If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Indiana University School of Medicine, Indianapolis, INIndiana University Center for Aging Research, Indianapolis, INRegenstrief Institute, Inc, Indianapolis, IN
Indiana University School of Medicine, Indianapolis, INIndiana University Center for Aging Research, Indianapolis, INRegenstrief Institute, Inc, Indianapolis, IN
Indiana University School of Medicine, Indianapolis, INIndiana University Center for Aging Research, Indianapolis, INRegenstrief Institute, Inc, Indianapolis, IN
To describe emergency department (ED) utilization among long-stay nursing home residents with different levels of dementia severity.
Design
Retrospective cohort study.
Setting
Public Health System.
Participants
A total of 4491 older adults (age 65 years and older) who were long-stay nursing home residents.
Measurements
Patient demographics, dementia severity, comorbidities, ED visits, ED disposition decisions, and discharge diagnoses.
Results
Forty-seven percent of all long-stay nursing home residents experienced at least 1 transfer to the ED over the course of a year. At their first ED transfer, 36.4% of the participants were admitted to the hospital, whereas 63.1% of those who visited the ED were not. The median time to first ED visit for the participants with advanced stage dementia was 258 days, whereas it was 250 days for the participants with early to moderate stage dementia and 202 days for the participants with no dementia (P = .0034). Multivariate proportional hazard modeling showed that age, race, number of comorbidities, number of hospitalizations in the year prior, and do not resuscitate status all significantly influenced participants’ time to first ED visit (P < .05 for all). After accounting for these effects, dementia severity (P = .66), years in nursing home before qualification (P = .46), and gender (P = .36) lost their significance.
Conclusions
This study confirms high rates of transfer of long-stay nursing home residents, with nearly one-half of the participants experiencing at least 1 ED visit over the course of a year. Although dementia severity is not a predictor of time to ED use in our analyses, other factors that influence ED use are readily identifiable. Nursing home providers should be aware of these factors when developing strategies that meet patient care goals and avoid transfer from the nursing home to the ED.
Though individuals with dementia live in and transition dynamically across the continuum, nursing homes remain a significant site for the provision of care and an important node for transitional care for this population.
Many transitions of nursing home residents to the emergency department (ED) are believed to be either unnecessary or preventable; they can be particularly burdensome for people with dementia. Because of the high costs and poor quality of care involved in unnecessary transfers, they have become a target of policymakers and a focus of a Centers for Medicare and Medicaid Services demonstration project.
The optimizing patient transfers, impacting medical quality, and improving symptoms: Transforming institutional care approach: Preliminary data from the implementation of a centers for Medicare and Medicaid services nursing facility demonstration project.
When nursing home residents with dementia transfer to the hospital for acute care, many of them will be seen first in the ED. Most EDs are designed most efficiently to provide episodic and emergent care to patients. Nursing home residents have complex comorbidities that can challenge the busy ED provider trying to differentiate between acute and chronic illness in a patient during a time-limited encounter. Likewise, the often chaotic emergency environment can be difficult for any patient with a health crisis to manage; it is exponentially more difficult to navigate for those with dementia. Persons with dementia may not be able to articulate their symptoms, participate in care planning, assist in care coordination, or even understand what is happening to them as they transition across sites of care. Indeed, dementia is a significant independent risk factor for several patient-centered outcomes including hospital admission from the ED, return to the ED within 30 days, and mortality after an ED visit.
There is a progressive spectrum of cognitive impairment, ranging from mild cognitive impairment to advanced dementia. Though there is 1 study indicating that more severely cognitive-impaired individuals may be less likely to use the ED than less cognitive-impaired patients, the question of how dementia severity influences ED utilization among long-stay nursing home residents remains largely unanswered.
In this work, we describe ED utilization among long-stay nursing home residents with different levels of dementia severity. To accomplish this, we analyzed a merged dataset that members of our study team have used previously to investigate health care utilization among older adults.
In our study, we evaluate for differences in disposition decisions and discharge diagnoses as well as in time to first ED visit for long-stay nursing home residents with varying dementia severities.
Methods
Study Design and Data Source
This is an observational study based on existing data. For our analyses, we used a merged set of Medicare and Medicaid claims and resident-level minimum data set (MDS) files encompassing the years 1999 through 2009. Participants were identified initially through records at Wishard Health Services (now Eskenazi Health Services), a large, public hospital system in Indianapolis, Indiana. Health service utilization data were captured across sites of care using Medicare and Medicaid claims. The Indiana University Institutional Review Board and the Centers for Medicare and Medicaid Privacy Board approved this study.
In our analyses, we were interested in annual rates and time to first ED use by long-stay nursing home residents stratified by the severity of their dementia. For study purposes, we defined a long-stay nursing home resident as a person aged 65 years or older who accumulated 90 or more consecutive days of nursing home residence between January 1, 1999 and December 31, 2008. Persons were qualified to join our study cohort if they were a long-stay nursing home resident between January 1, 2000 and December 31, 2008. We restricted our analysis of ED use to those ED visits that occurred between January 1, 2000 and December 31, 2009 in the first 365 days after a person's study qualification. We also examined participants' hospitalization and ED use in the year before their enrollment in the study.
Measures of Interest
Dementia severity was defined using data from the Cognitive Performance Scale (CPS) contained in the MDS assessment closest to each participant's qualification as a long-stay resident. We defined residents having CPS scores of 0–1 as not demented, those having CPS scores of 2–4 as having mild to moderate dementia, and those having CPS scores of 5–6 as having severe dementia, consistent with prior work.
These codes and conditions were as follows: arthritis (714.0 and 715.0), cancer (140.0-172.0 and 174.0-239.0), coronary artery disease (410.0, 411.0, 412.0, 413.0, and 414.0), congestive heart failure (428.0 and 398.91), chronic obstructive pulmonary disease (491.0, 492.0, and 496.0), diabetes (250.0), hypertension (401.0), liver disease (570.0-573.0), renal disease (585.0), and stroke (433.1 and 434.1). These conditions were obtained from 1999 until the qualification date of each resident. Counts of activities of daily living impairments were calculated using data from the MDS assessment completed closest to study qualification. “do not resuscitate” or “do not hospitalize” statuses were obtained from the MDS assessment completed closest to time of study qualification or carried forward if completed at an earlier MDS assessment. Discharge diagnoses were categorized using the Clinical Classification Software available through the Agency for Healthcare Research and Quality.
Statistical Analysis
The study data were analyzed in the following manner. First, our study participants were divided into 3 dementia severity groups: those without dementia, those with early to moderate dementia, and those with advanced dementia. Descriptive demographic and health characteristics, including comorbidities, were comparatively examined for these groups, and their differences tested using one-way analysis of variance models for continuous variables and χ2 tests for categorical variables. We calculated annual rates of ED use from the number of ED visits in the first 365 days after study enrollment (censored for death/nursing home discharge) among those participants who entered this study in any given year. These rates were then standardized to each 1000 nursing home bed days. We obtained the 95% confidence intervals for the annual ED rates under the assumption that counts of ED visits followed a Poisson distribution. We compared the percentages of participants with ED admissions within 365 days after qualification between the groups using χ2 tests. We compared disposition decisions for patients at the time of their first ED visit within 365 days using Fisher exact tests. We examined differences between diagnoses for patients who were discharged from the ED, first identifying the 10 most common discharge diagnoses among each dementia severity group, then examining those conditions that have been judged to be potentially avoidable among nursing home residents, and then looking at select conditions that had large differences in percentages between the three dementia severity groups. Fisher exact tests were also used to compare the dementia severity groups on these selected conditions. Cox proportional hazards regression models were used to compare the time from the study qualification to first ED visit between the dementia severity groups after adjusting for prespecified covariates including participant age, race, gender, number of comorbidities, number of hospitalizations in year before study qualification, years in nursing home until qualification, and do not resuscitate status. Participants without an ED visit were censored at the minimum of 365 days, their date of death, or their nursing home discharge date following the qualification date. We examined the distributions of time to first ED visits in patients with different levels of dementia severity. Log-rank test was used to compare the time distributions of the 3 dementia severity groups. SAS v 9.4 (SAS Institute, Cary, NC) was used for our analyses; P value of ≤ .05 was considered statistically significant.
Results
Table 1 shows the characteristics of the 4491 long-stay nursing home residents who were identified and followed in this study. Among this population, the majority of residents had early to moderate stage dementia (70.9%), whereas smaller percentages of residents had either no dementia (21.5%) or advanced stage dementia (7.5%). Participants with advanced dementia were older, more likely to be women, and more likely to be black than participants with early to moderate stage or no dementia. Study participants without dementia had the greatest number of comorbidities, whereas study participants with advanced dementia had the least number of comorbidities. Participants with advanced dementia had the highest number of ED visits in the year before the study enrollment and also had the highest percentage of do not resuscitate orders recorded.
Table 1Demographic and Clinical Characteristics of Older Adults With Long-Term Nursing Home Stays by Cognitive Levels by Dementia Severity Levels
Characteristics
Dementia (Advanced Stage)
Dementia (Early-Moderate Stage)
No Cognitive Impairment
P Value
N
Column
N
Column
N
Column
Age at qualification (years), mean (SD)
338
81.8 (7.9)
3186
80.0 (7.8)
967
77.5 (7.5)
<.0001
Female gender, n (%)
338
243 (71.9%)
3186
2058 (64.6%)
967
670 (69.3%)
.0018
Race, n (%)
338
3177
965
.0002
Black
-
157 (46.4%)
-
1225 (38.6%)
-
315 (32.6%)
-
Other
-
4 (1.2%)
-
41 (1.3%)
-
14 (1.5%)
-
White
-
177 (52.4%)
-
1911 (60.2%)
-
636 (65.9%)
-
Months of observation, mean (SD)
338
8.9 (4.4)
3186
8.2 (4.7)
967
7.0 (5.0)
<.0001
Arthritis, n (%)
338
154 (45.6%)
3186
1868 (58.6%)
967
650 (67.2%)
<.0001
Cancer, n (%)
338
74 (21.9%)
3186
995 (31.2%)
967
347 (35.9%)
<.0001
Coronary artery disease, n (%)
338
136 (40.2%)
3186
1648 (51.7%)
967
579 (59.9%)
<.0001
Congestive heart failure, n (%)
338
123 (36.4%)
3186
1557 (48.9%)
967
533 (55.1%)
<.0001
Chronic obstructive pulmonary disease, n (%)
338
82 (24.3%)
3186
1429 (44.9%)
967
512 (52.9%)
<.0001
Diabetes, n (%)
338
132 (39.1%)
3186
1552 (48.7%)
967
553 (57.2%)
<.0001
Hypertension, n (%)
338
251 (74.3%)
3186
2778 (87.2%)
967
866 (89.6%)
<.0001
Liver disease, n (%)
338
23 (6.8%)
3186
320 (10.0%)
967
137 (14.2%)
<.0001
Renal disease, n (%)
338
36 (10.7%)
3186
540 (16.9%)
967
218 (22.5%)
<.0001
Stroke, n (%)
338
62 (18.3%)
3186
658 (20.7%)
967
154 (15.9%)
.0044
Number of comorbidities, mean (SD)
338
3.2 (2.0)
3186
4.2 (2.1)
967
4.7 (2.0)
<.0001
Falls, n (%)
338
4 (1.2%)
3186
111 (3.5%)
967
23 (2.4%)
.0244
Number of ED visits in year before qualification, mean (SD)
338
0.9 (2.2)
3186
0.7 (1.4)
967
0.8 (1.4)
.0095
Number of hospitalizations in year before qualification, mean (SD)
For this binary ADL categorization, 0 to 1 impairment(s) were treated as having no ADL impairment, whereas 2 or more impairments were treated as having ADL impairments.
An “established” long-stay resident was one who had amassed 90 days of nursing home residence prior to January 1, 2000.
-
223 (66.0%)
-
2823 (88.6%)
-
881 (91.1%)
-
ADL, activities of daily living; SD, standard deviation.
∗ For this binary ADL categorization, 0 to 1 impairment(s) were treated as having no ADL impairment, whereas 2 or more impairments were treated as having ADL impairments.
† An “established” long-stay resident was one who had amassed 90 days of nursing home residence prior to January 1, 2000.
Figure 1 displays the annual rates of ED use per 1000 nursing home bed days for patients who qualified for study inclusion during a given study year. Participants enrolled in calendar year 2000 (21.8% of the sample) represent a mixture of patients: some who were long-stay residents who had already accumulated 90 days of nursing home residence by the beginning of that year and others who became new long-stay nursing home residents during that year. From 2001 through 2008, participants were only enrolled in the study if they newly qualified as long-stay nursing home residents. There appears to be a trend toward increasing ED visits over time with a rate of 4.0 ED visits per 1000 nursing home bed days in 2001 and a rate of 5.9 ED visits per 1000 nursing home bed days in 2008, though the upper 95% confidence intervals for these rates overlap.
Fig. 1Annual rates and upper 95% confidence bounds for ED use per thousand nursing home bed-days between 2000 and 2008.
Nearly one-half (47%) of all long-stay nursing home residents experienced at least 1 transfer to the ED over the course of a year. There were not statistically significant differences in overall transfer rates by severity of dementia (P = .50).
At participants’ first ED transfer, 36.4% of participants were admitted to the hospital, whereas 63.1% of participants were not. The remaining 0.52% died. A higher percentage of persons with advanced stage dementia were discharged after an ED visit (69.8%) when compared with persons with early to moderate stage dementia (62.9%) and no dementia (61.0%), though this difference was not statistically significant (P = .22). Among participants who visited the ED, persons with advanced stage dementia were significantly more likely than persons with early to moderate stage or no dementia to receive a diagnosis of a urinary tract infection (P < .05), whereas those with no dementia were significantly less likely than persons with early to moderate stage or advanced dementia to receive a diagnosis of an injury/poisoning (P < .01) (Table 2).
Table 2Comparisons Between ED Diagnoses by Dementia Status in Those With ED Visits Within 365 Days of Qualification Date
CCS Category
ED Diagnosis
Dementia (Advanced Stage)
Dementia (Early to Moderate Stage)
No Dementia
P Value
N
%
N
%
N
%
1
Infectious and parasitic diseases
5
3.0
51
3.4
10
2.2
.473
1.1.2
Septicemia (except in labor)
5
3.0
44
2.9
8
1.8
.390
2
Neoplasms
3
1.8
23
1.5
10
2.2
.544
3
Endocrine; nutritional; and metabolic diseases and immunity
9
5.3
72
4.8
19
4.2
.797
3.3
Diabetes mellitus with complications
2
1.2
26
1.7
4
0.9
.435
3.8
Fluid and electrolyte disorders
3
1.8
19
1.3
7
1.6
.670
4
Diseases of the blood and blood-forming organs
1
0.6
23
1.5
11
2.4
.242
4.1
Anemia
0
0
17
1.1
11
2.4
.038
5
Mental illness
5
3.0
46
3.1
10
2.2
.660
5.4
Delirium, dementia, and amnestic and other cognitive disorders
4
2.4
22
1.5
3
0.7
.191
6
Diseases of the nervous system and sense organs
7
4.1
109
7.3
29
6.5
.300
6.4
Epilepsy; convulsions
3
1.8
29
1.9
2
0.4
.055
6.4.2
Convulsions
3
1.8
29
1.9
2
0.4
.055
6.7
Eye disorders
1
0.6
25
1.7
10
2.2
.419
7.2
Diseases of the heart
10
5.9
130
8.7
47
10.5
.191
7.2.11
Congestive heart failure; nonhypertensive
4
2.4
34
2.3
9
2.0
.911
7.4
Diseases of arteries; arterioles; and capillaries
0
0
12
0.8
8
1.8
.082
8.1
Respiratory infections
8
4.7
76
5.1
25
5.6
.919
9.10
Gastrointestinal hemorrhage
5
3.0
19
1.3
6
1.3
.225
10.1.4
Urinary tract infections
18
10.7
75
5.0
22
4.9
.015
13
Diseases of the musculoskeletal system and connective tissue
6
3.6
77
5.2
34
7.6
.087
16
Injury and poisoning
37
21.9
289
19.4
58
12.9
.003
16.6
Open wounds
13
7.7
55
3.7
14
3.1
.037
17
Symptoms; signs; and ill-defined conditions and factors
The median time to first ED visit for participants with advanced stage dementia was 258 days, whereas it was 250 days for participants with early to moderate stage dementia and 202 days for participants with no dementia (P = .0034). Table 3 shows the results from the Cox proportional hazards regression model for factors predicting time to first ED visit in the year after study qualification. In this model, age, race, number of comorbidities, number of hospitalizations in the year prior, and do not resuscitate status all significantly influenced participants’ time to first ED visit (P < .05 for all), however, dementia severity (P = .66), years in nursing home until qualification (P = .46), and gender (P = .36) did not.
Table 3Estimated Hazard Ratios From Cox Proportional Hazards Regression Model
Variables
HR
95% CI
P Value
Dementia level-overall
-
.6615
Dementia (advanced stage) vs no cognitive impairment
1.05
0.87–1.27
.6165
Dementia (early-moderate stage) vs no cognitive impairment
0.98
0.87–1.09
.6568
Age at qualification (years)
0.98
0.98–0.99
<.0001
Female gender
1.05
0.95–1.15
.3610
Race-overall
-
.0061
Black vs white
1.13
1.03–1.24
.0104
Other vs white
1.46
1.04–2.06
.0294
Do not resuscitate order
0.91
0.82-1.00
.0488
Number of years in nursing home until qualification
1.02
0.97–1.08
.4642
Number of hospitalizations in year prior to qualification
Our study confirms high rates of transfer of long-stay nursing home residents; nearly one-half experienced at least 1 ED visit over the course of a year with nearly two-thirds being treated and discharged from the ED, which is consistent with earlier studies.
Many transfers of frail nursing home residents to the ED are thought to be avoidable, with optimal nursing home care, prevention practices, and appropriate advance care planning. Reducing ED transfers of nursing home residents will continue to be an important topic as health systems and nursing home providers transition to new payment models and change practices to prepare for new readmission metrics.
As interventions are developed to reduce these transfers, a deeper understanding of the role of resident characteristics, such as degree of cognitive impairment, is needed to help with targeting approaches. In this study, degree of dementia severity was not associated either with likelihood to transfer to the ED or with having that transfer result in a hospitalization. Age, race, number of comorbidities, number of hospitalizations in the year prior to study entry, and do not resuscitate status all influenced the time to first ED visit. These results help us both to understand patterns in ED transfer and disposition among long-stay nursing home residents with and without dementia and begin to identify potential targets for future work to address the acute care needs of this vulnerable population.
As dementia is a progressive illness that affects patients’ cognition, functional abilities, and health care utilization, it is not surprising to find that our participants with advanced dementia were older, more functionally impaired, and more likely to have visited the ED in the previous year than patients with less severe or no cognitive impairment. The fact that long-stay residents with advanced dementia had less comorbidity than participants with early to moderate stage dementia was unexpected and may suggest that long-stay nursing home residents with advanced dementia are “survivors” who are healthy enough to have lived long enough to develop advanced dementia. Despite having fewer comorbidities, residents with advanced dementia transferred to the ED at similar rates compared with early to moderate dementia and no dementia. These findings suggest that other factors beyond comorbidity and dementia severity may influence ED use in this population. Residents with advanced dementia are less able to describe their symptoms and are challenging to assess when there is a change in status, which may increase their likelihood to be transferred to the ED for further work-up.
The disposition decisions and ED disposition diagnoses for these patients may provide us with some insight into the issue of “avoidability” of some of these transfers. Admission rates were similar for patients seeking care in the ED across dementia severities, with over 60% of patients being discharged from the ED after evaluation. ED visits may be burdensome particularly for residents with dementia, as they often involve invasive tests and procedures as well as transfer to an unfamiliar environment. It is important to examine whether high discharge rates from the ED, including over 70% of residents with advanced dementia, represent transfer episodes that could have been prevented with more robust assessment and management resources in the nursing home. Final diagnoses for participants with advanced dementia did not differ significantly from the diagnoses for other participants, apart from a slightly higher proportion of urinary tract infections and injury/poisonings. These findings suggest that patients with advanced dementia do not “need” admission any more frequently than patients with less severe or no dementia and also are not presenting with significantly different acute medical conditions than other participants. The lower rate of admission among older adults with advanced dementia could also reflect that emergency providers are engaging patient surrogates in conversations around goals of care to develop treatment plans that avoid hospitalization.
Our data reveal a profound difference in time to first ED visit between patients with and without dementia. However, for patients with dementia diagnoses, dementia severity does not have a significant effect on time to first ED over the course of a given year. In our Cox proportional hazard modeling, we found that age, race, number of comorbidities, number of hospitalizations, and do not resuscitate status influence patients' time to first ED visit. Nursing home providers should be aware of these factors when developing strategies that meet patients’ care goals and avoid transfer from the facility to the ED.
From this work, we conclude that although dementia is associated with early ED visits, dementia severity does not have a significant influence on ED utilization, rate of admission to the hospital from the ED, or time to first ED utilization among a population of long-stay nursing home residents. Given that a large prior study found that surrogates of nursing home residents with advanced dementia prefer care focused on comfort,
it is worth questioning why patients with advanced dementia have seemingly similar patterns of ED utilization as those patients with early to moderate dementia and no dementia. The fact that these patterns of ED use do not differ suggests possible opportunities for proactively defining goals of care, including limiting types of treatments if desired. This type of approach is the hallmark of a patient-centered care planning process and is an important component of some interventions that seek to decrease avoidable hospitalizations among long-stay nursing residents.
The optimizing patient transfers, impacting medical quality, and improving symptoms: Transforming institutional care approach: Preliminary data from the implementation of a centers for Medicare and Medicaid services nursing facility demonstration project.
The Interventions to Reduce Acute Care Transfers (INTERACT) quality improvement program: An overview for medical directors and primary care clinicians in long term care.
There are several limitations worth noting in this work. This cohort was drawn from one state and may not be representative of all long-stay nursing home residents. Still, we have gathered information on a large cohort of patients and followed their care patterns over an extended period of time across settings. Information about resident or surrogate goals of care is not available in these datasets, beyond code status. These are important data that may influence the clinical decision to transfer a nursing home resident to the ED and are necessary to determine “appropriateness” and so should be a focus of a future investigation.
Given the expense to the health system and burden to the nursing home resident of ED transfers, there will be a continued focus on identifying and preventing avoidable transfers. Future work should continue to explore patient and provider factors associated with transfers that may be amenable to intervention, as well as use qualitative and quantitative methods to further define truly “avoidable” ED transfers.
Acknowledgments
The authors would like to gratefully acknowledge the contributions of Roberta Ambuehl for her assistance in data management for this study.
References
Prince M.
Bryce R.
Albanese E.
et al.
The global prevalence of dementia: A systematic review and meta-analysis.
The optimizing patient transfers, impacting medical quality, and improving symptoms: Transforming institutional care approach: Preliminary data from the implementation of a centers for Medicare and Medicaid services nursing facility demonstration project.
The Interventions to Reduce Acute Care Transfers (INTERACT) quality improvement program: An overview for medical directors and primary care clinicians in long term care.
This work was funded by a grant to Drs. LaMantia and Unroe from the Regenstrief Institute, Inc. Dr. LaMantia receives support from grant 5K23 AG043498, and Dr. Unroe receives support from grant 5K23 AG048323 from the National Institute on Aging. The views in this article are those of the authors and do not necessarily represent the views of the National Institute on Aging.
00111-0&id=gr1.jpg){kind=link}