Journal Logo


Nationwide Analysis of Changes in Medical Utilization Before and After a New Panic Disorder Diagnosis

Chang, Hu-Ming MD; Pan, Chun-Hung MD, MS; Chen, Pao-Huan MD; Chen, Yi-Lung MD; Su, Sheng-Siang MS; Tsai, Shang-Ying MD; Chen, Chiao-Chicy MD, PhD; Kuo, Chian-Jue MD, PhD

Author Information
doi: 10.1097/PSY.0000000000000810



Somatization is associated with increased health care service utilization and substantial societal costs (1–4). The term somatization, interchangeably used with medically unexplained symptoms or functional somatic syndromes, generally describes a tendency to experience somatic symptoms in response to psychological stress (5–7). Several studies have investigated the cognitive and behavioral characteristics of somatization, including a false illness belief, catastrophic interpretation of somatic symptoms, illness worry and health anxiety, selective attention to bodily symptoms, and reassurance-seeking behaviors (8–10). Although patients with somatization exhibit frequent care-seeking behaviors, medical reassurance failed to reduce their fear of illness or did so only temporarily (11). The failure of reassurance was attributable to dissatisfaction regarding the quality of their health care, which further reinforced their desire to seek better health care. Thus, identifying such patients and decreasing their medical utilization are crucial.

Among the patients with somatization, those with panic disorder (PD) had the highest risk of exhibiting multiple medically unexplained symptoms and high medical utilization (12,13). Compared with healthy control patients, patients with PD had a four times greater risk of having been previously diagnosed with medical conditions (14). PD is associated with high emergency department (ED) resource utilization, evidenced by the Epidemiologic Catchment Area study that found the proportion of PD patients who ever visit ED to be higher than the corresponding proportion for the major depression and control groups (15). Moreover, the presentation of panic attacks in ED, which mimic life-threatening conditions, could promote superfluous examinations and medical costs. Estimated annual per capita costs of PD were higher than those of the other mental disorders including mood, anxiety, and substance use disorders (16). Unnecessary examinations could result in overdiagnosis of trivial illnesses, reinforcing patients’ belief that some illness is present that remains underdiagnosed, thereby reinforcing care-seeking behavior.

The pattern of service utilization for patients with PD has been widely studied in literature (17–20). Before diagnosis of PD, the patients have the highest rates of visits to cardiology, family medicine, and emergency medicine (21). Although treatment of PD could decrease the frequency of panic attacks and visits to EDs, whether treatment of PD decreases medical utilization in other outpatient and inpatient departments is unknown. Research using self-reporting questionnaires demonstrated that patients with somatization exhibited no decrease in medical utilization during the 1 year after being assessed compared with the preceding year (2). One critical question is whether PD diagnosis is associated with a subsequent decrease in non-ED outpatient and inpatient visits. Second, the literature for medical utilization of patients with PD seldom scrutinized medical examination utilization, which is central in PD research. Using the criteria from the Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition), nearly half of patients with PD also fulfill the requirements for hypochondriasis (14). Repetitive medical examinations might be one of the primary presentations of PD, although this might not serve as effective reassurance and could reinforce care-seeking behavior. The fact that medical examinations are not routinely arranged during each visit represents a challenge for the research on medical examination, and a large cohort might be required to overcome this difficulty. Third, somatization might mask real psychiatric illness and result in delayed diagnosis and treatment. The change in psychiatric comorbidity after PD diagnosis is of interest.

This study aimed primarily to examine the change in medical utilization pattern for patients with PD before and after their diagnosis. We used a national cohort to examine the detailed medical utilization pattern including specialty visits, medication used, and most notably, medical examinations. Using a case-crossover design, we compared medical utilization within 1 year before and after PD diagnosis. Second, we examined the change in psychiatric comorbidity after PD diagnosis, which might have been masked by somatic symptoms. We hope that this study will fill the knowledge gap in the research related to medical utilization for PD.


Data Source

The National Health Insurance program in Taiwan was launched on March 1, 1995, and 99.9% of Taiwan’s population was enrolled at the end of 2014 (21). The National Health Insurance Research Database (NHIRD) contains registration files and medical claims data based on the National Health Insurance program and is representative of Taiwan’s entire population. The identities of NHIRD enrollees were fully encrypted by the Bureau of National Health Insurance to preserve patient anonymity. Investigators had to sign an agreement guaranteeing patient confidentiality to use the database for research purposes. A waiver of informed consent was granted upon institutional review because of the minimal privacy risk to individual patients.

The data were extracted from the Longitudinal Health Insurance Database 2005, which comprises a representative cohort of 1,000,000 patients randomly sampled from the 2005 registry of the NHIRD. Approximately 25.68 million individuals’ data are documented in this registry, with no significant differences between the sex distribution (χ2 = 0.008, df = 1, p = .93) for patients in the Longitudinal Health Insurance Database 2005 and those in the original NHIRD. The database includes patients’ demographic characteristics, diagnoses, medical expenditures, and prescription claims data. The study was approved by the Ethics Committee of the Taipei City Hospital and conformed to the provisions of the Declaration of Helsinki.

Study Cohort

The definition of PD is based on the International Classification of Diseases, Ninth Revision, Clinical Modification. Patients with newly diagnosed PD from January 1, 2000, to December 31, 2012, were selected as study participants. The index date was defined as the date when participants were diagnosed as having PD. All patients were subject to a full 1-year observation period before and after the index date. Patients diagnosed with PD between January 1, 1996, and December 31, 1999 (n = 406), were excluded from the establishment of the population at risk. A total of 8722 patients diagnosed as having PD from January 1, 2000, to December 31, 2012, were selected as the study cohort (n = 8722; Figure 1).

Study flow diagram. ICD-9 = International Classification of Diseases, Ninth Revision.

Case-Crossover Design

The case-crossover design used the study participants as their own controls. The case-crossover design examines the effects of brief exposure or acute events (22). By adopting the self-matching process, the case-crossover design can address control-selection bias in case-control studies. Our study assigned patients after the index date when a PD diagnosis was confirmed as cases; the same patients before the index date constituted the matched controls. A period of 1 year after the index date and 1 year before the index date were chosen as the case and control windows, respectively (Figure S1, Supplemental Digital Content,

Demographic and Clinical Characteristics

We collected the demographic profiles of these cases (controls) including sex, age (in which PD is diagnosed), employment status, level of urbanization, and Charlson comorbidity index. The levels of urbanization were categorized into level 1 (highly urbanized area), level 2 (moderately urbanized area), level 3 (newly urbanized area), level 4 (township area), and level 5 (rural area). The Charlson comorbidity index categorized comorbid physical conditions as 0, 1, or ≥2 according to the ICD-9 codes (23). We also collected the following variables before and after the PD diagnosis. Based on categorization used in other research, the medical utilization was categorized into outpatient, inpatient, and ED (2). Medical examinations in this study include electrocardiography, chest radiography, sonography, and endoscopy. The physical and psychiatric comorbidities were identified using the ICD-9 codes. Information on the concurrent medications was obtained from the Anatomical Therapeutic Chemical Classification system.

Statistical Analysis

Based on the case-crossover study design, we used conditional logistic regression analyses to explore the differences in medical utilization before and after PD diagnosis. The regression analyses were performed using SAS version 9.4 (SAS Institute, Cary, North Carolina). In the model for medical comorbidity, we conducted a multivariate analysis to examine the change of physical and psychiatric comorbidities. Variables with at least a very strong association (p < .001) were retained in the final explanatory model. Multivariate analyses were performed using backward variable selection. p < .001 was considered statistically significant.


Among the 8722 individuals with PD, 3145 were men (36.1%) and 5577 were women (63.9%). The mean (SD) age at the new diagnosis of PD was 45.5 (15.3) years. Most cases were identified in the department of psychiatry (58.2%), followed by the department of internal medicine (16.0%), family practice (9.1%), neurology (6.7%), and emergency medicine (3.7%). Detailed information on the demographic characteristics of PD cases is summarized in Table S1, Supplemental Digital Content,

Change in Medical Examinations Before and After PD Diagnosis

Table 1 shows the examinations used in 1 year preceding and following the date when PD was diagnosed. In the outpatient department, a significant decrease in the utilization of electrocardiography (risk ratio [RR] = 0.70) and radiography (RR = 0.89) after PD was observed. In the ED, a significant decrease in the number of electrocardiography (RR = 0.73) and radiography (RR = 0.80) after PD was observed.

Patterns of Medical Utilization Within 1 Year Before (Control Window) and 1 Year After (Case Window) the Index Date Among Patients With Newly Diagnosed Panic Disorder

Change in Service Utilization Before and After PD Diagnosis

Table 2 features the number of outpatient, inpatient, and ED visits in 1 year preceding the PD diagnosis date (control window) and in the 1 year following the PD diagnosis date (case window). In the outpatient services, the mean number of nonpsychiatric department visits decreased from 30.79 in the control window to 28.72 in the case window with statistical significance (RR = 0.989, p < .001), whereas the mean number of psychiatric department visits increased from 0.95 to 1.21 (RR = 1.193, p < .001). In the ED, the number of nonpsychiatric ED visits decreased (RR = 0.924). In the inpatient department, the number of nonpsychiatric hospitalizations decreased (RR = 0.892), whereas the number of psychiatric department hospitalizations increased (RR = 1.386). Using multivariate analysis, we observed that the visits to the departments decreased, including ED, gynecology, and ophthalmology (Table S2, Supplemental Digital Content, However, the visits to two departments increased after a new PD diagnosis, namely, psychiatry and neurology.

Patterns of Medical Utilization Within 1 Year Before (Control Window) and 1 Year After (Case Window) the Index Date Among Patients With Newly Diagnosed Panic Disorder

Change in Comorbid Psychiatric Illness Before and After PD Diagnosis

Table 3 shows the distribution of psychiatric comorbidities preceding and following PD diagnosis. In the multivariate conditional logistic regression analysis, only those with at least a very strong association (p < .001) were retained in the final regression model. After the diagnosis of PD, the number of individuals having comorbid depressive (adjusted RR [aRR] = 3.06) and bipolar disorder (aRR = 1.77) increased, whereas the number of individuals with comorbid nonpanic anxiety disorder (aRR = 0.69) decreased. Moreover, the distribution of physical comorbidities preceding and following PD diagnosis is provided in Table S3, Supplemental Digital Content,

Distribution of Psychiatric Comorbidity 1 Year After and 1 Year Before the Index Datea Among Patients With Newly Diagnosed Panic Disorder

Change in Concurrent Medications Before and After PD Diagnosis

Table 4 summarizes the distribution of concomitant medications preceding and following PD diagnosis. In particular, patients were more likely to use antidepressants (RR = 12.65) and benzodiazepines (RR = 11.63) after PD diagnosis compared with before PD diagnosis. Details of prescribed antidepressant types are provided in Table S4, Supplemental Digital Content,

Medications Used 1 Year After and 1 Year Before the Index Datea Among Patients Newly Diagnosed Panic Disorder

Long-Term Trend of Changes in Medical Utilization

For estimating long-term trends in medical utilization changes, we compared the patterns of medical utilization within 1 year before (control window) the index date and each year after the index date from the first year to the fifth year as the individual case windows. The results are given in Table S5, Supplemental Digital Content, We found the trends changes to resemble those in the main analysis (Table 2). For example, in the outpatient department, the number of nonpsychiatric department visits maintained a decreasing trend, whereas the number of psychiatric department visits increased 1 year after the baseline and gradually decreased later.


Strengths of This Study

To the best of our knowledge, this is the first study to analyze the changes in medical utilization among patients before and after a new PD diagnosis. Three major findings emerged from our study. First, patients used fewer medical examinations within 1 year after a PD diagnosis compared with during the 1 year preceding PD diagnosis. Second, the number of visits to nonpsychiatric inpatient and outpatient services decreased after PD diagnosis, whereas the number of visits to psychiatric inpatient and outpatient services increased. ED visits also decreased after PD diagnosis. Third, regarding psychiatric comorbidity, the proportions of depressive disorders and bipolar disorders identified increased, whereas that of nonpanic anxiety disorders decreased after PD was diagnosed. In addition, as revealed in Table S1 (Supplemental Digital Content,, an increasing trend occurred in the number of participants diagnosed with PD over a decade. This increase might have been partly due to an increased recognition of PD symptoms and improved identification of PD. Moreover, this trend corresponds with the increasing prevalence of common mental disorders in Taiwan (24) partly because of the adverse effect of rapid economic changes after industrialization.

Changes in Medical Examination Utilization

One major finding of our study is that medical examination utilization, including the utilization of electrocardiography and radiography, decreased after PD diagnosis. One possible explanation for this is a potential decrease in frequency of panic attacks, which usually require electrocardiograms and chest radiographs in EDs. According to an ED survey, 80% of patients with PD had atypical or nonanginal chest pain (25). We also noted that sonography and endoscopy demonstrated a similar trend of decrease after PD diagnosis. We know of no literature regarding the use of sonography or endoscopy associated with PD, although sonography or endoscopy could be used for common prevalent presentations of panic attacks, for example, a lump in the throat or difficulty swallowing. We also noted that the decrease in the utilization of examinations applied not only for EDs but also for inpatient and outpatient departments. A possible explanation for this is the decrease in somatic symptoms or a reduction in illness anxiety, followed by the diagnosis and treatment of PD. We also noted a trend whereby a decrease in examinations within primary care settings (outpatient departments and EDs) outweighed decreases in the inpatient department, corresponding with other research finding of patients with PD exhibiting high medical utilization, particularly in primary care settings. The problem of superfluous examinations requested by patients with somatization might be of clinical relevance, because excessive examination might not alleviate patients’ anxiety. A randomized controlled trial demonstrated that patients with chronic headache experienced no improvement in symptoms or reduction in health concerns after brain magnetic resonance imaging (26). The examination results might not have alleviated their worries and could have initiated a vicious cycle of doctor consultations for reassurance (9). To address this problem, examinations for patients with somatization should involve sufficient explanation and reassurance (27).

Changes in Specialists Consulted

We noted a change in the number of specialist visits after PD was diagnosed. In general, the number of visits to nonpsychiatric departments decreased after PD diagnosis, whereas the number of visits to psychiatric departments increased after PD diagnosis. The increase in visits to psychiatric departments might be for treatment of PD or other comorbid psychiatric disorders. The decrease in visits to EDs may be partly due to reduced frequency of panic attacks after treatment (28,29). The decrease in visits to nonpsychiatric outpatient and inpatient departments might reflect a decrease in somatic symptoms or illness anxiety. The result demonstrated that a diagnosis of PD might help to decrease medical utilization. Moreover, the number of visits to EDs and inpatient departments decreased, whereas that to outpatient departments slightly increased after PD diagnosis. The increase in the number of visits to the outpatient department is due to the large increase in visits to psychiatric departments.

Changes in Psychiatric and Physical Comorbidities

We noted some changes in the psychiatric comorbidity of PD after diagnosis. Bipolar and depressive disorders significantly increased, whereas nonpanic anxiety disorder decreased after adjustment for other variables. Our study did not imply that clinical care is unhelpful in reducing depressive or bipolar disorders after PD. The National Comorbidity Survey showed that 19% of individuals who received treatment of PD developed major depression and 45% of individuals who received no treatment of PD developed major depression (30). Nonpanic anxiety disorder decreased after PD diagnosis, suggesting that patients might be diagnosed as having unspecified anxiety disorder before they are identified as having PD. PD has high comorbidity with other psychiatric disorders, including depressive disorder, anxiety disorder, phobic disorder, and alcohol use disorder (31). However, somatization may mask patients’ psychopathology, causing a delay in true diagnosis. In keeping with other studies’ results, PD had a high comorbidity rate with depression, which might have been overlooked before the diagnosis of PD. The results of our study are in line with the National Comorbidity Survey, which showed that primary PD and panic attacks predicted a first onset of subsequent major depression (32). Moreover, we noted that the bipolar diagnosis significantly increased after PD diagnosed. Bipolar disorder had a higher comorbidity rate with anxiety disorders compared with depression, including PD (33). For psychiatrists, identification of bipolar disorder is also relevant for patients with newly diagnosed PD.

We noted changes in the medication used after PD diagnoses. The most significant change was associated with antidepressants and benzodiazepines, which are mainstream treatments for PD. Research has indicated that treatment with selective serotonin reuptake inhibitors could reduce ED visits along with costs of laboratory services in the subsequent 6-month period compared with the 6-month period before treatment with selective serotonin reuptake inhibitors (29). Increased benzodiazepine prescription increased the concern of substance use disorder. However, the occurrence of discontinuation symptoms when benzodiazepine is tapered or discontinued, which present as symptom reoccurrence or as rebound or emergence of withdrawal symptoms, does not necessarily imply physiological dependence (34). Furthermore, a survey on psychiatrists showed that the increase over prescribed dose is more often observed with benzodiazepines, whereas treatment for longer than intended and unsuccessful efforts to cut down (35).

One limitation of our study is that we did not obtain utilization patterns of nonpharmacological treatment—including psychotherapy and brief intervention, psychoeducation regarding disease concepts, and relaxation techniques—which represent a critical aspect of PD treatment. One study demonstrated that cognitive-behavioral therapy could reduce the panic and depressive symptoms and unnecessary visits to emergency care facilities among patients with PD (36). Although simply being informed of a PD diagnosis was not associated with a superior outcome for patients who visited an ED (37), brief intervention and exposure instruction in an ED could reduce depression and panic frequency in the subsequent 6 months (38,39). Identification of PD, initiation of pharmacotherapy, and brief psychoeducation in the ED is associated with continued treatment of PD in the subsequent 3 months (40). We emphasize that early identification with referral to psychiatric treatment is required for patients with PD in ED and primary care settings.


Several limitations of the study should be noted. First, PD and other psychiatric disorders might have been diagnosed by nonpsychiatric physicians, and diagnostic validity is unverifiable. Second, we did not examine the utilization of nonpharmacological treatment, for example, cognitive-behavioral therapy, which is extremely relevant to the treatment of PD. Third, we arbitrarily chose 1 year as a time window, and the change in long-term medical utilization was not assessed. Fourth, because this was an observation study, the association of PD diagnosis and decreased medical utilization requires further investigation. The change in medical utilization after diagnosis of other psychiatric illnesses with somatization (e.g., depression) might be of interest in further research. Fifth, we aimed to conduct a clinical epidemiological study by using the data from the NHIRD. This study includes only individuals who sought care in the health care system. Thus, the data do not represent whole-population epidemiological study.


This study demonstrated that diagnosis of PD is associated with a decrease in subsequent nonpsychiatric medical utilization. We suggest that early identification of PD and referral for appropriate treatment might be helpful for mitigating potentially unnecessary use of nonpsychiatric examinations and services. We also advocate further studies to replicate the analysis of the change in medical utilization after a PD diagnosis in other nationwide datasets.

Data Availability Statement: The National Health Research Institutes maintain the National Health Institutes Research Database (NHIRD), which is available to researchers upon application approval. Because the privacy of the insured patients is paramount to NHIRD use, the National Health Research Institutes encrypt the data to protect the identities of enrollees and health care providers (

Author contributions: Drs. Chang, Pan, and Kuo conceived and designed the study. Dr. Kuo acquired the data. Mr. Su performed the statistical analysis. Drs. Tsai and C.C. Chen provided administrative and material support. Drs. Chang and Kuo drafted the manuscript. Drs. P.H. Chen and Y.L. Chen made critical revisions to the manuscript for key intellectual content, and Drs. Tsai and C.C. Chen supervised the study.

The authors thank I-Shuan Wang, BS, with the Department of General Psychiatry, Taipei City Psychiatric Center, Taipei City Hospital, for data management and help with statistical analyses. Ms. Wang declares that she has no competing interests. This article was edited by Wallace Academic Editing.

Source of Funding and Conflicts of Interest: This research was supported by grants from the Ministry of Science and Technology, Taiwan (MOST 105-2314-B-532-006-MY3, MOST 108-2314-B-532-005) and Taipei City Hospital (10501-62-015 and 10601-62-002). The funding sources had no involvement in the study design, data collection, analysis, interpretation of data, writing of the report, or the decision to submit the manuscript for publication. The authors declare that they have no competing interests.


1. Barsky AJ, Orav EJ, Bates DW. Somatization increases medical utilization and costs independent of psychiatric and medical comorbidity. Arch Gen Psychiatry 2005;62:903–10.
2. Barsky AJ, Ettner SL, Horsky J, Bates DW. Resource utilization of patients with hypochondriacal health anxiety and somatization. Med Care 2001;39:705–15.
3. Barsky AJ, Orav EJ, Bates DW. Distinctive patterns of medical care utilization in patients who somatize. Med Care 2006;44:803–11.
4. Hiller W, Fichter MM. High utilizers of medical care: a crucial subgroup among somatizing patients. J Psychosom Res 2004;56:437–43.
5. Henningsen P, Zipfel S, Herzog W. Management of functional somatic syndromes. Lancet 2007;369:946–55.
6. Henningsen P, Zimmermann T, Sattel H. Medically unexplained physical symptoms, anxiety, and depression: a meta-analytic review. Psychosom Med 2003;65:528–33.
7. Lipowski ZJ. Somatization: the concept and its clinical application. Am J Psychiatry 1988;145:1358–68.
8. Brown RJ. Medically unexplained symptoms: a new model. Psychiatry 2006;5:43–7.
9. Rief W, Broadbent E. Explaining medically unexplained symptoms—models and mechanisms. Clin Psychol Rev 2007;27:821–41.
10. Deary V, Chalder T, Sharpe M. The cognitive behavioural model of medically unexplained symptoms: a theoretical and empirical review. Clin Psychol Rev 2007;27:781–97.
11. Salkovskis PM, Warwick HM. Morbid preoccupations, health anxiety and reassurance: a cognitive-behavioural approach to hypochondriasis. Behav Res Ther 1986;24:597–602.
12. Simon GE, VonKorff M. Somatization and psychiatric disorder in the NIMH Epidemologic Catchment Area study. Am J Psychiatry 1991;148:1494.
13. Katon WJ, Von Korff M, Lin E. Panic disorder: relationship to high medical utilization. Am J Med 1992;92:S7–11.
14. Furer P, Walker JR, Chartier MJ, Stein MB. Hypochondriacal concerns and somatization in panic disorder. Depress Anxiety 1997;6:78–85.
15. Markowitz JS, Weissman MM, Ouellette R, Lish JD, Klerman GL. Quality of life in panic disorder. Arch Gen Psychiatry 1989;46:984–92.
16. Batelaan N, Smit F, de Graaf R, van Balkom A, Vollebergh W, Beekman A. Economic costs of full-blown and subthreshold panic disorder. J Affect Disord 2007;104:127–36.
17. Katerndahl DA, Realini JP. Where do panic attack sufferers seek care? J Fam Pract 1995;40:237–43.
18. Kennedy BL, Schwab JJ. Utilization of medical specialists by anxiety disorder patients. Psychosomatics 1997;38:109–12.
19. Barsky AJ, Delamater BA, Orav JE. Panic disorder patients and their medical care. Psychosomatics 1999;40:50–6.
20. Deacon B, Lickel J, Abramowitz JS. Medical utilization across the anxiety disorders. J Anxiety Disord 2008;22:344–50.
21. National Health Insurance Administration, Ministry of Health and Welfare, ROC. National Health Insurance Annual Report 2014–2015. Taipei, Taiwan: Ministry of Health and Welfare; 2014.
22. Maclure M. The case-crossover design: a method for studying transient effects on the risk of acute events. Am J Epidemiol 1991;133:144–53.
23. Charlson M, Szatrowski TP, Peterson J, Gold J. Validation of a combined comorbidity index. J Clin Epidemiol 1994;47:1245–51.
24. Fu TS, Lee CS, Gunnell D, Lee WC, Cheng AT. Changing trends in the prevalence of common mental disorders in Taiwan: a 20-year repeated cross-sectional survey. Lancet 2013;381:235–41.
25. Fleet RP, Dupuis G, Marchand A, Burelle D, Arsenault A, Beitman BD. Panic disorder in emergency department chest pain patients: prevalence, comorbidity, suicidal ideation, and physician recognition. Am J Med 1996;101:371–80.
26. Howard L, Wessely S, Leese M, Page L, McCrone P, Husain K, Tong J, Dowson A. Are investigations anxiolytic or anxiogenic? A randomised controlled trial of neuroimaging to provide reassurance in chronic daily headache. J Neurol Neurosurg Psychiatry 2005;76:1558–64.
27. Hatcher S, Arroll B. Assessment and management of medically unexplained symptoms. BMJ 2008;336:1124–8.
28. Gorman JM, Liebowitz MR, Fyer AJ, Goetz D, Campeas RB, Fyer MR, Davies SO, Klein DF. An open trial of fluoxetine in the treatment of panic attacks. J Clin Psychopharmacol 1987;7:329–32.
29. Roy-Byrne PP, Clary CM, Miceli RJ, Colucci SV, Xu Y, Grudzinski AN. The effect of selective serotonin reuptake inhibitor treatment of panic disorder on emergency room and laboratory resource utilization. J Clin Psychiatry 2001;62:678–82.
30. Goodwin R, Olfson M. Treatment of panic attack and risk of major depressive disorder in the community. Am J Psychiatry 2001;158:1146–8.
31. Tilli V, Suominen K, Karlsson H. Panic disorder in primary care: comorbid psychiatric disorders and their persistence. Scand J Prim Health Care 2012;30:247–53.
32. Kessler RC, Stang PE, Wittchen H-U, Ustun TB, Roy-Burne PP, Walters EE. Lifetime panic-depression comorbidity in the National Comorbidity Survey. Arch Gen Psychiatry 1998;55:801–8.
33. Chen Y-W, Dilsaver SC. Comorbidity of panic disorder in bipolar illness: evidence from the Epidemiologic Catchment Area Survey. Am J Psychiatry 1995;152:280.
34. Nardi AE, Cosci F, Balon R, Weintraub SJ, Freire RC, Krystal JH, Roth T, Silberman EK, Sonino N, Fava GA. The prescription of benzodiazepines for panic disorder: time for an evidence-based educational approach. J Clin Psychopharmacol 2018;38:283–5.
35. Amrein R, Levitan M, Freire RCR, Nardi AE. Benzodiazepines in Panic Disorder. Panic Disorder: Springer; 2016:237–53.
36. Watkins PL. Manualized treatment of panic disorder in a medical setting: two illustrative case studies. J Clin Psychol Med Settings 1999;6:353–72.
37. Katerndahl DA. Predictors and outcomes in people told that they have panic attacks. Depress Anxiety 2003;17:98–100.
38. Swinson RP, Soulios C, Cox BJ, Kuch K. Brief treatment of emergency room patients with panic attacks. Am J Psychiatry 1992;149:944–6.
39. Dyckman JM, Rosenbaum RL, Hartmeyer RJ, Walter LJ. Effects of psychological intervention on panic attack patients in the emergency department. Psychosomatics 1999;40:422–7.
40. Wulsin L, Liu T, Storrow A, Evans S, Dewan N, Hamilton C. A randomized, controlled trial of panic disorder treatment initiation in an emergency department chest pain center. Ann Emerg Med 2002;39:139–43.

panic disorder; somatization; medical utilization; case-crossover; psychiatric service; aRR = adjusted risk ratio; ED = emergency departments; NHIRD = National Health Insurance Research Database; PD = panic disorder; RR = risk ratio

Supplemental Digital Content

Copyright © 2020 by the American Psychosomatic Society