| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to W. P. Achterberg, Cascade Zorgcentrum Rosendael, Indusdreef 5, 3564 GV Utrecht, The Netherlands. E-mail: wp.achterberg.gpnh{at}med.vu.nl
 |
Abstract |
|---|
 TOP Abstract Design and MethodsResultsDiscussionReferences |
|---|
Key Words: Psychosocial functioning • Mood disorders • Minimum Data Set
Admission to a nursing home involves adapting to other people and other activities, creating new social relationships, and finding resources for support while handicapped by various impairments. Well-being and satisfactory social functioning are not easy to establish in this context. Being successful in social engagement can be regarded as a critical component of quality of life for nursing home residents (Mor et al., 1995). It means that the resident has a high sense of initiative and involvement and can respond adequately to social stimuli in the social environment—participate in social activities and interact with other residents and staff. Previous research has associated low social engagement with increased mortality (Bennett, 2002; Kiely, Simon, Jones, & Morris, 2000) and cognitive decline (Bassuk, Glass, & Berkman, 1999).
In the United States, positive aspects of social functioning were included in the Resident Assessment Instrument (RAI) that was mandated by Congress (Morris, Hawes, & Fries, 1990). In constructing the Minimum Data Set (MDS) for the RAI, considerable attention has been paid to well-being and social functioning. The MDS items concerning social functioning focus on the engagement of residents in the social environment around them and aim to measure their sense of initiative and involvement. The Index of Social Engagement, which is constructed from the 6 MDS items on social engagement, reflects both social involvement and autonomy (Mor et al., 1995). The effect of several resident characteristics on social engagement have been studied; cognitive and ADL impairment were found to be related to low social engagement (Frijters et al., 2001; Mor et al., 1995; Schroll, Jónsson, Mor, Berg, & Sherwood, 1997), as were sensory and communication losses (Resnick, Fries, & Verbrugge, 1997). The effect of depression on social engagement is unknown. Several studies have found that (minor or major) depression has a much higher prevalence in older inpatient populations than in the community (Abrams, Teresi, & Butin, 1992; Falck, Pot, Braam, Hanewald, & Ribbe, 1999; Parmelee, Katz, & Lawton, 1992; Rovner, German, Brant, Burton, & Folstein 1991). Furthermore, it is frequently not recognized and/or treated. Depressed residents are likely to have more difficulty in engaging themselves in the new environment, as residents who are not socially engaged are more likely to be depressed. Successful treatment of depression in addition to or instead of offering social activities may be the key to successful social engagement.
Previous research has been based on cross-sectional data, with a mixture of residents who had been in the nursing home for a variable period of time (Frijters et al., 2001; Mor et al., 1995; Resnick et al., 1997; Schroll et al., 1997). There is little known about the predictors, course, and prevention of low social engagement. It is important to study the concept of low social engagement in newly admitted residents because this can facilitate the development of preventive strategies.
The aim of this study is to explore the effects of depression on social engagement in newly admitted Dutch nursing home residents.
|
Design and Methods |
|---|
|
TOPAbstractDesign and MethodsResultsDiscussionReferences |
|---|
In total, the physicians assessed 562 residents, 64.6% of whom were female. The mean age was 78.5 years (for females: 79.8, for males: 76.2), which is a representative sample compared with the national average. A relative large number of residents were admitted to psychogeriatric wards: 247 (44%), compared with the national average of 33.6% (Arcares, 1999; SIG Zorginformatie, 1998; see Table 1). Two hundred fifty four residents (45.2%) were admitted for long-term care. Two hundred twenty-six (40.2%) were admitted for rehabilitation (117 of them [20.8%] in a ward for specialized rehabilitation care), 81 (14.4%) were admitted for other reasons (such as terminal care, screening/observation or crisis intervention), and 1 (.2%) was not registered. As there are no national data available on the aims and type of ward described, it is not possible to make comparisons with our data. Because of missing values in independent variables, 543 residents were included in the multivariate logistic model.
|
Dependent Variable
We measured social engagement with the Index of Social Engagement (ISE), which is constructed from the following six (dichotomous) MDS items: (a) at ease interacting with others, (b) at ease doing planned or structured activities, (c) at ease doing self-initiated activities, (d) establishes own goals, (e) pursues involvement in life of facility, and (f) accepts invitations into most group activities. The ISE ranges from 0 (lowest level of social engagement) to 6 (highest level of social engagement). The items have shown moderately good inter-rater reliability (Sgadari et al., 1997). In the present sample, the six items demonstrated reasonable internal consistency (Cronbach's 
=.72).
The ISE measures a single construct that is correlated with actual participation in the activities in the nursing home and is distinct from measures for mood, behavioral problems, and conflicts in relationships (Mor et al., 1995).
The ISE has been dichotomized in two different ways: 0–2 versus 3–6 (sensitive for low social engagement; Resnick et al., 1997) and 0–4 versus 5–6 (sensitive for high social engagement; Schroll et al., 1997). Because this study is looking for factors associated with low social engagement, the 0–2 (low) versus 3–6 (not low) dichotomy is used.
Independent Variable
The MDS Depression Rating Scale (DRS) is a seven-item scale, with all items to be scored 0 (indicator not exhibited), 1 (indicator of this type exhibited at least once in the last 30 days and up to 5 days a week), or 2 (indicator exhibited daily or almost daily; Burrows, Morris, Simon, Hirdes, & Phillips, 2000). The scores range between 0 and 14. The mood items in the MDS 2.0 have good inter-rater reliability (Morris et al., 1997). In the present sample, the seven items demonstrated good internal consistency (Cronbach's =.87).
With a cut point of 3, it differentiates well between residents with few or many depressive symptoms. Compared with psychiatric criteria for depression (American Psychiatric Association, 1994) it has a high sensitivity (91%) and a lower specificity (69%; Burrows et al., 2000). Therefore, this scale has been used in this study to distinguish between residents with or without (easily observed) depressive symptoms. Thus, where the word depression is used, it means a relatively high level of depressive symptoms.
Control Variables
Cognitive function was measured according to the MDS-Cognitive Performance Scale (CPS), which is based on five MDS items. The CPS is a seven-category index, ranging from cognitively intact to very severely impaired (Morris et al., 1994). It has shown substantial agreement with the Mini-Mental State Examination in the identification of cognitive impairment in research settings (Hartmaier et al., 1995). The index is dichotomized by combining the three severe categories as low cognitive performance and the four other categories as high cognitive performance (Mor et al., 1995).
The activity of daily living (ADL) classification is based on six MDS items on self-performance of ADLs, each consisting of five categories, ranging from independent to totally dependent, and one item concerning urinary incontinence.
The seven-category (hierarchical) ADL index ranges from minor oversight to highly dependent. The ADL index is dichotomized, with the four highest scores classified as dependent, and the three lower scores classified as relatively independent (Mor et al., 1995).
Hearing and vision were measured according to two four-level ordinal MDS items that assess the ability to see or hear with environmental adjustments (such as glasses or hearing aids). Adequate and minimal impairment is dichotomized as no problems, and moderate impairment and severe impairment are dichotomized as problems with hearing or seeing.
The demographic variables are sex, ward type (somatic, psychogeriatric, or rehabilitation), and age (in four categories: <65, 65–74, 75–84, 
85).
Analysis
First, we applied univariate logistic regression analysis (SPSS, Version 10.1) to identify variables related to low social engagement. Subsequently, we studied correlations between dependent and independent variables by using chi square statistics. To determine the true effect of depression on social engagement, we entered possible confounders one by one in a multivariate logistic regression model. In the final equation, we calculated the 95% confidence intervals.
|
Results |
|---|
|
TOPAbstractDesign and MethodsResultsDiscussionReferences |
|---|
Cognition had a strong relation with low social engagement (p <.001) and a weaker relation with depression (p =. 045; see Table 2). Problems with cognition, hearing and seeing, and ward type were all associated with depression and low social engagement and could, therefore, be potential confounders (correlation with both dependent and independent variable) in the relation between depression and social engagement (see Table 2).
|
In univariate regression, cognition (odds ratio [OR] = 4.2), vision (OR = 2.2), hearing (OR = 1.5), and ADL (OR = 1.9) were significantly related to low social engagement; the demographic variables of sex and age were not.
Depression was highly associated with low social engagement (OR = 3.5, 95% CI: 2.3–5.3). When we controlled for possible confounders in the multivariate logistic model, it made little difference to the strength of this association (OR = 3.3, 95% CI: 2.1–5.1; see Table 3). In this model, low cognitive performance was predictive (OR = 3.5, p <.001), as was vision (OR = 1.7, p =.011; Table 3). The presented multivariate logistic model creates a predictive model with 66.4% overall correctly predicted (.5 = cut-off value).
|
|
Discussion |
|---|
|
TOPAbstractDesign and MethodsResultsDiscussionReferences |
|---|
The effect of depression on social engagement is found to be strong, and there are no confounding mechanisms. With our cross-sectional sample, we can not establish a causal relationship, but it is theoretically plausible that depressive symptoms (such as anxiety, withdrawal, and loss of interest) can act as obstacles in the receptiveness of a resident in responding to social stimuli. The relationship of depression and cognitive decline with social engagement is found to be stronger than the relationship between sensory or ADL impairment and social engagement. The strong univariate correlation between ward types and social engagement does not hold in the multivariate model, which means the differences between the wards are explained by different levels of depression and functional (mainly cognitive) impairments.
One limitation of this study might be that the participating residents were recruited in teaching nursing homes. Some physicians/homes included more patients than others, and there was an overrepresentation of psychogeriatric residents. The model controlled for cognitive status, so this is not likely to effect the validity of the results. Moreover, age and sex distributions are representative of all newly admitted nursing home residents in the Netherlands (Arcares, 1999; SIG Zorginformatie, 1998). It is not known whether our sample is representative for the percentage of residents on specialized rehabilitation wards, but in our multivariate logistic model we controlled for ward type. Another limitation is that this is a cross-sectional sample of newly admitted residents. The depression scores might, therefore, be higher and the social engagement scores lower because of the impact of institutionalization and may subsequently change over time. Longitudinal data are needed to study the longitudinal relationship between depression and social engagement.
One of the strengths of this study is that the data collection is solely intended for research purposes and is performed by trained physicians. In such a setting, there is little reason to dispute the reliability of the MDS data (Hawes, Phillips, Mor, Fries, & Morris, 1992; Ouslander, 1994; Teresi & Holmes, 1992). Moreover, the sample is relatively large (1.3% of all Dutch nursing home admissions) and is collected from a wide variety of nursing homes throughout the country (18% of all Dutch nursing homes). There are no data on the reliability of the MDS when physicians fill it out. Dutch nursing home physicians are responsible for the multidisciplinary care planning of all residents and are therefore very well informed about their functional and social abilities. We can not rule out the possibility that the social engagement items would be interpreted in a different way when they are conducted solely by a nurse, but we have no reason to believe this has created any bias.
Nursing homes should provide care and a customized living environment for frail residents. This should include, in addition to adequate treatment and personal care, facilitating psychosocial well-being and social interaction. Thorough assessment, immediately on admission, of these psychosocial needs is imperative. Depression should be recognized and treated. The MDS-depression algorithm (which contains 5 more items compared with the 12 items of the DRS and is triggered when at least one of them is positive) seems to provide too little specific information: In our sample, 68% of the residents have at least one DRS item scored positive. Such a sensitive trigger in the busy nursing home practice is in time likely to be ignored. Recognition and adequate staff reaction could be improved by using a more specific measure, like the DRS cut point of 3.
The results of this study also suggest that more attention should be paid to activities that are appropriate for residents with cognitive decline and (many) depressive symptoms. This will be a challenge, as depressed residents are not very inclined to respond to invitations to take part in (group) activities. For the well-being of residents with low social engagement and many depressive symptoms (in this sample 19.6% of all newly admitted!), it is essential that this challenge is taken on.
|
Footnotes |
|---|
1 Cascade Zorgcentrum Rosendael, Utrecht, The Netherlands.
2 Department of Nursing Home Medicine & EMGO-Institute, Amsterdam, The Netherlands.
3 Department of Nursing and Caring Research, Netherlands Institute of Primary Health Care, Utrecht.
Laurence G.Branch, PhD, Decision Editor
Received for publication June 7, 2002. Accepted for publication August 8, 2002.
|
References |
|---|
|
TOPAbstractDesign and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. L. Piven, R. A. Anderson, C. S. Colon-Emeric, and M. Sandelowski Certified Nursing Assistants' Explanatory Models of Nursing Home Resident Depression West J Nurs Res, October 1, 2008; 30(6): 653 - 672. [Abstract] [PDF]
|
|
|
|
 |
|
 A. Won, K. L. Lapane, S. Vallow, J. Schein, J. N. Morris, and L. A. Lipsitz Long-term effects of analgesics in a population of elderly nursing home residents with persistent nonmalignant pain. J. Gerontol. A Biol. Sci. Med. Sci., February 1, 2006; 61(2): 165 - 169. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Kolanowski, L. Buettner, M. Litaker, and F. Yu Factors that relate to activity engagement in nursing home residents American Journal of Alzheimer's Disease and Other Dementias, January 1, 2006; 21(1): 15 - 22. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
