| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
Correspondence: Address correspondence to Ann L. Gruber-Baldini, PhD, Division of Gerontology, Department of Epidemiology & Preventive Medicine, University of Maryland School of Medicine, Baltimore, MD, 21201. E-mail: abaldin{at}epi.umaryland.edu
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Depression • Assisted living • Residential care • Nursing homes • Dementia
The one large multistate study of depression in RC/AL facilities (Watson, Garrett, Sloane, Gruber-Baldini, & Zimmerman, 2003) found significant depressive symptomatology among 13% of RC/AL residents, with only 18% of these patients taking antidepressant medication. Depressive symptoms were more than twice as common among RC/AL residents with mild or moderate dementia than among those without dementia, and depressed residents were at greater risk of discharge to nursing home and death.
To our knowledge, few studies in long-term care have compared depression across RC/AL and nursing home settings, specifically for patients with dementia, and none have examined the association between depression and staff perception of depression, treatment, or adequacy of care in these settings. Given the high prevalence of dementia and depression in long-term care, and concern regarding underdetection and undertreatment, predictors of recognition and treatment are important to examine. In this article we examine the prevalence of depression in residents with dementia in long-term care, and we describe staff and facility characteristics associated with depression and the characteristics of residents with dementia who are depressed.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Measures
Depression
A supervisory staff member rated participants for depression by using items on the Cornell Scale for Depression in Dementia (CSDD; Alexoupoulos, Abrams, Young, & Shamoian, 1988). This modified administration format of the CSDD includes 19 items asking about symptoms over the prior 7 days, each rated on a 3-point scale, with higher scores indicating greater depression (range = 0–38). Scores of 7 or greater indicate symptomatology consistent with clinically significant depression (Alexoupoulos et al.; Vida, Des Rosiers, Carrier, & Gauthier, 1994).
Care Provision
For each resident, supervisory staff reported whether and how depression was formally assessed (by a mental health professional or by use of a written, standardized instrument) and treated (professional and other nonpharmacological); if detected and treated, how successful treatment was considered to be; and whether depression was perceived to be currently present (see Tables 1 and 2 for coding details). The research staff recorded antidepressant medication use (all regular prescription and nonprescription medications administered at least 4 of the past 7 days) verbatim from the residents' medication administration records and coded the information by using American Hospital Formulary Service system criteria (McEvoy, 2001). The administrator reported depression training as the percentage of supervisory and direct care staff who received formal training in depression assessment and treatment in the past year. The staff rated perception of training regarding how well trained they felt in depression assessment and treatment. The administrator for the facility answered queries concerning the use of mental health professionals in formal care planning and perceived facility ability to treat depression.
|
|
Analyses
Descriptive and regression results present bivariate analyses predicting differences in setting (RC/AL vs nursing home) and current levels of depression, using a CSDD cutpoint of 
7. Regression models we fit using generalized estimating equations (Liang & Zeger, 1986) assuming a binomial distribution and logit linking function for the dichotomous outcome of depression and other dichotomous outcomes, and a Gaussian distribution for analyses of setting (RC/AL vs nursing home) on continuous variables. An exchangeable correlation structure was specified to account for clustering within facilities. Odds ratios (ORs) are presented with 95% confidence intervals (CIs). Multivariate regressions (controlling for age, gender, race, cognition, comorbidity, and functional status) examine predictors of current depression; interactions of predictors with setting were also tested. Values of facility predictors were ascribed to all residents within that facility.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Depression and Care by Setting
Table 1 displays the depression prevalence for the 238 RC/AL residents and 109 nursing home residents and relevant components of care. Prevalence was not significantly different between RC/AL and nursing homes; 25% of the residents in these settings (24% in RC–AL, 27% in nursing homes) had CSDD scores consistent with depression. The two most frequent items listed by residents were being anxious (48%) and being easily annoyed (48%), and more than 20% of the overall sample of individuals endorsed items about being sad, not responsive, agitated, having slow movements, and waking many times at night (results not shown). Nursing home participants were more likely to be treated by a mental health professional and to reside in facilities that include mental health professionals in formal care planning than those in RC/AL. We found no other significant differences between nursing homes and RC/AL facilities.
Predictors of Depression
Table 2 presents both the descriptive distributions by presence of depression and the results of regressions predicting current depression. Among currently depressed participants, 42% had been recognized as depressed by the staff supervisor; 54% were currently on an antidepressant medication. A substantial proportion of those not currently depressed were also on antidepressants (33%). Only 28% of participants with current depression had any formal mental health treatment.
Participants with depression were more likely to be severely or very severely cognitively impaired, display behavioral symptoms, and be in pain than those who were not depressed. In the unadjusted (but not in the adjusted) model, depressed participants were more likely to have low activity. Residents in for-profit facilities were more likely to be depressed. Treatment (professional, nonpharmacologic, and antidepressant medication) was more common among those with depressive symptoms, as was staff perception of current depression. Depression was lower for those whom staff perceived success in treatment.
Behavioral symptoms and for-profit ownership showed statistically significant interactions with facility type (RC/AL vs nursing home). For behavioral symptoms, the association with depression was much stronger (OR = 7.33, 95% CI = 3.04–17.68) in RC/AL than in nursing homes (OR = 2.78, 95% CI = 1.38–5.58; interaction p =.030). Conversely, the association of depression with for-profit status was stronger in nursing homes (OR = 9.62, 95% CI = 3.65–25.35) than in RC/AL facilities (OR = 1.22, 95% CI = 0.61–2.47; interaction p =.001).
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Depression was more common for participants in for-profit nursing homes than for those in nonprofit homes and all RC/AL facilities. For-profit status in nursing homes has been found to be associated with more deficiencies (Harrington, Zimmerman, Karon, Robinson, & Beutel, 2000) and lower nurse staffing hours (Harrington & Swan, 2003). To our knowledge, an association between profit status and care in RC/AL has not been documented, and the meaning of for-profit status may differ across facility types.
Among six key domains of care, depression was the quality-of-life domain with the lowest perceived treatment success and staff rating of training adequacy (see introduction to this issue), although it is less common than behavioral symptoms, low activity, and low food and fluid intake. Also, the percent of participants who reside in facilities with no training provided for supervisors (21%) or care providers (35%), on depression was higher than lack of training for the other quality domains examined. However, level of training or perceived ability to assess or treat was unrelated to actual depression in these analyses and was surprisingly high. Admittedly, the percentage of staff with formal training in depression care was not validated, and the content of training was not specified. It is unclear why these staff members feel very well equipped to treat and assess depression, when depression is not easily treated and staff themselves report success in only 48% of the identified cases. Other studies in nursing home settings (not limited to dementia patients) have found very low rates of training for depression (Bagley et al., 2000), and, even when training is considered mandatory, there is poor compliance and the impact of knowledge beyond 1 month is minimal (Brooks, Renvall, Bulow, & Ramsdell, 2000; Cohen-Mansfield, Werner, Culpepper, & Barkley, 1997). Furthermore, although the perceived presence of depression was related to higher prevalence, it is worth noting that over one half of the participants with depression were undetected by staff.
About 54% of depressed and 33% of nondepressed participants were taking antidepressant medications, and the use of antidepressants was more prevalent among those who were depressed than other formal mental health treatments. Dose, frequency, and indication information about the medications was not recorded, and so we cannot determine the adequacy of treatment. A large portion of the 33% (those taking medications among nondepressed individuals) may indicate treatment success, but we cannot exclude the possibility of inappropriate use of these medications.
Some caveats are in order. The CSDD is a measure of depressive symptoms, not a clinical diagnosis, and caution should be used when the results are compared with more clinical studies. However, the use of the CSDD as a screening tool for patients with dementia has been advocated in a recent practice guideline (American Geriatrics Society & American Association for Geriatric Psychiatry, 2003), and its utility in predicting nursing home discharge and death in RC/AL settings is established (Payne et al., 2002; Watson et al, 2003). Typically, the CSDD is rated by a clinician after data are gathered through observation and caregiver interview. In this study, the nursing supervisor rated the CSDD items, for a number of reasons: (a) inability of moderately or severely demented patients to respond; (b) the four-state sample of individuals that precluded sending psychiatrists to all facilities; (c) lack of family proxies for many of residents; (d) potential differences in staffing (and availability of staff) across assisted living and nursing home facilities; and (e) 87% of supervisors reported high familiarity with the resident. Using a nursing supervisor established a consistent, clinically oriented respondent across the sites, but it would have been ideal to obtain depression information from multiple respondents. Thus, the CSDD rating likely provides an underestimate of the true prevalence of depression symptoms. Further, because CSDD ratings were provided by the same staff member who rated current detection and treatment, the degree of correspondence between these measures may be higher as a result of similar biases in underrecognition of depression and depression symptoms. Finally, the sample of individuals is not representative and so generalizability is limited.
Despite these limitations, this study suggests that undetected depression among residents with dementia is high in both long-term-care settings, particularly in for-profit nursing homes. The role of improved depression training and involvement of mental health professionals in long-term care should be further investigated.
|
Footnotes |
|---|
1 Department of Epidemiology and Preventive Medicine, University of Maryland School of Medicine, Baltimore.
2 Cecil G. Sheps Center for Health Services Research, University of North Carolina, Chapel Hill.
3 School of Social Work, University of North Carolina, Chapel Hill.
4 Regenstrief Institute and Indiana University Center for Aging Research, Indianapolis.
5 Department of Psychiatry, University of North Carolina, Chapel Hill.
6 Department of Epidemiology, University of North Carolina, Chapel Hill.
7 Alzheimer's Association, National Office, Chicago, IL.
Decision Editor: Richard Schulz, PhD
Received for publication June 25, 2004. Accepted for publication February 1, 2005.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
