Over the past 20 years, hospital-based emergency departments (EDs) have played an increasingly larger and more critical role in managing the medical care of older adults. Older adults have distinct patterns of service use and care needs in the ED.[1, 2] Older adults seen in the ED present with higher-acuity conditions that are more complex because they are accompanied by multiple comorbid conditions, atypical presentations of common diseases, and medical complications that result in traumatic injury.[3, 4] The result is often high ED resource use in the form of more-advanced imaging, laboratory and urine testing, time-consuming care coordination, long ED stays, and more-frequent hospital admissions. Older adults are also at high risk of adverse health outcomes after an ED visit; during the first 3 months after ED discharge, approximately 5% of older people will die, 20% will be hospitalized, and 20% will have a return outpatient ED visit.[6-8]
The needs of older adults in the ED are superimposed on wider, systemic problems with the emergency care system. The lack of primary care access is a major contributor to ED use for acute care needs. Older adults who enter the ED system face ED crowding, a major problem recognized by the Institute of Medicine and present in more than 90% of EDs. At the same time, standards of care within the ED and across medical care have changed, with greater reliance on high-cost imaging and greater intensity of services delivered.
This article explores demographic changes in the use of EDs by adults aged 65 and older from 2001 to 2009 using a nationally representative sample of ED encounters. Specifically, it describes broad demographic changes in the use of EDs by older adults and changes in the severity of illness, presenting complaints, and intensity of services. It also explores quality over time using measures of ED crowding, medication administration, and 72-hour return visits.
- Top of page
- Supporting Information
Data were used from the 2001–2009 National Hospital Ambulatory Medical Care Survey (NHAMCS). NHAMCS data are collected annually and represent a national probability sample of visits to hospital-based EDs of noninstitutional, general, and short-stay hospitals. Data from free-standing EDs and urgent care centers are not included. The Centers for Disease Control and Prevention (CDC) National Center for Health Statistics have collected data since 1992. The NHAMCS data are generated through a multistate estimation procedure whose goal it is to produce unbiased estimates of ED visit rates and specific services across the United States.
The 2001–2009 NHAMCS dataset includes information from 322,745 ED encounters representing 1.05 billion ED visits. NHACMS is publicly available data and contains no patient identifiers, so the institutional review board at the George Washington University found the study not to be human subject research.
The overall goal was to describe trends in care for older adults who presented to U.S. EDs over a 9-year period. First, changes in ED use for older adults (divided into age categories: 65–74, 75–84, and ≥ 85) were compared with those of younger individuals. The number of ED visits in each age category was tabulated according to year, then the raw and percentage differences between 2009 and 2001 data were computed.
The sample was then restricted to older adults (≥ 65), and the number of individuals was tabulated according to demographic characteristics (race and ethnicity, sex), payment source, disposition, and hospital-level variables (geographic location, profit status, urban vs nonurban). Similarly, the raw and proportional differences in visit numbers between 2009 and 2001 were calculated. Changes in principal reason for visits for older adults were then examined, and the 10 top reasons for visits were tabulated according to year to determine whether the types of medical complaints had changed over time in older adults.
Next, changes over time for resources used in the ED such as imaging tests (computed tomography (CT), magnetic resonance imaging (MRI), ultrasound), electrocardiograms (EKG), cardiac monitoring, and laboratory tests (blood tests and urinalysis), intravenous fluid administration, medication use, and procedures were tabulated. Several of the variables related to testing changed over the study period, and all data available for the specific test are presented and the first year the data were available is compared with the last for raw and proportional differences. For example, CT and MRI were combined before 2004 and reported separately in 2005 and after.
Changes in quality of care were then explored. For medication use, the proportions of encounters when a medication was given in the ED or at discharge and when any inappropriate medication was given according to the 2002 Beers criteria (a listing of inappropriate medications for older adults) were computed. How measures of ED crowding such as waiting time, length of stay, and rates of leaving without being seen had changed for older adults over the study period were then explored. Lastly, the number of older adults who were seen and returned for care in the same ED (72-hour return rate, a measure of quality) and the 72-hour return and admission rate were tabulated.
This article uses descriptive statistics and raw and proportional differences across years. Hypothesis testing was not performed because this study was exploratory. All data were tabulated using Stata, version 10 (Stata Corp., College Park, TX) and accounted for the weighted, complex-survey design.
- Top of page
- Supporting Information
From 2001 to 2009, NHAMCS included data on 47,554 separate ED encounters by older adults (≥ 65) representing 156 million U.S. ED visits. Over the course of the study period, visits for older adults increased from 15.9 million per year in 2001 to 19.8 million visits per year in 2009, a 24.5% increase, whereas overall ED visits increased from 107.5 million to 136.1 million per year, a 26.6% increase. The greatest increase was seen in individuals aged 55 to 64, whose visits increased 57.9% from 2001 to 2009. Increases in other age groups were 27.3% in those aged 65 to 74, 18.3% in those aged 75 to 84, and 30.3% in those aged 85 and older (Table 1).
Table 1. Demographic Characteristics of Individuals Aged 65 and Older: Emergency Department Visits and Hospital Use (2001–2009)
|Million Visits||Million Visits (%)|
|≥ 85||3.3||4.3||1.0 (30.3)|
|< 18||26.2||33.2||7.0 (26.7)|
|Aged ≥ 65||15.9||19.8||3.9 (24.5)|
|Demographic, individuals aged ≥ 65|
|Race and ethnicity|
|White, non-Hispanic||10.3||13.1||2.8 (27.2)|
|Black, non-Hispanic||1.7||2.2||0.5 (29.4)|
|Source of payment|
|For profit||1.6||1.7||0.1 (6.2)|
|Government, nonfederal||2.2||2.0||−0.2 (−9.1)|
In the population aged 65 and older, Hispanics experienced the lowest growth in visits, 11.1% from 2001 to 2009, versus 27.2% in whites and 29.4% in blacks. From 2001 to 2009, there were increases in older adults with private insurance (16.7%), Medicare (32.2%), and other sources of payment (25.0%) and a large decrease in the number of those with Medicaid (−37.5%).
The number of individuals who were treated and discharged increased less than the number of those admitted to the hospital (20.2% vs 33.2%), whereas the number admitted to the intensive care unit (ICU) increased the most (131.3%) (Table 1). The most common chief complaints over the study period were chest pain, shortness of breath, and abdominal pain. The ranking of the top complaints changed some over the study period (Table 2).
Table 2. Ten Most Common Reasons (Number of Visits) for Visits by Individuals Aged 65 and Older to U.S. Emergency Departments (2001–2009)
|1||Chest pain (1,349,525)||Shortness of breath (1,223,105)||Chest pain (1,359,379)||Chest pain (1,215,453)||Chest pain (1,176,215)||Chest pain (1,318,621)||Chest pain (1,353,867)||Shortness of breath (1,437,473)||Chest pain (1,510,829)|
|2||Shortness of breath (1,048,250)||Chest pain (1,180,384)||Shortness of breath (1,265,599)||Shortness of breath (925,594)||Shortness of breath (1,120,668)||Shortness of breath (1,125,170)||Shortness of breath (1,104,180)||Chest pain (1,408,241)||Shortness of breath (1,393,543)|
|3||Abdominal pain (528,435)||Accident, NOS (682,542)||Abdominal pain (749,446)||Accident, NOS (683,246)||Abdominal pain (786,200)||Abdominal pain (759,985)||Abdominal pain (831,031)||Abdominal pain (837,974)||Abdominal pain (821,189)|
|4||Accident, NOS (508,187)||General weakness (584,746)||Accident, NOS (665,793)||General weakness (599,911)||General weakness (676,487)||Accident, NOS (714,714)||General weakness (639,671)||Accident, NOS (813,900)||General weakness (774,243)|
|5||Vertigo (501,512)||Vertigo (576,388)||General weakness (655,062)||Vertigo (501,639)||Accident, NOS (490,095)||General weakness (550,646)||Accident, NOS (523,281)||General weakness (740,033)||Vertigo (607,894)|
|6||Dyspnea (452,054)||Abdominal pain (534,642)||Dyspnea (504,662)||Dyspnea (372,210)||Vertigo (468,513)||Vertigo (490,183)||Vertigo (494,593)||Dyspnea (509,595)||Back pain (481,807)|
|7||Headache (280,613)||Dyspnea (414,003)||Vertigo (496,577)||Back pain (316,654)||Nausea (333,077)||Nausea (378,478)||Dyspnea (385,477)||Vertigo (465,914)||Accident, NOS (436,218)|
|8||Back pain (270,419)||Nausea (330,041)||Fever (369,528)||Nausea (304,851)||Back pain (331,639)||Back pain (374,202)||Cough (332,984)||Nausea (410,065)||Syncope (406,308)|
|9||Nausea (258,476)||Cough (307,775)||Nausea (301,219)||Cough (292,186)||Syncope (321,269)||Dyspnea (350,671)||Nausea (321,078)||Back pain (371,229)||Psychological symptoms (391,920)|
|10||Cough (224,994)||Vomiting(290,864)||Leg pain (298,544)||Hip pain (244,385)||Dyspnea (314,023)||Psychological symptoms (327,144)||Fever (310,426)||Psychological symptoms (365,000)||Dyspnea (373,900)|
The greatest growth in ED visits was in the western United States, where visits per year increased from 2.8 million in 2001 to 5.0 million in 2009 (78.6% increase), followed by the South (27.3% increase). In older adults, the greatest increase in visit rates occurred in nonprofit hospitals (33.1%). The number seen in government hospitals shrank 9.1%.
With regard to providers, there was minimal (< 1%) change in the proportion of encounters in which a resident or intern was involved, which remained constant at approximately 10%, but dramatic growth in the proportion of encounters with a physician assistant, which from 2001 to 2009 increased from 5.1% to 6.9%, and a nurse practitioner, which increased from 1.0% to 3.1% over the study period.
There was rapid growth in the intensity of service use for older adults over the study period (Table 3). From 2001 to 2009, the overall proportion of individuals receiving radiographic imaging grew 5.7 percentage points, from 61.8% receiving any imaging in 2001 to 67.5% in 2009. There was little growth in the use of plain X-ray and ultrasound imaging between 2001 and 2009. By comparison, the proportion of individuals receiving CT grew 14.5 percentage points, from 12.2% in 2001 to 26.7% in 2009. EKG use increased 6.2 percentage points, from 40.4% in 2001 to 46.6% in 2009, and the use of cardiac monitoring increased 7.7 percentage points from 17.7% in 2001 to 25.4% in 2009. Laboratory tests also became more common, with urinalysis growing substantially more than blood tests, with growth rates of 10.9 percentage points (from 21.5% in 2001 to 32.4% in 2009) and 9.0 percentage points (from 58.1% in 2001 to 67.1% in 2009), respectively.
Table 3. Use of Resources for Individuals Aged 65 and Older in U.S. Emergency Departments (EDs) (2001–2009)
|Tests Ordered||2001||2009||Increase 2001–2009%|
|Patients Receiving Service, %|
|Any computed tomography||12.2||26.1||13.9|
|Magnetic resonance imaging||—||1.2||0.4a|
|Received medicines in ED||72.4||75.8||3.4|
|Received prescription at discharge||—||29.4||–1.9a|
|Received intravenous fluids in ED||36.0||46.1||10.1|
With regard to medication use, the growth in number of individuals receiving medications in the ED increased 3.4 percentage points, from 72.4% in 2001 to 75.7% in 2009, lower than the growth in overall visit rates (Table 3). In addition, data regarding prescriptions given at discharge were available from 2005 to 2009; over this period, there was a 1.9% decrease in the number of patients receiving prescriptions at discharge. There was also large growth in the number of patients receiving intravenous fluids (10.1%) and procedures (5.9%).
There were changes in a number of quality measures over the study period (Table 4). In 2001, 9.6% of older adults received medications deemed inappropriate according to the 2002 Beers criteria, compared with 4.9% in 2009. The rate dropped from 2004 to 2007 and leveled off in the 4–5% range. Data regarding prescriptions given at discharge were available from 2005 to 2009; over this period, prescribing of potentially inappropriate medications (PIMs) at discharge decreased by 1.4 percentage points, from 5.7% to 4.3%.
Table 4. Changes in Quality of Care and Emergency Department (ED) Crowding Indices for Individuals Aged 65 and Older Seen in U.S. EDs: 2001–2009
|Potentially inappropriate medications, %|
|Beers criteria violated in ED||9.6||4.9||–4.7|
|Beers criteria violated at discharge||—||4.3||1.4a|
|Admission numbers, proportions, and ED lengths of stay|
|Admitted to inpatient floor, %||36.2||38.7||2.5|
|Admitted to inpatient floors, million (%)||5.8||7.7||1.9 (33.2)|
|Average length of ED visit forindividuals admitted to inpatient floor, minutes (%)||327.9||329.0||1.1 (0.3)|
|Average length of admission for individuals on inpatient floor, days (%)||—||5.7||–0.3 (–5.5)a|
|Admitted to ICU, million (%)||0.5||1.1||0.6 (131.3)|
|Average length of ED visit for individuals admitted to ICU, minutes (%)||244.4||286.4||42.0 (17.2)|
|Average length of admission for individuals in ICU, days (%)||—||5.5||–1.5 (–21.4)a|
|Admitted for observation, million (%)||0.2||0.6||0.4 (250.0)|
|Average length of ED visit for individuals being observed, minutes (%)||353.6||356.9||3.3 (0.9)|
|Average length of admission for individuals being observed, days (%)||—||4.5||–1.0 (–17.5)a|
|Number left without being seen, million (%)||0.2||0.2||0.1 (43.8)|
|Length of visit for all older adults, minutes (%)||233.3||258.3||25.0 (10.7)|
|Patients, million (%)||0.5||0.8||0.3 (60.0)|
Standard measures of ED crowding, including waiting times and length of stay, did not appear to change appreciably throughout the study period in older adults, with the exception being the number of individuals who left without being seen, which increased 43.8%. Admission rates grew from 36.2% in 2001 to 38.7% in 2009, an increase of 2.5 percentage points. The number of individuals who had been seen in the same ED in the previous 72 hours increased 60.0% over the study period. Of individuals seen in the prior 72 hours, the proportion admitted at the return visit increased 2.2 percentage points, from 2.0% in 2001 to 4.2% in 2009. Full data from each year from 2001 to 2009 are included in Online Appendices 1 to 3.
- Top of page
- Supporting Information
There have been several changes in older adults’ use of EDs in the United States in the past decade. ED use increased 25% for adults aged 65 and older from 2001 to 2009. According to the Census Bureau, the population of adults aged 62 and older increased 21% from 2000 to 2010, from 41 million to almost 50 million people. The data are not directly comparable because the dates and ages do not perfectly overlap, but it suggests that ED use for older adults is outpacing population growth. Using 2000 and 2010 Census figures for the population aged 65 and older, visits in 2001 increased from 45.4 per 100 older adults in 2000 to 49.2 visits per 100 older adults in 2009. The greatest increase in use was in the oldest old (≥ 85) which probably mirrors demographic changes of an aging population, although there was even greater increase in use of individuals aged 55 to 64, which may reflect the aging of the baby boom generation and suggests that there may be even greater ED use for this group in the next decade. The greatest growth in ED use was in the western United States, followed by the South, which somewhat reflects migration patterns in retirement for older adults, although the growth of ED visits in the West (close to 80%) was much higher than the population growth in this group. According to U.S. census data from 2000 and 2010, growth in the population aged 65 and older was greatest in the West (23.5%), followed by the South (19.7%), Midwest (9.2%), and Northeast (5.9%). The current study confirms also a continued rise in the use of EDs and emergency medical services by older adults that has been documented in previous work.[16, 17]
Little change was found in the types of symptoms that brought older adults to the ED, with chest pain, shortness of breath, and abdominal pain being the three most common complaints. Many of the symptoms in the 10 most-common complaints of older adults are potentially serious and can represent life-threatening emergencies. This has public policy implications. It appears that shifting many of these encounters to lower-resource primary care sites may not be possible, indicating the unique role of the ED in acute medical care for older adults, but several new care models, such as the geriatric patient-centered medical home, the Program for All-inclusive Care of Elders, and the Interventions to Reduce Acute Care Transfers II programs may have the potential to alter the frequency of some lower-acuity ED visits or the number of return visits in the future.[18-20] In addition, accountable care organizations might create financial disincentives to have high rates of ED use.
One of the most dramatic findings was the increase in resource intensity of ED visits in older adults. This was observed at several levels. First, admission to the hospital for older adults increased more than individuals who were treated and released. In addition, there was an almost doubling in the use of ICU services, a strong indicator of higher rates of critical illness in this population. There were also dramatic increases in the use of advanced imaging tests, such as CT and MRI, along with increased use of laboratory tests, EKGs, cardiac monitoring, and urinalysis. Part of this may reflect changes in approaches to diagnosing ED patients in general, which has increasingly relied on objective testing rather than clinical judgment. It was not possible to assess whether increases in testing were evidence-based, although it does suggest an older, sicker population that is increasingly requiring more-intense resources to manage health problems. This may be a manifestation of improved technology outside the ED, such as chemotherapy, transplantation, and other advances that can prolong life, but require more resources to manage in the event of a complication or worsening of illness.
Another notable finding in this study is the proportion of older adults who received PIMs in the ED. Prior research using similar methods found higher rates of PIM use than found in the current study, which may be lower because the other study restricted analyses to discharged individuals, whereas as the current analysis included all individuals for whom data were available.
There were not large changes in quality measures related to ED crowding over the study period. These trends differ from other studies demonstrating increases in waiting times and lengths of stay using NHAMCS, which could indicate that increases in crowding in U.S. hospitals may be affecting older adults less than other groups.[22, 23] Finally, increases in return visits for older adults and rates of return and readmission within a short period of time (within 72 hours) were found. This may indicate a reduction in quality of care of older adults over the study period because more are returning for problems that were not addressed in the initial visit, although it is possible that this may be another indicator of greater severity of illness, indicating that even individuals who have been discharged are requiring greater downstream services.
There are several limitations of this study. One is the amount of missing data in some of the NHAMCS variables—more than 10% for some. In addition, some of the definitions changed over the study period; for example, as noted, CT and MRI were combined before 2005, which may result in an overestimate of early CT use. Also, NHAMCS is a probabilistic sampling of visits in specific EDs and is not complete. It is possible that these estimates are not completely accurate, although the consistency and the rigor of the methods and sampling framework are designed for this purpose.
The Beers criteria were used to judge PIM prescription. The Beers criteria have been criticized as inflexible and were not originally established for individuals in the ED. More-suitable criteria might be the 2012 Beers criteria or the Screening Tool of Older People's potentially inappropriate Prescriptions/Screening Tool to Alert doctors to the Right Treatment criteria.[24-26] Finally, these data are based on retrospective chart review, so it is difficult to know whether severity of illness is increasing or whether these changes can be attributed to evolving use patterns for testing, hospitalization, and ICU use (care standards). While we cannot directly measure severity, the plurality of findings that suggest a sicker patient population seems indisputable.
- Top of page
- Supporting Information
Conflict of Interest: Dr. Pines received a grant from the National Priorities Partnership on Aging that provided salary support for his role in this manuscript.
Author Contributions: JMP designed and obtained funding for this study. PMM and LBB acquired and analyzed the data. JMP prepared the first draft of this manuscript. JMP, PMM, JKC, LBB, and KER provided critical revisions to the manuscript, guided additional study questions that arose during the development of the manuscript, and helped to interpret the results.
Sponsor's Role: The sponsor had no role in the design, methods, subject recruitment, data collections, analysis, or preparation of the manuscript.