| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to Sheryl Zimmerman, PhD, Cecil G. Sheps Center for Health Services Research and the School of Social Work, University of North Carolina at Chapel Hill, 725 Martin Luther King Jr. Blvd., Chapel Hill, NC 27514. E-mail: Sheryl_Zimmerman{at}unc.edu
 |
Abstract |
|---|
 TOP Abstract Design and MethodsResultsDiscussionReferences |
|---|
Key Words: Long-term care • Residential care • Staff practices • Training • Observation
This article focuses on care and quality-of-life issues that have practice and policy relevance. It characterizes the current state of dementia care in RC/AL and nursing homes (based on report and observation); describes the characteristics and quality of life of residents with dementia in RC/AL and nursing homes (examining quality of life both cross sectionally and longitudinally, using multiple quality-of-life measures from the perspectives of residents and staff, and from observation); determines how dementia care (including special care units for residents with dementia) relates to resident quality of life; and considers the guidance this information provides for practice and policy.
|
Design and Methods |
|---|
|
TOPAbstractDesign and MethodsResultsDiscussionReferences |
|---|
16 beds of the "new-model" type (those proliferating under the recent surge of assisted living that provide nursing care and cater to an impaired population); and (c) "traditional" facilities with 16 beds, not meeting new-model criteria. Details of the CS-LTC and the facility typology can be found elsewhere (Zimmerman et al., 2001). The Dementia Care project enrolled a purposive sample of 45 facilities. For efficiency, facilities with fewer than 2 eligible residents (in smaller facilities) or 13 eligible residents (in other facilities) were excluded from study. Facilities were enrolled in a manner that maintained stratification across states and by facility type and that maximized the number of residents from smaller facilities. Twenty-two facilities (33%) declined to participate. These facilities did not differ from participating facilities in reference to type, size, or state. The final sample included 14 (31%) RC/AL facilities with < 16 beds; 11 (24%) traditional facilities; 10 (22%) new-model facilities; and 10 (22%) nursing homes. Twelve facilities were from North Carolina, and all other study states had 11 facilities. Given the purposive nature of facility selection, the descriptive data presented in this study are best used to formulate hypotheses.
Residents were randomly selected in each facility from among those aged 65 years or older who had a diagnosis of dementia. They were ineligible if they had a primary diagnosis of Huntington's disease, alcohol-related dementia, schizophrenia, manic-depressive disorder, or mental retardation. To provide similar representation across facility types, a maximum of 4 residents per smaller facility and 19 per larger facility were enrolled. A total of 575 eligible residents or their families were approached for consent. Of these, 421 (73%) agreed to participate, 66 (11%) declined, and 88 (15%) were unable to provide consent and had family who were unreachable. Additional information about the design is provided in the introduction to this issue (Zimmerman, Sloane, Heck, Maslow, & Schulz, 2005).
Data Collection
Data collection was conducted between September 2001 and February 2003. Data collectors observed the physical environment of all facilities and characteristics of a random sample of participating residents in each facility (79%) and conducted interviews with each resident participant (95% response rate), his or her most involved family member (84% response rate), the direct care provider who knew the resident best (98% response rate), the supervisor (position above a direct care provider) who knew the most about the resident (89% response rate), and the facility administrator (to obtain facility-level data; 100% response rate). In 4% of cases—usually in smaller facilities—the direct care provider and supervisor were the same individual. Ninety-four percent of direct care providers were nurse or personal care aides, and 78% of supervisors were registered nurses (RNs) or licensed practical nurses (LPNs).
Measures
Data were collected to assess care provision (facility-level and resident-level) and resident quality of life.
Facility-Level Care Provision
Dementia care measured at the facility level applied to all participants within a facility (or unit, if the facility included both an area designated for dementia care and an area not so designated). Administrators provided information about facility demographics (facility type, age, profit status, affiliation with another level of care or a chain of facilities, number of beds, presence of dementia-specific unit) and case-mix related to dementia diagnosis and each of six ADL impairments (eating, dressing, walking, transferring, bathing, and continence). The administrator also reported on several aspects of staffing, including the stability of care provider–resident assignments, whether the facility provided care based on a universal worker perspective (where staff fill multiple roles) and/or a specialized worker perspective (where staff have specialized roles), the number of nurses and nursing or personal care aides (overall and contract), staff turnover (at the administrator, nursing, and aide level), and the extent to which the facility sought to hire workers with experience in dementia care. Four measures of facility policies and practices were obtained, based on the Policy and Program Information Form (POLIF; Moos & Lemke, 1996): policy choice (7 items), leniency of admissions (24) and discharge policies (24), and acceptance of problem behavior (16). The latter three measures were assessed separately for dementia-specific and non-dementia-specific areas when applicable, and all were scored from 0 to 100, reflecting the percentage of items endorsed. Involvement in formal care planning of professional staff (averaged across administrator, physician, nurse, activity specialist, social service worker, mental health specialist, clergy, and dietician) and aides was scored from never to weekly.
Facility-level items regarding formal staff training, resident assessment, and treatment practices were ascertained for each of the following six domains of care: depression, pain, behavioral symptoms, ambulation, nutrition, and hydration. (The brief reports included in this issue further detail these measures.) Reports of the proportion of supervisory and direct care staff who received formal training in each domain within the past year were categorized as 0 (none), 1 (some; 1–74%), or 2 (most; 75%); scores were then summed across the six domains to yield a 0 through 12 summary score. Similarly, an assessment variable was created, consisting of the sum of domains for which the facility used professional assessment (e.g., mental health professionals for depression) and written, standardized assessment. A professional treatment variable was created as the average across the six areas of the percent of residents receiving ongoing, professional treatment for impairment in that area; "other/informal treatment" was created to be the number of areas for which the facility used other treatments; and perceived treatment success was the number of areas in which the administrator felt the facility did "quite a bit" or "extremely well" treating their residents. Except for perceived success, assessment and treatment were obtained separately for dementia-specific and non-dementia-specific areas. Three additional measures of treatment included the proportion of study participants who had received an antipsychotic or sedative hypnotic medication at least 4 out of the last 7 days (reported by the supervisor), the extent to which the facility provided and encouraged activity participation in 10 domains, separately by care area (e.g., exercise, personal care, social, meal preparation, work-oriented; Zgola, 1987), and the use of stimuli in seven areas (e.g., craft or household items).
Finally, data collectors systematically observed the environment, using the Therapeutic Environment Screening Survey for RC/AL (TESS–RC/AL), a measure derived from the Therapeutic Environment Screening Survey for Nursing Homes (TESS-NH; Sloane et al., 2002). Observations were used to compute two scales (separately for dementia-specific and non-dementia-specific areas): the Special Care Unit–Environmental Quality Score (SCU-EQS), which ranges from 0 to 38 and assesses 18 components relevant for individuals with dementia (e.g., orientation and memory cues); and the Assisted Living–Environmental Quality Score (AL-EQS), which ranges from 0 to 30 and assesses 15 components (some of which are in the SCU-EQS) more characteristic of assisted living environments (e.g., resident autonomy and homelikeness).
Resident-Level Care Provision
Dementia care measured at the resident level referred specifically to the study participants or their staff care providers, using data from interviews with staff and family members, and direct observation. The supervisor reported whether the resident had received cholinesterase inhibitors at least 4 of the past 7 days. Direct care providers reported their approaches to dementia care, their work satisfaction, and their work stress. The Approaches to Dementia measure was used to assess staff attitudes; it contains 19 items, summed to form a total score as well as person-centered and hope subscores (Lintern, Woods, & Phair, 2000b). The Work Stress Inventory was used to assess the frequency of 45 staff stressors related to work events, resident care, relationships with coworkers and supervisors, workload and scheduling, and physical design (Schaefer & Moos, 1993). Work satisfaction was measured using the 21-item Staff Experience Working with Demented Residents measure, which assesses satisfaction of one's own expectations, coworkers and supervisors, work environment, and resident care (Åström, Nilsson, Norberg, Sandman, & Winblad, 1991). These measures are described in detail elsewhere in this issue (Zimmerman, Williams, et al., 2005). In addition, family members reported the amount of time they spent each week visiting or talking on the phone with the resident.
Direct observations of study residents were conducted at 5-min intervals during three 1-hr observation periods (chosen to exclude mealtimes), from which four measures of resident care were derived. Communication was measured as the percent of observations during which the resident received any verbal communication from a staff member, physical contact was the percent of observations during which the resident had any physical contact with another person, and personal detractors and positive person work were similarly measured as the percent of observations in which any personal detractors (staff behaviors that demean or depersonalize) or positive person work (positive interactions between staff and resident) were noted (Bradford Dementia Group, 1997). Whether the resident was ever observed in restraints (full or partial bedrails, trunk, wrist, ankle, or chair restraints) also was noted, and during the first observation each hour, residents were assessed to determine if they appeared ungroomed, unkempt, or unclean and whether appearance was ever inappropriate with respect to time of day, season, or place.
Resident Characteristics
The supervisor provided information on several resident characteristics, including demographics (age, gender, race, marital status) and length of stay. The presence of behavioral symptoms of dementia during the past 2 weeks was measured using the Cohen-Mansfield Agitation Inventory (CMAI; Cohen-Mansfield, 1986), functional status was measured as the number of activities in which the resident needed supervision or assistance using the seven items from the Minimum Data Set Activities of Daily Living (MDS-ADL) scale (Morris, Fries, & Morris, 1999), residents were classified as being depressed if they scored 7 on the Cornell Scale for Depression in Dementia (CSDD; Alexopoulos, Abrams, Young, & Shamoian, 1988), and comorbidity was the number of chronic conditions (out of 11) reported by the supervisor. Finally, cognitive deficit was categorized based on the Mini-Mental State Exam (MMSE; Folstein, Folstein, & McHugh, 1975) score administered to the resident, or (if the MMSE was not available) the Minimum Data Set–Cognition Scale (MDS-COGS; Hartmaier, Sloane, Guess, & Koch, 1994) reported by the supervisor. Cognition was scored as follows: mild (MMSE 17–30 or MDS-COGS 0–1), moderate (MMSE 10–16 or MDS-COGS 2–4), severe (MMSE 3–9 or MDS-COGS 5–6), or very severe (MMSE 0–2 or MDS-COGS 7–10).
Quality of Life
Quality of life was assessed by the resident (three measures), direct care provider (six measures, one of which was longitudinal), and through observation (three measures); measures are described in detail elsewhere (Sloane et al., 2005, this issue). Unless otherwise noted, higher scores indicate better quality of life. Residents with a MMSE of 10 or greater completed the Dementia Quality of Life (DQOL; Brod, Stewart, Sands, & Walton, 1999), the Quality of Life in Alzheimer's Disease Activity measure (QOL in AD–Activity; Albert et al., 1996), and the Quality of Life–Alzheimer's Disease (QOL-AD; Logsdon, Gibbons, McCurry, & Teri, 2000), as modified by Edelman, Fulton, Kuhn, and Change (2005, this issue) for use in long-term care settings.
Care providers completed proxy versions of the QOL in AD–Activity, and the QOL-AD, as well as the positive and negative affect portions of the QOL in AD (higher negative affect scores indicate poorer quality of life), and the Alzheimer Disease Related Quality of Life (ADRQL; Rabins, Kasper, Kleinman, Black, & Patrick, 2000). Additionally, the proxy version of the QOL-AD was readministered 6 months after initial data collection. We computed raw change as the difference between baseline and 6 months, with positive scores indicating improvement in quality of life; to account for regression to the mean, we estimated adjusted change as the residual from regression of raw change on the baseline value, which therefore has a sample mean of exactly zero (Cronbach & Furby, 1970). Because change is influenced by baseline status, we based all statistical comparisons on adjusted values.
Using the observational procedures described above, we recorded behaviors in accordance with the Dementia Care Mapping (DCM) protocol (Bradford Dementia Group, 1997) and a modification of the Philadelphia Geriatric Center Affect Rating Scale (PGC-ARS; Lawton, Van Haitsma, & Klapper, 1996). DCM-derived measures included the percent of observations with a Type I Behavior Category Code (DCM % BCC Type I), considered to be "good" behaviors such as conversation or creative expression, and the mean Well- and Ill-Being (WIB) score, with anchors of +5 and –5, indicating the state of well-being. (See Brooker, 2005, and Sloane et al., 2005, this issue, for a more in-depth discussion). The PGC-ARS was coded to record the predominant emotion at each observation (scored from 0 for anxiety, fear, or sadness, to 3 for high pleasure) and summarized at the resident level as the percent of the highest possible score.
Analyses
We used simple descriptive methods (means and standard deviations for continuous measures, percentages for categorical measures) to describe the components of dementia care. For measures assessed separately in dementia-specific and non-dementia-specific care areas, comparisons used generalized estimating equations (GEE: Diggle, Heagerty, Liang, & Zeger, 2002) applied to linear (continuous) or logistic (dichotomous) models, specifying an exchangeable correlation matrix to account for resident clustering within facility.
We estimated means and standard errors of the quality-of-life measures according to facility type and resident characteristics, dichotomized at the sample median or at commonly accepted cutpoints; we adjusted the standard errors for clustering using Taylor series expansion methods (Woodruff, 1971). We tested the statistical significance of these associations using linear mixed models with random effects specified as follows: for care provider reported quality-of-life measures, models included random effects for facility and care provider (nested within facility); for resident-reported measures, models include a random effect for facility; and for directly observed measures, models include random effects for facility and observer.
We estimated the association between care and quality of life using partial Pearson correlation coefficients, adjusting for facility type; resident age, gender, race, marital status, length of stay; and cognitive, ADL, number of comorbid conditions, depressive, and behavioral symptoms. To maximize the sample size for resident-reported quality-of-life analyses in the presence of covariate missingness, we did not adjust associations for age, race, marital status, or length of stay (none of which was significantly related to quality of life). We tested the statistical significance of these associations using linear mixed models, controlling for these resident characteristics and with random effects as specified above.
|
Results |
|---|
|
TOPAbstractDesign and MethodsResultsDiscussionReferences |
|---|
|
Table 2 separates care by dementia-specific area for all resident-level variables. It shows that supervisors (of the residents enrolled in this study) who worked in non-dementia-specific areas had slightly more experience but that positive person work and physical contact were witnessed more often in dementia-specific areas (22% vs 17%, and 9% vs 6% of observations, respectively). Overall, 13% and 20% of resident participants were in restraints and ungroomed during at least one observation, and 29% were taking a cholinesterase inhibitor. On average, families spent almost 7 hr per week visiting or talking with the resident.
|
|
|
|
The remaining rows on Table 4 indicate the diversity with which policies and practices relate to quality of life. While virtually all policies and practices under study related to quality of life, they did so inconsistently across different measures. For example, having more flexible admission, discharge, and acceptance of problem behavior policies related positively to care provider report of QOL in AD–positive affect, and involving professional staff in care planning related positively to resident reported QOL-AD and observed affect (PGC-ARS). The strongest association for resident-reported quality of life was witnessed for facility use of antipsychotic and sedative hypnotic medications (negatively associated with QOL-AD, p <.01). The one facility-level component of care that related to quality of life as assessed by residents, care providers, and observation was the provision of professional treatment for the six care areas under study; it was negatively related to quality of life as reported by residents (DQOL) and observed (DCM BCC Type I codes), but positively related to care provider reports (QOL in AD–positive affect).
In addition, one component of resident-level care related to quality of life across all three sources (see Table 5). Residents who were observed to be ungroomed reported their own quality of life to be worse (QOL-AD), as did care providers (QOL in AD–activity) and observation (DCM BCC Type I codes). Residents who had staff who espoused more dementia-sensitive attitudes (especially hope) rated their quality of life higher on two measures (DQOL and QOL in AD–activity). Observed interactional style (more communication, positive person work, and physical contact, fewer personal detractors) was positively related to care provider-rated and observed quality of life. Finally, family involvement was related to higher QOL in AD–activity, as rated by care providers.
|
Discussion |
|---|
|
TOPAbstractDesign and MethodsResultsDiscussionReferences |
|---|
The present study of care provided to 421 residents in 45 facilities was designed to do just that. At follow-up, residents fared better in facilities with specialized workers, with more staff training in more areas (supervisor and direct care staff), and that encouraged activity participation more frequently. Specifically, mean raw change was –1.3 (SD 7.4) versus –3.0 (SD 8.2) in facilities with specialized workers compared to those without. In facilities in which 75% or more of supervisors were trained in at least five of the six domains, raw mean change was –1.0 (SD 8.3) versus –3.4 (SD 7.5); comparable figures for care provider training were –0.4 (SD 7.6) vs –3.5 (SD 7.9). In facilities that encouraged activities once a day or more, it was –1.9 (SD 7.8) versus –2.6 (SD 8.0). Interestingly, these are all facility-level variables, and none of the resident-level components of care related to change in quality of life. On the one hand, such findings are promising because they imply that facility-wide change can impact resident well-being; on the other, they call into question the degree to which individualized care is benefiting residents with dementia. It must be acknowledged, however, that this study may not have had sufficient power to detect some of these associations—but even if it did, they likely would have been small, and, similar to the ADRQL study, of questionable clinical relevance. Nonetheless, considering this acknowledged limitation, the fact that three components of care related to change over time may highlight the utility of turning attention to the areas of specialized workers, staff training, and encouragement of activity participation. In one area this attention may spark debate, as many states promote the practice of universal workers (Mollica, 2002), and the expanded use of specialized workers may change demands for care.
The authors of the ADRQL study concluded that the ADRQL is sensitive to change over time (although noted that such sensitivity might be limited) and appropriate for use as an outcome measure in intervention studies. In the current study, change in quality of life measured with the QOL-AD was significantly different over an even shorter period of time (6 months as opposed to 2 years); also, it detected a significant relationship with cognitive and affective status (such that a more favorable status at baseline related to relatively better quality of life at follow-up) and was markedly lower for residents immediately before the time of discharge or death compared to those who remained in the facility through 6 months (raw change –4.7 [SD 7.7] vs –1.7 [SD 7.9]). Finally, the fact that the QOL-AD detected differences among components of care further merits its consideration as an outcome measure. It might be a particularly useful measure if interest was in how the resident rated his or her quality of life, as the patient version of the QOL-AD can be reliably and validly completed by those with a MMSE score as low as 10; the degree to which this version is sensitive to change is unknown, however (Logsdon, Gibbons, McCurry, & Teri, 2002).
If one were of a different mindset, one would recognize that a longitudinal study comparing care to quality of life in a cohort of current residents (as opposed to a new admissions cohort) may be insensitive to the effects the care environment had exerted since the time of admission. In such a case, a cross-sectional comparison of care to quality of life, adjusting for resident status, might best indicate this relationship. In making those comparisons, this study found many associations of care to outcomes; given the multiple comparisons, it is best to focus on the detected patterns.
The 11 measures used in this study define quality of life differently and from three different vantage points. None constitutes a gold standard, although many suggest that the resident's point of view should take priority (Brod et al., 1999). In this study, we could conduct analyses for at most 120 resident reports, and significant associations with so modest a sample are worthy of discussion. From the resident's perspective, quality of life was higher for those in facilities that more frequently involved more staff in care planning and whose care providers felt more hope (e.g., that residents can make decisions, that they will not inevitably go "down hill," and that feeling attached to residents need not be avoided). Also, quality of life was lower in facilities that provided more treatment, including antipsychotic and sedative hypnotic medications, and when residents themselves were ungroomed. Other authors have found a relationship between anxiolytic treatment and reduced quality of life, and the likely explanation is that more intense treatment is used (not entirely successfully) for residents who are more impaired (a relationship that persisted despite controlling for resident status in these analyses; Gonzalez-Slavador et al., 2000).
A limitation of relating care provider assessments of quality of life to outcomes is that such assessments are influenced by caregiver factors (Gonzalez-Slavador et al., 2000; Karlawish, Casarett, Klocinski, & Clark, 2001; Winzelberg, Williams, Preisser, Zimmerman, & Sloane, 2005, this issue). Thus, it may come as no surprise that residents with whom workers communicate more and toward whom they display positive person work (e.g., enable the resident to do what he or she couldn't otherwise do) tend to be rated more highly. One finding to note is that these same interpersonal components are related to observational indicators of quality of life, such that these residents display more positive affect, behaviors, and general well-being. Thus, to the extent that workers have the time and can feel and act positively toward residents, quality of life is likely to be improved. Further, these attitudes relate to worker satisfaction as well, and so all parties may benefit when positive interactions are maximized (Zimmerman, Williams, et al., 2005, this issue). Finally, contrary to conventional wisdom, more stability in staff–resident assignment was related to worse staff ratings of quality of life (but not to resident or observer ratings of quality of life). Whether stability is affecting care provider attitudes (and hence ratings), or whether it is actually affecting resident quality of life is not known. A recent study showed no clear superiority of permanent versus rotating staffing, and so this area merits further attention (Burgio, Fisher, Fairchild, Scilley, & Hardin, 2004).
Finally, it would be remiss to not stress the fact that (a) resident appearance was related to at least one measure of quality of life as rated by residents, care providers, and observation, and (b) facility type (RC/AL vs nursing home) and number of dementia beds were not related to any quality-of-life measures. Grooming may be an inherent indicator of dignity and, as such, may be an implicit marker of poor quality of life. As far as setting of care, there is increasing evidence that the quality of care in nursing homes has been improving (Feldman & Kane, 2003) and no overwhelming indication that special dementia care is related to better outcomes (Phillips et al., 1997). Thus, while RC/AL developed in part so that older adults could avoid nursing home placement, the tide may have turned, and these settings may be less different than some consider—and equally suitable (although perhaps not equally affordable) for the care of residents with dementia (Kane & Wilson, 2001; Zimmerman et al., 2003). At minimum, it is likely that such gross categorizations of care (RC/AL, nursing home) do not relate to differences in care that are affecting resident quality of life.
What then do these myriad findings suggest? They certainly suggest directions for hypothesis generation and further exploration and evaluation. While causal attribution is not possible, the findings suggest that facilities should consider (and studies should evaluate) using a specialized worker perspective, train all staff in domains central to dementia (depression, pain, behavioral symptoms, ambulation, nutrition, and hydration), and encourage activity participation (related to change in quality of life over time). They suggest that attention be paid to resident grooming (related to quality-of-life ratings by residents, staff, and observation). They suggest that facilities should involve staff in care planning, encourage care providers to feel more hope, and avoid antipsychotic and sedative hypnotic medications, if possible (related to resident perceptions of quality of life). They suggest that staff should communicate more, and positively, with residents (related to care provider rating and observed quality of life) and that rotating worker assignment be further explored (related to care provider rating). To the extent that all of these areas are under the control of the facility, and can be implemented with few new resources, all are worth consideration and evaluation to improve the quality of life of long-term care residents with dementia. In fact, the Alzheimer's Association is undertaking a national educational campaign, the Campaign for Quality in Residential Care, to implement and evaluate many of these components of care. Thus, the growth of evidence-based practice to improve the quality of life for residents with dementia in long-term care is evident, and promising.
|
Footnotes |
|---|
1 Cecil G. Sheps Center for Health Services Research and the School of Social Work, University of North Carolina at Chapel Hill.
2 Cecil G. Sheps Center for Health Services Research and the Department of Family Medicine, University of North Carolina at Chapel Hill.
3 Cecil G. Sheps Center for Health Services Research and the School of Public Health, University of North Carolina at Chapel Hill.
4 Alzheimer's Association, National Office, Chicago, IL.
5 Department of Biostatistics, School of Public Health, University of North Carolina at Chapel Hill.
6 Department of Sociology and Anthropology, University of Maryland–Baltimore County.
7 Regenstrief Institute, Inc. and Center for Aging Research, Indiana University, Indianapolis.
8 Cecil G. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill.
Decision Editor: Richard Schulz, PhD
Received for publication September 20, 2004. Accepted for publication March 3, 2005.
|
References |
|---|
|
TOPAbstractDesign and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. N. Rahman and J. F. Schnelle The Nursing Home Culture-Change Movement: Recent Past, Present, and Future Directions for Research Gerontologist, April 1, 2008; 48(2): 142 - 148. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Moyle, M. Mcallister, L. Venturato, and T. Adams Quality of life and dementia: The voice of the person with dementia Dementia, May 1, 2007; 6(2): 175 - 191. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
