MODELS OF GERIATRIC CARE, QUALITY IMPROVEMENTAND PROGRAM DISSEMINATION

A Geriatric Emergency Service for Acutely Ill Elderly Patients:Pattern of Use and Comparison with a Conventional EmergencyDepartment in Italy

Fabio Salvi, MD,� Valeria Morichi, MD,� Annalisa Grilli, MD,� Raffaella Giorgi, MD,w

Liana Spazzafumo, MD,z Stefano Polonara, MD,§ Giuseppe De Tommaso, MD,w

Alessandro Rappelli, MD,� and Paolo Dessı-Fulgheri, MD�

The current disease-oriented, episodic model of emergencycare does not adequately address the complex needs of olderadults presenting to emergency departments (EDs). Dedi-cated ED facilities with a specific organization (e.g., geri-atric EDs (GEDs)) have been advocated. One of the fewGED experiences in the world is described and its outcomescompared with those of a conventional ED (CED). In asecondary analysis of a prospective observational cohort of200 acutely ill elderly patients presenting to two urban EDsin Ancona, Italy, identifiers and triage, clinical, and socialdata were collected and the following outcomes considered:early (30-day) and late (6-month) ED revisit, frequent EDreturn, hospital admission, and functional decline. Death,functional decline, any ED revisit and any hospital admis-sion were also considered as a composite outcome. Oddsratios and 95% confidence intervals (CIs) were calculated.Overall, GED patients were older and frailer than CEDpatients. The two EDs did not differ in terms of early, late,or frequent ED return or in 6-month hospital admission orfunctional decline. The mortality rate was slightly but sig-nificantly lower in the GED patients (hazard ratio 5 0.47,95% CI 5 0.22–0.99, P 5.047). The data suggest noninfe-riority and, indirectly, a slight superiority for the GEDsystem in the acute care of elderly people, supportingthe hypothesis that ED facilities specially designed forolder adults may provide better care. J Am Geriatr Soc56:2131–2138, 2008.

Key words: geriatric health services; emergency medicine;outcomes; aged

Elderly people are an ever-increasing population in over-crowded emergency departments (EDs).1 Their com-

plex medical and social needs require more time andresources than those of younger adults.1,2 Older adults arefrequently admitted1–3 and when discharged from the EDface adverse health outcomes such as ED return, hospital-ization, functional decline, and death.1,2,4–7

It is widely agreed that the current disease-oriented,episodic model of emergency care does not adequately ad-dress the complex needs of older patients.8,9 The aim of EDsis to provide acute intervention and timely health care to allpatients with emergent or urgent problems. When a med-ically complex older person with reduced mobility, im-paired memory, or poor social support presents to the ED,the system experiences crisis, slows down, and becomes in-efficient. Unfamiliarity of the ED staff with the manage-ment of geriatric patients further complicates the clinicalapproach.1,10

Geriatric EDs (GEDs), modeled on trauma or pediatriccenters and endowed with suitable equipment, an appro-priate environment, and dedicated clinical and nursingprotocols, have been advocated.1,8,9 Two such facilities arefound at the Nassau University Medical Center (Long Is-land, NY) and the Hadassah University Medical Center(Jerusalem, Israel), which have opened a geriatric emer-gency room staffed by geriatricians rather than emergencyphysicians.1,8 A geriatric emergency service was opened 12years ago at the Italian National Institute for Research andCare on Aging (INRCA), Ancona, Italy. This is a ‘‘hybrid-ized ED and observation unit’’ (as later suggested)8 man-aged by geriatric staff with several years’ clinical experiencein emergency medicine. Unfortunately, no data are avail-able to document the effectiveness of this novel approach.

Here patient characteristics and 6-month mortality, EDreturn, hospitalization, and functional decline are describedin a sample of geriatric patients from two Italian EDs todetermine the noninferiority of a GED compared with aconventional ED (CED).

Address correspondence to Fabio Salvi, MD, Clinica di Medicina Interna,Azienda Ospedali Riuniti, via Conca n. 71, 60131 Ancona, Italy.E-mail: f.salvi@univpm.it

DOI: 10.1111/j.1532-5415.2008.01991.x

From the �Department of Internal Medicine, University ‘‘Politecnica delleMarche,’’ Ancona, Italy; wGeriatric Emergency Service and zStatistic andBiometry Center, Department of Gerontological Research, Italian NationalInstitute for Research and Care on Aging, Ancona, Italy; and §EmergencyDepartment, Azienda Ospedali Riuniti, Ancona, Italy.

JAGS 56:2131–2138, 2008r 2008, Copyright the AuthorsJournal compilation r 2008, The American Geriatrics Society 0002-8614/08/$15.00

METHODS

Study Design and Population

This is a secondary analysis of a prospective observationalcohort of 200 acutely ill ED patients aged 65 and older whowere enrolled from two urban EDs in an Italian city with100,000 inhabitants (Ancona) to test the validity of theIdentification of Seniors At Risk (ISAR) screening tool.11

The local ethics board approved the study. The CED waspart of a 633-bed tertiary-care academic hospital (AziendaOspedali Riuniti); the GED was a hybridized ED with a six-bed observation unit (OU) designed for elderly nontraumapatients and staffed by geriatricians within the 214-bed ac-ademic-affiliated INRCA hospital.1 Patients aged 65 andolder were enrolled in June 2006 from the GED (n 5 100)and July 2006 from the CED (n 5 100), taking care thatnone presenting to the ED in the course of the study periodwas recruited again. Those who had cognitive impairment(defined as a score of �5 on the Short Portable MentalStatus Questionnaire (SPMSQ)) and no proxy, those too illto respond, and trauma patients were excluded. The refusaland exclusion rate was less than 10%. No data are availablefor excluded and nonscreened patients or for those whorefused to participate.

Data Collection and Measures

A trained research assistant (VM) screened patients pre-senting to the ED Monday to Friday from 9:00 AM to 6:00PM using a standard information sheet explaining the studyprotocol to patients and proxies. Interviews were conductedwith patients; an accompanying family member or otherinformant was interviewed in the case of patients with cog-nitive impairment. Written consent including access tomedical records was obtained.

Identifiers were collected using a standardized datacollection sheet that included age, sex, marital status, livingstatus (a relative with moderate to severe dementia or im-paired mobility living with a patient was defined as dis-abled), arrival (emergency medical service, privateambulance, own transport; daytime vs nighttime), triage(four-level), final diagnosis, and patient disposition (dis-charge, admission, or admission to an intensive care unit(ICU)). While in the ED, patients also underwent a briefgeriatric assessment by the research assistant using theCharlson Comorbidity Index,12,13 the SPMSQ for cognitivefunction,14 and the Katz activities of daily living (ADL)scale15 for functional status before the current event.Length of ED stay (LOS) and, in case of admission, lengthof in-hospital stay (LOiHS) were also recorded.

Thirty-day and 6-month follow-up telephone inter-views were conducted with the person (patient or proxy)originally interviewed at the ED to collect data on mortal-ity, number of ED visits and hospital admissions, and cur-rent Katz ADL dependence (telephone ADL scales havebeen validated against face-to-face interviews16). No pa-tient was lost to follow-up. Date of death was confirmedaccording to governmental administrative databases.

Outcomes

Outcomes were early (within 30 days) and late (within 6months) unscheduled ED revisit, frequent ED return (�3

unscheduled ED visits over 6 months), hospital admission(within 6 months), 6-month functional decline (defined asloss of �1 ADLs), and 6-month mortality. Finally, death,functional decline, or any ED revisit or hospital admissionwere considered as a composite outcome.

Data Analysis

Descriptive statistics were used for age, sex, marital status,living status, arrival by ambulance, daytime or nighttimearrival, triage (ordinal ranking from 1 (emergent) to 4(nonurgent)), and patient disposition. The Student t test andthe chi-square test were used to compare the two EDs. Theunadjusted positive odds ratio (OR) was calculated for eachoutcome for GED patients. Positive OR adjusted for pos-sible confounders (age; sex; living status (alone or with dis-abled relative vs others); admission at the time ofrecruitment; and Charlson Comorbidity Index, SPMSQ,and ADL scores) were calculated using multiple logistic re-gression analysis. Because admission at the time of recruit-ment reduced the 30-day period in which a patient risked anearly ED revisit, the actual postdischarge interval (calcu-lated as 30 days minus LOiHS) was added as a covariate forthis outcome. Summary measures were reported as OR andcorresponding 95% confidence interval (CI). Statistical an-alyses were performed using SPSS statistical software ver-sion 14 (SPSS Inc., Chicago, IL).

RESULTS

The characteristics of the 200 patients (mean age80.3 � 7.4) are reported in Table 1. Overall, the studypopulation was old (28.5% were aged �85) and frail, assuggested by high levels of comorbidity, disability, and cog-nitive impairment; nine (4.5%) were currently in long-termcare and 116 (58%) were admitted at the time of recruit-ment.

Patients characteristics according to ED setting (CEDvs GED) are also listed in Table 1. GED patients were olderand medically and socially frailer, as suggested by the highernumber of widow(er)s, paid caregivers, and ambulance ar-rivals, but no significant differences emerged in terms oftriage, comorbidity, admission at the time of enrollment,ICU admission (7 CED vs 9 GED patients), and LOiHS. Incontrast, LOS was significantly shorter for CED than forGED patients, although the latter measure included timespent in the OU. LOS distribution was unimodal (median5.3 hours, range 0.8–24.0) in the CED and bimodal (me-dian 14 hours, range 0.4–24.0) in the GED (Figure 1).

Diagnoses are reported according to ED setting in Ta-ble 2. Acute coronary syndromes and heart failure weremore frequent in the CED patients, whereas transient is-chemic attack or stroke and exacerbation of chronic ob-structive pulmonary disease (COPD) were more frequentin the GED. Diagnosis of pneumonia and arrhythmiaand overall diagnosis distribution did not differ betweensettings.

Early ED Return (30 Days)

At 30 days, 13 patients (6.5%), five of them GED patients,had died, and six (3%) had been in the hospital since thetime of recruitment; of the remaining 181 patients, 48(26.5%) had required one or more ED revisits, and 24

2132 SALVI ET AL. NOVEMBER 2008–VOL. 56, NO. 11 JAGS

(13.3%) had been admitted to the hospital. Early ED returnwas not different between the EDs (Table 3). Female sexand greater comorbidity were significant risk factors(OR 5 2.28, 95% CI 5 1.04–5.00, P 5.04; OR 5 1.21,

95% CI 5 1.01–1.46, P 5.04, respectively). Moreover, theactual postdischarge interval (30 daysFLOiHS) increasedthe risk of early ED return (OR 5 1.07, 95% CI 5 1.00–1.15, P 5.04).

Table 1. Baseline Characteristics of the Studied Population According to Emergency Department (ED) Setting

Characteristic Total (N 5 200)

Conventional ED

(n 5 100)�Geriatric ED

(n 5 100)� P-Value

Age, mean � SD 80.3 � 7.4 78.1 � 7 82.5 � 7.2 o.001w

Age, n (%)

65–74 47 (23.5) 34 13 o.001z

75–84 96 (48) 49 47

85 57 (28.5) 17 40

Sex, n (%)

Female 115 (57.5) 47 68 .004z

Male 85 (42.5) 53 32

Marital status, n (%)

Single 10 (5) 5 5 o.001z

Married 110 (55) 70 40

Widowed 80 (40) 25 55

Living situation, n (%)

Alone 26 (13) 12 14 .03z

Nondisabled relative 131 (65.5) 74 57

Disabled relative 8 (4) 5 3

Paid caregiver 26 (13) 8 18

Institution 9 (4.5) 1 8

Arrival, n (%)

Emergency medical service 36 (18) 12 24 .03z

Private ambulance 84 (42) 40 44

Own transport 80 (40) 48 32

Arrival time

Daytime (8 AM–8 PM) 169 (84.5) 87 82 .44z

Nighttime (8 PM–8 AM) 31 (15.5) 13 18

Triage code, n (%)

1 (emergent) 1 (0.5) 0 1 .18z

2 (urgent) 51 (25.5) 29 22

3 (semiurgent) 136 (68) 68 68

4 (nonurgent) 12 (6) 3 9

Charlson Comorbidity Index, mean � SD 3.4 � 2 3.3 � 2.3 3.4 � 1.7 .73w

Charlson Comorbidity Index, n (%)

0–2 74 (37) 40 34

�3 126 (63) 60 66 .46z

Short Portable Mental Status Questionnaire, mean � SD 3.9 � 4 2.5 � 3.3 5.2 � 4.2 o.001w

Short Portable Mental Status Questionnaire, n (%)

0–4 137 (68.5) 83 54 o.001z

�5 63 (31.5) 17 46

Activities of daily living performed, mean � SD 3.8 � 2.3 4.3 � 2 3.2 � 2.5 .001w

Activities of daily living performed, n (%)

4–6 123 (61.5) 71 52

�3 77 (38.5) 29 48 .009z

Disposition, n (%)

Discharged 84 (42) 47 37 .20z

Admitted 116 (58) 53 63

Length of in-hospital stay, days, mean � SD 10.3 � 6.9 10 � 6.6 10.5 � 7.2 .74w

Length of stay in the ED, hours, mean � SD 9.5 � 7.6 6.2 � 4.5 12.8 � 8.5 o.001w

�Because n 5 100, percentages are not reported in these columns because they are the same as n.wResults of t-test.zResults of chi-square analysis.

SD 5 standard deviation.

GERIATRIC ED FOR ACUTELY ILL ELDERLY PATIENTS 2133JAGS NOVEMBER 2008–VOL. 56, NO. 11

Late and Frequent ED Return

Late (6-month) ED return was recorded for 93 (46.5%)patients, 24 (12%) had three or more ED revisits, meetingthe definition of frequent users. Comorbidity was a mod-erate risk factor for 6-month and frequent ED return(OR 5 1.23, 95% CI 5 1.02–1.48, P 5.03; OR 5 1.36,95% CI 5 1.02–1.81, P 5.04, respectively), whereas fe-male sex was a strong risk factor only for 6-month EDreturn (OR 5 2.8, 95% CI 5 1.35–5.82, P 5.006). In con-trast, dementia and disability reduced the probability offrequent ED visits (OR 5 0.79, 95% CI 5 0.66–0.96,P 5.02; OR 5 0.7, 95% CI 5 0.51–0.94, P 5.02, respec-tively). Late and frequent ED return rates were not signifi-cantly different between the two EDs (Table 3).

Hospitalization and Functional Decline (6 Months)

At 6 months, 65 (32.5%) patients had been admitted atleast once, 24 had been admitted more than once, and sevenhad been admitted four times. None of the variables weresignificantly associated with outcome in the logistic regres-sion model, including ED setting.

Functional decline at 6 months was observed in 40 ofthe 141 (28.4%) surviving subjects; 20 surviving patientswith an ADL score of 0 at recruitment could not undergofurther functional decline and were excluded. Age, cogni-tive impairment, and admission at the time of recruitmentwere risk factors for 6-month functional decline (OR 5 1.1,95% CI 5 1.02–1.17, P 5.008; OR 5 1.3, 95% CI 5

1.12–1.52, P 5.001; OR 5 2.45, 95% CI 5 1.04–5.78,P 5.04, respectively). There were no differences in any ofthese measures between the EDs (Table 3).

Mortality and Composite Outcome

At 6 months, 39 patients (19.5%), 19 of them GEDpatients, had died. Despite the greater frailty of the GEDpatients, mortality rates were not significantly different be-tween the EDs, although the Cox regression model adjust-ing for age; sex; living status; admission at the time ofrecruitment; and Charlson Comorbidity Index, SPMSQ,and ADL scores showed a lower, although barely significantrisk for GED patients (hazard ratio 5 0.47, 95% CI 5 0.22–0.99, P 5.047; Figure 2).

Finally, the composite outcome was recorded for 146 of190 (76.8%) patients. (Ten surviving patients with neitheradmission nor ED return were excluded because they hadan ADL score of 0 at the ED index visit.) Again, there wasno difference between the EDs before or after adjustmentfor the confounders (Table 3).

DISCUSSION

The management of acutely ill, medically complex, and so-cially frail elderly patients presenting to the ED is a clinicalchallenge in terms of diagnosis, decision-making, and ther-apeutic planning1 and involves environmental factors.9

Clinical staff education in geriatric emergency medicine,evidence-based protocols for common geriatric syndromes,interdisciplinary approaches, and ideally, structural modi-fications of ED facilities, as well as provision of dedicatedspaces, have been proposed.8,9 To the best of the authors’knowledge, the geriatric emergency service operating atINRCA, Ancona, is the first chronological attempt any-where to address these problems, although no data about itseffectiveness were available. This study was devised tocompare the user characteristics and 30-day and 6-month

Figure 1. Distribution of length of emergency department (ED) stays in the two ED settings under study.

2134 SALVI ET AL. NOVEMBER 2008–VOL. 56, NO. 11 JAGS

patient outcomes from the INRCA GED and from a CED inthe same town.

Data analysis showed the GED patients to be older andmore disabled and cognitively impaired (hence frailer) butto have similar comorbidity levels, diagnoses, and ICU ad-mission rates. The expected differences in terms of more-complex management, higher admission rates, and worseoutcomes than with the less-comorbid, less-disabled, andless-demented CED patients were not observed. There wereno significant differences between the two EDs in admissionrates at the index ED visit; in early, late, and frequent ED

return; or in 6-month hospital admissions and functionaldecline. The mortality rate was lower in GED patients, al-though the difference was barely significant. These datasuggested a noninferiority and possibly a slight superiorityof the GED to the CED, in line with the hypothesis that anED system specially designed for older patients may providebetter care.8,9

The Ancona INRCA includes a research center and anacademic-affiliated hospital with two geriatric wards (oneadjacent to the GED) and cardiology (with a six-bed ICU),neurology (with a stroke unit), nephrology (with a hemo-dialysis service), and general and specialist surgery units.Postacute care and rehabilitation units and 24-hour radi-ology and laboratory services are also available. An anes-thesia and resuscitation service, not equipped with beds,supports the cardiac ICU, the surgery units, and the GED.The INRCA GED was opened 12 years ago and supportsthe hospital’s activity, with more than 5,000 patients seeneach year. Its aims are to visit and treat elderly patients withacute nontrauma illness, relapse, and complications of theirmultiple chronic conditions while minimizing inappropri-ate admissions and optimizing disposition to specialty orpostacute care wards or referral to home-based care. Ge-riatric staff with emergency care experience manage it, andits organization and facilities are specially designed for se-nior patients. At triage, particular attention is given to pa-tients with suspected delirium or atypical presentation.Subjects with mobility problems can await evaluation inreclining chairs17 or lying on beds. Modified diagnostic andtherapeutic algorithms are adopted for geriatric patientswith abdominal or chest pain, urinary retention, syncope orfall, symptoms of cerebrovascular attack, and dehydration.(For example, a rehydration protocol by nasogastric tuberather than intravenous fluids is being evaluated.) Familymembers are allowed to stay on to provide assistance topatients with functional impairment or to supply additionalbackground and history for those with cognitive impair-ment. The medical records of former patients, reportingdata such as existing conditions, last known medications,allergies, and dates of previous ED visits, are stored in anelectronic database. Radiology and laboratory requestsfrom the GED have priority, and a dedicated laboratoryanalyzer has also become available. At discharge, patientsundergo evaluation through an abbreviated comprehensivegeriatric assessment for functional and cognitive impair-ment. They are also screened using ISAR for risk of func-tional decline, hospitalization, and ED return.11 Whenindicated, an interdisciplinary team including a socialworker coordinates the transition to home care (by tele-phone contact with the primary care physician or commu-nity services) or a nursing home, with the possibility ofbeing directly or shortly admitted to INRCA postacute careor rehabilitation units.

Appropriate structural modifications for ED facilitiesto meet the needs of the growing number of older adultsvisited have recently been proposed;9 these include skylights to provide natural lighting, hospital recess lamps to bedimmed in the evening to follow diurnal patterns, large-faced clocks and calendars to potentially reduce the risk oftime disorientation, sound-proof curtains between openbays to preserve privacy and provide a more acoustic-friendly environment; stretchers with pressure-reducing

Table 2. Diagnoses of the Studied Population Accordingto Emergency Department (ED) Setting

Diagnosis

Total

Conven-

tional ED

(n 5 100)

Geriatric

ED

(n 5 100)

P-

Valuen (%) n

Cardiovascular 63 (31.5) 38 25

Heart failure 19 (9.5) 12 7

Arrhythmias 16 (8) 9 7

Acute coronary syndromes 9 (4.5) 7 2

Syncope/presyncope 7 (3.5) 5 2

Other cardiovascular (e.g.,peripheral artery disease,deep vein thrombosis)

12 (6) 5 7

Neurological 31 (15.5) 10 21

Transient ischemic attack,stroke, or cerebrovascularaccident symptoms

12 (6) 3 9

Altered consciousness orconfusion

6 (3) 1 5

Vertigo 4 (2) 2 2

Other neurological (e.g.,epilepsy)

9 (5) 4 5

Gastrointestinal 31 (15.5) 14 17

Intestinal (sub)occlusion 9 (4.5) 2 7

Lower gastrointestinalproblem

8 (4) 5 3

Upper gastrointestinalproblem

6 (3) 1 5 .38�

Liver problem 4 (2) 3 1

Abdominal pain 4 (2) 3 1

Respiratory 22 (11) 10 12

Pneumonia 11 (5.5) 6 5

Chronic obstructivepulmonary diseaseexacerbation

9 (4.5) 2 7

Other respiratory 2 (1) 2 0

Genitourinary 12 (6) 6 6

Hematology 10 (5) 6 4

Other diagnosis 12 (6) 6 6

Undefined diagnosis 19 (9.5) 10 9

Asthenia or anorexia 5 (2.5) 1 4

Multiple diagnoses 4 (2) 3 1

Fever 3 (1.5) 0 3

Other undefined 7 (3.5) 6 1

�Result of chi-square analysis for overall diagnosis distribution between the

two EDs.

GERIATRIC ED FOR ACUTELY ILL ELDERLY PATIENTS 2135JAGS NOVEMBER 2008–VOL. 56, NO. 11

mattresses or reclining chairs to lie on while awaiting EDevaluation, wall and hallway support rails, and aisle light-ing and nonskid tiles to aid in mobility and decrease the riskof falls. The INRCA GED possesses some of the abovecharacteristics. It is endowed with an emergency room, aclinic for semi- and nonurgent visits, and six observationbeds (2 per room). Each room has windows and nonskid tilefloors; although beds are separated by curtains, low crowd-ing ensures a sufficiently quiet environment. Further struc-

tural changes are being planned to improve quality ofservice and number of patients visited.

The greatest advantage of the INRCA GED is probablyits hybrid structure between an ED and an OU.8 Here, pa-tients with unclear diagnoses, atypical presentations, ormobility problems can safely and comfortably await theclinical evolution and be observed for up to 24 hours untiladmission or discharge.

A retrospective observational cohort study of 8,385 el-derly patients admitted to an OU, mainly for chest pain,dehydration, syncope, and COPD exacerbation, docu-mented the effectiveness of a protocol-driven OU for thesepatients.18 It reported a slightly different LOS (90 minuteslonger) and higher rates of hospital admission (26% vs18.5%) and ED return (20% vs 10%) than for youngersubjects, but ED returns occurred mainly for problemsother than those treated during the first visit, and 74% ofpatients were discharged after their stay in the OU, avoidingin-hospital associated risks. Rates of mortality or of adverseoutcomes were not greater. Average LOS (15.9 hours) wassimilar to that at the INRCA GED (12.8 hours), whereasadmission and 30-day ED return rates were significantlylower (26.1% vs 63% and 9.4% vs 26%, respectively).Nonetheless, it should be considered that observation in theINRCA GED is often related to mobility problems orgreater need for nursing rather than to presenting com-plaints that enable protocol-driven management. Overall,functioning of the INRCA GED is similar to that of a stan-dard OU; patients, lying in observation beds, are evaluated

Table 3. Results of Tables of Contingency and Multiple Logistic Regression Models According to Outcome and Emer-gency Department (ED) Type

Outcomes

Conventional ED Geriatric ED

n

Unadjusted OR

(95% CI) P-Value

Adjusted OR

(95% CI) P-Valuen (%)

Early ED revisit: 30 days� 181 0.96 (0.5–1.9) 1.00 1.06 (0.5–2.3) .88

Yes 25 (27) 23 (26)

No 68 (73) 65 (74)

ED revisit: 6 months 168z 0.72 (0.4–1.3) .35 0.66 (0.3–1.4) .25

Yes 51 (59.3) 42 (51.2)

No 35 (40.7) 40 (48.8)

Frequent ED return 161 1.2 (0.5–2.9) .83 1.11 (0.4–3.1) .84

Yes 11 (13.8) 13 (16)

No 69 (86.2) 68 (84)

Admission 167z 0.75 (0.4–1.4) .43 0.63 (0.3–1.3) .20

Yes 36 (42.4) 29 (35.4)

No 49 (57.6) 53 (64.6)

Functional decline 141 1.3 (0.6–2.7) .57 0.65 (0.3–1.6) .34

Yes 20 (26) 20 (31.2)

No 57 (74) 44 (68.8)

Composite outcomew 190 1.44 (0.7–2.9) .39 0.95 (0.4–2.1) .89

Yes 73 (73.7) 73 (80.2)

No 26 (26.3) 18 (19.8)

Note: Confounding variables were age; sex; living status; admission at the time of recruitment; and Charlson Comorbidity Index, Short Portable Mental Status

Questionnaire, and activity of daily living scores.�Actual postdischarge interval (defined as 30 daysFlength of in-hospital stay) was considered instead of admission at recruitment for early ED return.wThe composite outcome included the occurrence of any one of the following: death, functional decline, any ED revisit, or hospital admission.z Seven and six patients who were dead at 6 months had had ED revisit and admission, respectively, within 30 days.

OR 5 odds ratio; CI 5 confidence interval.

Figure 2. Cox-regression survival analysis according to theemergency department (ED) setting. Solid line indicates patientsvisited in the geriatric emergency department (GED); dashed lineindicates patients visited in the conventional emergency depart-ment (CED).

2136 SALVI ET AL. NOVEMBER 2008–VOL. 56, NO. 11 JAGS

three times for disposition (where possible within 18 hours)as each physician begins his or her shift and then proceedsto aid the other doctor working in the ED. At the time ofthis study, the CED lacked an OU; this could explain theshorter LOS and its different distribution between the EDs,as well as some differences in outcome and prognosis.

Few similar facilities exist in the world.1,8 The NassauUniversity Medical Center (Long Island, NY) and the Had-assah University Medical Center (Jerusalem, Israel) haveequipped a geriatric emergency room staffed by geriatricians,but the medical literature provides no details about internalstructure, functioning, or outcomes. The Montreal JewishGeneral Hospital (Canada) has tested an ED geriatric con-sultation team staffed by a geriatrician, a full-time nurse cli-nician, and part-time physical and occupationaltherapists.19,20 A hospital in Turin (Italy) has reserved twoto four Emergency Medicine beds managed by geriatricnurses and geriatricians for older patients.21 Unfortunately,all reports on these cases are descriptive, and none include acontrol group.22

Limitations of the Study

The fact that the INRCA hospital, unlike Azienda OspedaliRiuniti, lacked a resuscitation ward may have introduced afirst pre-ED selection bias. Another limitation may be thatthis was a convenience sample of 200 elderly ED patientsthat excluded patients too ill to collaborate and those withcognitive impairment and no available informant. None-theless, the rate of urgent visits in the sample was higherthan that of the general Italian ED population (25.5% vs8.4%) and was similar in the two EDs, suggesting that thesample was representative of elderly ED patients. Elderlypatients are known to use EDs appropriately,1,2,23 becauseemergent and urgent visits are more frequent than semi-urgent and nonurgent ones, at least in the United States andCanada.24,25 Furthermore, in the current study, diagnosesand number of ICU admissions were similar for the twoEDs. These findings, together with data adjustment for co-morbidity, strengthens the significance of the slight differ-ence in mortality, although the effect of other variablescannot be excluded.

Third, this sample does not include patients presentingfor trauma. Falls and trauma are frequent presenting com-plaints in elderly people,26 and their exclusion may haveintroduced another selection bias, although because theINRCA hospital lacks an orthopedic unit, whereas the Azi-enda Ospedali Riuniti has two, this may have ensuredgreater patient homogeneity.

Finally, although comparing different care systems re-quires a study design free of any selection biases and con-founders, this was a secondary analysis of a study designedto validate the ISAR screening tool at two Italian ED.27

Further studies with appropriate design (trials enrollingpatients matched for age, sex, severity and presenting com-plaint) are therefore warranted.

CONCLUSION

The complexity and specificity of the problems of elderly pa-tients call for a new model of emergency care, which is achallenge for EDs worldwide. Internal structure and careprocess modifications specially aimed to older patients have

been proposed to improve the quality of emergency care.1,8–10

This study described and compared the patient characteristicsand 6-month outcomes of a GED and a CED for the first time.The fact that management of the frailer and older GED pa-tients did not result in significantly different health outcomessuggests the noninferiority and, indirectly, a slight superiorityof this emergency care model, as confirmed by the slightly butsignificantly lower mortality rate among the GED patients.The GED approach, modeled on pediatric EDs and traumacenters,28,29 is an attempt to address the special needs of olderpatients. Further data are needed to devise an appropriate andeffective emergency care system for elderly people.

ACKNOWLEDGMENTS

Conflict of Interest: The editor in chief has reviewed theconflict of interest checklist provided by the authors and hasdetermined that the authors have no financial or any otherkind of personal conflicts with this manuscript.

F.S. is the recipient of a research grant for this studyfrom University ‘‘Politecnica delle Marche.’’

Author contributions: F.S., V.M., and G.D.T. conceivedthe idea and designed the study. V.M., A.G., and R.G. con-ducted the study and performed data collection under thesupervision of F.S., G.D.T., and S.P. F.S. and V.M. managedthe data. F.S. and L.S. performed the statistical analyses.F.S., V.M., A.G., and L.S. interpreted the data. F.S., V.M.,and L.S. drafted the manuscript. A.G., A.R., and P.D.F.contributed to its revision. F.S. takes responsibility for themanuscript as a whole.

Sponsor’s Role: This study was supported by a researchgrant from the University of Ancona ‘‘Politecnica delleMarche’’ (F.S.), which did not have any role in the design,execution, analysis and interpretation of data, or writing ofthe manuscript.

REFERENCES

1. Salvi F, Morichi V, Grilli A et al. The elderly in the emergency department: A

critical review of problems and solutions. Intern Emerg Med 2007;2:292–301.

2. Aminzadeh F, Dalziel WB. Older adults in the emergency department: A sys-

tematic review of pattern of use, adverse outcomes, and effectiveness of in-

terventions. Ann Emerg Med 2002;39:238–247.

3. George G, Jell C, Todd BS. Effect of population ageing on emergency depart-

ment speed and efficiency: A historical perspective from a district general

hospital in the UK. Emerg Med J 2006;23:379–383.

4. Friedmann PD, Jin L, Karrison TG et al. Early revisit, hospitalization, or death

among older persons discharged from the ED. Am J Emerg Med 2001;19:

125–129.

5. McCusker J, Cardin S, Bellavance F et al. Return to the emergency department

among elders: Patterns and predictors. Acad Emerg Med 2000;7:249–259.

6. Caplan GA, Brown A, Croker WD et al. Risk of admission within 4 weeks of

discharge of elderly patients from the emergency departmentFthe DEED

Study. Age Ageing 1998;27:697–702.

7. McCusker J, Bellavance F, Cardin S et al. Prediction of hospital utilization

among elderly patients during the 6 months after an emergency department

visit. Ann Emerg Med 2000;36:438–445.

8. Adams JG, Gerson LW. A new model for emergency care of geriatric patients.

Acad Emerg Med 2003;10:271–274.

9. Hwang U, Morrison RS. The geriatric emergency department. J Am Geriatr

Soc 2007;55:1873–1876.

10. Schumacher JG. Emergency medicine and older adults: Continuing challenges

and opportunities. Am J Emerg Med 2005;23:556–560.

11. McCusker J, Bellavance F, Cardin S et al. Detection of older people at increased

risk of adverse health outcomes after an emergency visit: The ISAR screening

tool. J Am Geriatr Soc 1999;47:1229–1237.

12. Charlson ME, Pompei P, Ales KL et al. A new method of classifying prognostic

comorbidity in longitudinal studies: Development and validation. J Chronic

Dis 1987;40:373–383.

GERIATRIC ED FOR ACUTELY ILL ELDERLY PATIENTS 2137JAGS NOVEMBER 2008–VOL. 56, NO. 11

13. Murray SB, Bates DW, Ngo L et al. Charlson Index is associated with one-year

mortality in emergency department patients with suspected infection. Acad

Emerg Med 2006;13:530–536.

14. Pfeiffer E. Short Portable Mental Status Questionnaire for the assessment

of organic brain deficit in elderly patients. J Am Geriatr Soc 1975;23:433–441.

15. Katz S, Ford AB, Moskowitz RW et al. Studies of illness in the aged. The index

of ADL: A standardized measure of biological and psychological function.

JAMA 1963;185:914–921.

16. Shinar D, Gross CR, Bronstein KS et al. Reliability of the activities of daily

living scale and its use in telephone interview. Arch Phys Med Rehabil

1987;68:723–728.

17. Wilber ST, Bruger M, Gerson LW et al. Reclining chairs reduce pain from

gurneys in older emergency department patients: A randomized controlled

trial. Acad Emerg Med 2005;12:119–123.

18. Ross MA, Compton S, Richardson D et al. The use and effectiveness of an

Emergency Department Observation Unit for elderly patients. Ann Emerg

Med 2003;41:668–677.

19. Gold S, Bergman H. A geriatric consultation team in the emergency depart-

ment. J Am Geriatr Soc 1997;45:764–767.

20. Sinoff G, Clarfield AM, Bergman H et al. A two-year follow-up of geri-

atric consults in the emergency department. J Am Geriatr Soc 1998;46:

716–720.

21. Cotroneo AM, Stralla U. La Geriatria e l’Urgenza. Una realta del III Millennio.

Scientifiche Melograno Servizi, Torino, Italy: 2003.

22. Hastings SN, Heflin MT. A systematic review of interventions to improve

outcomes for elders discharged from the emergency department. Acad Emerg

Med 2005;12:978–986.

23. Strange GR, Chen EH, Sanders AB. Use of emergency departments by elderly

patients: Projections from a multicenter data base. Ann Emerg Med 1992;

21:819–824.

24. McCaig LF, Nawar EW. National hospital ambulatory medical care survey:

2004 emergency department summary. Adv Data 2006;372:1–32.

25. Li G, Lau JT, McCarthy ML et al. Emergency department utilization

in the United States and Ontario, Canada. Acad Emerg Med 2007;14:

582–584.

26. Aschkenasy MT, Rothenhaus TC. Trauma and falls in the elderly. Emerg Med

Clin N Am 2006;24:413–432.

27. Salvi F, Morichi V, Grilli A et al. Predictive validity of the Identification of

Seniors At Risk (ISAR) screening tool in elderly patients presenting to two

Italian emergency departments. Aging Clin Exp Res, in press.

28. Li M, Baker M, Ropp L. Pediatric emergency medicine: A developing subspe-

cialty. Pediatrics 1989;84:336–342.

29. Committee on Trauma. Resources for Optimal Care of the Injured Patient:

1999. Chicago: American College of Surgeons, 1999.

2138 SALVI ET AL. NOVEMBER 2008–VOL. 56, NO. 11 JAGS