| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to Debra Dobbs, PhD, Cecil G. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill, 725 Martin Luther King Jr. Blvd., Chapel Hill, NC 27599-7590. E-mail: Debra_Dobbs{at}unc.edu
 |
Abstract |
|---|
 TOP Abstract Research Design and MethodsResultsDiscussionReferences |
|---|
Key Words: Nursing homes • Assisted living • Residential care
|
Research Design and Methods |
|---|
|
TOPAbstractResearch Design and MethodsResultsDiscussionReferences |
|---|
We randomly selected residents from among those aged 65 years or older who had a diagnosis of dementia. A total of 575 eligible residents were approached for enrollment. Of these, 421 (73%) agreed to participate, 66 (11%) refused, and 88 (15%) were unable to provide consent and had family members who were unreachable.
Data Collection
Data collection occurred between September 2001 and February 2003. For each resident, we conducted on-site interviews with the resident, the direct care provider who provided the most hands-on care and knew the most about the resident's care, health, mood, and daily activities, and the supervisor (i.e., staff member above a direct care provider level who knew the most about the resident). The facility administrator provided facility-level data, and the family provided information about their level of involvement in care. Further details about the Dementia Care sample and data collection procedures can be found in the introduction to this issue.
Measures
Activity involvement
We measured activity involvement using the Patient Activity Scale–Alzheimer's Disease (PAS-AD; Albert et al., 1996), which was reported for each resident by the direct care provider (n = 400) as well as by self-report for residents (n = 99) scoring 10 or higher on the Mini-Mental State Exam (MMSE; Folstein, Folstein, & McHugh, 1975). We selected the PAS-AD because it includes activities judged to be within the capacity of demented individuals who receive supervision and aid in daily activities (Albert et al.; Logsdon, Gibbons, McCurry, & Teri, 1999). The 15 items include 5 that involve travel outside of the facility (e.g., going shopping, to church, for a car ride, to the movies, and to see family and friends) and 10 that can be carried out in the facility (e.g., being with pets, exercising). Each activity is rated for opportunity (yes/no) and engagement during a one-week time frame. Response categories for engagement are frequently (
3 times/week = 2), occasionally (1–2 times/week = 1), or never (0). Responses are aggregated into a summary activity measure, (range = 0–30), with higher scores indicating more activity. Because norms for the frequency of such activities among demented people do not exist, Albert and colleagues recommends defining "higher" and "lower" activity based on those above and below the median of the distribution. Hence, lower activity involvement is defined here as less than 9.0 for both care provider and resident respondents, which is the same cutpoint used by Albert and colleagues. Internal consistency for the PAS-AD was very good (
= 0.79 and 0.80 for care provider and resident, respectively) and interrater reliability (care provider only, n = 18 pairs) was excellent (0.95 intraclass correlation coefficient for continuous measure and K = 1.00 for dichotomous measure of lower activity involvement).
Resident characteristics
We categorized dementia severity as mild, moderate, severe, and very severe based on scores from the MMSE and Minimum Data Set Cognition Scale (MDS-COGS; Hartmaier, Sloane, Guess, & Koch, 1994). MMSE category ranges are 18, 11–17, 3–10, 0–2 respectively; MDS-COGS cutpoints are 0–1, 2–3, 5–8, 9–10. We measured depression using the Cornell Scale for Depression in Dementia (CSDD; Alexopoulos, Abrams, Young, & Shamoian, 1988); behavioral symptoms with the Cohen-Mansfield Agitation Inventory (CMAI; Cohen-Mansfield, 1986); and pain using the Philadelphia Geriatric Center Pain Intensity Scale (PGC-PIS; Parmelee, Katz, & Lawton, 1991). We measured immobility by direct observation (Williams et al., 2005, this issue). We measured low food intake using the Structured Meal Observation (SMO; Reed, Zimmerman, Sloane, Williams, & Boustani, 2005, this issue). We measured functional status using the Minimum Data Set–Activities of Daily Living (MDS-ADL; Morris, Fries, & Morris, 1999) as a count of the number of disabilities (range = 0–7).
Facility characteristics
We obtained facility type, ownership (nonprofit, for-profit), bed size, and activity provision on a facility level from the administrator. We asked administrators to what degree the facility provides and encourages resident participation in 10 activities common to long-term care (exercise, personal care, social, housekeeping, meal preparation, crafts, work-oriented, special events, sensory, and intellectual; Zgola, 1987), and we coded responses as either not/rarely (less than one day/week) or regularly.
We asked supervisors three resident-level questions related to assessment: whether or not the resident's ability to participate and preferences for participation were assessed by an activity director; or by a written assessment; and how involved family members were in determining resident activities (from 0 = not at all to 4 = extremely). Supervisors also reported whether anything was done to encourage involvement in activities that the resident preferred and was able to do (yes/no); how well they feel the facility has been able to involve the resident in activities suited to his or her abilities and preferences (from 0 = not at all to 4 = extremely); and how well trained they feel in identifying residents' preferences and abilities to participate in activities, and helping residents participate in activities, as well as to actually help residents participate in activities. Finally, families reported their own involvement in care (number of hours/week spent visiting with or talking to the resident for social reasons).
Analysis
We computed simple descriptive statistics separately for RC/AL facilities and nursing homes. We used generalized estimating equations (GEE; Diggle, Heagerty, Liang, & Zeger, 2002) for the statistical comparison of these characteristics by setting, applied to linear or logistic (for continuous and binary characteristics, respectively) models and an exchangeable correlation structure with facility as the clustering variable. P values were based on score statistics (Boos, 1992). To examine the association between resident and facility characteristics and care provider report of activity involvement, we estimated odds ratios and 95% confidence intervals using a separate binary logistic regression model for each characteristic, controlling for clustering using GEE empirical standard error estimates and an exchangeable correlation matrix. We estimated adjusted odds ratios controlling for gender, race, age, cognitive status, number of comorbid conditions, and ADL dependencies. We repeated analyses using linear regression with the continuous PAS-AD as the dependent variable; results were very similar, and only the logistic regression results are reported. We also tested interactions of predictors with setting.
|
Results |
|---|
|
TOPAbstractResearch Design and MethodsResultsDiscussionReferences |
|---|
10 on MMSE) to report their own activity preferences and involvement. Based on supervisor report, slightly more nursing home residents exhibited lower activity involvement (56% vs 43%); this situation was reversed when reported by the 99 residents who could self-report (43% vs 55%); neither difference was statistically significant. As shown in Table 1, RC/AL residents enjoyed more activities than did nursing home residents based on staff report (10.7 vs 9.3, p =.025); no such difference was found for the 99 residents who could self-report (12.4 vs 13.0, p =.319). Based on staff report, the activities with the highest mean for both RC/AL and nursing home residents were (not shown) listening to radio, tapes or watching TV (1.49 and 1.62, respectively, on a scale of 0–3). Going shopping had the lowest mean (0.16 and 0.04), and 86.6% of RC/AL residents and 95.5% of nursing home residents had not gone shopping at all in the last week. Staff reported a high percentage of residents in both RC/AL facilities and nursing homes getting together with family and friends at least once in the last week (78.5% and 70.5%), but relatively few (33.6% of RC/AL residents and 16.2% of nursing home residents) had spoken on the telephone. Further, 31.5% of RC/AL residents versus 12.0% of nursing home residents had been outside often in the last week. The differences between RC/AL and nursing home residents for this finding was significant (p = 0.021).
|
Table 2 shows the distribution of characteristics related to lower activity involvement and the associated odds ratios and 95% confidence intervals. Lower activity involvement was more common in those with severe or very severe cognitive impairment, but this association was limited to nursing home residents and remained significant with adjustment for other resident characteristics (OR = 3.83; 95% CI = 2.69–5.45). Behavioral symptoms, depression, and ADL impairment were other resident characteristics associated with lower activity involvement, but the effects diminished in the adjusted model. Family involvement in assessing activities (OR = 0.86; 95% CI = 0.75–0.98), family social involvement (OR = 0.92; 95% CI = 0.87–0.97), and staff encouragement of activity involvement (OR = 0.32; 95% CI = 0.15–0.69) were all related to more activity involvement. Aside from cognitive impairment, there were no significant interactions between resident or facility characteristics and facility type (all were p >.05).
|
|
Discussion |
|---|
|
TOPAbstractResearch Design and MethodsResultsDiscussionReferences |
|---|
There were significant differences by facility type for activity involvement among residents who were more cognitively impaired: A higher percentage of nursing home residents compared to RC/AL residents with more severe cognitive impairment had lower activity involvement. Recognizing that these are adjusted activities, nursing home providers may gain insight from RC/AL facilities about how they are engaging these individuals in activities. Perhaps it is related to the social model of care philosophy that RC/AL facilities incorporate in their care practices.
Increased resident activity participation was associated with two measures of family involvement: the amount of time the family reports being socially engaged with the resident and the family's degree of involvement in assessing resident preferences as reported by the supervisor. Nursing home families were more likely to be involved in the assessment process (2.0 vs 1.6, p =.037). There is indication that it may be worthwhile to include families in the assessment process. In addition, when staff reported encouraging resident participation, the odds were higher that residents were more involved in activities. Of course, in a cross-sectional study such as this, a causal ordering of events cannot be established. It is possible that staff encouragement and family involvement correlates with more social residents. The fact that family involvement and staff encouragement relate to activity involvement could be tested to target resident participation in some of the activities with low involvement mentioned in this article (going outside, shopping, and talking on the telephone with family and friends). That this effort might be worthwhile is supported by reports that these are viewed by many residents as key to quality of life (Dobbs, 2004).
Nonetheless, one limitation of this study is worth reporting. It relied on staff data for the outcome variable (because only a small number of residents were able to respond for themselves). While the measure used was designed for proxy report, and while proxies are useful when participants cannot respond for themselves, there is no gold standard against which to compare their reports.
|
Footnotes |
|---|
1 Cecil G. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill.
2 Cecil G. Sheps Center for Health Services Research and the School of Social Work, University of North Carolina at Chapel Hill.
3 Regenstrief Institute, Inc. and Indiana University Center for Aging Research, Indianapolis, IN.
4 Cecil G. Sheps Center for Health Services Research and the Department of Epidemiology, School of Public Health, University of North Carolina at Chapel Hill.
5 Cecil G. Sheps Center for Health Services Research and the School of Medicine, University of North Carolina at Chapel Hill.
6 Alzheimer's Association, National Office, Chicago, IL.
Decision Editor: Richard Schulz, PhD
Received for publication June 24, 2004. Accepted for publication April 15, 2005.
|
References |
|---|
|
TOPAbstractResearch Design and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. Robison, L. Curry, C. Gruman, M. Porter, C. R. Henderson Jr., and K. Pillemer Partners in Caregiving in a Special Care Environment: Cooperative Communication Between Staff and Families on Dementia Units Gerontologist, August 1, 2007; 47(4): 504 - 515. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
