| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
a Centre on Aging, University of Victoria, British Columbia, Canada
b Department of Sociology, University of Victoria, British Columbia, Canada
Neena L. Chappell, Director, Centre on Aging, Sedgewick Building, Rm. A104, University of Victoria, P.O. Box 1700, Victoria, BC, V8W 2Y2, Canada E-mail: nlc{at}uvic.ca.
 |
Abstract |
|---|
 TOP Abstract Review of the LiteratureMethodsResultsDiscussionReferences |
|---|
Methods. Data were drawn from the Intermediate Care Facility Project. The sample (
) included residents with dementia, aged 65 and older, in intermediate care facilities throughout the province of British Columbia, Canada. Longitudinal data included 6 outcomes: cognitive function, behavioral problems of agitation and social skills, physical functioning, and quality of life measured through affect and expressive language skills. Separate multiple linear regression equations were estimated, relating each of these outcomes to 5 dimensions of care: preadmission and admission procedures, staff training and education, nonuse of physical and chemical restraints, flexible care routines and resident-relevant activities, and the environment.
Results. The results showed there is virtually no clustering of dimensions along SCU/non-SCU lines. Neither SCU status nor the individual dimensions were highly predictive of outcomes. Residents' affect at t1 emerged as a characteristic that was significantly correlated with other outcomes.
Discussion. This Canadian research can be added to the few but growing number of rigorous studies that suggest SCUs are not homogeneous and do not necessarily provide better care than non-SCUs. Moreover, it raises questions about the benefits of "best practice" dimensions of care, regardless of SCU status.
AS the demographic shift continues, the proportion of elderly persons and of old elderly persons in society is increasing. The risk of dementia rises rapidly with age among the oldest age groups, and thus the number of individuals with more severe levels of dementia has increased. Long-term care facilities, as a result, now tend to house higher proportions of residents who suffer from advanced dementia than was true in the past. To meet this demand, providers of care have of necessity been experimenting with management in this area. This experimentation has resulted in the identification, rethinking, and implementation of several dimensions of care, such as assessment and care planning, staff training, flexible care routines, individualized care plans, and the environmental setting.
It is commonly thought that such dimensions of care, adequately implemented and in the right combination, will result in optimal outcomes for residents suffering from dementia (Holmes et al. 1994
). Although these dimensions of care are in theory expected to result in superior outcomes—and thus have become the philosophy behind the creation of special care units (SCUs)—their effects on outcomes are not fully known. In this study we empirically examined whether such "best practices" dimensions of care or a subset of them are, in fact, more likely to be implemented within SCUs. We also examined whether specific dimensions or combinations of dimensions are related to enhanced outcomes over time for dementia sufferers in both types of long-term care facilities (SCUs and non-SCUs).
|
Review of the Literature |
|---|
|
TOPAbstractReview of the LiteratureMethodsResultsDiscussionReferences |
|---|
). Much of the research on institutional dementia care has focused on SCUs, because they are perceived to epitomize the dimensions of care that should lead to optimal outcomes. For example, in a survey of 1,497 facilities with SCUs, Leon 1994(p. S83) found that SCUs typically possessed six features: modified physical environments, physically separated units with controlled on–off access, limited admissions to residents with diagnoses of dementia, extra staffing, designated unit leadership, and specialized staff training and programming. Until recently, however, the limited scientific evidence that did indicate a link between dimensions of care and outcomes did not go beyond case studies and descriptive reports (e.g., Gutman and Killam 1989; Lahaie and Theroux 1992; Ohta and Ohta 1988; U.S. Office of Technology Assessment 1987). Four widely cited articles (Mace 1987; Pynoos and Stacey 1986; Rabins 1986; Ronch 1987) were not based on outcomes from research.
Recently, researchers have empirically established the effects of SCU residence on a range of resident outcomes. For example, Bellelli and associates 1998 reported decreases in behavioral disturbances, and in the use of both psychotropic drugs and physical restraints, in an assessment of SCUs in Italy. In another longitudinal study, Bianchetti, Benvenuti, Ghisla, Frisoni, and Trabucchi 1997 likewise reported that behavioral problems decreased among residents 6 months after they were transferred from a traditional nursing home to an SCU, largely because of the removal of physical restraints. The residents' functional ability and cognitive status, however, did not improve. Martichuski, Bell, and Bradshaw 1996 evaluated the effect of small-group activity and caring staff on resident outcomes in SCUs and reported a decrease in the use of both physical restraints and psychotropic drugs but no change in the level of physical functioning or in the frequency of negative behaviors (except frowning, which declined). Among positive behaviors, significant changes in walking with others and singing were observed. Although these studies lack comparison groups, they show that the effects of SCU residence on outcomes are variable and, importantly, not always in the desired direction.
More elaborate studies, incorporating more sophisticated methodologies on larger samples, have similarly been unable to show SCUs to be generally more effective than non-SCUs. In fact, in a review, Maslow 1994 noted that the few positive findings appear to contradict the conviction that SCUs are of benefit to residents, families, and staff (see also Martin, Gwyther, and Whitehouse 1994). In response to the low level of scientific knowledge about the impact and nature of SCUs, the U.S. National Institute on Aging funded 10 multiyear collaborative SCU evaluation studies, which were initiated in 1991 (Ory 1994). Although substantive findings from those studies are not yet published (Teresi, Grant, Holmes, and Ory 1998), other studies have shown quite clearly that SCUs are not necessarily advantageous to residents, compared with non-SCUs. For example, Phillips and colleagues 1997 reported no significant difference in speed of decline for nine measures of functional status among residents of SCUs and non-SCUs in 800 facilities. Furthermore, Saxton, Silverman, Ricci, Keane, and Deeley 1998 comparison of residents of a specialized Alzheimer's disease facility with a matched sample of those in a traditional nursing home showed similar rates of decline in cognitive functioning and activities of daily living (ADLs), but a striking absence of decline in mobility among those in the specialized facility.
The similarity of outcomes in residents of SCUs compared with non-SCUs is hardly surprising if the dimensions of care in each of the two types of facilities are similar. Holmes, Teresi, Ramirez, and Goldman 1997 compared staff time spent on behalf of residents in SCUs and non-SCUs and found differences among only 3 of 10 staff groups. They also found that staff in non-SCUs spent more time on occupational and physical therapy and therapeutic recreation. Grant, Kane, and Stark 1995 reported that SCUs offered more dementia-specific staff training, environmental design, and programming features than did non-SCUs, but that most non-SCUs offered some such features. In addition, traditional units in facilities without a SCU were more likely to report dementia-specific features than did traditional units in facilities with a SCU. In comparing SCUs with non-SCUs, Grant, Kane, and Potthoff 1996 reported that SCUs scored higher for their training methods and staff stability, but not for training content; staff categories trained; staffing patterns across shifts, units, and days of the week; or staff turnover. It seems clear that SCUs have not yet effectively incorporated all of the dimensions of care that, theoretically, would produce superior outcomes for residents with dementia and that traditional units are, at the same time, incorporating some of these dimensions.
Methodological limitations in much of the previous research may also prevent an accurate determination of the differential effects of SCUs versus non-SCUs. Much of the research to date has examined a single dimension rather than several, and most employed cross-sectional rather than longitudinal designs. An independent review of the literature for the study reported here identified five dimensions of care that should lead to high quality of care as measured by a number of outcomes. These dimensions include assessment and diagnosis; staff specialization and ongoing education; nonuse of restraints (both physical and chemical); flexible care routines, including client-relevant activities; and specialized environmental design and adaptation. These dimensions are similar to the six components of care Holmes and associates 1994 derived from concept mapping: the environment, activities programming, staffing, training and special assignment, rational care planning and family involvement, and structural characteristics.
These dimensions of care have received varying degrees of attention in the literature. A considerable amount of research is available on staff education. Appropriate staff education tends to reduce the use of physical restraints, and there are no apparent adverse effects to their removal (C. Cohen, Neufield, Dunbar, Pflug, and Breuer 1996; Cruz, Abdul-Hamid, and Heater 1997; Dunbar, Neufeld, White, and Libow 1996; Evans and Strumpf 1989; Evans, Strumpf, Allen-Taylor, Capezuti, Maislin, and Jacobsen 1997). For example, residents do not have more falls, or more serious falls, when physical restraints are removed. Karlsson, Bucht, and Sandman 1998 reported that staff with less knowledge exhibited the least negative attitudes toward physical restraint use. Chemical restraints, normally in the form of psychotropic drugs, are viewed as only modestly beneficial and their adverse effects can be significant. The use of psychotropic drugs is related to more falls among residents (Blake et al. 1988; Ebly, Hogan, and Fung 1997). Some research has indicated that educational programs can result in a decrease in amount and inappropriate use of psychopharmacology (Avorn et al. 1992; Ray et al. 1993).
There is controversy over whether there is a relationship between the removal of physical restraints and the use of pharmacological restraints. The removal of physical restraints is related to a decrease in the use of antipsychotic medications and a decrease in agitated behaviors (Werner, Cohen-Mansfield, Koroknay, and Braun 1994; Werner, Koroknay, Braun, and Cohen-Mansfield 1994). However, Sloane and colleagues 1991 argued that SCUs are successful at decreasing the use of physical restraints, but not chemical restraints—that the use of physical and chemical restraints must be examined separately. They are different phenomena.
Other research has suggested that staff education can also be effective in preventing aggressive behavior, physically nonaggressive behavior, and verbally agitated behavior among cognitively impaired, institutionalized elderly persons (Landreville, Bordes, Dicaire, and Verreault 1998; Palmer and Withee 1996). However, as Landreville and associates 1998 stated, effectiveness is not yet conclusive.
Flexible care routines and resident-relevant activities, also referred to as individualized care planning, are considered one of the important components of quality care (Morgan and Stewart 1997). Involving family members is considered a necessary part of ensuring client-centered activities, for familes' past knowledge of, and current involvement with, the resident (Anderson, Hobson, Steiner, and Rodel 1992; Hansen, Patterson, and Wilson 1988). The specific type of activities that should be encouraged is difficult to document, and certain activities may be more suited to residents of SCUs and others better suited to residents of non-SCUs (Grant and Potthoff 1997).
Physical environmental features are an integral component of optimal care. Units designed with the needs of the demented resident in mind—for example, with a continuous wandering loop and fewer residents than in a traditional unit—are related to positive outcomes such as increased social interaction (Kovach, Weisman, Chaudhury, and Calkins 1997). Benefits of specialized design extend to staff, who are less stressed in well-designed facilities (Lyman 1989), and family caregivers, who show high rates of satisfaction when compared with family caregivers of residents in other types of facilities (Grant and Sommers 1998). Lighting, noise, and temperature have varying influences on the observed behaviors of nursing home residents with Alzheimer's disease (Cohen-Mansfield and Werner 1995). For example, an increase in strange movements was observed in the dark, but a decrease in pacing, verbal agitation, repetitious mannerisms, and other behaviors was observed during noisy periods. Unlocked doors can also decrease agitation (Namazi and Johnson 1992), although the benefits of controlled access to and from a unit (e.g., preventing disoriented residents from wandering out of the building) probably outweigh the disadvantages (Leon 1996). There remain considerable differences of opinion concerning the ideal environment, and any given environmental design, as well as the anticipated effect on outcomes, is conditioned by the underlying philosophy of the institution (Maas, Swanson, Specht, and Buckwalter 1994).
Findings to date are not conclusive about the dimensions that lead to high quality of care for those suffering from dementia. Zimmerman and associates 1997 pointed out that the lack of consensus among unit coordinators on underlying philosophical principles (e.g., support for cognitive functioning, maximization of independent functioning, and promotion of safety and security) implies diverse needs of residents and inherent contradictions in philosophies. They noted further that there is a weak relationship between espoused philosophy and observed care, except in restraint use. Similarly, Williams and Rees 1997 indicated that care received is not always equal to care planned and delivered. In other words, the differences sometimes believed to exist between SCUs and non-SCUs may not. There may not be uniformity either across SCUs or across non-SCUs.
Despite the lack of support in the scientific literature for SCUs, they have received strong governmental support in British Columbia, Canada, where one SCU was established in 1976, 14 in the 1980s, and 37 between 1990 and 1995. Existing literature suggests that SCUs may not be fundamentally different from non-SCUs in terms of care delivered and that SCUs may not influence outcomes for residents with dementia more effectively than non-SCUs. In the current study we had two fundamental purposes. First, we explored the extent to which SCUs in Canada actually do embody the dimensions of care that should lead to the best outcomes. Second, we examined best care practices—referred to here as dimensions of care—in relation to six key resident outcomes, independent of the type of facility (SCU or non-SCU).
|
Methods |
|---|
|
TOPAbstractReview of the LiteratureMethodsResultsDiscussionReferences |
|---|
argument that 6 months is probably too short for evaluation of the impact of long-term facility stays. Specially trained research personnel collected data on newly admitted residents from the residents themselves, medical charts, family, and staff. The study included 77 intermediate care facilities, with 51 SCUs and 101 integrated units; a unit was defined by the presence of one nursing station. One facility declined to participate. Some facilities had both SCUs and integrated units. A unit was defined as an SCU if it served primarily dementia residents; identified itself as a specialized unit for the care of residents with dementia; was separated from the rest of the facility by closed doors; and satisfied at least one of the following criteria: (a) the unit had staff trained specifically in dementia care, (b) unit activities were designed with the dementia residents in mind, or (c) assessment and care planning included an evaluation of potential residents for their needs and their "fit" to the programs of the SCU. The mean number of beds in the SCUs was 30, and the mean percentage of residents with dementia was 96%. The mean number of beds in the integrated units was 53, and the mean percentage of residents with dementia was 58%. We screened for eligibility all newly admitted residents to the participating facilities who, according to the contact person at the facility (typically the director of nursing, or DoN), suffered from moderate or severe dementia.
To be eligible, residents were required to have a primary or secondary diagnosis of either Alzheimer's disease or vascular dementia, have confirmatory evidence of dementia in their medical charts, be unlikely to die or move from the unit in the 12 months following their admission, have the ability to communicate in English, and be at least 65 years of age. Because SCUs tend to have residents with higher levels of dementia, and to allow comparisons across facility types, we screened only residents identified by DoNs as having at least moderate dementia. DoN assessments of level of dementia were not used in analyses for this article. Their assessments were used solely to provide a pool of participants with probable moderate or severe dementia. Scientifically valid levels of cognitive function were subsequently obtained with the Multi-focus Assessment Scale-Revised (MAS-R, described below; see Reid and Chappell in press, for an analysis of the relationship between staff assessments and the MAS-R). Fifty-four percent of all newly admitted residents with moderate or severe dementia were eligible. The most common reason for noneligibility was a lack of a specific diagnosis of dementia in the medical chart. Signed consent was obtained from a responsible family member or guardian for 88% of the eligible sample.
Prior to collection of outcome data, all long-term care institutions in the province that cared for persons with dementia completed a comprehensive survey. In addition to data on facility characteristics such as number of beds and number and severity of dementia of residents, information on the five identified dimensions of care was gathered from DoNs. For the outcomes phase, a representative sample stratified by unit (SCUs and non-SCUs) was obtained on the two dimensions that an expert steering committee judged to be the most important of the five dimensions (staff education and flexible care), excluding facilities in the far north of the province because of cost. A post facto analysis revealed no significant difference between units included in the study and those excluded (including those in the far north) on staff education, flexible care, use of restraints, or assessment. Because of the inadequacy of the information on the environment in the initial mail-out survey, the 
-
(TESS; see Measures section) was used in the outcomes phase. Comparable data, therefore, do not exist for this dimension for the purpose of comparing those included in and those excluded from the study.
Measures
Outcomes for residents were chosen from the literature (Kane 1997; Warshaw 1997): behavioral problems—agitation and social skills; functioning; quality of life—affect and expressive language skills. We used cognitive functioning as an outcome in this study to assess change, not because we believed it to be reversible. To assess change over time in outcomes and to associate them with the dimensions of care, we measured outcomes at admission and at 12 months. The Cohen-Mansfield Agitation Inventory (
; Cohen-Mansfield and Marx 1989) is a 14-item questionnaire that measures the occurrence of three types of agitation during the previous 2 weeks, including (a) aggressive behaviors (hitting, kicking, pushing, scratching, tearing things, cursing or verbal aggression, and grabbing); (b) physically nonaggressive behaviors (pacing, inappropriate robing or disrobing, repeating sentences or questions, trying to get to a different place, handling things inappropriately, acting generally restless, and having repetitious mannerisms); and (c) verbally agitated behaviors (complaining, constantly requesting attention, showing negativism, repeating sentences or questions, and screaming). Using a 5-point scale (
), nursing staff indicated the number of occurrences of each specified behavior during the preceding 2 weeks. Scales were summed to create a total agitation score (range 14–70; 
).
A measure of ADLs was obtained with the Minimum Data Set (MDS), Item E. Item E monitors the level of independence of a resident on a 5-point scale (
) for each of nine ADLs. These include bed mobility, transfer between surfaces, movement between locations, dressing, eating, toilet use, personal hygiene, walking, and bathing. Summed scale values can range from 0 (independent) to 36 (completely dependent; 
for t1 and .87 for t2).
Mood was measured with the affect scale of the Feeling Tone Questionnaire (FTQ); reliability statistics on the FTQ were not yet available when this research was initiated. The FTQ measures mood in persons with communication difficulties. In this research, interviewers asked 16 mood-life satisfaction questions (e.g., "Do you have any pain?" "Do you sleep well?") and rated the affect expressed by the respondent. The interviewer rated each response on a 5-point affect score (1 = "very positive," 5 = "very negative"). The scores were added, and scale totals ranged from 16 to 80, with 80 representing the most extremely negative affect possible (
= .89 for t1 and .95 for t2).
Trained research assistants administered the Multi-focus Assessment Scale-Revised (MAS-R; Crockett, Coval, Tuokko, Buree, and Koch 1991) to assess cognitive and behavioral functioning. The MAS-R is a revision of the original MAS (Tuokko, Crockett, Holliday, and Coval 1987) and consists of seven scales (early memory and present orientation measure cognitive function; social behavior skills; auditory capability; visual comprehension; expressive language skills; and accessability). Social behavior skills (0–11 point scale) and expressive language skills (i.e., ability to effectively communicate on the basis of speech, language production, and comprehension assessment; 3–15 point scale) were considered outcomes because they can potentially be affected by quality of care. Alphas for social behavior skills for 
. Alphas for expressive language skills for 
. For all MAS-R scales, a higher score means higher competence, except for the expressive language skills, where a higher score means poorer language skills. Crockett and colleagues 1991 reported interrater reliability to be high (.96 or greater for all scales).
Cognitive function was measured by the MAS-R on two scales. The first of these measures early memory and the second measures present orientation. Each consists of 11 items. Early memory refers to memories acquired at a relatively young age (e.g., "When was WWII?" "How many years of schooling did you have?"). Present orientation measures the ability to acquire and retain current information (e.g., "Who is the Prime Minister of Canada?" "How many people share your room?"). Alphas for 
.
Change scores (t1 minus t2) were calculated to produce each outcome score. For agitation, physical functioning, affect, and expressive language skills, a negative sign (-) signifies decline, and a positive sign (+) signifies improvement. For cognitive function and social skills the reverse is true; a negative sign (-) signifies improvement, and a positive sign (+) signifies decline. Each of the outcome variables at t1 was considered an independent variable in the equation where it was not a component part of the dependent variable.
Multiple indicators of each of the five dimensions were collected. The preadmission and admission dimension was measured with three questions on the use of standardized versus unstandardized forms and whether admission criteria were in written form. These items were combined into one measure (range 0–3). For example, a facility that used standardized forms during both preadmission and admission and also admitted residents according to written admission criteria would score 3. Higher scores indicated better care delivered.
Staff training and education was measured with a series of questions on the provision or nonprovision of types of training and not length of training time or quality of instruction and training materials. DoNs were asked whether each of care aides, licensed practical nurses (LPNs), registered nurses (RNs), and support staff received each of the following types of training or education (yes or no): general care for residents with dementia, management of inappropriate behaviors, role of the family, stress reduction techniques, safety issues, and off-site dementia training. Because only 40% of facilities had LPNs, it was necessary to standardize for number of staff types employed. For each type of education, a proportion was calculated with the number of staff types receiving each type of education as the numerator and total staff types employed by the facility as the denominator. (For example, the denominator for any given type of education was 3 for a facility that did not employ LPNs but did employ RNs, care aides, and support staff.) These were multiplied by 100 to produce a percentage score, and the six variables were summed to create an education dimension. Scale range was 0–600 (
).
Nonuse of physical restraints was measured with a question asking whether the facility used any of 11 specified physical restraints (ankle cuff, bed rails, Dutch doors, geri-chair, isolation, lap belt, posey vest, seclusion, sheet restraint, wrist restraint, and wheelchair tray) during the previous year for the purpose of behavioral management. Nonuse of pharmacological restraints was measured identically, with questions concerning the use of 17 commonly used psychotropic drugs in three categories (antidepressants: Aventyl, Desyrel, Elavil, Luvox, Paxil, Prozac, and Sinequan; anxiolytics: Ativan, Restoril, Rivotril, Serax, and Xanax; neuroleptics: Haldol, Loxapac, Mellaril, Orap, and Risperdal) during the previous year specifically for the management of behavioral difficulties. The purpose was not to distinguish which drugs were medically necessary or not, but to determine how many of these drugs were used as chemical restraints in any given facility. A confirmatory factor analysis produced one factor that included 10 of 17 drugs (Desyrel, Elavil, Luvox, Prozac, Restoril, Rivotril, Xanax, Loxapac, Mellaril, and Orap) and 10 of 11 physical restraints (lap belts were excluded). Facilities received a score between 0 and 20 (
). For example, if a facility used 5 of the 10 drugs and 3 of the 10 physical restraints, it received a score of 8. The factor analysis produced a scale that included both physical and chemical restraints, contrary to Sloane and colleagues 1991, who argued that they are separate dimensions. Unlike the other four dimensions, a higher score indicated poorer quality of delivery.
Flexible care routines and resident-relevant activities were treated as one dimension because of considerable conceptual overlap. Two questions measured this dimension: whether facilities integrated activities in day-to-day care and whether support staff received instruction in activation techniques. Less meaningful data were elicited than had been expected, thus limiting the development of extensive categories for this dimension. Possible scores ranged from 0 (the facility does neither of these) to 2 (the facility does both). (Several other questions were dropped because they were not scalable and did not relate singly or summed to any of the outcome measures.) This dimension is referred to as flexible care hereafter.
To measure the environmental dimension, research assistants completed a modified version of the TESS in each of the participating facilities. The TESS is designed to "evaluate the appropriateness of a nursing home unit for residents with dementing disorders" (Sloane and Mathew 1990). Instrument C was the most current version of the TESS available at the time of study commencement. The instrument consists of scales designed to assess general design, maintenance, lighting, noise, residents' rooms, and programming orientation. Scales were summed to calculate a total environmental score for each facility. Possible total scores ranged from 1 to 154 (
).
Because the five dimensions addressed issues more or less objectively, an additional measure of quality of care captured the subjective, intangible factors that may influence outcomes. Experts working in the area of dementia and long-term care, but not employed directly by facilities involved in dementia care, were consulted. They were asked to provide a list of "gold standard" facilities. That is, they were asked where they themselves would choose to go or where they would place a loved one, should the need arise. Five such panels were convened, each representing a contiguous set of health units and together covering all facilities in the province. They were able to select gold standard facilities without difficulty. Of the 152 units included in the study, 62 were considered gold standard by the panel.
Table 1 shows variable means, standard deviations, coding, and directions of variables.
|
Multiple linear regression models were then estimated with each of the outcome change scores (t1 - t2) as the dependent variables. Control variables included four facility characteristics not included in the dimensions—ownership (private or public), number of intermediate care facility (ICF) beds, percentage of the resident population with dementia, and percentage of demented residents with severe dementia—and two resident characteristics not included in the dimensions—age and sex (see Porell, Caro, Silva, and Monane 1998). Because the change score (t1 - t2) was correlated with the t1 score, the t1 score was forced into the equation first to determine its contribution to explaining the variance in the dependent variable. Assumptions of linearity, collinearity, and normality were met, with one exception. Change in social skills was skewed with only 0.3% scoring 0 and 65.8% scoring the maximum of 11. Transformations such as squaring all values did not significantly change the distribution. Results for social skills should, therefore, be interpreted with this in mind.
Missing data were nonproblematic; 0.6%–5% for all outcome variables except affect. For affect it was 13% (
). Analyses rerun with only those cases for which there were complete data revealed the same results as when substitution of the mean was used. Therefore, results from the latter procedure are reported.
|
Results |
|---|
|
TOPAbstractReview of the LiteratureMethodsResultsDiscussionReferences |
|---|
The average age of participants at t1 was 82 years. Sixty-seven percent of the remaining sample was female. Twenty-one percent of the participants were born in British Columbia, 46% elsewhere in Canada, 18% in the British Isles, and 14% elsewhere in the world. English was the first language of 85%. Approximately half (52%) completed high school or higher (this is comparable to educational attainment for all British Columbians 65 years and older). The median household income for these residents was between $15,000 and $19,999 (similar to the Canadian median for persons aged 65 and older).
One-sample t tests showed that the mean change scores for the five outcomes plus cognitive functioning were significantly different than 0 for five of the six measures (agitation being the sole exception; see Table 2 ). In all five instances residents declined.
|
, p < .01; data not shown in table).
|
, p < .05), affect (
, p < .05), and cognitive function (
, p < .01) among residents. Facility type was also significantly correlated with three of the outcomes. SCU residents were more likely to experience decline in physical functioning (, p < .01), expressive language skills (
, p < .01), and social skills (, p < .05). The remaining four dimensions were related to one outcome only: the greater the use of restraints, the less the decline in expressive language skills (
, p < .05); the more training and education staff received, the less the increase in agitation (
, p < .05); the better the environment, the greater the improvement in mood (
, p < .05); and the more flexible care provided, the less decline in social skills (, p < .05).
|
To ascertain whether the bivariate relationships between the dimensions and outcomes were explained or conditioned by other factors, we performed multiple regression analyses (Table 5 ). It can be noted first that t1 of the dependent variable was statistically related to change in that variable for all of the outcomes. This is hardly surprising. Of more interest is the fact that t1 ranged from a beta of -.733 for affect (i.e., affect at t1 was highly predictive of the change that took place in affect over a 1-year period) to a beta of .228 for social skills (i.e., social skills at t1 explained only 5% of the variance in the change that took place in social skills over a 1-year period). In other words, t1 varied in its predictability of change depending on the outcome examined. For social skills, cognitive function, physical functioning, expressive language skills, and agitation, residents doing better at t1 deteriorated more over the course of the year, suggesting those doing worse had less distance to fall. For affect, those doing worse at t1 deteriorated more by t2.
|
For agitation, affect at t1 was the only significant predictor. Residents with better affect at t1 experienced less of an increase in agitation at t2.
For social skills, cognitive function at t1, agitation at t1, physical functioning at t1, affect at t1, expressive language at t1, and flexible care were significant predictors. Residents with better affect at t1, better cognitive function at t1, less agitation at t1, better physical functioning at t1, and better expressive language skills at t1 and residents in facilities having less flexible care showed less decline in social skills by at t2.
Other than physical functioning at t1, none of the independent variables predicted decline in physical functioning over 1 year.
Residents in facilities with poorer assessment procedures, in facilities chosen as gold standard by the panel, in public facilities, and in facilities with fewer beds deteriorated less in affect.
Finally, for expressive language skills, residents who had better cognitive function at t1, better affect at t1, and worse physical functioning and who were women exhibited less decline in expressive language at t2.
Staff education, use of restraints, whether the unit was an SCU, percentage of residents with dementia or severe dementia, and age of resident were all unrelated to any of the outcomes in the multivariate analyses.
|
Discussion |
|---|
|
TOPAbstractReview of the LiteratureMethodsResultsDiscussionReferences |
|---|
; Phillips et al. 1997; Williams and Rees 1997). There is no uniform distinction in terms of the implementation of various dimensions of quality of care in SCUs compared with non-SCUs. In other words, many non-SCUs are implementing a similar quality of care to that found within SCUs. Many SCUs are similarly implementing best practice dimensions of care. These findings are especially important in light of a general assumption that SCUs necessarily provide better care.
We used the panel selection, an original approach, to tap a subjective dimension of care that eludes objective measurement. Panel selection was related to affect only. Individuals in facilities chosen by the panels as gold standard tend to show less decline in positive affect, suggesting that these panels assessed the mood of the residents with some accuracy.
Turning to the individual dimensions, none is highly predictive of resident outcomes. Staff education and the use of restraints are unrelated to any outcome. Environmental features and flexible care are related to one outcome each: change in cognitive function and change in social skills, respectively. In the latter instance, the effect is opposite to that predicted. Assessment is related to two outcomes: change in cognitive function and change in affect; those with worse assessment procedures show less deterioration in their residents. In other words, the individual dimensions have little overall predictive value and, in some instances, are related opposite to what would be expected.
Outcome characteristics of the residents at t1, other than the one being examined as the dependent variable, are most predictive of change over time for four of the outcomes. Most notable is affect at t1, which predicts all outcomes examined here, except physical functioning. That is, over time, affect predicts outcomes at t2, although other outcomes at t1 do not influence affect. To the extent that resident affect is alterable and better affect among residents directly impacts their cognitive function, agitation, social skills, and expressive language, it should be of interest to care providers.
In addition, residents in smaller facilities and public facilities deteriorate less over time in affect. This may contribute to the generally better public image of such institutions in Canada. Women decline less in expressive language than men, consistent with gender differences in this skill earlier in life.
A word of caution must be sounded about the measurement used here. Measurement of the dimensions could be strengthened. For example, staff education consisted of a reporting of whether training was provided. Evidence of actual staff skills would be a stronger measure. Similarly, flexible care was a restricted measure based on summary reports of facility representatives. Evidence of actual resident-specific care patterns and activities would be an improvement. To measure use of restraints, we asked facility representatives about the facility's general approach instead of using evidence of restraints from resident charts or observation. That is, more direct patient-specific measurement could reveal stronger relationships between care provided and resident outcomes.
Also, these data refer only to newly admitted residents suffering from dementia. Furthermore, this resident sample included only those who had medical chart confirmation of their dementia diagnosis. Those lacking chart confirmation were excluded. In addition, although units (SCUs and non-SCUs) were chosen randomly, residents could not be randomly assigned to SCUs and non-SCUs. The sample, therefore, was a convenience sample of necessity. Nevertheless, the fact that residents in SCUs could differ systematically from those in non-SCUs and in ways (unknown) that influence the results must be acknowledged.
This research suggests that measurement of quality of care needs further development. Apparently, so many dimensions go together to result in the care provided to individuals that measures of single dimensions are inadequate. Furthermore, as others have also noted, when facilities implement best practices as currently known on one dimension, they may not implement best practices on other dimensions. In other words, most facilities do well implementing high quality of care on some dimensions; very few do so along all dimensions. This fact may be preventing strong predictive value because it may require high scores on many, or most, of the dimensions in order to see a difference in the outcomes for the residents.
Nevertheless, this study does not support the common belief that dimensions of care cluster within SCUs. It also adds to a growing body of literature that suggests the existence of weak relationships between SCUs and quality outcomes and, more generally, between facility characteristics and quality outcomes. Although the proportion of variance explained in this Canadian study is extremely small, it is this very deficiency (of statistical significance) that adds to its importance. It confirms research in the United States that neither SCUs nor dimensions of care that are believed to reflect best practices are related to resident outcomes.
|
Acknowledgments |
|---|
Received for publication June 25, 1999. Accepted for publication February 18, 2000.
|
References |
|---|
|
TOPAbstractReview of the LiteratureMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. J. Jalbert, L. A. Daiello, and K. L. Lapane Dementia of the Alzheimer Type Epidemiol. Rev., November 1, 2008; 30(1): 15 - 34. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Hyde, R. Perez, and B. Forester Dementia and Assisted Living Gerontologist, December 1, 2007; 47(suppl_1): 51 - 67. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Colin Reid and N. L. Chappell Staff Ratios and Resident Outcomes in Special Care Units: Do Activity Aides Make a Difference? Journal of Applied Gerontology, February 1, 2003; 22(1): 89 - 103. [Abstract] [PDF]
|
|
|
|
|
|
J. B. Tornatore and L. A. Grant Burden Among Family Caregivers of Persons With Alzheimer's Disease in Nursing Homes Gerontologist, August 1, 2002; 42(4): 497 - 506. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
