Concurrent Psychiatric Diagnoses by Age and Race Among Persons With Bipolar Disorder

We identified all patients who had a diagnosis of bipolar disorder at any time during fiscal year 2000 (October 1, 1999, to September 30, 2000) at a Department of Veterans Affairs (VA) facility in western Pennsylvania from the local administrative database by using ICD-9 codes (296.0x, 296.1x, and 296.4x to 296.8x, inclusive) (5). For these patients, we ascertained demographic characteristics (age and race) and inpatient and outpatient ICD-9 psychiatric diagnosis variables during the same fiscal year (2000) from the VA National Patient Care Database (NPCD). The NPCD is an administrative database that includes all inpatient and outpatient encounters in the VA health care system. In the VA, providers and professional coders determine ICD-9 codes by recording them on a patient encounter form as part of the administrative data routinely collected by the VA health care facilities, which in turn are included in the local and national VA electronic databases. We ascertained the following diagnoses occurring at any time during fiscal year 2000: schizophrenia (code 295.xx), depressive disorders (codes 296.2x to 296.3x, 300.4, 309.1, and 311), alcohol use disorders (codes 291.xx, 303.xx, and 305.xx), drug use disorders (292.xx and 304.xx), posttraumatic stress disorder (PTSD) (code 309.81), and other anxiety (non-PTSD) disorders (code 300.0x).

To evaluate whether diagnostic patterns differed by age-race groups, we identified four groups: older whites (60 years or older), younger whites (younger than 60 years), older African Americans, and younger African Americans. Chi square tests of proportions were used to compare the frequencies of key diagnoses among these age-race groups.

A total of 878 patients were identified as having a diagnosis of bipolar disorder in fiscal year 2000. Of those, 816 had complete data on race or ethnicity and age. One-hundred eighty-seven (23 percent) were aged 60 years or older, 90 (11 percent) were women, and 123 (15 percent) were African American. Patients from other minority groups (three patients) were excluded, because the vast majority of patients were either white or African American. One-hundred seventy-one patients (21 percent) were older whites, and 15 (2 percent) were older African Americans.

Of the 813 patients, 252 (31 percent) had a record of a primary diagnosis of schizophrenia, and 436 (54 percent) had a record of a primary diagnosis of depressive disorder in the same period (fiscal year 2000)—disorders that are each considered mutually exclusive of bipolar disorder, although often considered rule-out diagnoses in the differential diagnostic process. In addition, 244 patients (30 percent) had a diagnosis of alcohol use disorder; 111 (14 percent), drug use disorder; 114 (14 percent), PTSD; and 92 (11 percent), another anxiety disorder.

Among the four age-race groups, older African Americans were the most likely to receive a diagnosis of schizophrenia during the same fiscal year compared with the other three groups (Table 1). A diagnosis of a depressive disorder was more common among younger patients than older patients.

Almost a third of our sample of patients with a diagnosis of bipolar disorder also received a diagnosis of schizophrenia at some point over the 12-month observation frame. Likewise, more than half had a diagnosis of a depressive disorder over this time frame. Both these disorders are technically considered mutually exclusive of bipolar disorder and thus represent instability in the primary diagnosis. We found that among patients with a diagnosis of bipolar disorder in a routine care setting, older African Americans were the most likely of the four groups to have received a diagnosis of schizophrenia during the same period.

The higher rates of schizophrenia diagnoses among African Americans compared with whites in previous studies (6,7) have been largely attributed to clinician bias in diagnosing schizophrenia among African Americans rather than to higher rates of this illness among African Americans compared with whites (3). Such bias may be due to the perception among providers of inpatient mental health services that African Americans are more violent or dangerous, which would lead to a diagnosis of psychosis (8).

Nonetheless, the elevated occurrence of schizophrenia diagnoses among older African Americans with bipolar disorder suggests that the inconsistent diagnosis of schizophrenia in this group may have been exacerbated by inconsistent treatment experiences in the past. For example, older African Americans might be seeing providers who have less experience in considering cultural contextual factors that need to be taken into account in the determination of the presence or absence of psychosis (8). Furthermore, historical changes in diagnostic criteria, including the narrowing of the diagnosis of schizophrenia, may also have contributed to the diagnostic instability among older patients. In particular, pre-DSM-III and DSM-III-R differential diagnostic practice may have yielded diagnoses of schizophrenia without a ruling out of bipolar disorder. Thus older African Americans could have carried pre-DSM-III-R diagnoses of schizophrenia from earlier stages in their course of their treatment. Moreover, some clinicians might also "label" patients with an existing diagnosis without conducting a comprehensive assessment.

The risk of multiple diagnoses may also be greater because older patients may receive inconsistent care as a result of weak social support systems. Both older and African-American patients may also have fewer economic resources than their younger, white counterparts. In these situations, patients often do not achieve remission, because of inconsistent care, and they subsequently receive different diagnoses.

Finally, factors at the health-system level may also contribute to inconsistency of diagnoses. Older African Americans might be less likely to receive care from a single provider over time because of changes in eligibility for public health insurance, or they may receive care at facilities that do not have the financial means to ensure comprehensive follow-up care. Thus a lack of continuity of care could result in providers' being less aware of patients' medical history, leading to inconsistency of diagnoses. A lack of institutional support in these settings for formal mechanisms for reviewing diagnoses and improving diagnostic accuracy in routine care may also contribute to diagnostic inconsistency, especially among older African Americans. Given that little information is available about the extent to which diagnoses are formally revised, it is important to delineate the determinants of inconsistent diagnoses in order to improve continuity of care for patients with axis I disorders.

This study had limitations that warrant consideration. Foremost is that our diagnostic data were derived from administrative data sets and not confirmed by more formalized diagnostic procedures. Nevertheless, we wanted to examine how psychiatric diagnoses were assigned in a routine care situation—that is, without the formal structured assessments that are available in research settings. Our sample also included few women, partly because this was a veteran population. Thus we were very limited in our ability to examine gender-related differences in concurrent psychiatric diagnoses. In addition, the results of the study may not be generalizable beyond the VA. However, a majority of patients with bipolar disorder receive care from publicly funded providers, such as the Veterans Health Administration and Medicaid. Hence our sample may be representative of patients with bipolar disorder who are seen at other types of publicly funded facilities.

Despite these limitations, our findings on greater instability in the assignment of axis I diagnoses among older African Americans are of considerable clinical significance, because such diagnostic instability can lead to additional disparities in mental health treatment for this group. Because diagnostic inconsistency can adversely affect decisions about treatment, clinicians need to consider the potential impact of concurrent, mutually exclusive diagnoses observed in routine care settings, especially among older African-American patients.

This study was supported by Health Services Research and Development Service Investigator-Initiated Award IIR 02-283-2, the Center for Health Equity Research and Promotion, grant MH-30915 from the Mental Health Intervention Research Center, and the Department of Veterans Affairs Mental Illness Research, Education, and Clinical Center. Dr. Kilbourne is funded by a Career Development Award from the VA Health Services research and development program.

Dr. Kilbourne is affiliated with the Center for Health Equity Research and Promotion at the Department of Veterans Affairs (VA) Pittsburgh Healthcare System and with the departments of medicine and psychiatry at the University of Pittsburgh. Dr. Haas is with the department of psychiatry at the University of Pittsburgh and with the Mental Illness Research, Education, and Clinical Center of the VA Pittsburgh Healthcare System. Dr. Mulsant is with the department of psychiatry of the University of Pittsburgh and with the Geriatric Research, Education, and Clinical Center of the VA Pittsburgh Healthcare System. Dr. Bauer is with the VA Providence Medical Center in Providence, Rhode Island, and with the department of psychiatry and human behavior at Brown University in Providence. Dr. Pincus is with the RAND-University of Pittsburgh Health Institute in Pittsburgh. Send correspondence to Dr. Kilbourne at VA Pittsburgh Center for Health Equity Research and Promotion, Mailstop 151-C, University Drive C, Pittsburgh, Pennsylvania 15240 (e-mail, [email protected]).