with panic disorder with agoraphobia were more likely to be receiving such treatment (χ²=5.70, df=1, p=.02).

Predictors of treatment

A series of logistic regression analyses was conducted to examine demographic and clinical predictors of receiving at intake psychotropic medication, psychotherapy, and concurrent medication and therapy. Demographic variables, including age, gender, race, marital status, education, and insurance, along with clinical variables (presence of agoraphobia, age of panic disorder onset, number of comorbid anxiety diagnoses, and comorbid major depressive disorder), were included as predictors. Table 3 displays the results of these analyses, including the tests of the full models, as well as the regression coefficients, Wald statistics, odds ratios, and 95% confidence intervals for significant predictors. In terms of pharmacologic treatment, results showed that non-Hispanic Caucasians were more likely than those from other racial or ethnic backgrounds to be receiving such treatment. Furthermore, those with agoraphobia were over three times as likely as those without to receive medication and those with comorbid major depressive disorder were twice as likely as those without to receive medication. Out of the three insurance categories, the only significant predictor of receiving psychotherapy was having Medicare, Medicaid, or public assistance insurance. In terms of concurrent medication and therapy, those who had at least a four-year college degree were five times as likely as those without such a degree to be receiving such treatment.

Reasons for not seeking psychiatric treatment

Data from the treatment not received form were available for 53 patients who were not receiving medication at intake and 85 who were not receiving psychotherapy at intake. No significant differences were found between persons with panic disorder with agoraphobia and those with panic disorder without agoraphobia on reasons for not seeking treatment; thus responses were collapsed across groups and are displayed in Table 4. The most commonly reported reasons for not seeking pharmacotherapy were not believing in using medication for emotional problems, lack of recommendation by the primary care provider, and concern about side effects. For psychotherapy, the most common reasons were not believing that one has a problem, not believing in using psychotherapy for emotional problems, and perceiving treatment to be inconvenient or being too busy for treatment.

Despite the degree of impairment and disability typically associated with panic disorder, the study presented here found that 38% of primary care patients with the disorder were not receiving psychiatric treatment. These results are fairly consistent with previous studies that have found 27% to 37% of individuals with panic disorder do not receive treatment (8,18,33,47). The nontreatment rate in this study was much higher for patients with panic disorder without agoraphobia. Furthermore, this study revealed a number of notable differences between treatments received by those with versus without agoraphobia, which has not been closely examined in prior studies.

Over half (56%) of primary care patients with panic disorder were receiving psychotropic medication, with those with agoraphobia being much more likely (67%) than those without agoraphobia (36%) to be receiving medication. These rates are comparable to those reported in other studies in primary care settings (42% to 64%) (8,22,48). The most commonly received psychotropic medications for both groups were SSRIs or SNRIs, with 34% of the sample using a SSRI or SNRI. The mean SSRI or SNRI doses reported were at or above the suggested therapeutic doses (46), except for paroxetine, which was below the suggested therapeutic dose. It is unclear whether those taking paroxetine were in fact receiving an inadequate dose or whether these individuals had not yet titrated up to the suggested therapeutic dose. Significantly more patients with agoraphobia (35%) were receiving benzodiazepines than those without agoraphobia (8%), with 26% of the whole sample using such a medication. This rate is higher than past findings for primary care patients with panic disorder (7% to 17%) (8,22). Given the variability in how benzodiazepines are used, the mean dose in our sample was not determined; therefore, it is unclear whether such medications were being used at the suggested dose or in the appropriate manner determined to be efficacious. Other medications, such as TCAs and other antianxiety medications (for example, buspirone and meprobamate), were used infrequently.

Results showed that those with agoraphobia (38%) were more likely to be receiving psychotherapy than those without agoraphobia (24%). However, in both groups psychotherapy was underutilized, with only 33% of the sample as a whole receiving therapy at the time of intake. These rates are in line with previous studies, which have found that only 27% to 44% of primary care patients with the disorder receive therapy (8,17,48). These findings suggest that medications are the first line of treatment. The most commonly received psychotherapy for both groups in our study used supportive and psychodynamic techniques. Despite empirical evidence supporting the efficacy of cognitive-behavioral therapy for panic disorder (30,31), cognitive-behavioral techniques were infrequently utilized, with only 15% of those without agoraphobia and 27% of those with agoraphobia receiving therapy that used cognitive techniques and even fewer receiving exposure therapy techniques. These findings are similar to those in past research showing that cognitive-behavioral therapy is underutilized (24,48), suggesting a need for better treatment dissemination. Furthermore, making such interventions more accessible to primary care patients and adapted to their particular needs may be beneficial.

A number of factors predicted likelihood of receiving psychiatric treatment. Patients with agoraphobia were over three times as likely as those without agoraphobia to receive medication, which is not surprising given that agoraphobia is associated with greater disturbance (49) and may be more likely to be detected and treated. Comorbid major depressive disorder increased the odds of receiving medication, which has also been found in previous studies of panic disorder and anxiety disorders in general (23,48). It is possible that primary care providers are more familiar with depressive symptoms and their treatment, making it more likely that patients with comorbid depression receive medication. The results also showed that persons from ethnic minority groups were less likely to be receiving medication, which is consistent with other studies (23,48). It is unclear whether these findings are due to differences in patient treatment preferences or biases on the part of the treatment provider. Surprisingly, having Medicare, Medicaid, or public assistance insurance was the only significant individual predictor of receiving psychotherapy. It is possible that these patients have worse functioning than those with private insurance, while having greater means to pay for services than those with no insurance, which could lead to increased likelihood of using therapy. It is also possible that these patients were enrolled in public mental health programs and receiving therapy from a social worker or case worker in such programs. Education level predicted likelihood of receiving concurrent medication and therapy, with those who had a four-year college degree or more being over five times as likely as those who did not to be receiving such treatment. Past studies have also found a relationship between education level and greater mental health service use (33,50,51). One explanation for this finding is that education is often related to socioeconomic status, and therefore, those with higher education levels may be more likely to have the resources to pay for services. It is also possible that those with higher education levels may have better access to the available services or different views about treatment.

Primary care patients with panic disorder reported several barriers to seeking treatment, including not believing in using medication or therapy for emotional problems and not receiving a recommendation by their primary care provider. Some patients were also concerned about medication side effects or felt psychotherapy was inconvenient. Although these findings highlight the importance of providers' making treatment recommendations, it appears that the majority of treatment barriers lie within the patient. Thus interventions aimed at addressing these barriers, such as providing patient education about the disorder and effective treatment options, may be beneficial.

A number of limitations should be considered when interpreting the findings. The sample was primarily female and Caucasian, and data collection occurred in only one geographic region of the United States. Future studies should include sites in other regions and a more diverse sample. Another consideration is that our assessment of psychiatric treatment did not specify the symptoms for which the patient was receiving treatment. It is possible that the focus of treatment may have been on comorbid conditions. Indeed the sample had high rates of comorbidity; however, these rates are comparable to those reported in other research (18). No retrospective information was obtained from participants on the length of time on a particular medication before intake. Furthermore, we obtained only a limited amount of medication adherence data from self-reports. Thus it is possible that although a proportion of the sample was receiving treatment known to be effective, they may not have received the treatment for an adequate duration of time or they may have had low treatment adherence. Although reasons for not receiving treatment were examined, data were obtained only from patients and only from a subset of the sample, because this measure was added partway through the study. Therefore, the study did not examine providers' reasons for not using particular treatments. So it is unclear whether these findings would have differed had data been available from the providers and from the whole sample. Future research should attempt to elucidate these issues.

The findings suggest that there are a number of differences in the treatment received by primary care patients with panic disorder with versus without agoraphobia. Across the board, many patients with panic disorder were not receiving psychiatric treatment. When treatment was received, psychotropic medication—particularly SSRIs or SNRIs for those with panic disorder either with or without agoraphobia and benzodiazepines for those with agoraphobia—appeared to be the treatment of choice. Psychotherapy was underused for both groups, and for those who did receive therapy, empirically supported interventions, such as cognitive-behavioral therapy, were rarely used. A number of treatment barriers were identified that could potentially be addressed by providers. Overall, the findings suggest a need for creating processes for better treatment dissemination, making interventions more accessible or adapting them to the needs of this population, and addressing treatment barriers.

The Primary Care Anxiety Project is supported by an unrestricted grant from Pfizer Pharmaceuticals, Inc. Dr. Weisberg's time and effort was supported by a Patient-Oriented Career Development Award (K23 MH069595) from the National Institute of Mental Health.

Dr. Keller is a consultant for or has received honoraria from CENEREX, Cephalon, Cypress Bioscience, Cyberonics, Forest Laboratories, Janssen, JDS Pharmaceuticals, Organon Pharmaceuticals, Novartis, Pfizer, and Wyeth. He has also received grant or research support from Pfizer and is a member of the advisory boards for Abbott Laboratories, Bristol-Myers Squibb, CENEREX, Cyberonics, Cypress Bioscience, Forest Laboratories, Janssen Pharmaceuticals, Novartis, Organon Pharmaceuticals, and Pfizer. Dr. Weisberg has received honoraria from AstraZeneca and has received grant or research support from Pfizer. Dr. Marcks reports no competing interests.