General-Medical Conditions in Older Patients With Serious Mental Illness
Objective: The burden of medical comorbidities was compared between older (≥60 years) and younger patients with serious mental illness. Methods: Patients (N = 8,083) diagnosed with schizophrenia, schizoaffective disorder, or bipolar disorder in 2001 were identified from VA facilities in the mid-Atlantic region. Medical comorbidities were identified by an ICD-9-based clinical classification algorithm. Results: Older, versus younger, patients were more likely to be diagnosed with cardiovascular or pulmonary conditions, and less likely to be diagnosed with substance-use disorders or hepatic conditions. Conclusions: More aggressive detection and management of general-medical comorbidities in older patients with serious mental illness is paramount. (Am J Geriatr Psychiatry 2005; 13:250-254)
It is well known that serious mental illnesses contribute to significant personal and societal costs.1 Much of the costs may be attributed to general medical comorbidity.2,3 Although individual studies have documented the burden of medical comorbidity among individuals with a particular psychiatric diagnosis, including schizophrenia4 and bipolar disorder,5 such studies were often limited to younger patient populations or limited in sample size and generalizability (e.g., based on convenience samples of patients from a few clinics). Because general-medical conditions may affect the treatment course and prognosis of older patients with serious mental illness, it is important to ascertain the prevalence of general-medical conditions among older as well as younger patients with serious mental illness in order to customize treatment strategies. We assessed the prevalence of medical comorbidity in a large sample of patients diagnosed with serious mental illness and compared the prevalence of general-medical comorbidities between older (defined as ≥60 years old) and younger patients diagnosed with serious mental illness. We hypothesized that older patients would be more likely to be diagnosed with general-medical conditions than patients under age 60.
METHODS
Data Sources
We conducted a retrospective analysis of data from fiscal year (FY) 2001 (October 1, 2000 through September 30, 2001) from the National Patient Care Database (NPCD) of the Veterans Health Administration (VHA). The NPCD includes all inpatient and outpatient encounters to the VHA, including primary care and specialty mental health outpatient and inpatient visits. As many as 10 ICD-9 (International Classification of Disease, 9th Revision) codes for each inpatient encounter and up to 15 ICD-9 codes for each outpatient encounter can be recorded. Service providers determine ICD-9 diagnoses, and professional coders review and assign ICD-9 codes as part of the administrative data routinely collected by the VHA healthcare facilities. We ascertained currently active psychiatric and general medical diagnoses recorded in FY 2001. These diagnoses are based on problem lists, which reflect medical problems that are currently active, regardless of when they were initially recognized. We used a previously established algorithm to identify psychiatric and general medical comorbidity diagnoses from ICD-9 codes based on whether the patient had an ICD-9 code for the given condition from either one inpatient or two separate outpatient visits.6 Two separate outpatient visits are required to confirm a diagnosis so as to account for “rule-out” diagnoses.6
Patient Sample
We included all patients with a diagnosis of any one of three serious mental illnesses (defined as schizophrenia, schizoaffective disorder, or bipolar disorder) in FY 2001. Patients who received care in any facility within the VHA Stars and Stripes Integrated Services Network ([VISN 4]; including Pennsylvania, Delaware, and areas of West Virginia, Ohio, New Jersey, and New York) were included. The vast majority of veterans with mental disorders receive their care through the VHA, given the generous benefits offered by the VHA, and because most of these patients are disenfranchised, often precluding them from attaining private health insurance.
Patients were considered to have a diagnosis of a serious mental illness if they had an ICD-9 code for schizophrenia (ICD-9 codes: 2950-29565 or 2958-29595, inclusive), schizoaffective disorder (2957- 29575), and bipolar disorder (2960-29616, 2964-29689, or 30113) recorded from either one inpatient or two outpatient encounters. Although the use of one inpatient or two outpatient codes to identify patients with mental disorder maximizes specificity while retaining sensitivity, we are also aware that many patients may receive overlapping psychiatric diagnoses. Hence, we used the following decision hierarchy (in order of inclusion) to identify unique patient groups: schizophrenia, schizoaffective disorder, and bipolar disorder. For example, if patients were diagnosed with schizophrenia and were also diagnosed with one of the other mental disorders, then they were considered to have a diagnosis of schizophrenia. This classification method is based on the assumption, that schizophrenia is considered to be the most debilitating disorder.1
Ascertainment of Variables
For each patient, we identified general-medical and substance- use comorbidities diagnosed in FY 2001 based on ICD-9 code algorithms developed by the Agency for Healthcare Research and Quality (AHRQ) Clinical Classifications Software and modified by the VHA Health Economics Resource Center.7 The most frequently occurring general-medical comorbidities were included and grouped into the following categories: cardiovascular (hypertension, congestive heart failure, peripheral vascular disease, stroke/ transient ischemic attack, and ischemic heart disease), endocrine (diabetes, hyperlipidemia, thyroid disorders, and obesity), hepatitis, musculoskeletal (lower back pain, hip fracture, osteoarthritis), pulmonary (chronic obstructive pulmonary disease, asthma, and pneumonia), cancers, accidents or injuries, and substance use disorders (alcohol or drug use disorders). Additional socio- demographic and utilization data were collected from the NPCD and included age, gender, race/ethnicity, age, and marital status. This study was reviewed and approved by local institutional review boards. Subjects were not required to obtain informed consent, given that this was a secondary analysis of existing data.
Statistical Analysis
Descriptive statistics were applied to estimate the percentage of patients with each general-medical condition and with any condition in one of the aforementioned comorbidity categories (cardiovascular, endocrine, etc.). Bivariate analyses (chi-square tests) were used to compare the prevalence of each general-medical condition having at least one condition for each comorbidity category by age-group. We also ascertained a total count of general-medical comorbidities by summing up the individual conditions and compared the proportion of patients having 0, 1, 2, or 3-or-more conditions by age-group with the Jonckheere-Terpstra test. For bivariate analyses, age was categorized as ≥60 years and <60 years because in the VHA, patients ≥60 years are often seen in separate geriatric clinics. Two-tailed tests were used to determine significance, with α set at 0.05, and the Hochberg method was used to adjust for multiple comparisons.8 The Hochberg method adjusts the p values by means of the following formula: α/(number of total comparisons - rank order of original p value by decreasing value + 1). The advantage of the Hochberg method is that it provides more protection from Type II errors than the Bonferroni correction, which is a more conservative method for determining statistical significance.8 To assess whether patient demographics and number of primary care contacts might have influenced the probability of diagnosis (i.e., patients receiving more frequent health services are more likely to have medical conditions detected), we ran multivariable logistic- regression analyses for each comorbidity category, adjusting for patient factors and number of primary care visits within the past year.
TABLE 1. Prevalence of General-Medical Comorbidity by Age-Groups in Patients With Serious Mental Illness (N = 8,083)
RESULTS
A total of 8,083 patients within the Stars and Stripes Veterans Integrated Service Network received a diagnosis of one of the three mental disorders in FY 2001. Of the 8,083 patients, the mean age was 54 years (standard deviation [SD]: 12; range:19-100 years); 27.5% (N = 2,224) were ≥60 years old; 6.9% (N = 554) were women; and 21.8% (N = 1,763) were African American. These demographic characteristics are comparable to those of VHA patients across the United States.5 Overall, 61% (N = 4,932) were diagnosed with schizophrenia; 9% (N = 705) were diagnosed with schizoaffective disorder; and 30% (N = 2,446) were diagnosed with bipolar disorder.
The most common comorbid conditions among our population included cardiovascular conditions (30.7%), endocrine conditions (29.0%), and substance use disorders (21.2%; Table 1). Overall, older, versus younger, patients were more likely to be diagnosed with general medical comorbidities, including cardiovascular, endocrine, and pulmonary conditions, and cancers (Table 1). In contrast, older patients were less likely to be di\agnosed with a substance use disorder, accidents/injuries, or hepatic conditions than their younger counterparts (Table 1). Older, versus younger, patients were also more likely to be diagnosed with three or more conditions, and similar results were obtained when substance use disorder was included in the total comorbidity count (Table 1). Also, multivariable analyses in which we controlled for patient factors and number of primary care contacts suggests that these potential confounders did not explain any of the differences in comorbidity diagnoses between older and younger patients (data available from the author).
DISCUSSION
Our study found a substantial burden of medical comorbidity in patients diagnosed with serious mental illness. Moreover, older patients with serious mental illness, compared with their younger counterparts, were more likely to be diagnosed with cardiovascular, endocrine, or pulmonary conditions and were more likely to experience a greater burden of medical comorbidity overall. Younger, compared with older patients with serious mental illness, were more likely to be diagnosed with hepatic or substance use disorders and accidents/injuries. These findings remained robust even when adjusting for patient factors and number of primary-care contacts.
To date, no studies have comprehensively assessed the burden of medical comorbidity in a large sample of patients with serious mental illness. Underlying causes of the increased burden of general- medical conditions in older compared with younger patients with serious mental illnesses include the aging process as well as extrapyramidal- and endocrine-related side effects from long-term psychopharmacologic medication use. More recently, it has been suggested that the recurrent mood and psychotic episodes and recurrent stress associated with serious mental illness may directly contribute to the increased burden of general-medical comorbidities, particularly cardiovascular and endocrine conditions, because of “allostatic load.”9 Allostatic load is the “wear-and-tear” on the body and brain resulting from chronic overactivity of physiological systems involved in adaptation to environmental challenges that are triggered by the mental illness itself. Chronic overactivity of physiological systems in turn may speed up the aging process and lead to increased probability of cardiovascular and endocrine disease. If untreated, these general-medical conditions can potentially lead to functional decline and premature mortality among older patients with serious mental illness.
At the same time, there are some comorbidities that may disproportionately occur in younger versus older populations with serious mental illness. Our results suggest that younger compared with older patients are more likely to be diagnosed with substanceuse disorders, hepatic conditions, and accidents/ injuries, all of which could eventually lead to adverse outcomes over time (e.g., permanent disability, liver failure). Nonetheless, compared with their younger counterparts, older patients tend to use less alcohol and illicit drugs and are less likely to be engaging in risky behavior that would lead to hepatic conditions (e.g., from hepatitis C or cirrhosis of the liver).7 Although care for comorbid conditions in patients with serious mental illness has often focused on the treatment of coexisting substance-use disorders,10 few treatment strategies have been developed to manage coexisting medical conditions in these patients. Therefore, the observed increased prevalence of general-medical as opposed to substance-use comorbidities in older compared with younger patients suggests the need for strategies that focus on improving access to and quality of general-medical services for older patients as well as coordinating general-medical and psychiatric care for this group.
Despite our comprehensive assessment of comorbid conditions in a large sample of patients with serious mental illness, there are limitations to this study that warrant consideration. Foremost is the fact that our data were derived from administrative datasets, and, hence, diagnoses were not confirmed by more formalized diagnostic procedures. The prevalence of medical comorbidity may have been underestimated because we relied on secondary diagnostic data rather than a comprehensive clinical exam. Also, we were not able to compare comorbidity prevalences with a non-psychiatrically ill control group. This study’s findings may only apply to the VHA patient population, and the low percentage of women in our sample may have limited the generalizability of this study, as well. Nonetheless, the VHA’s disproportionately older patient population is often considered to be representative of what the general population in the United States will look like 10-20 years from now.
Despite these limitations, our study has important clinical implications. We found a substantial burden of general-medical comorbidity in patients diagnosed with serious mental illness, despite the use of a more conservative method for identifying medical comorbidity diagnoses (e.g., requiring two separate outpatient diagnoses or one inpatient diagnosis). Older patients with serious mental illness are especially burdened by general- medical comorbidity, as opposed to substance-use disorders. Assessing whether medical conditions are being adequately detected and treated in older patients with serious mental illness, and whether treatment models that integrate general-medical and psychiatric care could improve the quality of care for these patients, should be considered in future research endeavors.
References
1. Murray CJ, Lopez AD: Evidence-based health policy: lessons from the Global Burden of Disease Study. Science 1996; 274:740-743
2. Sokal J, Messias E, Dickerson FB, et al: Comorbidity of medical illnesses among adults with serious mental illness who are receiving community psychiatric services. J Nerv Ment Dis 2004; 192:421-427
3. Jeste DV, Gladsjo JA, Lindamer LA, et al: Medical comorbidity in schizophrenia. Schizophr Bull 1996; 22:413-430
4. Simon GE, Unutzer J: Healthcare utilization and costs among patients treated for bipolar disorder in an insured population. Psychiatr Serv 1999; 50:1303-1308
5. Peele PB, Xu Y, Kupfer DJ: Insurance expenditures on bipolar disorder: clinical and parity implications. Am J Psychiatry 2003; 160:1286-1290
6. Lurie N, Popkin M, Dysken M, et al: Accuracy of diagnoses of schizophrenia in Medicaid claims. Hosp Community Psychiatry 1992; 43:69-71
7. Yu W, Ravelo A, Wagner TH, et al: Prevalence and costs of chronic conditions in the VA healthcare system: Med Care Res Rev 2003; 60:1468-1678
8. Hochberg Y: A sharper Bonferroni procedure for multiple tests of significance. Biometrika 1988; 75:800-802
9. McEwen BS: Sex, stress, and the hippocampus: allostasis, allostatic load, and the aging process. Neurobiol Aging 2002; 23: 921-939
10. Drake RE, Mueser KT, Brunette MV, et al: A review of treatments for people with severe mental illnesses and co-occurring substance use disorders. Psychiatr Rehabil J 2004; 27:360-374
Amy M. Kilbourne, Ph.D., M.P.H.
Jack R. Cornelius, M.D., M.P.H.
Xiaoyan Han, M.S.
Gretchen L. Haas, Ph.D.
Ihsan Salloum, M.D., M.P.H.
Joseph Conigliaro, M.D., M.P.H.
Harold A. Pincus, M.D.
Received October 7, 2004; revised November 11, December 7, 2004; accepted December 21, 2004. From the VA Pittsburgh Healthcare System, Pittsburgh, PA (AMK,JRC,XH,GLH,IS,JC), the Dept. of Medicine, Univ. of Pittsburgh, Pittsburgh, PA (AMK,JC), the Dept. of Psychiatry, Univ. of Pittsburgh, Pittsburgh, PA (AMK,JRC,GLH,IS,HAP), and RAND; Univ. of Pittsburgh Health Institute, Pittsburgh, PA (HAP). Send correspondence and reprint requests to Amy M. Kilbourne, Ph.D., M.P.H., VA Pittsburgh Center for Health Equity Research and Promotion (151-C), University Drive C, Pittsburgh, PA 15240. e-mail: Amy.Kilbourne@med.va.gov
2005 American Association for Geriatric Psychiatry
Copyright American Psychiatric Press, Inc. Mar 2005