Background
Cognitively impaired nursing home residents are a large and growing segment of the Medicare population who often require acute care services such as emergency department (ED) visits and hospitalizations. In 2009, 68% of all nursing home residents had some degree of cognitive impairment (CI) [
1], with the number of persons with CI expected to triple by 2050 [
2]. Further, cognitively impaired nursing home residents are at increased risk for potentially preventable hospitalizations and ED visits [
3‐
5]. Thus, the population of cognitively impaired nursing home residents will continue to grow rapidly and will likely continue to require frequent acute care services.
Unfortunately, CI is usually progressive, frequently leading to feeding problems and dysphagia [
6]. When this occurs, providers and families face the difficult decision regarding feeding tube (FT) placement for artificial hydration and nutrition. Although there is no evidence that the use of FTs improves clinical outcomes or survival in this patient group [
7‐
11], the practice is incredibly common with as many as 34% of nursing home (NH) residents with severe CI having a FT [
12]. This may be due in part to the perception by physicians and families that FT placement has minimal risks or burdens [
13,
14]. However, studies suggest that NH residents with FTs are at increased risk for hospitalizations [
15] and ED visits [
3,
4], compared to those without FTs. One recent study found that 26% of tube fed NH residents had a potentially preventable ED visit [
4]. Frequent hospital transfers in this already vulnerable population often leads to greater cognitive and functional decline, delirium, iatrogenic complications and death [
16,
17].
Studies to date on the relationship between ED visits and hospitalizations in NH residents with FTs have focused on those with advanced dementia. Little is known about whether the risk of acute care utilization in NH residents with FTs varies by the severity of their CI and whether these ED visits and hospitalizations are potentially preventable. Such information may give families and providers a clearer picture of the likely burdens associated with FT use based on cognitive status and allow them to make a more informed care decision about initial feeding tube placement. Thus, our objective was to determine the rates of ED visits and hospitalizations by differing levels of CI among a national random sample of NH residents with FTs. Further, we examined the rates of potentially preventable ED visits and hospitalizations by determining the rates of Ambulatory Care Sensitive (ACS) ED visits and ACS hospitalizations.
Results
In our study, we found 3479 NH residents had feeding tubes, suggesting that nearly 70,000 NH residents in the US were tube-fed in 2006. Table
1 shows the sociodemographic characteristics of NH residents with FTs by level of CI. Twenty-nine percent (1003/3479) of NH residents with FTs had Very Severe CI and were considered “comatose” or having “no discernible consciousness”. Since our analytic sample represents a random 5% sample of Medicare beneficiaries, this result suggests that over 20,000 US NH residents with FTs were “comatose” or had “no discernible consciousness” in 2006. Most participants were female (54%), and almost 20% were African American. NH residents with FTs in this study sample also had profound ADL impairment with 42% reporting dependence in all ADLs, representing nearly 30,000 US NH residents who are completely dependent in all ADLs (toileting, transferring, eating, dressing and bathing).
Table 1
Characteristics of nursing home residents with feeding tubes and characteristics by level of cognitive impairment (CI)
1
*
Sociodemographics
| | | | | |
Sex | | | | | |
Male | 46 | 48 | 49 | 46 | 40 |
Female | 54 | 52 | 51 | 54 | 60 |
Race/Ethnicity | | | | | |
White | 73 | 86 | 79 | 70 | 58 |
Black | 19 | 10 | 16 | 20 | 28 |
Other | 8 | 4 | 5 | 10 | 14 |
Age group | | | | | |
65-75 | 27 | 36 | 25 | 24 | 25 |
76-85 | 47 | 47 | 49 | 47 | 44 |
86+ | 26 | 17 | 25 | 30 | 31 |
Married | 40 | 44 | 41 | 39 | 36 |
Resident diagnoses & conditions
| | | | | |
CVA | 38 | 18 | 36 | 49 | 51 |
Stage 2+ pressure ulcer | 37 | 29 | 35 | 33 | 48 |
ADL impairment level2
| | | | | |
Independent-limited ADL | 9 | 28 | 8 | 1 | 0 |
Assistance (0–12) |
Extensive ADL assistance (13–17) | 32 | 49 | 44 | 30 | 5 |
Total dependence in most ADLs (18–19) | 17 | 11 | 21 | 25 | 12 |
Total dependence in all ADLs (20) | 42 | 13 | 26 | 43 | 83 |
CHESS category | | | | | |
Mild (0) | 20 | 15 | 19 | 22 | 26 |
Moderate (1–3) | 75 | 85 | 75 | 73 | 69 |
Severe (4–5) | 4 | <1 | 6 | 5 | 5 |
Treatments & preferences
| | | | | |
Do not resuscitate order | 33 | 24 | 33 | 37 | 39 |
Any urinary catheter use | 48 | 39 | 45 | 53 | 57 |
Table
2 shows the results of our overdispersed Poisson regression that modeled the rates of ED visits and ACS ED Visits among this national random sample of NH residents with feeding tubes. Ten percent of NH residents with FTs required an ED visit and 2.3% had a potentially preventable ACS ED visit in 2006. Twenty-four percent (79/335) of all ED visits by NH residents with FTs were considered potentially preventable and due to an ACS condition. Overdispersed poisson models revealed that NH residents with moderate/severe CI had significantly higher rates of ED and ACS ED visits compared to those with mild/moderate CI [IRR = 1.42 (1.16-1.74) and IRR = 1.80 (1.35-2.41), respectively]. In contrast, NH residents with very severe CI had similar adjusted rates of ED visits and ACS ED visits as NH residents with mild/moderate CI [IRR = 1.05 (.85-1.30) and IRR = 1.17 (.87-1.57), respectively].
Table 2
Rates of ED Visits and ACS ED visits according to the characteristics of nursing home residents with feeding tubes
Total | 335 (9.6) | | | 79 (2.3) | | |
Severity of cognitive impairment1
| | | | | | |
No/Mild | 82 (10.4) | 1.00 (.83-1.20) | .92 (.76-1.11) | 24 (3.1) | 1.79 (1.37-2.33)*** | 1.84 (1.40-2.41)*** |
Mild/Moderate | 105 (9.5) | Ref | Ref | 21 (1.9) | Ref | Ref |
Moderate/Severe | 65 (11.2) | 1.35 (1.10-1.65)** | 1.42 (1.16-1.74)** | 14 (2.4) | 2.02 (1.50-2.71)*** | 1.80 (1.35-2.41)*** |
Very severe | 83 (8.3) | .89 (.73-1.08) | 1.05 (.85-1.30) | 20 (2.0) | 1.42 (1.07-1.89)* | 1.17 (.87-1.57) |
Table
3 shows the results of our over dispersed Poisson regression that modeled the rates of hospitalizations and ACS hospitalizations. The hospitalization rate for NH residents with FTs in 2006 was 16% and the ACS hospitalization rate was 7%. Forty-four percent (242/553) of all hospitalizations in this study sample were potentially preventable and due to an ACS condition. In contrast to our ED results, severity of CI did not have a significant effect on rates of hospitalizations.
Table 3
Rates of hospitalizations and ACS hospitalizations according to the characteristics of nursing home residents with feeding tubes
Total | 553 (15.9) | | | 242 (6.9) | | |
Severity of cognitive impairment1
| | | | | | |
No/Mild | 138 (17.6) | 1.02 (.88-1.91) | 1.13 (.96-1.32) | 54 (6.9) | .99 (.82-1.20) | 1.18 (.97-1.44) |
Mild/Moderate | 167 (15.1) | Ref | Ref | 85 (7.7) | Ref | Ref |
Moderate/Severe | 94 (16.2) | 1.03 (.86-1.24) | .98 (.82-1.18) | 37 (6.4) | .95 (.75-1.19) | .89 (.71-1.12) |
Very severe | 154 (15.4) | 1.09 (.92-1.27) | .93 (.78-1.10) | 66 (6.6) | .98 (.81-1.20) | .87 (.70-1.09) |
Discussion
Examining a 5% random sample of Medicare beneficiaries in 2006, we found that approximately 25% of NH residents with FTs required an ED visit or hospitalization, with 44% of hospitalizations and 24% of ED visits being potentially preventable for an ACS condition. Moreover, nearly 30% of NH residents with FTs were considered “comatose” or without “discernible consciousness”, representing 20060 NH residents with FTs in the US. We further found that 42% of NH residents with FTs were dependent in all ADLs, representing 31,700 NH residents who are unable to eat, bathe, toilet, transfer, or dress.
Previous studies suggest that there are little or no benefits to FTs in cognitively impaired patients, particularly among those with advanced dementia. Evidence suggests that FTs in cognitively impaired patients do not prevent aspiration pneumonia, improve function, prevent or improve pressure ulcers, reduce the risk of infection, meaningfully improve nutritional status, decrease weight loss, improve wound healing, improve patient comfort, or decrease mortality – all the reasons commonly cited to support FT placement [
6,
7,
10,
11,
20,
41‐
47]. Among our national sample of tube-fed NH residents, we found that nearly half were classified as having at least severe or very severe CI (CPS 4–6, MMSE < 7).
Our results further indicate that the risks of hospitalization and ED visits vary by degree of CI. Tube-fed NH residents with severe CI (CPS 4–5, MMSE 5–7) are more likely to need ED evaluation than NH residents with moderate CI (CPS 2–3, MMSE 15–19) or very severe CI (CPS 6, MMSE 0–4). One possible explanation could be that NH staff, providers and families may recognize that ED visits may be especially burdensome for NH residents with profound CI and thus attempt to minimize ED visits for these residents. Alternatively, since NH residents with very severe CI are described as “comatose” or showing “no discernible consciousness”, these patients may be physically unable to become agitated. In contrast, patients with severe CI may become agitated during intercurrent illness, prompting transfers to the ED. Evidence suggests that a frequent precipitant for acute care transfers is dementia-related behavioral issues [
48], which are commonly exacerbated with acute medical illness.
In contrast to our ED visit results, we found that hospitalizations did not vary significantly by degree of CI. Our results may reflect the fact that the decision to hospitalize is often based on more data such as lab tests and radiology results which often do not account for CI. In contrast, the initial decision to transfer a NH resident for an ED evaluation is based more on overall clinical impression that may implicitly incorporate a patient’s cognitive status. Alternatively, the presence of CI may actually cloud the ability of nursing staff to appropriately recognize the nature and severity of a change in health status. As a result, nursing staff may err on the side of transferring residents to the hospital for conditions that could be potentially managed in the NH.
Our ACS results suggest that severity of CI has a significant effect on rates of potentially preventable, or ACS, ED visits, but little effect on ACS hospitalizations. NH residents with no/mild CI (CPS 0–1, MMSE = 22-25) and severe CI (CPS 4–5, MMSE = 5-7) have significantly higher rates of ACS ED visits than those with moderate CI (CPS 2–3, MMSE = 15-19) or very severe CI (CPS 6, MMSE = 0-4), the latter of which is the population typically studied in this body of literature. While families and caregivers may prefer more aggressive care for acute medical problems in the earlier stages of CI, our findings suggest that the very presence of even mild CI in a NH resident with a FT may increase the risk of a potentially preventable ED visit. Prior research suggests that family members of NH residents with dementia are often unsure about the length and expected course of the illness and such uncertainty makes it difficult to determine whether an acute illness is part of a downward trajectory or a temporary, reversible setback [
49].
For many cognitively impaired NH residents, a transfer to the hospital or ED likely represents a significant burden. This frail, vulnerable population is often easily confused, frustrated and frightened; tends to exhibit behavioral and psychiatric disturbances when under stress; and is often unable to communicate their needs or understand care instructions. Given the fast-paced nature of the ED and focus on rapid triage and treatment, many cognitively impaired patients may not understand why they are moved from their familiar surroundings to a strange setting with unfamiliar caregivers who must perform uncomfortable procedures such as venipuncture or FT reinsertion. Moreover, hospital transfers in this population frequently results in greater cognitive and functional decline, delirium, iatrogenic complications and death [
16,
17]. Thus, our finding that one in four tube-fed NH residents require hospitalization or ED visits likely represents a substantial burden for this vulnerable cognitively impaired population.
There is growing recognition of the unique and important role ED care providers can play in supporting early palliative care interventions along a patient’s disease trajectory, promoting quality of life, as well as reducing treatment costs [
50‐
54]. Recent studies indicate that palliative interventions in the ED may improve timely provision of care, improve care outcomes, increase direct referrals to hospice, decreased lengths of stay, improve patient and family satisfaction, and reduce intensive care utilization, as well as health care costs [
55‐
57]. Additional research is needed to evaluate the effect of such interventions with frail NH residents, such as those with feeding tubes and CI, who may frequently visit the ED.
Our results should be interpreted in light of our study’s limitations. First, we relied on administrative data and therefore lacked detailed clinical information regarding causes of ED and hospital visits. This also limited our ability to account for individual clinical characteristics. However, we were able to rely on validated scales and frequently used MDS resident-level variables to risk adjust our outcomes. Second, we were unable to determine the indication for FT placement. It is possible that some residents with moderate/severe or very severe CI received a FT for nutritional support after a stroke, trauma, or head or neck cancer for which there is some data to suggest FTs may lead to better outcomes [
14,
41,
58]. Third, by focusing only on residents with FTs we may have selected for a population with an inherent preference towards more aggressive care, including transfers to the acute care hospital. Despite these limitations, this is the first study to describe rates of hospitalization and ED visits, including for potentially preventable conditions, among a national random sample of NH residents with FTs and different levels of CI.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
CS conceptualized and designed the study; acquired, analyzed and interpreted the data; and drafted the manuscript. NS and PG participated in the study design and coordination; helped interpret the data; and draft the manuscript. SL helped conceptualize and design the study; analyze and interpret data; and draft the manuscript. All authors read and approved the final manuscript.