ES.1. Purpose and Research Questions
This study addresses acute care use by older residents of residential care facilities (RCFs) who have severe cognitive impairment, including Alzheimer's disease or dementia. The two main research questions are:
Does having severe cognitive impairment affect the risk of any hospitalization and emergency department (ED) use among people living in RCFs? Among people living in RCFs who had at least one ED visit, does having severe cognitive impairment affect the number of ED visits?
Does living in a special care unit or facility that only served people with Alzheimer's disease or dementia affect the risk of any hospitalization and ED use among RCF residents? Among RCF residents with at least one ED visit, does living in a special care unit or facility that only serves people with Alzheimer's disease or dementia affect the number of ED visits?
ES.2. Data and Methods
The study uses the resident file of the 2010 National Survey of Residential Care Facilities (NSRCF) merged with the characteristics of the facilities in which residents live. Although the survey is the most current detailed survey of RCFs, it does not allow for the merging of Medicare or Medicaid claims data. All data on residents are provided by staff who know the resident, in consultation with facility records.
The key independent or explanatory variable of interest in this study is severe cognitive impairment, defined as either having a diagnosis of Alzheimer's disease or dementia or exhibiting at least three of seven symptoms related to memory impairments, confusion, or problems with orientation that were available in the survey. The seven symptoms related to: (1) long-term memory (e.g., forgetting one's own age or marital status); (2) short-term memory (e.g., having difficulty remembering what the person ate for breakfast or something that was told to them a few minutes earlier); (3) difficulty remembering or experiencing periods of confusion; (4) knowing the location of one's own bedroom; (5) recognizing staff names or faces; (6) knowing that one is in a facility; and (7) knowing what season of the year it is. Residents who did not meet either criterion were considered not to be severely cognitively impaired.
The NSRCF collected limited data on resident use of acute care. Three resident-level outcome variables were defined and analyzed in this study--whether the resident had been hospitalized (been a patient in a hospital overnight or longer, excluding trips to the ED that did not result in a hospital stay) and whether the resident had been treated in a hospital ED during the past 12 months or since the resident moved into the current facility if his or her length of stay (LOS) had been less than 1 year. The third outcome is the number of times the resident had been treated in a hospital ED over the same period (the number of hospitalizations was not available in the NSRCF).
Using logistic regression analysis of data about residents in the 2010 NSRCF, this study found that severe cognitive impairment was a marginally significant negative predictor of any hospitalization or any ED use (pp
However, when the analysis includes a variable for residence in a dementia-specific setting--that is special care units or facilities that only serve residents with Alzheimer's disease--the findings change. For any hospitalization and ED use, severe cognitive impairment was no longer a significant variable when residence in a special care unit or facility that only served people with Alzheimer's disease was entered into the equation. Residents in a special care unit or facility that only served people with Alzheimer's disease were less likely to have any hospital or ED use. This finding suggests that for severely cognitively impaired residents, living in a special care unit or a facility that only served people with Alzheimer's disease makes a significant difference because they were less likely to be hospitalized or visit the ED compared to similarly cognitively impaired people who did not reside in these types of settings.
Although not the main focus of this paper, this study also sheds light on general factors associated with hospital and ED use in RCFs. Relatively few variables were statistically significant predictors of any hospital or ED use. Only the number of chronic conditions, congestive heart failure, number of limitations in activities of daily living (ADLs), LOS in the facility, bed size, and hospital bed supply were statistically significant variables. Variables that are statistically significant in the equations estimating the number of ED visits among residents who had at least one ED visit include the number of limitations in ADLs, LOS, and local market supply of hospital and nursing home beds. Notably, most of these variables have to do with the health condition or disability of the resident rather than the characteristics of the facility.
Indeed, for the equations estimating any hospital or ED use, none of the potential policy or organizational variables--whether the resident is a Medicaid beneficiary, whether the facility participates in Medicaid, RCF ownership type, chain status, whether the facility is part of a continuing care retirement community, direct care staffing ratio, or the amount of monthly charges--were statistically significant predictors (at the p
This research contributes to the scant literature on the effect of Alzheimer's disease and cognitive impairment on hospitalization and ED use among people living in RCFs. More broadly, it also helps illuminate the relationship between people living in RCFs and the acute care system. Although this study found that having severe cognitive impairment or living in a special care unit or facility that only serves people with Alzheimer's disease seems to reduce the risk of hospitalization and ED use, it was not able to address the question of whether that reduction was appropriate or how that reduction was achieved. Data to estimate the prevalence of potentially avoidable hospitalization or ED use were not available. A finding of the study was that few policy or RCF organizational variables had any statistically significant effect; in general, only a few, largely health status, variables seem to be important in predicting hospitalization and ED use. Given the importance of RCFs in serving people with disabilities, especially cognitive impairment, more research is needed to address these issues.
|The Full Report is also available from the DALTCP website (http://aspe.hhs.gov/office_specific/daltcp.cfm) or directly at http://aspe.hhs.gov/daltcp/reports/2014/RCFdementia.cfm.|