Historically, little attention has been paid to comorbid psychiatric conditions of persons with schizophrenia. Schizophrenia typically involves a severe and persistent mental illness that is more severe than other conditions; therefore, other psychiatric problems that are comorbid with schizophrenia are considered “lesser” problems and receive less attention (
1). This is an unfortunate oversight, because comorbid conditions complicate conceptualization and treatment of psychotic symptoms (
2,
3) and may inhibit their improvement (
4). In recent years, however, there has been increased awareness of comorbid psychiatric conditions among persons with schizophrenia (
5). This change is likely due to recent advances in treatment and less restrictive diagnostic criteria regarding psychiatric comorbid conditions (
6).
Among individuals with schizophrenia, anxiety disorders occur at rates greater than those in the general population (
1,
5,
7,
8), with some researchers even suggesting that comorbid anxiety and schizophrenia could represent a specific subtype (
9–
11). A recent meta-analysis reported that more than a third of patients with schizophrenia were diagnosed as having a comorbid anxiety disorder (
7), and some national samples have reported even higher anxiety disorder prevalence rates. For example, a report from the Epidemiologic Catchment Area survey from the National Institute of Mental Health reported that nearly half of individuals with schizophrenia had comorbid panic disorder (
12). Comorbid anxiety of individuals with schizophrenia has the potential to limit recovery and quality of life, especially in areas such as social adjustment and work functioning (
13–
17), and has been linked to increased suicide risk (
18).
On the basis of meta-analytic data, initial steps have been taken to establish population-based prevalence rates for anxiety disorders among persons with schizophrenia (12.1% for obsessive-compulsive disorder, 14.9% for social anxiety, 10.9% for generalized anxiety disorder, 9.8% for panic disorder, and 12.4% for posttraumatic stress disorder [PTSD]) (
7). However, there was substantial variation in study population and design within the meta-analysis, potentially contributing to the variance in prevalence rates seen across studies. Further, the extant studies primarily relied on self-report or interview data, with little use of medical record data. Thus it was unclear from the study at what rates these conditions co-occur in real-world clinical situations. As other researchers have noted, clinical studies would provide important information on detection rates (
7).
A health system for veterans who have served honorably in the military, the Veterans Health Administration (VHA) comprises the largest integrated health care system in the United States and maintains numerous patient registries designed to evaluate trends in patient care. One such registry, the National Psychosis Registry, contains information related to the care of VHA patients with psychotic disorders and is maintained at the national Serious Mental Illness Treatment Resource and Evaluation Center (SMITREC). Previously, SMITREC used the registry to study comorbid anxiety of VHA patients with schizophrenia. The study demonstrated that having comorbid psychiatric problems is moderately common (15.3% PTSD comorbidity and 10.9% other anxiety disorder comorbidity during fiscal year [FY] 2011 [
19]), although such evaluations have not examined the characteristics of VHA patients with comorbid disorders or the potential clinical effects of such comorbidity in terms of clinical outcomes, including service utilization.
The overarching goal of this study was to utilize the National Psychosis Registry to evaluate the characteristics of VHA patients with comorbid schizophrenia and anxiety disorders. More specifically, this study was designed to assess the prevalence of comorbid anxiety disorders among VHA patients with schizophrenia, evaluate patient characteristics associated with a comorbid anxiety disorder, and assess the relationships between comorbid anxiety and service utilization.
Methods
Study population
The study population consisted of all individuals with schizophrenia who were included in the U.S. Department of Veterans Affairs (VA) National Psychosis Registry during FY 2011 (October 1, 2010–September 30, 2011). The National Psychosis Registry (
19) maintains diagnostic, demographic, utilization, and medication records for all VHA patients with diagnoses of schizophrenia or other serious mental illnesses and provided this information for all VHA patients with schizophrenia who were included in this study. Institutional review board approval from the VA Ann Arbor Health Care System was obtained for this study.
Treatment of PTSD has received increased attention within the VHA in recent years (
20,
21). Because of potential treatment differences between VHA patients with PTSD compared with those with another anxiety disorder, it is important to consider these groups separately. Thus comorbid anxiety diagnoses were grouped as PTSD or non-PTSD for purposes of the study.
Measures
Demographic and clinical characteristics.
Demographic variables included age, sex, marital status, and race (white versus other). Medical comorbidity was measured with the Charlson Comorbidity Index (
22), a measure of the effects of serious general medical conditions on ten-year mortality, with higher scores representing more severe medical comorbidity. Conditions are assigned scores according to their severity, and total scores are summed. For example, diabetes is assigned a score of 1, and AIDS is assigned a score of 6. More complete information can be found in other publications (
22). In addition, an indicator was created for service-connected disability status (yes or no), defined as a physical or psychiatric condition directly related to military service that affects one’s ability to hold gainful employment.
Psychiatric diagnoses.
All individuals in this study had a schizophrenia spectrum disorder (ICD-9 codes 295.0–295.9). Primary anxiety disorders of interest were PTSD (ICD-9 code 309.81), generalized anxiety disorder (300.02), obsessive-compulsive disorder (300.3), social anxiety disorder (300.23), panic disorder (300.01), agoraphobia (300.21 and 300.22), other or unspecified phobia (300.20 and 300.29), and anxiety not otherwise specified (300.00 and 300.09). Other diagnoses of interest included depression (293.83, 296.2, 296.3, 296.90, 296.99, 298.0, 300.4, 301.12, 309.0, 309.1, and 311.x), personality disorders (301.0, 301.20, 301.22, 301.4–301.7, 301.81–301.83, and 301.9), and substance use disorders (292.x, 304.0–304.9, and 305.2–305.9, with the exception of codes where the fifth digit is a 3, signifying remission).
Treatment variables.
Inpatient bed section codes were used to characterize psychiatric and general medical hospital admissions and length of stay for each admission. Indicators were created for receipt of outpatient mental health care, psychotherapy treatment, and anxiety medication (alprazolam, buspirone, chlordiazepoxide, clonazepam, clorazepate, diazepam, estazolam, flurazepam, lorazepam, midazolam, oxazepam, prazepam, quazepam, temazepam, and triazolam).
Analyses
The prevalence of any anxiety and specific anxiety diagnoses among patients with schizophrenia was calculated for FY 2011. Demographic, clinical, and treatment variables were compared for patients with no anxiety disorder, PTSD, or non-PTSD anxiety disorders. We used a chi square test for categorical variables and the Wilcoxon rank-sum test to compare means for continuous variables. Patients with missing demographic and clinical data were not included in the final sample. Because of the large sample size, differences were considered significant if there was a raw difference greater than 5% between percentages, a relative difference greater than 5% between means (for measures without percentages), and the alpha rejection level for statistical comparison was p<.001. These criteria were designed to evaluate both clinical (5% difference) and statistical (p<.001) significance.
Results
Aim 1: assess the prevalence of anxiety disorders
During FY 2011, 87,006 individuals were diagnosed as having schizophrenia in the VHA. Of these, 23.8% had a comorbid anxiety disorder, as indicated in
Table 1. Of the VHA sample with schizophrenia, 15.2% had a diagnosis of PTSD and 8.6% had a diagnosis of a non-PTSD anxiety disorder. Patients with diagnoses of both PTSD and another anxiety disorder (3.0%, N=2,659) were counted in the PTSD group. Among those with non-PTSD anxiety disorders, specific diagnoses were anxiety not otherwise specified (8.0%), generalized anxiety disorder (2.3%), obsessive-compulsive disorder (1.2%), panic disorder (1.1%), agoraphobia (.6%), social anxiety disorder (.2%), and other or unspecified phobia (.1%).
Aim 2: evaluate patient characteristics
Table 2 presents information regarding patient characteristics. Among the three patient groups (no comorbid anxiety, comorbid PTSD, and comorbid non-PTSD anxiety disorder), there were no significant differences in terms of sex (94.1%, 89.8%, and 92.0% of the groups were male, respectively) or age (57.9, 55.4, and 56.0 years, respectively). Patients without an anxiety disorder were less likely than those with PTSD to be married (21.3% versus 30.6%, χ
2=539.6, df=1, p<.001), but there was no significant difference compared with the non-PTSD anxiety group (26.0%). There was also no significant difference in marital status between the PTSD and non-PTSD groups. Patients with a non-PTSD anxiety disorder were more likely to be white (73.8%) compared with those without an anxiety disorder (59.0%, χ
2=621.6, df=1, p<.001) or PTSD (57.7%, χ
2=541.1, df=1, p<.001). There were no significant differences in white racial makeup between the no–anxiety disorder and PTSD groups. Patients in the PTSD group were the most likely to have a service-connected disability (72.9%) compared with those with no anxiety disorder (57.9%, χ
2=1,035.5, df=1, p<.001) or non-PTSD anxiety (56.7%, χ
2=568.53, df=1, p<.001). There were no significant differences in rates of service-connected disability between the no–anxiety disorder and non-PTSD anxiety disorder groups.
Patients without a comorbid anxiety disorder had significantly lower rates of other comorbid mental disorders compared with patients with comorbid anxiety disorders. Specifically, 20.6% of patients with no anxiety disorder had depression, compared with 47.7% of those with PTSD (χ2=4,304.5, df=1, p<.001) and 46.8% of those with non-PTSD anxiety (χ2=2,588.8, df=1, p<.001). Only 3.7% of patients with no anxiety disorder had a personality disorder, compared with 11.2% of those with PTSD (χ2=1,319.4, df=1, p<.001) and 10.8% of those with non-PTSD anxiety (χ2=811.1, df=1, p<.001). There were significant differences among all three groups with regard to substance use disorders. Patients with PTSD had higher rates of substance use disorder diagnoses (37.1%) compared with those with non-PTSD anxiety (29.8%, χ2=126.6, df=1, p<.001) and those without an anxiety disorder (21.0%, χ2=1,579.2, df=1, p<.001). Substance abuse diagnosis rates between patients with no anxiety and a non-PTSD anxiety disorder were also different (χ2=276.9, df=1, p<.001). General medical comorbidity showed a similar pattern, with patients with PTSD having the highest scores (1.4) compared with patients in non-PTSD anxiety (1.3, z=−4.9, p<.001) and no-anxiety groups (1.2, z=14.3, p<.001). General medical comorbidity rates between patients with non-PTSD anxiety and no anxiety were also different (z=5.25, p<.001).
Aim 3: assess comorbid diagnoses and service utilization
Table 3 presents data regarding diagnoses and service utilization. Patients with comorbid PTSD or non-PTSD anxiety had nearly double the rates of service utilization compared with those without comorbid anxiety. Over a fifth of VHA patients with PTSD (23.2%) or a non-PTSD anxiety disorder (21.8%) had at least one psychiatric hospitalization during FY 2011, a nonsignificant difference between the groups. However, patients without comorbid anxiety had significantly lower psychiatric hospitalization rates (11.9%) than both those with PTSD (χ
2=1,181.3, df=1, p<.001) and those with non-PTSD anxiety disorders (χ
2=584.6, df=1, p<.001). VHA patients with PTSD and non-PTSD anxiety had similar average lengths of psychiatric hospitalization (PTSD, 6.2 days; non-PTSD, 6.3 days), although both were longer than the average psychiatric hospitalization of patients without comorbid anxiety (4.4 days; versus PTSD, z=23.8, p<.001; versus non-PTSD anxiety, z=33.8, p<.001). A similar pattern was found regarding general medical hospitalizations, with 27.7% of those with PTSD and 25.8% of those with a non-PTSD anxiety disorder having a general medical stay (nonsignificant difference), compared with 18.9% of those without an anxiety disorder (no comorbid anxiety versus PTSD). There was no difference between the medical hospitalization rates of those with PTSD and non-PTSD anxiety. However, when compared with patients with no anxiety disorder, those with PTSD had more frequent hospitalizations (χ
2=522.68, df=1, p<.001; no comorbid anxiety versus non-PTSD anxiety; χ
2=199.24, df=1, p<.001). Lengths of hospital stay were not significantly different across the three groups (no anxiety, 16.7 days; PTSD, 16.0 days; and non-PTSD anxiety, 16.6 days).
Receipt of outpatient anxiety medication (benzodiazepines and buspirone) was most common among patients with non-PTSD anxiety (56.1%). This was higher than that of patients with PTSD (41.5%, χ2=408.9, df=1, p<.001) and those with no anxiety disorder (24.9%, χ2=3,236.6, df=1, p<.001). Rates of outpatient anxiety medication were also significantly different between groups of patients with PTSD and those with no anxiety disorder (χ2=1,509.1, df=1, p<.001). Compared with patients without an anxiety disorder, a greater percentage of patients with PTSD had at least one mental health visit during FY 2011 (88.9% versus 96.5%, χ2=716.1, df=1, p<.001) as well as at least one psychotherapy visit (66.3% versus 83.3%, χ2=1,497.0, df=1, p<.001). Also compared with those without an anxiety disorder, a greater percentage of patients with non-PTSD anxiety had at least one mental health visit in FY 2011 (88.9% versus 94.6%, χ2=231.4, df=1, p<.001) as well as at least one psychotherapy visit (66.3% versus 79.7%, χ2=554.0, df=1, p<.001). There were no significant differences between PTSD and non-PTSD anxiety groups in terms of percentages who had attended a mental health visit (96.5% versus 94.6%, respectively) or a psychotherapy visit (83.3% versus 79.7%, respectively).
Discussion
Anxiety disorders among persons with schizophrenia represent an area of clinical concern because of their effects on functioning, psychiatric symptoms, and suicide risk. Recent meta-analytic data (
7) suggest that comorbid anxiety disorders among persons with schizophrenia are common, although the available literature is plagued with methodological inconsistencies. In addition, there is little information regarding the extent to which comorbid anxiety disorders are being diagnosed in clinical settings. The aims of this study were to utilize the National Psychosis Registry to evaluate the prevalence of anxiety disorders among VHA patients with schizophrenia, to examine patient characteristics associated with anxiety disorders, and to assess the effects of comorbid anxiety on service utilization.
With the exception of patients with PTSD, VHA patients had lower rates of anxiety disorder diagnoses than have been found in past research. In addition, many specific anxiety conditions were diagnosed at low rates, and there was a very high diagnosis rate of anxiety not otherwise specified. These results suggest that there may be issues of underdetection of comorbid anxiety among VHA patients with schizophrenia, with certain diagnoses (generalized anxiety disorder or obsessive-compulsive disorder) being particularly overlooked. Also, there may be a tendency for VHA providers to offer more nonspecific diagnoses (anxiety not otherwise specified) rather than determine the specific anxiety disorder that VHA patients with schizophrenia are experiencing. Regardless of cause, it may be important to improve screening procedures aimed at detecting and properly defining these conditions among VHA patients with schizophrenia to support proper treatment and overall symptom management.
Some evidence suggests that VHA patients with comorbid anxiety disorders had greater severity of psychiatric symptoms, compared with those without comorbid anxiety disorders: higher rates of depression, personality disorders, and substance use disorders. Among patients with comorbid anxiety, there were some group differences: those with PTSD were more likely to have a high-level service-connected disability, as well as higher rates of substance use disorders and general medical comorbidity, compared with the non-PTSD anxiety group.
Despite some demographic differences, VHA patients with schizophrenia and comorbid anxiety represented a group comparable in regard to service utilization with those with schizophrenia and PTSD. VHA patients with comorbid PTSD and non-PTSD anxiety disorders were hospitalized for mental health concerns nearly twice as frequently and for longer stays than those without comorbid anxiety. Further, patients with comorbid anxiety disorders were also medically hospitalized at higher rates, although for similar lengths of time, compared with patients without anxiety disorders. Comorbid anxiety disorders were also related to increased outpatient service utilization (increased percentage with at least one mental health or psychotherapy visit and increased percentage who received antianxiety medication during FY 2011, compared with those with no anxiety). These results suggest that although there might be demographic differences between these groups, there might be similar treatment needs with regard to the co-occurrence of anxiety among individuals with schizophrenia within the VHA system.
The results of this evaluation have several implications for service delivery to VHA patients with schizophrenia. Reduced rates of comorbid anxiety diagnoses among VHA patients with schizophrenia relative to outside samples suggests the need for increased screening for the presence of comorbid anxiety by VHA providers working with patients with schizophrenia. Increased rates of negative outcomes of care such as general medical and psychiatric hospitalization suggest the need for outpatient interventions targeting anxiety for this group, as well as increased access to anxiety-targeting specialty mental health clinics (such as PTSD clinics). This group may additionally benefit from increased access to other specialty mental health care given their increased rates of other serious psychiatric comorbidity (including substance use disorders, depression, and personality disorders).
Strengths of this study include the use of a large nationwide sample, use of medical record data to evaluate how comorbid anxiety was diagnosed and treated, and the examination of a clinical issue by evaluating an entire clinical population (all VHA patients with schizophrenia). However, this study also had its limitations. Diagnoses were ascertained from medical records rather than direct assessment of patients, and some individuals may have had anxiety disorders that were not reflected in their medical records. Thus direct comparison of anxiety disorder rates with those in other studies may not be appropriate, and lower rates might not necessarily mean poorer detection (
23). In addition, the method of comparing diagnosis and service receipt within the same period may be of concern, because increased service receipt may have been associated with increased clinical scrutiny that aided in the detection of comorbid anxiety. Regardless, the rates presented here were determined from a large, nationally representative sample and offer insight into how anxiety conditions accompanying schizophrenia are being diagnosed and treated in clinical care.
Conclusions
Comorbid anxiety disorders among patients with schizophrenia have traditionally been neglected and negatively affect recovery and functioning. In this study, the presence of anxiety disorders was found to be related to greater psychiatric severity and increased use of general medical and mental health services. These results have important implications for clinical practice, and within the VHA efforts to improve appropriate screening and treatment of comorbid anxiety in schizophrenia might be beneficial. Future work in this area could focus on clarifying which treatments are more effective for those with schizophrenia and comorbid anxiety, as well as better establish anxiety screening and service offerings for individuals with schizophrenia in the VHA.
Acknowledgments and disclosures
The authors report no competing interests.