Abstract
This study examines emergency department (ED) visits by nursing home (NH) residents aged 65 and over, and factors associated with hospital admission from the ED visit using data from the 2001–2008 National Hospital Ambulatory Medical Care Survey. Cross-sectional analyses were conducted on patient characteristics, diagnosis, procedures received, and triage status. On average, elderly NH residents visited EDs at a rate of 123 visits per 100 institutionalized persons. Nearly 15% of all ED visits had ambulatory care sensitive condition diagnoses. Nearly half of these visits resulted in hospital admission; chronic obstructive pulmonary disease, congestive heart failure, kidney/urinary tract infection, and dehydration were associated with higher odds of admission. Previous studies suggested that adequate medical staffing and appropriate care in the NH could reduce ED visits and hospital admissions. Recent initiatives seek to reduce ED visits and hospitalizations by providing financial incentives to spur better coordination between NH and hospital.
Emergency departments (EDs) are used by nursing home (NH) residents at rates almost 5 times higher at ages 65–74 and almost 3 times higher at ages 75 and older than their noninstitutionalized counterparts (McCaig & Burt, 2005). Almost half of ED visits made by NH residents resulted in hospitalizations (Wang, Shah, Allman, & Kilgore, 2011). Further research is needed to characterize whether ED visits by this population, particularly those eventually hospitalized, can be reduced or avoided. This study uses nationally representative data to examine characteristics of ED visits made by elderly NH residents as well as factors associated with hospital admission from the ED. This study specifically focuses on the relationship of ED visits and subsequent inpatient hospitalization with ambulatory care sensitive conditions (ACSCs).
Description of the Problem
Prior research found that NH residents present at EDs with greater medical acuity and complexity than their noninstitutionalized counterparts (Wang et al., 2011). When NH residents are transferred to EDs, their care is often fragmented with little continuity of care between the NH and hospital (Ackermann, Kemle, Vogel, & Griffin, 1998). Information critical to effective emergency care such as the mental status and medical conditions of NH residents is often not available to ED staff (Ackermann et al., 1998; Zimmer, Eggert, Treat, & Brodows, 1988). In addition, obtaining a complete and accurate medical history is often difficult for ED staff when cognitive problems exist among these patients (Vladeck, Miller, & Clauser, 1993). Thus, lack of information and the frail health of these patients pose challenges to the hospital and ED staff. If the ED visit results in admission to the hospital, the stay itself may be hazardous for the elderly NH resident. Studies have reported that the elderly commonly experience functional decline, delirium, iatrogenic illnesses, and other adverse events during hospital stays (Covinsky et al., 2003; Creditor, 1993; Ebersole & Hess, 2001; Inouye et al., 1999). In addition, there are cost and efficiency concerns associated with hospitalization of NH residents. A study showed that the average spending for NH resident’s hospitalization with any ACSC was $10,140. Among ACSC hospitalizations, asthma, pneumonia, and diabetes were the most costly conditions (Grabowski, O’Malley, & Barhydt, 2007). The average cost of ED visits for people age 65 and up was $1,306 (Agency for Healthcare Research and Quality, 2009). The cost for ED visits could be higher for NH residents because of their greater medical acuity and complexity.
Due to the problems that NH residents encounter in the ED, some researchers have examined the extent of unnecessary ED visits and hospital admissions from EDs among the NH population (Irvine, Van Buren, & Crossley, 1984; Kerr & Byrd, 1991; Stark, Gutman, & McCashin, 1982). However, most of the data are outdated or derives from a single ED or NH. One study used 2004 national data to examine potential preventable ED visits by NH residents, but it used reason for visit, instead of the actual diagnosis to classify potentially preventable ED visits (Caffrey, 2010). In addition, the study did not have information on ED visits, nor whether such visits resulted in a subsequent inpatient admission or not.
The Institute of Medicine (IOM) has recommended use of hospitalizations for ACSCs as an indicator for whether appropriate care was provided to avoid hospitalization (Institute of Medicine [IOM], 1993). The IOM’s ACSCs indicator has been used increasingly to examine the appropriateness of ED visits by NH residents and subsequent hospitalizations (Carter, 2003; Carter, Datti, & Winters, 2006; Carter & Porell, 2005; Grabowski et al., 2007; Intrator, Castle, & Mor, 1999; Intrator, Zinn, & Mor, 2004). A study used 2000–2002 National Hospital Ambulatory Care Survey (NHAMCS) data to compare ACSCs among ED visits made by older adults inside and outside of NHs. The study found higher risk for ACSCs and hospitalizations for NH residents compared with their noninstitutionalized counterparts (Carter et al., 2006). A more recent study used 2005–2008 NHAMCS data to characterize ED use by NH residents, but it did not examine characteristics associated with hospital admissions from EDs (Wang et al., 2011).
Purpose
We use national data from 2001 to 2008 to examine ED use and subsequent hospital admission among NH residents. The large sample size permits more detailed analyses with two specific aims: (1) evaluate and characterize ED visits made by NH residents, including whether the ED visit resulted in hospital admission: and (2) examine whether presence of ACSCs was associated with hospitalization from the ED.
Method
Data Source
The study is based on nationally representative samples of ED visits collected in the 2001–2008 NHAMCS. The scope of NHAMCS is patient visits to EDs and outpatient departments of non-Federal, short stay, or general hospitals. For the purpose of this study, we only use data from the ED component of the NHAMCS.
NHAMCS employs a four-stage probability sample design involving samples of geographic primary sampling units (PSUs), hospitals with EDs within PSUs, emergency service areas (ESAs) within EDs, and patient visits within ESAs. The final sampling stage involves a systematic random sample of ED visits during a randomly assigned 4-week reporting period. The U.S. Census Bureau, acting as the data collection agency for the survey, provided training to field representatives throughout the nation who, in turn, oversaw data collection at the hospital. Hospital staff were instructed to complete the information requested on the patient record form. During 2001–2008, data were collected for 287,803 sampled ED visits from non-Federal, short stay, and general hospitals. During this time period, the average number of participating hospital EDs was 367. The unweighted response rate for each year ranged from 85.3 to 91.9% (McCaig & Burt, 2003, 2004, 2005; McCaig & Nawar, 2006; National Center for Health Statistics [NCHS], 2010; Nawar, Niska, & Xu, 2007; Niska, Bhuiya, & Xu, 2010; Pitts, Niska, Xu, & Burt, 2008).
Measurement
The information regarding patient residence on the patient record form was collected in two different manners across the years. From 2001 to 2004, NH residents were identified by the question “Does the patient reside in a NH or other institution?” From 2005 to 2008, the NH residence was among several response choices for patient residence. Although it is possible that some ED visits with positive responses to the question in 2001–2004 may have been in an institution other than a NH (e.g., prison, mental hospital), we limited this possibility by focusing our analysis exclusively on ED visits among the elderly (i.e., those aged 65 or older). We also compared 2001–2004 estimates with 2005–2008 combined estimates for potential differences due to question wording and found no statistical differences in estimates. In our analysis, ED visits among patients whose place of residence was marked as unknown were assumed to reside in the community based on the similarity of characteristics for these cases with those responding “No” in 2001–2004 or “private residence” in 2005–2008. Basic characteristics of ED visits in 2001–2004 versus 2005–2008 were generally very similar (data not shown).
Statistical Analysis
We conducted two types of analyses in the study. The descriptive analysis characterizes ED visits by NH residents and the second analysis models the likelihood of hospitalization among NH residents’ ED visits. The descriptive analysis includes the average annual ED visit rates among elderly NH residents per 100 institutionalized persons during 2001–2008. ED visit rates were calculated by dividing the estimated count of ED visits among the elderly NH residents by estimates of the civilian institutionalized population in each cohort of interest for the same years, overall, by age (65–74, 75–84, 85+), sex, and race (White, non-White). For example, the visit rate for males was obtained by taking the weighted ED visits made by males divided by the male institutionalized population estimated by the Census Bureau. The institutionalized population estimates were derived by subtracting estimates of the civilian noninstitutionalized population from estimates of the civilian resident population as of each year during 2001–2008. Each set of annual population estimates are projections based on Census 2000 data developed by the Population Division, U.S. Census Bureau using the July 1 set of state population estimates (U.S. Census Bureau).
The descriptive analysis also summarizes the services ordered or provided during ED visits and outcomes of ED visits by elderly NH residents. We calculated the mean number of diagnostic or screening tests performed based on responses to a number of check boxes (ranging from 21 to 30 during 2001–2008) indicating whether any of the specific diagnostic or screening tests were ordered or performed. We also calculated the percent of ED visits among elderly NH visits that had each of the nine most frequent diagnostic or screening tests performed. In addition, we estimated the mean number and the percent of the five most frequent procedures performed from a list of possible procedures indicated (11 procedures for 2001–2006 and 13 procedures for 2007–2008). We also summarized information on triage status (nonurgent, semiurgent, urgent, emergent, or unknown/no triage), and disposition status (no follow-up planned, return if needed, return/refer to other physician or clinic, transfer to other facility, or admit to hospital). Finally, we compared the mean number of diagnostic tests and procedures as well as the percentage of subsequent hospitalization between ED visits with and without ACSCs.
To examine the relationship between ACSCs and hospitalization following ED visits, a multivariate logistic regression model was performed. The dependent variable for the model was whether the visit resulted in admission to the hospital. Patients who were dead on arrival or died in ED were excluded from the logistic regression. The main independent variable was the presence of ACSCs. Presence of ACSCs was determined by the first-listed diagnosis as coded by the International Classification of Diseases, 9th Revision, Clinical Modification codes as developed by Millman et al (IOM, 1993). The first-listed diagnosis is the diagnosis to which the ED visit has been attributed. The list of 13 ACSCs included grand mal seizure disorders, severe ear, nose, and throat infections, chronic obstructive pulmonary disease (COPD), bacterial pneumonia, asthma, congestive heart failure (CHF), hypertension, angina, cellulitis, diabetes, hypoglycemia, kidney/urinary tract infection, and dehydration.
In the multivariate model, we also controlled for several possible confounders based on the Andersen model of health services use (Aday & Andersen, 1974). The predisposing components include variables that describe individuals’ tendency to use services. The enabling components represent the means that are available to individuals for the use of services, including both resources to the individual and family as well as attributes of the community. The need components refer to illness level. The need components are hypothesized to be the most immediate determinants of health service use while the predisposing components are the most distant determinants. In this article, predisposing factors included age (65–74 years, 75–84 years, 85 years, and over), sex, race (White, non-White). Enabling factors included expected payment source (private insurance, Medicare, Medicaid, other) and metropolitan statistical area (MSA) status (yes, no). In addition to ACSCs, need factors included number of procedures (0–1, 2, or more), number of ED diagnoses (0–1, 2, or 3), and triage information (<15 min, 15–60 min, 1–2 hr, 2 or more hours, unknown/no triage). We also included individual year indicators to examine year effect.
All estimates presented have been weighted to account for the complex sample design of NHAMCS. The weight includes four basic components: inflation by the reciprocal of the probability of selection at the provider and visit level, adjustment for nonresponse, a population weighting ratio adjustment, and weight smoothing. The sum of the visit weights is an unbiased estimate of the annual number of visits. Detailed information on estimation for NHAMCS data is described elsewhere (NCHS, 2010).
SUDAAN was used to perform the statistical analyses in this study (RTI, 2005). The standard errors presented in tables and used in tests of significance for this article were approximated. SUDAAN computes standard errors using a first-order Taylor approximation of the deviation of estimates from their expected values. Statements of differences in estimates are based on statistical tests (e.g., chi-square tests or students-t) with significance at the p < .05 level.
The study was reviewed by National Center for Health Statistics (NCHS) Research Ethics Review Board. Waivers of the requirements to obtain informed consent of patients and patient authorization for release of patient medical records by health care providers have been granted.
Results
ED Visit Rates
During 2001–2008, a total of 4,970 (unweighted) ED visits were made by NH residents aged 65 and over which on average is 2.0 million visits annually. Table 1 presents the average annual rate of ED visits for elderly NH residents by selected characteristics. The overall visit rate was 123.2 per 100 persons per year. Although the difference was not significant, the rate of ED use by elderly nursing home residents was higher in 2005–2008 than in 2001–2004 (128.3 per 100 persons per year compared with 118.1 per 100 persons, p = .15). The visit rate for NH residents 85 years and older was lower than the rate of residents aged 65–74 (113.0 vs. 153.2 per 100 persons per year, p < .001). Male NH residents had a higher visit rate than female residents (154.5 vs. 111.6 per 100 persons per year, p < .001). Non-White NH residents had a higher visit rate than White residents (175.6 vs. 117.0 per 100 persons per year, p < .001).
Average Annual Rate of Emergency Department (ED) Visits for Elderly Nursing Home Residents by Selected Characteristics: United States, 2001–2008.
Note. aInstitutionalized population estimates are derived by subtracting estimates of the civilian noninstitutionalized population from estimates of the civilian resident population as of July 1 of each year 2001–2008. Each set of annual population estimates are projections based on Census 2000 data developed by the Population Division, U.S. Census Bureau using the July 1 set of state population estimates. More information can be obtained from the Census website at www.census.gov.
Patient Characteristics
Table 2 describes attributes of ED visits made by NH residents. The average age of elderly ED patients from NHs was 83 (data not shown). The oldest (85 years and older) age group accounted for 44.1% of NH resident ED visits. Among ED visits made by NH residents, female residents comprised 66.1% of ED visits and White residents comprised 85.0% of the ED visits. Nearly 86% of ED visits made by elderly NH residents had Medicare as a payment source for the ED visits and about a third lived in the South.
Patient and Emergency Department (ED) Visit Characteristics Among Elderly Nursing Home Residents Who Visited the ED: United States, 2001–2008.
Note. aFigure does not meet standards of reliability and precision.
Health Status, Length of Visit, and Disposition
Among ED visits made by NH residents, 61.2% had two or three diagnoses related to the visit (Table 2). Approximately a third of ED visits were made by NH residents who had emergent conditions (needed to be seen in less than 15 min) and 37.4% of the ED visits were made by those with urgent conditions (needed to be seen between 15 and 60 min). Approximately 15% had at least one ACSC. The most frequent ACSC was kidney/urinary tract infection, followed by CHF and dehydration. About 48.7% of all ED visits resulted in admission to the hospital and 33.5% of all ED visits resulted in return or referral to other physician or clinic. ED visits with ACSCs had a higher percentage of subsequent hospital admission than those without ACSCs (54.6% vs. 47.7%, p = .01, data not shown). On average, elderly NH residents spent 4.7 hr in the ED (Table 3).
Services Ordered or Provided During Emergency Department (ED) Visits by Elderly Nursing Home Residents: United States, 2001–2008.
Diagnostic Tests and Procedures
On average, NH residents had 6.1 diagnostic or screening tests provided during the ED visit, and 61.9% of ED visits had five or more tests performed (Table 3). Complete blood count (CBC) and X-ray were the most common tests followed by electrocardiogram (EKG), blood urea nitrogen/creatinine, and glucose testing. On average, 22.2% of the visits had two or more procedures performed. The most common procedures were intravenous (IV) fluids and bladder catheter. ED visits with ACSC diagnoses had a higher average number of diagnostic or screening tests performed than those without ACSCs (7.0 vs. 6.0, p < .001, data not shown). No significant differences were observed on the average number of procedures performed or time spent in ED.
Characteristics Associated With Hospitalization
Table 4 presents the unadjusted percentages and adjusted odds ratios (aOR) for hospital admission subsequent to ED visit. Factors that were significantly associated with hospital admission in the unadjusted analyses were also significant in the multivariate analysis. After adjusting for patient sociodemographic and clinical characteristics, NH residents with two or more procedures performed in ED visits were more than twice as likely to be admitted to the hospital compared with those having less than two procedures performed (aOR = 2.54, 95% confidence interval [CI] = [2.13, 3.04]). The likelihood of hospitalization also increased as the number of ED diagnoses increased. Compared with NH residents with nonurgent conditions, those who needed to be seen in less than 2 hr were more likely to be hospitalized, and the odds of admission increased with urgency. Several of the selected ACSCs were significantly associated with hospital admission including COPD (aOR = 1.87, 95% CI = [1.00, 3.48]), CHF (aOR = 3.66, 95% CI = [2.29, 5.86]), kidney/urinary tract infection (aOR = 1.65, 95% CI = [1.21, 2.25]), and dehydration (aOR = 3.16, 95% CI = [1.89, 5.27]).
Hospital Admission From the Emergency Department (ED) Visit, by Patient Characteristics: United States, 2001–2008.
Note. −2 log-likelihood ratio: 634.7.
Degrees of freedom: 35.
R 2: .12.
*p < .05 (for multivariate model). **p < .05 (for unadjusted model).
aFigure does not meet standards of reliability and precision.
Discussion
The large sample combining 8 years of national data permitted investigation of services provided during ED visits and the relationship between individual ACSCs and hospital admission from the ED. During 2001–2008, the average annual number of ED visits made by elderly NH residents was 2.0 million. The overall visit rate of 123.2 per 100 institutionalized population indicated that on average NH residents had more than one ED visit per year. The visit rate was lowest among those aged 85 and over, and higher among male or non-White NH residents. Visit rates by age and sex are consistent with a study that used Centers for Medicare & Medicaid Services administrative claims and NH resident assessment data to examine ED use by NH residents (Stephens, Newcomer, Blegen, Miller, & Harrington, 2012). Evidence suggests that African Americans were more likely to be admitted to NHs with poor quality which may contribute to the higher visit rate than Whites (Grabowski, 2004).
Some of the characteristics of ED visits made by elderly NH residents shown in this study were similar with those reported in previous studies (Ackermann et al., 1998; Wang et al., 2011). Approximately two thirds of the ED visits were made by women and the most common laboratory tests were CBC, X-ray, and EKG. ED length of stay was almost 5 hr.
About one in five of the ED visits had ACSC diagnoses and half of all the ED visits led to admission to the hospital. ED visits with ACSCs had a higher percentage of subsequent hospital admission than those without ACSCs. Our study showed that the presence of COPD, CHF, kidney/urinary tract infection, and dehydration was associated with hospitalizations subsequent to ED visits after controlling for other covariates. The increased likelihood of hospital admission associated with presence of any of these ACSCs ranged from 2 to 3 times higher than the likelihood for ED visits without these ACSCs.
In addition to ACSCs, certain additional needs and enabling factors were significantly associated with hospitalizations subsequent to ED visits. Significant factors included all of the other need component variables (i.e., number of procedures, number of diagnoses, and triage status) and two enabling component characteristics (i.e., non-MSA status and other insurance as payment source). None of the variables under the predisposing components were significant in the model. The results support the hypothesis that the need components are the most proximate determinants of health services use and the predisposing components are the most distal.
These findings regarding ACSCs reinforce the results of previous studies which suggest that many of the ED visits and subsequent hospitalizations can be prevented if NH residents receive appropriate care (e.g., early detection and treatment of chronic conditions) or on-site evaluation in the NH (Ackermann et al., 1998; Rector, Spector, Shaffer, & Finch, 2005).
Other studies indicated that ED use by NH residents may be reduced by routine use of advance directives, treatment of substance abuse problems, and more frequent use of physician assistants or nurse practitioners within the NH (Kayser-Jones, Wiener, & Barbaccia, 1989; Kerr & Byrd, 1991). Another study showed that some services provided in the ED would not have been necessary had there been adequate communication with NH staff (Ackermann et al., 1998).
Although not all ACSCs are indicative of avoidable hospitalizations, it is possible that some of the clinical symptoms leading to complications were not recognized early enough in the disease process for intervention treatments to be successful, particularly in the case of residents hospitalized due to dehydration (Carter, 2003). A previous study found that NHs with nurse practitioners or physician assistants on staff, the provision of IV therapy, and the operation of certified nurse assistant training programs appeared to reduce ACSCs hospitalizations (Intrator et al., 2004). In addition, significantly lower hospitalizations among NH residents were demonstrated among nursing facilities with nurse practitioners or physician assistants, special care units, and more physicians (Buchanan et al., 1990; Garrard et al., 1990; Intrator et al., 1999). Moreover, recent evaluations of the Evercare model of intensive primary care provided in NHs by the Centers for Medicare & Medicaid Services found that the hospitalization rate for Evercare enrollees was half that of controls, and that ED use for Evercare enrollees was half than that of controls (Kane, Keckhafer, & Robst, 2002). Although these studies were not able to randomly assign NHs to different models of primary care delivery, they suggest a correlation between adequate on-site medical evaluation and staffing and a reduction in ED visits and hospitalizations made by NH residents.
Promoting new models of primary care and care coordination which may reduce overall health care costs is one of the goals of the 2010 Patient Protection and Affordable Care Act (Goodson, 2010). These provisions reflect the increasing popularity of health care delivery models such as accountable care organizations (ACOs) and patient-centered medical homes with policy makers and administrators. Care coordination that reduces ED visits and hospital admissions among the elderly are emphasized in recent initiatives such as the NH Value-Based Purchasing demonstration, the Bundled Payment initiative, and the Shared Savings program (Centers for Medicare & Medicaid Services, 2009, 2012; U.S. Department of Health & Human Services, 2012). The NH Value-Based Purchasing demonstration assesses NH performance based on selected performance measures and the payment pool for each state will be determined based on Medicare savings that result from reductions in Medicare expenditures, primarily from hospitalizations. Bundled payments and the Shared Savings Program give providers new incentives to coordinate care, improve the quality of care, and save money from Medicare. The bundled payment initiative aligns payments for services delivered across an episode of care, rather than paying for services separately. By bundling payment across providers for multiple services, providers will have a greater incentive to coordinate and ensure continuity of care across settings. The Medicare Shared Savings Program allows participating ACOs to share with the federal government in the savings achieved by providing more efficient care while meeting performance standards on quality of care. The Shared Savings Program is designed to improve quality of care and patient outcomes by promoting accountability for the care, requiring care coordination, and encouraging investments in infrastructure and redesigned care process.
The strengths of this study include the use of multiple years of national data to study the use of ED services by elderly NH residents. NHAMCS provides detailed data on the provision of services in hospital EDs including diagnostic or screening services, procedures, triage status, and disposition of the ED visits. In addition, because NHAMCS also provides information on patient demographic characteristics and clinical conditions, we were able to take these characteristics and conditions into account when evaluating factors associated with hospital admission following an ED visit. The large sample size allowed for testing of each individual ACSC which has not been reported in previous national studies. We were also able to report a detailed list of services and procedures provided during the ED visit.
Our study is subject to some limitations. First, our findings assume in 2001–2004 that all ED visits made by the elderly residing in a NH or other institution were from a NH rather than another institution, such as a prison or mental hospital. However, the number of visits from institutions other than NHs should be small. According to the 2000 Census, only 5.1% of institutionalized persons aged 65 and over were in institutions other than NHs (U.S. Census Bureau). Although it is possible that some ED visits with positive responses to the question in 2001–2004 may have been in an institution other than a NH (e.g., prison, mental hospital), we limited this possibility by focusing our analysis exclusively on ED visits among the elderly (i.e., those aged 65 or older). We compared 2001–2004 estimates with 2005–2008 estimates for potential differences due to question wording and found no statistical differences in estimates. Also, the ED visits from 2001 to 2004 versus 2005 to 2008 had very similar basic characteristics (e.g., patient demographics and number of diagnosis). In addition, we evaluated the model using indicator variable for 2001–2004 versus 2005–2008 in place of the yearly indicator to examine potential issues with question wording differences. The results were essentially the same as those reported which adds to our confidence in our findings despite slightly different strategies used to measure the dependent variables in the two time periods. The number of ED visits and the visit rate made by elderly NH residents fluctuated across the years, but the differences were not significant. Also, our estimates of ED visit rates among elderly NH residents may be lower bound estimates as we assumed that none of the ED visits among patients with unknown place of residence resided in a NH. Finally, the analysis was limited by the lack of information on the characteristics of the NH and the residents. These data were not available in NHAMCS, which was primarily designed to measure characteristics of ED visits.
In conclusion, this study presents characteristics of ED visits by elderly NH residents at the national level. Nearly half of ED visits made by elderly NH residents resulted in hospital admission, and selected ACSCs were positively associated with hospital admission outcomes. The types of ACSCs associated with hospital admissions from EDs among NH residents suggest further research is needed on their clinical management within NHs. Previous studies suggested that having adequate medical evaluation and staffing as well as appropriate care in the NH may help reduce ED visits and hospital admissions among NH residents. Recent initiatives seek to reduce ED visits and hospital admissions among the elderly by providing financial incentives to promote accountability for the care and care coordination. More research is needed to investigate whether increased monitoring by NHs, timely access to care, or interdisciplinary resident-centered care could have avoided hospitalizations for ACSCs noted in our study. Transferring NH residents to hospitals can be very disruptive and disorienting, particularly for older people with dementia. In addition, hospitalizations often result in unintended problems such as bedsores, infections, and delirium. At the same time, it is important to recognize that the majority of hospitalizations of NH residents are not classified as ACSCs, and it is possible that some hospitalizations classified as ACSCs cannot be avoided. Future research should better identify (1) the relationship of NH characteristics with ED use and hospitalization, and (2) opportunities to improve primary care provided in NHs.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.