Trends in Mental Health Admissions to Nursing Homes, 1999–2005 Catherine Anne Fullerton, M.D., M.P.H. Thomas G. McGuire, Ph.D. Zhanlian Feng, Ph.D. Vincent Mor, Ph.D. David C. Grabowski, Ph.D.
Objective: The study examined 1999–2005 data on first-time nursing home admissions of individuals with mental illness, dementia, or both to identify trends and characteristics. Methods: The Minimum Data Set was used to estimate the number and percentage of persons newly admitted to nursing homes who had mental illness (schizophrenia, bipolar disorder, depression, or an anxiety disorder), dementia, or both from 1999 to 2005. Data from 2005 were used to compare demographic characteristics and comorbid conditions of the three groups and treatments received. Results: The number of individuals admitted with mental illness increased from 168,721 in 1999 to 187,478 in 2005. The 2005 number is more than 50% higher than the number admitted with dementia only (118,290 in 2005). The increase was driven by growth in admissions of persons with depression—from 128,566 to 154,262 in 2005. Persons admitted with depression had higher rates of comorbid conditions than those admitted with dementia or with neither dementia nor mental illness. They also had high rates of antidepressant treatment and high rates of receipt of training in skills required to return to the community. Conclusions: Current trends show that the proportion of nursing home admissions with mental illness, in particular depression, has overtaken the proportion with dementia. These changes may be related to increased recognition of depression, availability of alternatives to nursing homes for persons with dementia, and increased specialization among nursing homes in the care of postacute, rehabilitation residents. In light of these trends, it is critical to ensure that nursing homes have resources to adequately treat residents with mental illness to facilitate community reintegration. (Psychiatric Services 60:965–971, 2009)
M
ental illness is common among nursing home residents. The high prevalence of dementia among residents is well known, but major mental illness other than dementia is also common. Indeed, an estimated 560,000 nursing
home residents in 2002 had a mental illness other than dementia—a number that dwarfs the 51,000 (1) individuals in beds at psychiatric hospitals in that year. Mental illness is one factor—and sometimes the decisive factor—contributing to nursing home
Dr. Fullerton, Dr. McGuire, and Dr. Grabowski are affiliated with the Department of Health Care Policy, Harvard Medical School, 180 Longwood Ave., Boston, MA 02115 (email:
[email protected]). Dr. Fullerton is also with the Department of Psychiatry, Cambridge Health Alliance, Cambridge, Massachusetts. Dr. Feng is with the Center for Gerontology and Health Care Research and Dr. Mor is with the Department of Community Health, both at Brown University, Providence, Rhode Island.
PSYCHIATRIC SERVICES
' ps.psychiatryonline.org ' July 2009 Vol. 60 No. 7
placement (2). Addressing the needs of nursing home residents with behavioral symptoms associated with dementia, such as decreasing the rate of use of physical and chemical restraints, has been a principal focus of research in nursing homes (3,4). However, less is known about individuals in nursing homes who have major mental illness. Estimates of the rate of significant depressive symptoms among nursing home residents range from 10% to 44% (5–7). Estimates of the prevalence of schizophrenia in nursing homes range from 4% to 13% (8,9). Among persons with mental illness, a diagnosis of schizophrenia or bipolar disorder was found to be associated with a greater likelihood of admission to a nursing home over a three-year period (10), and depression was shown to increase the risk of nursing home admission of elderly persons both from the community and after an acute hospitalization (11–13). Depression and behavioral symptoms among nursing home residents are associated with high rates of functional impairment, disability, poor health outcomes, increased rates of hospitalization and mortality, and greater emergency service use (14–18). High rates of comorbid general medical conditions, deficits in activities of daily living and instrumental activities of daily living, cognitive impairment, and lack of available family members predict admission to nursing homes among older adults with mental illness (19,20). The nursing home sector has undergone a remarkable transformation over the past two decades with re965
spect to payers and patient mix. Until the early 1980s, nursing homes largely provided custodial care to long-stay residents. The short-stay, postacute side of the nursing home market was negligible. Medicare, the primary payer for postacute services, accounted for only 1.6% of total nursing home expenditures in 1980 (21). A series of policy changes, however, expanded the postacute side of the market. By 2006 Medicare accounted for 15.7% of all nursing home expenditures. Further, federal standards for Medicaid reimbursement were repealed, and social work services were bundled into the nursing home’s per diem rate. Growth in the population of short-stay nursing home residents and decreased reimbursement for social work may have led to important changes in the relative proportions of nursing home residents admitted with mental illness and dementia. Little is known about recent trends in admission of individuals with mental illness to nursing homes. This observational study used longitudinal data on the census of new nursing home admissions to describe trends in mental illness. We regard this as a first step in directing research and policy attention to the largest population of vulnerable persons with mental illness in institutionalized settings. More specifically, we examined admissions to nursing homes from 1999 to 2005 of individuals diagnosed as having mental illness, dementia, or both. We compared the demographic characteristics, comorbidities, and treatment of individuals with and without mental illness.
Methods Data We used the Centers for Medicare and Medicaid Services (CMS) national registry of nursing home resident assessments from the Minimum Data Set (MDS) to examine the prevalence of mental illness among new nursing home admissions. The MDS resident assessment instrument contains nearly 400 data elements, including cognitive functioning, physical functioning, psychosocial well-being, diagnoses, and treatment variables. In describing annual trends, we examined all first-time admissions from 1999 to 966
2005 (N=7,364,470) and short-stay and long-stay admissions from 1999 to 2004 (N=6,368,159). We used 2005 data (996,311 first-time admissions) to examine demographic characteristics, comorbid conditions, and treatments received. Institutional review board permission was obtained for this study from Brown University. Measures Mental illness diagnosis. We used the MDS full assessment form to identify a behavioral health diagnosis on admission. We grouped individuals into four overall categories: mental illness only (section I1dd, I1ee, I1ff, or I1gg indicated on admission MDS), dementia only (including Alzheimer’s and other dementia) (section I1q or I1u indicated), both mental illness and dementia, and neither. Mental illness only was further categorized into four hierarchical categories: schizophrenia, bipolar disorder, depression, and anxiety disorders (in this order). For example, an individual with a diagnosis of both schizophrenia and an anxiety disorder was categorized in the schizophrenia group. Because the MDS does not include the primary reason for admission, a mental illness diagnosis can represent either the primary reason for admission or a comorbid condition. Demographic characteristics, comorbidity, and mental health treatment. Age, sex, race-ethnicity, marital status, and educational status are documented on the MDS full assessment form. We categorized race-ethnicity as white, black, Hispanic, and other (American Indian or Alaskan Native and Asian or Pacific Islander). Marital status was classified as never married, married, widowed, and separated or divorced. Educational status was categorized as less than high school (no schooling, grade 8 or less, and grades 9 to 11), graduated from high school (a high school, technical school, or trade school, and possibly some college), graduated from college, graduate degree, or unknown (not reported). We identified the following comorbid conditions: diabetes, other endocrine, cardiovascular, musculoskeletal, neurological excluding dementia, pulmonary, sensory, and other diseases. PSYCHIATRIC SERVICES
Limitations in activities of daily living were calculated with use of a scale ranging from 0 to 28, with higher values indicating greater disability. We used CMS’s definition of cognitive impairment as having impaired shortterm memory (MDS item B2a=1) and not being independent in regard to daily decision making (MDS item B4>0). Treatment variables included both medication treatment and other treatments provided within the past seven or 14 days. Medication treatment variables included average number of medications and any receipt of antipsychotics, antidepressants, or anxiolytics-hypnotics in the past seven days. Other treatments consisted of residence in an Alzheimer’s and dementia special care unit (in the past 14 days), receipt of skills training to return to the community (in the past 14 days), and evaluation or treatment by a mental health specialist (in the past seven days). We also examined episodes of restraint (in the past seven days). Short stay versus long stay. We followed the method of Mor and colleagues (22) to categorize length of stay. Residents discharged within 90 days were categorized as short-stay residents. Those who remained in the nursing home at least 90 days were categorized as long-stay residents. Analysis We examined overall trends from 1999 to 2005 in the number of persons admitted with mental illness, dementia, or both in short-stay and long-stay populations. We calculated overall prevalence rates of schizophrenia, bipolar disorder, depression, and anxiety disorders among individuals admitted to nursing homes during this period. We describe the demographic characteristics, comorbid conditions, and treatments by mental illness category. Statistical comparisons for categorical variables were made with an overall chi square test.
Results In 2005 the number of persons with a mental illness newly admitted to a nursing home exceeded the number with dementia. Of the 996,311 persons newly admitted to U.S. nursing homes in 2005, 19% (N=187,478)
' ps.psychiatryonline.org ' July 2009 Vol. 60 No. 7
PSYCHIATRIC SERVICES
Figure 1
Proportions of persons newly admitted to nursing homes with dementia, mental illness, or both, 1999–2005, by length of stay Dementia only Mental illness only Both Overall
Percentage of admissions
20 18 16 14 12 10 8 6 4 2
Percentage of admissions
1999
2000
2001
2002
2003
2004
2003
2004
2003
2004
2005
Short-stay residents
20 18 16 14 12 10 8 6 4 2 1999
Percentage of admissions
were admitted with mental illness only, 12% (N=118,290) had dementia only, 6% (N=64,669) had both a mental illness and dementia, and 63% (N=625,874) had neither mental illness nor dementia. The relative proportions of persons newly admitted to nursing homes with mental illness and dementia, as well as the absolute number, reflect important changes over recent years (Figure 1). In 1999 the number of new admissions with dementia exceeded the number with mental illness. By 2005 the number with mental illness exceeded the number with dementia. Residents admitted with mental illness increased from 168,721 in 1999 to 187,478 in 2005, and the number admitted with dementia fell from 208,505 to 118,290. The percentage of individuals admitted with both mental illness and dementia was roughly constant. In 1999 long-stay residents accounted for 38.5% of all nursing home admissions. By 2005 this proportion was 25.3%. Trends similar to those described above can be seen among both short-stay and long-stay residents (Figure 1). As expected, the percentage of residents admitted with dementia only was higher in the long-stay population. From 1999 to 2004, the proportion of residents admitted with only dementia decreased in both the short-stay population (by 4.4 percentage points) and the longstay population (by 3.4 percentage points). In contrast, the proportion of residents admitted with only mental illness increased in both the shortstay population (by 4.7 percentage points) and the long-stay population (by 3.2 percentage points). In both cases, the trends were stronger in the short-stay population. Finally, the proportion of residents admitted with both dementia and mental illness increased in the long-stay population (by 3.8 percentage points). Between 1999 and 2005 the number of individuals admitted with depression increased 4.5 percentage points, from 11.0% (N=128,566) to 15.5% (N=154,262) (Figure 2). Over the same period, there was a .1 percentage point increase in residents admitted with bipolar disorder, from .4% (N=4,597) to .5% (N=5,299). Ad-
2000
2001
2002
Long-stay residents
30 28 26 24 22 20 18 16 14 12 10 8 1999
2000
2001
missions of residents with an anxiety disorder declined by .2 percentage points, from 2.5% (N=29,221) to 2.3% (N=22,513). The percentage of newly admitted residents with schizophrenia remained roughly constant at .5%. Thus the large increase in admissions of persons with mental illness from 1999 to 2005 was primarily due to the increase in residents with a diagnosis of depression. As shown in Table 1, individuals ad-
' ps.psychiatryonline.org ' July 2009 Vol. 60 No. 7
2002
mitted with mental illness or dementia differed from other nursing home residents (and from one another) in their demographic characteristics, comorbid conditions, and treatments received. Compared with those who had neither mental illness nor dementia, those with a mental illness were generally younger and white (except for those with schizophrenia). Those admitted with schizophrenia, an anxiety disorder, or dementia were 967
Figure 2
Precentage of admissions
Proportions of persons newly admitted to nursing homes with mental illness excluding dementia, 1999–2005 16 14 12
Depression Anxiety Schizophrenia or bipolar
10 8 6 4 2 0 1999
2000
2001
more poorly educated, and those with schizophrenia or bipolar disorder were less likely to be married. Individuals admitted with schizophrenia or bipolar disorder had lower activities of daily living scores and higher levels of cognitive impairment than those with depression, an anxiety disorder, or neither mental illness nor dementia. Individuals with anxiety, depression, or dementia were more likely to have cardiovascular, musculoskeletal, or neurological comorbidities. Individuals in each category of mental illness were more likely to have pulmonary disease than those with dementia and those without dementia or mental illness. Individuals with schizophrenia or dementia were on fewer medications than those with other mental illnesses and those without dementia or mental illness. As expected, individuals with depression had higher rates of antidepressant treatment. Individuals with schizophrenia or bipolar disorder, and to a lesser extent those with dementia, had higher rates of antipsychotic use. Those with an anxiety disorder had higher rates of anxiolytic-hypnotic use. Those with schizophrenia and dementia were less likely than those with bipolar disorder, depression, or an anxiety disorder to receive training in skills required to return to the community and more likely to be physically or chemically restrained. Finally, those with mental illness or dementia were more likely to receive evaluation or therapy by a licensed mental health professional than those without mental illness or dementia. 968
2002
2003
2004
2005
Because of the substantial increase over time in individuals admitted with depression, we examined demographic, clinical, and treatment variables for 1999 and for 2005. Age, gender, and racial-ethnic make-up of new admissions with depression were largely the same. An increase was found in the proportion of depressed individuals who were married (31.7% to 35.2%), and a decrease was noted in the proportion who were widowed (55.1% to 50.2%). Functional status was similar. With the exception of diabetes and lung disease, both of which increased, all other comorbid conditions decreased from 1999 to 2005. In 2005 individuals were taking more medications than in 1999 (12.2 compared with 10.4 medications); a greater proportion used psychotropic medications (antipsychotics, 11.0% in 2005 and 8.3% in 1999; antidepressants, 81.8% and 77.8%; and anxiolytics, 34.0% and 32.5%); and they experienced fewer episodes of restraint (1.3 and 2.4). Finally, a larger percentage of depressed residents received skills training to return to the community in 2005 (58.7% compared with 41.8% in 1999), and a smaller proportion were evaluated by a mental health professional in 2005 (5.4% compared with 10.4%). Given the large numbers involved, all overall differences were statistically significant at a p
