Psychosomatics 40:422-427, October 1999
© 1999 The Academy of Psychosomatic Medine
Effects of Psychological Intervention on Panic Attack Patients in the Emergency Department
John M. Dyckman, Ph.D.,
Robert L. Rosenbaum, Ph.D.,
Rosarie J. Hartmeyer, Ph.D., and
Lawrence J. Walter, M.A.
Received October 9, 1998; revised December 2, 1998; accepted January 19, 1999. From the Departments of Psychiatry, Kaiser Permanente Medical Centers, Vallejo, Oakland, and Martinez, California; and the Division of Research, Kaiser Permanente Medical Care Program, Oakland, California. Address correspondence and reprint requests to Dr. Dyckman, Psychiatry Clinic, Kaiser Permanente Medical Center, 1761 Broadway, Suite 100, Vallejo, CA 94589.
|
ABSTRACT
|
For patients initially seen in the emergency department (ED) for panic attack, this study evaluated the effect of two brief psychological interventions in the ED on later utilization of emergency, psychiatric, and nonpsychiatric medical department services. Each of the two intervention groups received usual ED care, a brochure on panic disorder, and a referral to treatment at the psychiatry department; one of the two groups also received 20–30 minutes of contact with a representative from the psychiatry department. Both intervention groups were compared with a historical control group. The contact condition reduced ED use after the initial visit to the ED, although all three groups had more visits to the psychiatry department and to all nonpsychiatric departments. This decrease was statistically significant (P=0.0017) when compared with the brochure condition but not when compared with the historical control group (P=0.0672). The decrease seen in ED use is an important therapeutic and economic finding.
Key Words: Panic Attacks • Psychological Intervention • Emergency Department
|
INTRODUCTION
|
Panic disorder is a potentially debilitating condition with an estimated 12-month prevalence of 2.3% and a lifetime prevalence of 3.5%.1 Panic symptoms mimic many other diseases, especially cardiac illness, so the emergency department (ED) is a frequent point of entry into the health care system for these patients. In fact, panic patients are much more likely to be seen initially in a medical setting than in a mental health clinic.2 Many panic patients visit the ED repeatedly before seeking psychiatric treatment, and this situation is frustrating and costly for everyone involved. The prevalence of panic symptoms in patients who arrive at the ED with somatic complaints but without diagnosable organic disease ranges from 6.7%3 to as high as 18%.4,5 Although panic attack is not coded in the DSM-IV as a separate disorder, panic attack is prerequisite to a diagnosis of panic disorder.6
Increasingly over the past decade, panic symptoms have been diagnosed by primary medical providers,7 and at least one study has shown that early intervention with exposure instructions in the ED can reduce depression, agoraphobic avoidance, and panic frequency.8 For these reasons and because we observed that some patients use the ED as their exclusive locus of treatment, we designed this study to examine whether a brief, inexpensive psychiatric intervention done in the ED by paraprofessionals would increase compliance with referral to the psychiatry department and reduce later medical utilization for symptoms of panic disorder.
|
METHODS
|
Study Setting
The study was conducted simultaneously at two outpatient hospitals, about 45 miles apart, of a large health maintenance organization (HMO) from 5/1/95 through 4/30/96. Both facilities had psychiatric clinics with ongoing cognitive-behavioral treatment for patients affected with panic disorder.
Study Population
The study included patients who had not previously been seen in psychiatry who arrived and were seen in the ED between 2:00 A.M. and 6:00 P.M., had a diagnosis of definite or probable panic episode, and had been health plan members for  1 month (and thus had access to service) in both the 12-month periods before and after the index ED visit.
Diagnoses, demographics, and location of participants are summarized in Table 1.
Design
This study was a quasi-experimental design. To establish three groups, recruitment proceeded stepwise: the control group was selected retrospectively (archivally) from the medical records of 775 adult patients seen in the ED. The two intervention groups were chosen prospectively during the intervention study phase. This study looked at the effect of adding one of two simple interventions to usual care. The patients in all three groups received usual care from their ED physician that might include anti-anxiety medication at the discretion of the physician.
The patients in Group B (brochure) received from ED personnel the usual ED care and a brochure explaining panic disorder, its symptoms, and treatment available in the psychiatry department.
In addition to receiving the usual ED care as well as the brochure and psychiatric referral, the patients in Group C (contact) received immediate consultation: After the ED physician diagnosed panic attack, the ED clerk contacted the psychiatry department to request an immediate consultation with a specially trained educator from the psychiatry department who talked with the patient in the ED for about 20–30 minutes; instructed the patient in anxiety-management techniques (diaphragmatic breathing or attentional refocusing, depending on the patient's needs and response); and ended the interviews by giving the patient both the brochure and a referral to the psychiatry department.
Patients who arrived at the ED between 2 A.M. and 6 P.M. (inclusive) and were diagnosed by the treating ED physician as having a panic attack were assigned to intervention groups arbitrarily but not randomly: Patients seen on Tuesday, Thursday, or Saturday were assigned to the brochure group; those who were seen on Monday, Wednesday, or Friday were assigned to the contact group. This calendar-based assignment to intervention groups was designed to accommodate the availability of research assistants. The hours under examination were selected to reflect actual time of arrival at the ED, so that all available patients in the contact group could be captured.
The historical control group provides a baseline comparison for course and utilization under conditions of "usual care" (i.e., in the absence of systematic instruction or treatment). To select the historical control group, the ED logs at the two facilities were searched for patients who had been seen in the prior 12 months and whose computer records contained the key words, "stress," "dyspnea," "SOB (shortness of breath)," "hyperventilation," "noncardiac chest pain," "anxiety," or "panic." The complete medical records of these patients were then examined to see if the episode in question met the DSM-IV criteria for panic attack (PA) and whether the treating physician listed "anxiety attack" or "panic episode" as part of the discharge diagnosis. Rating was done either by graduate students or by a postdoctoral psychologist. By using the DSM-IV criteria, ED panic episodes were coded as either definite PA; probable PA (physician's diagnosis of "anxiety" or "panic," either without specific information about onset or with <4 of 13 symptoms documented); or not PA (psychiatric problem other than PA). Two of the authors (J.M.D. and R.J.H.) also served as criterion raters and read the charts independently until all raters reached an 80% agreement rate. Because selection of patients in the control group depended entirely on medical chart records of ED visits, the smaller number of documented symptoms for this group yielded a greater rate of "probable panic" diagnoses than for either the brochure group or the contact group.
We examined three dependent outcome measures: 1) ED visit rate, 2) psychiatry department visit rate, and 3) rate of visits to all nonpsychiatric medical departments (including ED). In keeping with epidemiologic practice, data for each study subject were obtained for the 12-month periods before and after the index ED visit. Data on doctor visits were obtained directly from computerized registration records and so captured services actually received in all departments and facilities of the HMO.
Statistical Analyses
A series of comparisons was made to control for possible confounding factors. Because chi-square tests showed no statistically significant differences between the facilities for group size, ratio of men to women, or number of panic symptoms, data from both facilities were combined for analyses. This approach was chosen also because chi-square analysis showed no statistically significant differences in sex ratio by group (overall ratio of women to men was about 2:1).
Because differences in mean age by group were statistically significant (F(2,351)=12.93, P<0.0001) and because age was also positively associated with rates of visits to the ED and to all nonpsychiatric medical departments (including ED), we included age as a covariate in comparing utilization rates among groups. Mean visit rates for each group were adjusted for age by directly standardizing to the age distribution of the pooled groups.
Initial (i.e., preintervention) visit rates for each group pair were compared by using Poisson regression, a method useful in modeling counting processes such as outpatient visits, which may violate the normality assumptions of linear regression.9,10 This modeling technique had the additional advantage of allowing for unequal follow-up time, as not all patients were members of the health plan continuously during the observation period. The dependent variable in the model was defined as the natural logarithm of the number of events for each person (i.e., visits to ED; to the psychiatry department; or to all medical departments, including ED, but excluding the psychiatry department) for a person divided by the number of months (expressed as a proportion of 1 year) for which the person was actively enrolled in the health plan during the 12-month observation period before the index ED visit. Independent variables consisted of intervention group and age. After adjusting for age, we reported the results as rate ratios comparing incidence density for each group (i.e., mean visit rate per 100 members per member-year) with each of the other groups.
To analyze group differences for change in utilization over time (and thus to explore the differences produced by the interventions), the Generalized Estimating Equation of Liang and Zeger10 was used to perform Poisson regression while taking into account the correlation between repeated measures. In this model, independent variables consisted of intervention group; time period (i.e., the 12-month periods before and after the index ED visit); a time-by-group interaction term; and age. The exponentiated coefficient on the time-by-group term can be interpreted as a rate ratio that estimates the change in visit rate over time for the first group compared with the second group. Rate ratio >1.0 thus indicates that visit rate increased for the first group, compared with the second group, from the period before the index ED visit to the period after the index ED visit.
|
RESULTS
|
Table 2 presents mean visit rate for each dependent variable by group for the 12-month periods before and after the index ED visit. For all groups, visits to other departments (psychiatry, nonpsychiatric medical) increased after the index ED visit, as did ED visits made by the brochure group (Table 2).
View this table:
[in this window]
[in a new window]
|
TABLE 2. Age-adjusted mean number of visits per patient per member-year for each group before and after index visit to emergency department (ED)
|
Preintervention Utilization Rates
Table 3 shows initial visit rates for each dependent variable for each group pair. Statistically, initial visit rates for all types of visits were significantly higher in the contact group than in the control group. The contact group also had more ED visits and marginally more visits to the psychiatry department than did the brochure group (P=0.0566). No statistically significant differences in visit rates were seen between the brochure and control groups.
View this table:
[in this window]
[in a new window]
|
TABLE 3. Age-adjusted rates of visit use compared between control and intervention groups for 12-month period before index emergency department (ED) visit
|
Changes in Utilization
Age-adjusted rate ratios for change in ED visit rate over time for the 3 pairs of compared groups were 0.675 for contact vs. control (P=0.0672), 0.397 for contact vs. brochure (P=0.0017), and 1.673 for brochure vs. contact (P=0.0417).
ED use by the contact group decreased by about 30% more than for the control group (i.e., 1.0 – 0.675= 0.325); however, this difference lacked statistical significance (P=0.0672). Mean ED use for the contact group declined even more—about 60% more than for the brochure group (P=0.0017), in which ED visit rates actually increased after the index ED visit. ED visit rates for the brochure group increased by 67% more than for the control group (i.e., 1.673-1.0=0.673).
Mean rate of visits to all nonpsychiatric medical departments (including ED) increased for all groups, but no statistically significant group differences were found for change in these visit rates over time.
Compliance With Psychiatric Treatment
Treating ED physicians referred all study subjects for follow-up treatment in the psychiatry department for their panic symptoms. The number of psychiatry department visits increased substantially and significantly for all three groups after the index ED visit (Table 2). About 65% of the referred patients in Groups B and C actually made and kept an appointment in the psychiatry department. The proportion of compliant patients did not vary significantly by intervention group. Rates of compliance for the brochure group (64.1%) and for the contact group (65.6%) were almost identical (chi-square=0.03, NS).
Among the patients in the brochure and contact groups who complied with referral to the psychiatry department, rate ratio (0.298, P=0.0001) was significant, showing that the reduction in ED use for the contact group, compared with the brochure group, was most pronounced for the patients who complied with referral: ED visit rates for the contact group showed 70% fewer ED visits than for the brochure group. In addition, the patients in the contact group who came to the psychiatry department were likely to stay in treatment longer than the patients in the brochure group: mean number of psychiatry department visits was 7.87 for the contact group, whereas mean number of clinic visits for the brochure group was 4.62 (t(75)=2.845, P<0.006).
|
DISCUSSION
|
This study used interventions that were designed to change the observed tendency of some patients to use the ED as their exclusive locus of treatment for panic attack. Our results show that ED visits prompted by panic episodes are associated with a statistically significant increase in later psychiatric and nonpsychiatric medical utilization, but that patients who received brief face-to-face contact in the ED with a representative from the psychiatry department had a statistically significant reduction of ED use. The finding that patients given this contact had fewer later ED visits is an important economic and therapeutic finding consistent with our original hypothesis, that is, that providing disorder-specific information and intervention when patients seek symptomatic relief maximizes the effectiveness of treatment for panic episodes and reduces reliance on the ED. Thus, these results indicate it can be helpful to provide immediate, brief psychological intervention in the ED to patients arriving there because of panic episodes. To determine the strength of the intervention, however, further research should deliberately match initial rate of ED use among groups or randomize assignment to groups.
Only slightly more than one third of patients in the historical control group sought available psychiatric treatment for their condition. Even when cardiac disease or other medical conditions have been ruled out, some panic patients are skeptical of their ED physician's determination that their symptoms are not due to occult organic disease and so continue to visit the ED. In contrast, almost two thirds of patients in the brochure and contact groups sought psychiatric treatment, suggesting that even brief education in the ED can be a powerful tool in helping patients locate and use appropriate treatment for panic episodes. Elsewhere, the principal author has described a model of panic disorder that can be easily and concisely presented to patients.11 Interestingly, contact did not increase compliance with referral any more than providing usual ED care and brochure.
The observed increase in ED use among patients in the brochure group is surprising and invites further study. One hypothesis, although speculative, is that having a brochure to give to patients somehow alters ED physicians' rapport with some patients, causing these patients to feel "unheard" about their symptoms and thus more likely to "try again" when experiencing anxiety. Perhaps a more positive explanation can be found in the concept of "role induction," discussed by Hoehn-Saric et al.12 in their use of educative films for patients wait-listed for psychotherapy. This concept suggests that education about psychotherapy can increase patients' expectation of change, their role in producing change, and the likelihood of positive outcome. Accordingly, contact in the ED with a representative from the psychiatry department may have fostered positive expectations about psychological treatment of physical panic symptoms and thus decreased reliance on the ED.
The observation that the contact group had higher initial rates of ED use raises the possibility that effects were due to regression to the mean: that is, the likelihood that any group scoring highly on a measure will, when measured again, show scores that are closer to the population mean. Because ours was a quasi-experimental study, patients were not selected or matched by their initial rate of ED use; therefore, the decrease in ED use cannot be unequivocally attributed to the intervention. Nonetheless, we are intrigued that as ED use decreased and psychiatry department use increased, no differential increase was seen in overall use of nonpsychiatric medical departments (including ED); this finding suggests that the intervention in the ED produced the specific targeted effect of facilitating treatment of panic symptoms in the appropriate venue.
|
ACKNOWLEDGMENTS
|
Preliminary results from this study were presented at the 13th Annual Meeting of the Society for the Exploration of Psychotherapy Intervention, Toronto, Ontario, Canada, April 24–26, 1996.
This research was funded by a grant from the Kaiser Permanente Innovation Program (Grant No. 940130). Institutional review and approval was granted January 19, 1995, by the Institutional Review Board of the Kaiser Foundation Research Institute.
The authors thank David Burnash, M.B.A., for design of the computer program used to screen medical records, and Joan Bisagno, Ph.D., Avvy Mar, M.A., and Suzanne Staccone, M.A., who served as raters and who provided emergency department interventions. The Medical Editing Department, Kaiser Foundation Research Institute, provided editorial assistance.
|
REFERENCES
|
-
Kessler RC, McGonagle KA, Zhao S, et al: Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States. Arch Gen Psychiatry 1994; 51:8–19[Abstract]
-
Katerndahl DA, Realini JP: Where do panic attack sufferers seek care? J Fam Pract 1995; 40:237–43[Medline]
-
Klein E, Linn S, Colin V, et al: Anxiety disorders among patients in a general emergency service in Israel. Psychiatr Serv 1995; 46:488–492[Abstract/Free Full Text]
-
Yingling KW, Wulsin LR, Arnold LM, et al: Estimated prevalences of panic disorder and depression among consecutive patients seen in an emergency department with acute chest pain. J Gen Intern Med 1993; 8:231–235[Medline]
-
Langewitz W, Kiss A, Saner R: Panic disorder in the emergency room (orig. in GERMAN). Ther Umsch 1995; 52:201–207[Medline]
-
American Psychiatric Association: Diagnostic and Statistical Manual, 4th Edition. Washington, DC, American Psychiatric Association, 1994
-
Gerdes T, Yates WR, Clancy G: Increasing identification and referral of panic disorder over the past decade. Psychosomatics 1995; 36:480–486[Abstract/Free Full Text]
-
Swinson RP, Soulios C, Cox BJ, et al: Brief treatment of emergency room patients with panic attacks. Am J Psychiatry 1992; 149:944–946[Abstract/Free Full Text]
-
McCullagh P, Nelder JA: Generalized Linear Models. New York, Chapman and Hall, 1983
-
Liang KY, Zeger SL: Longitudinal data analysis using generalized linear models. Biometrika 1986; 73:13–22[Abstract/Free Full Text]
-
Dyckman JM: A communications model of panic disorder. Anxiety Dis Pract J 1994; 1:77–82
-
Hoehn-Saric R, Frank JD, Imber SD, et al: Systematic preparation of patients for psychotherapy. I. Effects on therapy behavior and outcome. J Psychiatr Res 1964; 33:267–281[Medline]
Get information about faster international access.
Privacy Policy
Copyright © 1999
Academy of Psychosomatic Medicine.
All rights reserved.
Home
| Search
| Current Issue
| Past Issues
| Subscribe
| All APPI Journals
| Help
| Contact Us
|
TOP
ABSTRACT
INTRODUCTION
