This Article Abstract Full Text (PDF) Services Download to citation manager Citing Articles Citing Articles via HighWire PubMed PubMed Citation

How Good Is Assisted Living? Findings and Implications From an Outcomes Study

¹ Cecil G. Sheps Center for Health Services Research
² School of Social Work
³ Department of Family Medicine, University of North Carolina at Chapel Hill.
⁴ Department of Sociology and Anthropology, University of Maryland, Baltimore County.
⁵ Department of Epidemiology and Preventive Medicine, University of Maryland, Baltimore.
⁶ Department of Health Policy and Administration, University of North Carolina at Chapel Hill.

Address correspondence to Dr. Sheryl Zimmerman, Program on Aging, Disability, and Long-Term Care, Cecil G. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill, 725 Airport Rd., Campus Box 7590, Chapel Hill, NC 27599-7590. E-mail: Sheryl_Zimmerman{at}unc.edu

	Abstract

TOP Abstract Methods Results Discussion References

 
Objectives. The objective of this work was to determine 1-year medical outcomes, nursing home transfer, and functional change of assisted living (AL) residents and their relationship to care.

Methods. On-site interviews and observations regarding the status and care of 2,078 residents in 193 facilities across four states were conducted; follow-up was by telephone interview with care providers.

Results. Annual mortality and transfer rates were 14.4 and 21.3 per 100 residents. The probability of hospitalization and new/worsening morbidities over a standardized quarter per 100 residents was 12.7 and 22.7. Standardized change in function was notable among those who were transferred or died and small among others. Facility characteristics did not generally relate to medical outcomes and transfer, and those that related to functional change were small and occurred across multiple functions. Facilities that are affiliated with another level of care were more likely to transfer; nurse staffing was favorable for hospitalization but not transfer; and aide turnover was protective for mortality.

Discussion. No single component defines "good" AL care. Predictors and outcomes are inconsistent, and effect sizes are small. Therefore, practice and policy should not focus narrowly on any one area or restrict the type of care—this being welcome news that supports diversity to accommodate individual preferences.

ASSISTED LIVING (AL) is a term applied to a wide array of residential facilities for older adults. Broadly speaking, AL includes all group residential programs not licensed as nursing homes (NHs) that provide personal care in activities of daily living (ADLs) and can respond to unscheduled needs for assistance (Kane & Wilson, 1993). In a more restrictive sense, it refers to the values underlying the manner in which that care is provided, including providing a choice of services and lifestyles and the right to negotiate risk associated with that choice, minimizing the need to move, and ensuring that AL, with its residential emphasis, avoids the characteristics of an institutional setting. In addition, AL is meant to allow for variety, thereby avoiding a "one-size-fits-all" approach to care (Assisted Living Quality Coalition, 1998). Consumers and developers have embraced the concept of AL, and they now provide care for almost 1 million disabled and older adults in > 36,000 facilities nationwide (Mollica, 2002). Only a concept two decades ago, AL has accounted for approximately 80% of new projects in the senior housing industry and is projected to grow as much as 40% in the next 20 years (Adler, 1998; National Investment Conference, 2001).

The diversity inherent in the conceptualization of AL has resulted in diversity in its practice. AL facilities range from those with low services and low privacy (27%) to high services and high privacy (11%); overall costs (not specific to these facilities types) range from $1,338 to $7,130 per month (Hawes, Phillips, Rose, Holan, & Sherman, 2003). Categorized another way, AL facilities include those that are freestanding with 30 units (71% of facilities) or 30 units (14%), as well as facilities associated with congregate care (3%), a continuing care retirement community (7%), or skilled nursing care (5%) (Golant, 2004).

Although the diversity of AL enables consumer choice, it has raised concerns regarding the quality of care. AL facilities are understaffed and staffing requirements are minimal, while at the same time, they are responsible to care for residents with increasingly complex medical and physical conditions. Concerns about care are valid, as a recent report of the General Accounting Office (1999) found that 27% of surveyed facilities had been cited for five or more quality-of-care or consumer protection deficiencies or violations during 1996 and 1997. Because AL is not federally regulated, state regulators rallied in response to these concerns, and between 2000 and 2002, virtually every state developed regulatory models as variable as the states themselves (Mollica, 2002). The unfortunate consequence of these regulations is that although they are intended for consumer protection, they stifle some of the variability that is fundamental to AL. Even more unfortunately, they are being imposed absent data indicating which components of care are and are not related to resident outcomes.

Outcomes of most interest to policy-makers relate to the balance of care provision and need, as reflected in medical conditions (Kissam, Gifford, Mor, & Patry, 2003). The essential issue is that residents not be allowed to "age in place" if the facility is not able to provide adequate care. The matter is not that straightforward, however, as 19 states allow for the completion of negotiated risk agreements that expressly allow residents to accept certain risks associated with reduced care, so as to maximize their preferences and remain in the facility (Mollica, 2002). For consumers, aging in place may well be the outcome of greatest interest, as AL developed in part because they wanted to avoid NH placement (Kane & Wilson, 2001).

To date, the majority of dialogue related to the quality of AL has compared it with NH care (Golant, 2004). Only two studies have been published comparing outcomes within different types of AL facilities. Phillips, Munoz, Sherman, Rose, Spector, and Hawes (2003) reported on departures from AL but limited the examination to the 40% of AL facilities with > 10 beds that did not have any rooms housing three or more unrelated persons (privacy), offered assistance with at least two of three prescribed ADLs (service), and did not evidence both low privacy and low services otherwise defined. Whereas they did examine the effect of multiple facility characteristics, the analyses were an examination of limited outcomes in a prescribed type of AL facility. Hedrick and colleagues (2003), on the other hand, studied more diverse types of AL facilities (those known as adult family homes, adult residential care, and AL) across more outcomes but limited the examination of components of care to only facility type.

Given the growth and variability in AL and the increasing state regulation that is prescribing the nature of AL care, it is imperative that research provides data to influence these decisions. This article will inform the policy and practice debate by presenting the results of one of the largest outcome studies of AL to date, using data from 2,078 residents in 193 diverse AL facilities across four states who were followed for 1 year. The primary aim is to determine medical outcomes (mortality, morbidity, hospitalization), NH transfer, and change in function (in ADLs, cognition, behavior, affect, social functioning and withdrawal) and how they relate to the structure and process of care (facility type, administration, staffing, environmental factors, and facility policies) in AL facilities. Findings will determine whether some components of AL care are more favorable than others in reference to outcomes, thereby allowing for a more informed discourse on the relative merits of different types of AL care.

	METHODS

TOP Abstract Methods Results Discussion References

 
Design and Sample
Data for this study were derived from the Collaborative Studies of Long Term Care (CS-LTC). In keeping with the broad focus, the CS-LTC defined AL as facilities or discrete portions of facilities, licensed by the state at a non-NH level of care, that provide room, board, 24-hour oversight, and assistance with ADLs. A multilevel sampling frame was used to select facilities and residents for participation. Within each state, a region of counties was identified that represent the state. The sampling frame consisted of all licensed AL facilities within each region, stratified into three types to ensure that the sample reflected the broad range of AL: (1) facilities with < 16 beds, (2) facilities with 16 beds of the "new-model" type (i.e., meant to capture the type of facilities proliferating under the recent surge of AL and having criteria identified in a pilot study that differentiate them from other facilities [being built after January 1, 1987, and having at least one of the following components: at least two different monthly private pay rates; 20% of the residents requiring assistance with transfer; 25% of the residents who are incontinent daily; and either a registered nurse (RN) or licensed practical nurse (LPN) on duty at all times]), and (3) "traditional" facilities with 16 beds, not meeting new-model criteria.

The study was designed to select approximately equal numbers of residents from each facility type; as a result, more of the smaller facilities (n = 113) were studied compared with the other facility types (n = 40 traditional and 40 new-model facilities). Facilities were randomly selected for participation; the overall recruitment rate was 59%. Differences between participating AL facilities and others indicated that nonparticipating facilities had more owners working more hours in the facility, had more rate levels, and housed a slightly less impaired resident population. There were no differences in reference to proprietary status; affiliation with other long-term care facilities; facility age, size, or occupancy rate; and resident age, ethnicity, or race. Further details about the CS-LTC sampling and data collection procedures are available elsewhere (Zimmerman et al., 2001).

Baseline data were collected from October 1997 through November 1998; 2,078 residents from 193 AL facilities across Florida, Maryland, New Jersey, and North Carolina (all of which exhibited well-developed yet different types of AL industries) participated in this study. In smaller facilities, all residents 65 years or older were asked to participate; in larger facilities, residents were randomly chosen to a maximum of 20 subjects. Consent was obtained from residents or a responsible party, and the overall participation rate was 92%. Residents and primary care providers were interviewed on site to obtain information regarding resident status. Administrators were interviewed on site and observations conducted to document the structure and process of care. Follow-up data regarding medical events and NH transfer were collected through quarterly telephone contact with the staff caregiver most familiar with the resident; information regarding functional change was collected at 1 year. When a resident died or was discharged before 1 year, a rating of the function in the 2 prior weeks was obtained.

Facility-Level Measures
Modifiable facility-level components of care under study were those significant in NH research and of theoretical importance. In addition to basic administrative data (e.g., ownership, affiliation with another level of care, staffing), the administrator provided information on the process of care, responding to a modification of the Policy and Program Information Form (POLIF) (Moos & Lemke, 1996). The POLIF includes measures that assess philosophical policies in three broad areas: requirements for residents, individual freedom and institutional order, and provision of services and activities. Six scales were derived that substantially follow the algorithms developed and tested by Moos and Lemke: acceptance of problem behavior, policy choice, policy clarity, provision for privacy, resident control, and availability of social and recreational activities. New measures were developed to assess overall admission policies (a sum of the number of resident characteristics or conditions that are accepted for admission into a facility; =.84); admission policies specific to ADL functioning (focusing on a subset of conditions; =.77), overall provision of services (including on- and off-site service provision; =.80), and provision of health services (the subset of services related to medical care; =.74). Each aggregate measure calculates the percentage of component items that are positively scored; if > 25% of the items for a particular measure were missing, the measure was not calculated. For facilities missing some items but < 25% of items, a percentage of positively scored nonmissing items was calculated. Analyses were conducted relative to a 10-point change in the aggregate measure. Finally, data related to environmental characteristics were derived from observation using the Therapeutic Environment Screening Survey–Residential Care (TESS-RC). This measure is a refinement of the TESS–Nursing Home (TESS-NH) (Sloane et al., 2002) and is completed through a structured observation of the facility. The TESS-RC yields a summary score, the Assisted Living Environmental Quality Score (AL-EQS), a 15-item scale including items such as facility cleanliness, hominess, and privacy ( =.75). Neighborhood attractiveness was rated on a 4-point Likert scale, from very unattractive (e.g., other homes/buildings/landscapes in very bad repair/maintenance) to very attractive (e.g., very clean/exceptionally well maintained).

Resident Function (Baseline and 1-Year Outcomes)
Data related to six functional areas were collected at baseline and follow-up from care providers. Impairment in physical ADLs was measured with items replicating the Minimum Data Set ADL Self-Performance Index (MDS-ADL), a 10-item categorical scale of ADL dependency (Morris, Fries, & Morris, 1999). The MDS-ADL assigns a score based on dependency over the last 7 days in bed mobility, eating, locomotion, transfer, toileting, dressing, and personal hygiene; scores range from 0 (independence/no assistance) to 4 (total dependence). Ratings of cognitive impairment replicated items from the Minimum Data Set Cognition Scale (MDS-COGS). Scores of 2 on the MDS-COGS are indicative of cognitive impairment, and scores of 5 indicate severe impairment (Hartmaier, Sloane, Guess, & Koch, 1994). Behavioral problems were measured using the 14-item version of the Cohen-Mansfield Agitation Inventory (CMAI), a scale that identifies the frequency of reported agitated behaviors over the last 2 weeks (Cohen-Mansfield, 1986). Depression was assessed by care provider report on the Cornell Scale for Depression in Dementia (CSD-D), an observer-rated scale of depressive symptomatology designed to rate depression in persons with dementia; the Cornell scale consists of 19 items, each scored 0–2, with scores of 8 thought to be indicative of depression for people with dementia (Alexopoulos, Abrams, Young, & Shamoian, 1988). Questions regarding social function in the preceding 7 days were based on the Assisted Living Social Activity Scale (AL-SAS) (Zimmerman, Scott, et al., 2003). The AL-SAS items reflect participation (yes/no) in 11 activities including private activities, group activities, and outings. The analyses reported here include six other activities, as well. The resulting score was reversed so that a higher score indicates the number of activities in which the resident did not participate. Social withdrawal was assessed with an eight-item subscale of the Multidimensional Observation Scale for Elderly Subjects (MOSES), which measures contact with and interest in people, events, and activities over the last week, rating passivity and apathy on a 4-point scale (Helmes, Csapo, & Short, 1987).

Resident Medical Outcomes and NH Transfer
Whether or not transfer and a medical event occurred was documented per quarter over the year of follow-up. Medical events included mortality, hospitalization (excluding psychiatric), and new or worsening morbidity.

Resident-Level Control Variables
These data, collected from the resident or knowledgeable caregiver, include demographic information, diagnosis of Alzheimer's disease or other dementias, and co-morbidities, noting the presence/absence of 31 conditions.

Analyses
Longitudinal analyses were performed to assess the effects of facility characteristics on the incidence of mortality, hospitalization, morbidity, and NH transfer, as well as their effects on functional change over 1 year. The longitudinal models were fit separately for each facility characteristic (i.e., in any given model, only one facility characteristic was represented as an independent variable). All models accounted for clustering within the facility.

Mortality and NH transfer were viewed as endpoint events and thus assessed using Cox proportional hazards methods (Cox & Oakes, 1984). Hospitalization and morbidity, measured on a quarterly basis, were assessed using repeated measures analysis. Here, generalized estimating equations were employed to fit a Poisson regression model (Liang & Zeger, 1986). Relative risks associated with mortality, hospitalization, morbidity, and NH transfer were estimated using the facility characteristic as the exposure variable and resident-level covariates (i.e., age, gender, race, marital status, education, tenure in facility, MDS-ADL, dementia diagnosis, MDS-COGS, CSD-D, and co-morbidities) to adjust for case-mix differences among facilities. A relative risk > 1 indicates that the rate in the exposed group (i.e., for dichotomous variables, this is the presence of the characteristic, such as chain affiliation, and for continuous variables, this is either a per-unit increase or the value at 1 SD above the mean) is higher than in the referent group (i.e., absence or 1 SD below the mean); values lower than 1 indicate that the rate is lower in the exposed group than the referent.

Functional change was assessed using generalized estimating equations to estimate the difference between the baseline and the follow-up score, assuming a Gaussian distribution and identity link function. The outcome measures were standardized according to the baseline distribution of that measure, allowing the effect sizes to be compared between different functions. The estimated effect sizes are thus expressed in SD units (Samsa, Edelman, Rothman, Williams, Lipsomb, & Matchar, 1999). The traditional benchmarks for standardized effect sizes are.2 for small effects,.5 for moderate, and .8 for large effect sizes (Cohen, 1977). For every outcome variable, a negative sign indicates that the referent group had a greater decline than the other category(ies). Only those who survived in the facility to 1 year were included in the functional change analyses. Covariates were not included in these models because the change assessed was within the same resident at the two time points, thereby minimizing the potential confounding effects of case-mix differences among facilities.

	RESULTS

TOP Abstract Methods Results Discussion References

 
Structure and Process of Care
Table 1 shows the distribution of facility characteristics. The majority of facilities were for-profit (82.7%); one fourth (24.2%) were affiliated with a NH, hospital, or continuing care retirement community. Almost one-half (47.1%) had an RN or LPN on staff, the mean number of RN care hours per resident per day being.11 (SD =.31); also, the average annual turnover of aides was 67.7% (SD = 97.8). The majority of neighborhoods were attractive (52.2%) or very attractive (30.0%). The reported process of care area that was most highly endorsed was latitude in admission policies (e.g., an average of 72.6% of resident characteristics were allowed admission), and the area least endorsed was resident control (e.g., an average of 28.2% of items were endorsed that allowed for resident control).

	DISCUSSION

TOP Abstract Methods Results Discussion References

The most important finding of this work is that residents in smaller AL facilities (operationalized as those with fewer than 16 beds and that, in fact, had an average bed size of 8.9 [SD 3.6]) and "traditional" AL facilities (those that did not meet the definition of the "new-model" of AL care) did not fare worse in reference to medical outcomes, NH transfer, or functional decline compared with residents in new-model AL. In fact, functional outcomes of residents in new-model facilities were less favorable than those for residents in other facilities. They showed more ADL decline (.35-SD standardized decline) than those in smaller facilities and more social and withdrawal decline (.23- to.49-SD standardized decline) than residents in both other facility types. Also, the mortality findings showing that traditional facilities are superior to small facilities (rr = 0.66) further question the comparative merits of the new-model facilities. Whereas one might surmise that more decline among those in new-model facilities is due in part to an initial higher level of function (and therefore more room for decline), such is not the case (i.e., they are nonsignificantly less impaired than residents in smaller facilities in ADL function; Zimmerman, Gruber-Baldini, et al., 2003). Thus, these data speak against the romanticization of modern AL, a finding of critical importance for the practice and regulation of AL. Already, small facilities have folded under the weight of regulations (despite the fact that residents report more satisfaction in these facilities) (Sikorska, 1999); if regulatory efforts are not quickly informed by the relative merits of different types of AL care, then the future of AL will be determined not by the quality of care but by regulatory and market pressures (Zimmerman, Sloane, & Eckert, 2001). Currently, there are countless variants of AL across the United States catering to different clients, and each should be evaluated before it is regulated out of existence.

Reverting for a moment to the descriptive findings to make another point about differences across settings of care, the annual mortality rate in this study (14.4%), combined with the NH transfer rate (21.3%), is higher than the 19% departure rate (over 7 months) found by Phillips and colleagues (2003). This difference likely reflects the effect of the sampling frame on the results obtained as well as state differences. Other work by this investigative team found that residents in smaller facilities (those most likely to be excluded from the Phillips study) exhibited more cognitive, functional, and behavioral symptoms than those in other facility types (Zimmerman, Gruber-Baldini, et al., 2003). Thus, it cannot be overstated that in a field as diverse as AL, it is essential to consider the population from which the data were obtained when interpreting results.

Three points are especially clear in reference to functional change witnessed within this AL cohort: (1) disposition (i.e., death or discharge to a NH) related to the degree of change; (2) for those who died or were discharged, change was evidenced across more than ADLs; and (3) change among those who remained in AL over 1 year was small. Specifically, mean and standardized changes in all areas were greatest for those who were later transferred or died, and effect sizes were large for change in ADLs, affect, social function, and social withdrawal. Therefore, AL staff may be advised to attend to these indicators of quality of life as residents decline, and interventions may be warranted that might reduce rates of transfer or delay death. Some promising work has already been done in AL to reduce depression and learn how to promote physical activity (Jones, 2003; Mihalko & Wickley, 2003).

In AL, medical outcome indicators (mortality, morbidity, and hospitalization) and NH placement were affected by some components of care, but the effect was not consistent across indicators or reflected among the majority of components of care examined in this study. Similarly, functional change related to components of care, but not consistently so. The primary message, then, similar to the more macro finding of facility type, is that there is no one component of AL as it is currently practiced (across the four states in this study, selected to represent the variation in AL practice and regulation) that constitutes "bad" care. This point notwithstanding, the strength of a few findings merits further discourse:

Transfer, Hospitalization, and Mortality as Related to Policies, Administration, and Staffing
Facilities that have more restrictive admission policies, are affiliated with another level of care, or have RNs/LPNs on staff were more likely to transfer residents to NHs, controlling for resident characteristics (rr = 0.88, 1.93, and 1.50, respectively). In the first case, the fact that reported policies related to observed practices is an important verification, one that has been surprisingly absent from the literature (see, e.g., Chapin & Dobbs-Kepper, 2001). As for the second, it may come as no surprise that AL facilities affiliated with a NH or continuing care retirement community were more likely to transfer residents, and the ability to transition may, in fact, have been known by residents at the time of admission. The finding related to nursing is less intuitive, however, and runs counter to that reported by Phillips and colleagues (2003). Although the analytic techniques and variables across these studies differed, a discrepancy as marked as this calls for further examination.

What is likely less apparent to consumers, and is relevant for the evolving field of AL and workforce issues, is that facilities that have more RN care hospitalized their residents less (rr = 0.99), perhaps by avoiding the need for hospitalization (or favoring NH transfer). Although there are multiple roles for the nurse in AL (e.g., overseeing medication management, administering skin care treatments, conducting assessments), some states do not require that a nurse be on site (Mollica, 2002). An impediment to defining the role and scope of nursing practices in AL has been a lack of research on outcomes related to nursing care (Wallace, 2003), and this study demonstrates a tangible benefit of nursing care in AL.

The importance of staffing is further reflected in the finding that aide turnover was protective for mortality (rr = 0.96). Additional analyses revealed a decreased risk for residents in facilities with any turnover relative to the 28% of the sample who were in facilities with 0% turnover. Although this finding initially may seem to challenge common wisdom, some report that the most motivated, caring staff leave sooner than others because the work and working conditions are challenging and interfere with care provision (Tellis-Nayak & Tellis-Nayak, 1989). Alternatively, it is possible that innovations in care are not pursued in facilities with low turnover. Regardless of the explanation, outcomes of care should be carefully considered before deciding that a component connotes poor care at face value.

Behavioral Decline as Related to Administration, Staffing, and Admission Policies
Residents in facilities that are nonprofit, have an RN/LPN on staff, and have more restricted admission policies exhibited more standardized behavioral decline than those in other facilities (standardized decline averaged.27–.28 SD). Two of these findings may relate to the distribution of resident status at baseline. First, case mix in AL differs by proprietary status (Zimmerman, Gruber-Baldini, et al., 2003), with residents in nonprofit facilities having slightly lower rates of behavioral impairment. Although the findings reported here are standardized and reflect each resident's change relative to the baseline level, residents in nonprofit facilities have more room for additional decline. The finding related to admission policies can be similarly understood: Residents admitted into facilities with more restrictive policies have more room for subsequent decline. Perhaps what is most noteworthy in these findings is that nonprofit facilities did not discharge the residents as their behavioral impairment increased. This is an important finding for aging in place.

Other findings from these analyses are suggestive but not as marked, and the conclusions to be drawn are less clear. For example, regarding morbidity related to resident control, while effect sizes were small, every 10-point increase (of 100 points) in allowable control (e.g., resident involvement in regular meetings, in planning or policy) translated to a 5% increased risk of new or worsening morbidities, controlling for the resident's initial status. This relationship is not necessarily causal, and it is likely that it can be explained by other factors. For example, it could be that residents who are more involved in facility life are better known to staff, thereby making ascertainment of medical status more thorough. Alternately, negotiated risk may be in play, and facilities that allow more control may also allow residents more of a voice to refuse recommended care. Thus, just as presumed indicators of "good" or "bad" care should be judged on their merits, so, too, should presumed "bad" outcomes.

Environmental factors matter, and residents in facilities that are located in an attractive neighborhood experienced fewer incident or worsening morbidities (rr = 0.83) and were one-half as likely to be transferred to a NH (rr = 0.54). This extrinsic factor may serve as a proxy for resident differences not able to be analytically controlled (e.g., these residents may have a more privileged background and a better prognosis due to a life of better health habits), thereby constituting a community effect (Fennel, Miller, & Mor, 2000). Additional analyses indicate that compared with facilities in unattractive neighborhoods, residents in facilities in attractive neighborhoods had a significantly lower Medicaid case mix (i.e., 45% of residents in attractive neighborhoods compared with 66% lived in facilities that had > 5% of their residents on Medicaid; ² = 153.4, p <.001). Thus, facilities in attractive neighborhoods are less likely to transfer, in part because their residents are better able to continue paying the fees to live there. The policy implication herein is that aging in place can be facilitated if economic constraints are lessened. It is clear that the explanation is not entirely economic, however, as analyses run with both neighborhood and Medicaid case mix indicated that each had a significant independent effect (rr = 0.48 for neighborhood [p =.015] and 1.35 for Medicaid [p =.034]).

Finally, it is worth considering the components of care that did not make a consistent difference in terms of outcomes that have been (inconsistently) implicated in other work. It is possible that they were not significant because the number of events or change over 1 year was too small to allow their effect to be detected; if true, it is unlikely that change so small would have been clinically significant even if it were related to care. Further, if components of care become significant over time—over 2 years of follow-up, for example, during which time more clinically meaningful change would occur—then the stability of the care environment must be carefully considered. Other design issues need to be considered as well. Most notably, in this cross-sectional cohort, it is likely that facility effects were more notable at or shortly after the time of admission. Also, resident status was obtained by care provider report, the reliability and validity of which were not ascertained. Finally, other quality-of-life outcomes such as satisfaction were not part of this analysis, an omission that does not capture the multifaceted nature of quality of life in AL (Mitchell & Kemp, 2000).

In the end, it must be acknowledged that no single component or set of components defines "good" care in AL. Predictors and outcomes are not consistent, and effect sizes are small. Therefore, intervention and policy oversight should not focus narrowly on any one area to the exclusion of others, and there should be no expectation that a single change or set of changes will have global impact. Indeed, practice and policy need not restrict the type of AL available—this being an evidence-based conclusion likely to be welcomed because it supports one of the foundations of AL: to allow diversity to accommodate different styles and preferences.

	Acknowledgments

 
This research was supported by grants from the National Institute on Aging (RO1 AG13871, RO1 AG13863, K02 AG00970). The authors note appreciation for the cooperation of the facilities, residents, and families participating in the Collaborative Studies of Long-Term Care as well as Ms. Verita Custis Buie (University of Maryland, Baltimore) and Dr. Joan F. Walsh (University of North Carolina, Chapel Hill) who provided expert project coordination. Gratitude also is extended for data collection overseen by Betty Concha and conducted by Mary Alice McGurrin, Christine Schmitt, Ida Altman, Joan Bassler, Susan Baxter, Shirley Carter, Betty Dorsey, Diane Eagle, Susan Fallen, David Fallen, Jo Magness, Connie Nunamaker, Ronald Nunamaker, and Barbara Smith (University of Maryland, Baltimore) and for data management performed by Jane Darter (University of North Carolina, Chapel Hill).

	Footnotes

 
Decision Editor: Charles F. Longino, Jr., PhD

Received for publication September 21, 2003. Accepted for publication August 30, 2004.

	References

TOP Abstract Methods Results Discussion References

 

• Adler, S. (1998). Get ready for consolidation: Today's product in the pipeline will build tomorrow's big players. Contemporary Long Term Care, August, 39–45.
• Alexopoulos, G. S., Abrams, R. C., Young, R. C., Shamoian, C. A. (1988). Cornell Scale for Depression in Dementia. Biological Psychiatry, 23,271-284.[Medline]
• Assisted Living Quality Coalition. (1998). Assisted living quality initiative. Building a structure that promotes quality. Washington, DC: Public Policy Institute, American Association of Retired Persons.
• Chapin, R., Dobbs-Kepper, D. (2001). Aging in place in assisted living: Philosophy versus policy. The Gerontologist, 41,43-50.[Abstract/Free Full Text]
• Cohen, J. (1977). Statistical power analyses for the behavioral sciences. New York: Academic Press.
• Cohen-Mansfield, J. (1986). Agitated behaviors in the elderly. II. Preliminary results in the cognitively deteriorated. Journal of the American Geriatrics Society, 34,722-727.[Medline]
• Cox, D. R., Oakes, D. (1984). Analysis of survival data. London: Chapman and Hall.
• Fennell, M. L., Miller, S. C., Mor, V. (2000). Facility effects on racial differences in nursing home quality of care. American Journal of Medical Quality, 15,174-181.[Abstract/Free Full Text]
• General Accounting Office. (1999). Assisted living: Quality of care and consumer protection issues in four states. GAO-HEHS-99-27. Washington, DC: General Accounting Office.
• Golant, S. (2004). Do impaired older persons with health care needs occupy U.S. assisted living facilities? An analysis of six national studies. Journal of Gerontology: Social Sciences, 59B,S68-S79.
• Hartmaier, S., Sloane, P. D., Guess, H., Koch, G. (1994). The MDS Cognition Scale: A valid instrument for identifying and staging nursing home residents with dementia using the Minimum Data Set. Journal of the American Geriatrics Society, 42,1173-1179.[Medline]
• Hawes, C., Phillips, C. D., Rose, M., Holan, S., Sherman, M. (2003). A national survey of assisted living facilities. The Gerontologist, 43,875-882.[Abstract/Free Full Text]
• Hedrick, S., Sales, A. E. B., Sullivan, J. H., Gray, S. L., Tornatore, J., Curtis, M., et al. (2003). Resident outcomes of Medicaid-funded community residential care. The Gerontologist, 43,473-482.[Abstract/Free Full Text]
• Helmes, E., Csapo, K. G., Short, J. A. (1987). Standardization and validation of the Multidimensional Observation Scale for Elderly Subjects (MOSES). Journal of Gerontology, 42,395-405.
• Jones, E. D. (2003). Reminiscence therapy for older women with depression. Effects of nursing intervention classification in assisted-living long-term care. Gerontological Nursing, 29,26-33.
• Kane, R., Wilson, K. B., (Eds.). (1993). Assisted living in the United States: A new paradigm for residential care for frail older persons?. Washington, DC: Public Policy Institute/American Association of Retired Persons.
• Kane, R., Wilson, K. B. (2001). Assisted living at the crossroads: Principles for its future. Portland: Jessie F. Richardson Foundation.
• Kissam, S., Gifford, D. R., Mor, V., Patry, G. (2003). Admission and continued-stay criteria for assisted living facilities. Journal of the American Geriatrics Society, 51,1651-1654.[Medline]
• Liang, K. Y., Zeger, S. L. (1986). Longitudinal data analysis using generalized linear models. Biometrica, 73,13-22.
• Mihalko, S. L., Wickley, K. L. (2003). Active living for assisted living: promoting partnerships within a systems framework. American Journal of Preventive Medicine, 25,193-203.[Medline]
• Mitchell, J. M., Kemp, B. J. (2000). Quality of life in assisted living homes: a multidimensional analysis. Journal of Gerontology: Social Sciences, 55B,S117-S127.
• Mollica, R. (2002). State Assisted Living Policy 2002. Retrieved May 5, 2005, from http://www.nashp.org/Files/ltc_15_AL_2002.pdf.
• Moos, R. H., Lemke, S., (Eds.). (1996). Evaluating residential facilities: The multiphasic environmental assessment procedure. Thousand Oaks, CA: Sage.
• Morris, J. N., Fries, B. E., Morris, S. A. (1999). Scaling ADLs within the MDS. Journal of Gerontology: Medical Sciences, 54A,M546-M553.
• National Investment Conference. (2001). The case for investing in seniors housing and long term care properties. Annapolis, MD: National Investment Conference for the Senior Housing and Care Industries.
• Phillips, C. D., Munoz, Y., Sherman, M., Rose, M., Spector, W., Hawes, C. (2003). Effects of facility characteristics on departures from assisted living: Results from a national study. The Gerontologist, 43,690-696.[Abstract/Free Full Text]
• Samsa, G., Edelman, D., Rothman, M. L., Williams, G. R., Lipsomb, J., Matchar, D. (1999). Determining clinically important differences in health status measures. Pharmacoeconomics, 15,141-155.[Medline]
• Sikorska, E. (1999). Organizational determinants of resident satisfaction with assisted living. The Gerontologist, 39,450-456.[Abstract]
• Sloane, P. D., Mitchell, C. M., Weisman, G., Zimmerman, S., Long, K. M., Lynn, M., et al. (2002). The therapeutic environment screening survey for nursing homes (TESS-NH). Journal of Gerontology: Social Sciences, 57B,S69-S78.
• Tellis-Nayak, V., Tellis-Nayak, M. (1989). Quality of care and the burden of two cultures. The Gerontologist, 29,307-313.
• Wallace, M. A. (2003). Is there a nurse in the house? The role of nurses in assisted living: Past, present and future. Geriatric Nursing, 24,218-221, 235.[Medline]
• Zimmerman, S., Gruber-Baldini, A. L., Sloane, P. D., Eckert, J. K., Hebel, J. R., Morgan, L. A., et al. (2003). Assisted living and nursing homes: Apples and oranges? The Gerontologist, 43,107-117.[Abstract/Free Full Text]
• Zimmerman, S., Scott, A. C., Park, N. S., Hall, S. A., Wetherby, M. M., Gruber-Baldini, A. L., et al. (2003). Social engagement and its relationship to service provision in residential care/assisted living. Social Work Research, 27,6-18.
• Zimmerman, S., Sloane, P. D., Eckert, J. K., Buie, V. C., Walsh, J., Hebel, J. R., et al (2001). An overview of the Collaborative Studies of Long-Term Care. In S. Zimmerman, P. D. Sloane, & J. K. Eckert (Eds.), Assisted living: Needs, practices and policies in residential care for the elderly (pp. 117–143). Baltimore: Johns Hopkins University Press.
• Zimmerman, S., Sloane, P. D., Eckert, J. K. (2001). Emerging issues in residential care/assisted living. In S. Zimmerman, P. D. Sloane, & J. K. Eckert (Eds.), Assisted living: Needs, practices and policies in residential care for the elderly (pp. 317–331). Baltimore: Johns Hopkins University Press.

This article has been cited by other articles:

				J. C. Munn, D. Dobbs, A. Meier, C. S. Williams, H. Biola, and S. Zimmerman The End-of-Life Experience in Long-Term Care: Five Themes Identified From Focus Groups With Residents, Family Members, and Staff Gerontologist, August 1, 2008; 48(4): 485 - 494. [Abstract] [Full Text] [PDF]

				C. A. Giuliani, A. L. Gruber-Baldini, N. S. Park, L. A. Schrodt, F. Rokoske, P. D. Sloane, and S. Zimmerman Physical Performance Characteristics of Assisted Living Residents and Risk for Adverse Health Outcomes Gerontologist, April 1, 2008; 48(2): 203 - 212. [Abstract] [Full Text] [PDF]

				S. Zimmerman and P. D. Sloane Definition and Classification of Assisted Living Gerontologist, December 1, 2007; 47(suppl_1): 33 - 39. [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. A. Kane, J. Chan, and R. L. Kane Assisted Living Literature Through May 2004: Taking Stock Gerontologist, December 1, 2007; 47(suppl_1): 125 - 140. [Abstract] [Full Text] [PDF]

				C. A. Beel-Bates, B. Ingersoll-Dayton, and E. Nelson Deference as a Form of Reciprocity Among Residents in Assisted Living Research on Aging, November 1, 2007; 29(6): 626 - 643. [Abstract] [PDF]

				S. C. Stearns, J. Park, S. Zimmerman, A. L. Gruber-Baldini, T. R. Konrad, and P. D. Sloane Determinants and Effects of Nurse Staffing Intensity and Skill Mix in Residential Care/Assisted Living Settings Gerontologist, October 1, 2007; 47(5): 662 - 671. [Abstract] [Full Text] [PDF]

				S. Zimmerman, C. S. Williams, P. S. Reed, M. Boustani, J. S. Preisser, E. Heck, and P. D. Sloane Attitudes, Stress, and Satisfaction of Staff Who Care for Residents With Dementia Gerontologist, October 1, 2005; 45(suppl_1): 96 - 105. [Abstract] [Full Text] [PDF]

				L. C. Mead, J. K. Eckert, S. Zimmerman, and J. G. Schumacher Sociocultural Aspects of Transitions From Assisted Living for Residents With Dementia Gerontologist, October 1, 2005; 45(suppl_1): 115 - 123. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Services Download to citation manager Citing Articles Citing Articles via HighWire PubMed PubMed Citation