GERIATRICS/ORIGINAL RESEARCH

A Systematic Review and Qualitative Analysis to Inform theDevelopment of a New Emergency Department-Based

Geriatric Case Management ModelSamir K. Sinha, MD, DPhil, FRCPC, Edward S. Bessman, MD, Neal Flomenbaum, MD, Bruce Leff, MD

From the Division of Geriatric Medicine, University of Toronto, Toronto, Ontario, Canada (Sinha); the Division of Geriatric Medicine andGerontology, Center on Aging and Health (Sinha, Leff) and the Department of Emergency Medicine (Bessman), Johns Hopkins University School ofMedicine, Baltimore, MD; the Department of Emergency Medicine, Weill Cornell Medical College, Cornell University, New York, NY (Flomenbaum);

and the Department of Health Policy and Management, Johns Hopkins University Bloomberg School of Public Health, Baltimore, MD (Leff).

Study objective: We inform the future development of a new geriatric emergency management practice model.We perform a systematic review of the existing evidence for emergency department (ED)–based casemanagement models designed to improve the health, social, and health service utilization outcomes fornoninstitutionalized older patients within the context of an index ED visit.

Methods: This was a systematic review of English-language articles indexed in MEDLINE and CINAHL (1966 to2010), describing ED-based case management models for older adults. Bibliographies of the retrieved articleswere reviewed to identify additional references. A systematic qualitative case study analytic approach was usedto identify the core operational components and outcome measures of the described clinical interventions. Theauthors of the included studies were also invited to verify our interpretations of their work. The determinedpatterns of component adherence were then used to postulate the relative importance and effect of thepresence or absence of a particular component in influencing the overall effectiveness of their respectiveinterventions.

Results: Eighteen of 352 studies (reported in 20 articles) met study criteria. Qualitative analyses identified 28outcome measures and 8 distinct model characteristic components that included having an evidence-basedpractice model, nursing clinical involvement or leadership, high-risk screening processes, focused geriatricassessments, the initiation of care and disposition planning in the ED, interprofessional and capacity-buildingwork practices, post-ED discharge follow-up with patients, and evaluation and monitoring processes. Of the 15positive study results, 6 had all 8 characteristic components and 9 were found to be lacking at least 1component. Two studies with positive results lacked 2 characteristic components and none lacked more than 2components. Of the 3 studies with negative results demonstrating no positive effects based on any outcometested, one lacked 2, one lacked 3, and one lacked 4 of the 8 model components.

Conclusion: Successful models of ED-based case management models for older adults share certain keycharacteristics. This study builds on the emerging literature in this area and leverages the differences in thesemodels and their associated outcomes to support the development of an evidence-based normative and effectivegeriatric emergency management practice model designed to address the special care needs and thereby improvethe health and health service utilization outcomes of older patients. [Ann Emerg Med. 2011;57:672-682.]

Please see page 673 for the Editor’s Capsule Summary of this article.

Provide feedback on this article at the journal’s Web site, www.annemergmed.com.A podcast for this article is available at www.annemergmed.com.

0196-0644/$-see front matterCopyright © 2010 by the American College of Emergency Physicians.doi:10.1016/j.annemergmed.2011.01.021

ppaa

INTRODUCTION

Studies of emergency department (ED) utilization in the UnitedStates find that older adults, especially those older than 75 years,have among the highest visitation rates of any segment of the

population1 and that the burden of visitations will increase as the c

672 Annals of Emergency Medicine

opulation ages. Older adults are also more likely than others toresent to EDs with urgent or critical issues1 and are at least twices likely to be admitted to the hospital from the ED, although mostre still treated and discharged back into the community.2

The clinical heterogeneity of older patients contributes to the

linical challenges that emergency physicians and nurses face in

Volume , . : June

pdehttpEihdm

dnpbsniuimmiatesd

MS

PACwwSf“TpSw

cmopws

Sinha et al The Geriatric Emergency Management Nursing Model

providing their care. It has long been recognized that within thispopulation, common diseases more often present atypically;comorbidities can confound standard approaches, including theinterpretation of common diagnostic tests; polypharmacy isubiquitous; depleted physiologic reserves and impairedcognition must be anticipated; and traditional social supportsystems may be compromised.3 It is no wonder that suchpatients are more difficult and time-consuming to assess andmanage in the ED compared with younger patients.4,5

Furthermore, although older patients tend to receive the mostresource-intensive care of any age group within EDs, theirproblems are less likely to be accurately diagnosed, partlybecause the actualization of many of the earlier listed long-established principles related to working with older patients intocurrent routine practices has been slow. Consequently, they aremore likely to be discharged from EDs with unrecognized anduntreated problems.6-8

Other factors contribute to suboptimal care delivery.Providing optimal assessment and discharge planning tends tobe more complicated for older patients in ED settings, wheretime pressures and the need to maintain rapid patientthroughput are considered essential.9 A relative lack ofknowledge among ED professional staff in geriatric principles ofcare and practice may compound the strain of trying toadequately address the often complex and interrelated healthand social care needs of older patients.5 All these factors mayhelp account for the higher rates of adverse outcomes, includingED revisitation, hospitalization, functional decline, and death,that are experienced by older adults compared with others

Editor’s Capsule Summary

What is already known on this topicThe proportion of older patients visiting emergencydepartments (EDs) is increasing; these patients havegreater needs and unique postdischarge problems.

What question this study addressedThis study reviewed existing literature on managementof geriatric patients in emergency care and synthesizeda set of process, component, and outcome measurescomposing a geriatric emergency practice model.

What this study adds to our knowledgeA set of 8 characteristic components were derivedfrom the 18 studies that met inclusion criteria.Programs having more of these components tendedto produce better outcomes.

How this study is relevant to clinical practiceIt characterizes current efforts to improve the EDmanagement of geriatric patients.

within months of an index ED visit.2 r

Volume , . : June

The need for better ED care and management of olderatients has been recognized by several investigators, who haveeveloped and evaluated ED-based interventions designed tonhance care and reduce adverse outcomes. These interventionsave been previously reviewed and demonstrate mixed results athe patient or systems level.2,10-12 Hickman et al13 attributedhe disparate results of these studies as evidence of a lack ofrovider understanding of the key characteristics of effectiveD-based care models. Additionally, the variability in model

mplementation and evaluation criteria and methodology hasindered more effective comparisons, as well as the broaderevelopment, acceptance, and dissemination of a commonodel.The best models of care for older ED patients have not been

efined.14 Therefore, to inform the future development of aew geriatric emergency management practice model weerformed a systematic review of the existing evidence for ED-ased case management models designed to improve the health,ocial, and health service utilization outcomes foroninstitutionalized older patients within the context of an

ndex ED visit. To our knowledge, this review is unique in itsse of a systematic qualitative case study analytic approach to

dentify the core operational components and outcomeeasures of the described clinical interventions. Thisethodology further uses adherence analysis to facilitate the

dentification of specific model characteristics that may bessociated with clinical effectiveness. We propose and describehe recommended components and processes for the geriatricmergency management practice model, as well as a set oftandard research outcome measures for use in ongoing research,evelopment, and dissemination of this innovative care model.

ATERIALS AND METHODStudy Design

This systematic review, conducted in accordance withreferred Reporting Items for Systematic Reviews and Meta-nalyses15 criteria, examined articles indexed in MEDLINE andINAHL. A MEDLINE search (from 1966 to January 2010)as conducted for English-language articles containing the keyords “geriatric,” “older adults,” or “seniors,” or Medicalubject Heading (MeSH) terms “Geriatrics” or “Health Servicesor the Aged” AND key word “emergency,” or MeSH termsEmergencies,” “Emergency Service, Hospital,” or “Emergencyreatment” AND key word “assessment.” The same search waserformed in CINAHL with subject headings “Emergencyervice” AND “Aged.” Bibliographies of the retrieved articlesere reviewed to identify additional references.

Articles were included if they constituted descriptions oflinical interventions, with clearly defined assessment andanagement components, undertaken to improve outcomes for

lder adults within the context of an index ED visit. Therefore,rogram descriptions, observational studies, and clinical trialsere included. We excluded articles that did not report

ufficient quantitative information to judge outcomes (eg,

eferred to achieving a decreased hospitalization rate but did not

Annals of Emergency Medicine 673

ioitAd

uptver

mtma

R

3aaasrcwk

Fi

The Geriatric Emergency Management Nursing Model Sinha et al

quantify this), those whose only purpose was to describe or testrisk screening instruments, and those describing an ED-basedintervention limited to patients with a single diagnosis (eg, falls,delirium). Finally, we excluded articles that did not meetCochrane evaluation criteria for randomized controlled trialsand nonrandomized controlled trials16 or Meta-analysis ofObservations Studies in Epidemiology evaluation criteria forobservational studies.17

Data Collection and ProcessingEach study was reviewed by one of the authors (S.K.S.) in

detail and abstracted for the following data: setting, location andcountry of intervention, study design, characteristics ofintervention and delivery method, participant and providercharacteristics, patient inclusion and exclusion criteria, desiredoutcomes, outcome measures used, and results, including bothprimary and secondary outcomes.

Primary Data AnalysisOne author (S.K.S.) reviewed each article to focus specifically

on identifying the core operational components of theirdescribed clinical interventions. Such components include bothstructures and processes. This was accomplished by using asystematic qualitative case study analytic approach outlined byHuberman and Miles.18 This approach starts with listing allmodel components identified within each intervention. Furthereffort was then made to determine whether individual studieswere using the same terms to describe different components ordifferent terms to describe the same components. This processultimately enabled a set of categories, ie, “high-risk screening” or“post-ED discharge follow-up with patients,” to be conceptuallyspecified either inductively or deductively.18

After a comprehensive list of identified characteristiccomponents had been determined, each described clinicalintervention was then identified by one of the authors (S.K.S.)as either adhering or not to a particular characteristiccomponent. This was done by incorporating methodologicalrules for evaluating criteria adherence, similar to those Hedrickand Inui19 and others20,21 have used in adherence analyses.Adherence was largely determined by whether written evidenceof such adherence existed. For example, a study report thatalluded to the selection of higher-risk older patients but did notactually describe the screening procedure used for selectionwould have been listed as nonadherent with using a “high-riskscreening” selection process.

Because component adherence served as the basis of ourqualitative analysis, we surveyed the investigators of each of theinvolved studies to primarily assess their level of agreement withour interpretations of their study’s component adherence. Theseinvestigators were also asked to list any other distinct modelcharacteristics related to their interventions that we may nothave considered and to list any other studies that we may haveoverlooked. We successfully contacted 11 of the 18 authors. Allauthors agreed with the components that we identified and the

literature we surveyed. No additional suggestions were made a

674 Annals of Emergency Medicine

dentifying components or literature that we may haveverlooked. Two authors22,23 did provide us additionalnformation not published in their articles that demonstratedhat they did use high-risk screening tools in their interventions.ny uncertainty in adherence assignment was arbitrated byiscussion among the investigators (S.K.S. and B.L.).

The determined patterns of component adherence were thensed to postulate the relative importance and effect of theresence or absence of a particular component in influencinghe overall effectiveness of their respective interventions. Thearious outcome measures that were used to evaluate theffectiveness of the described interventions included in thiseview were further determined.

The heterogeneity in trial design, interventions, and outcomeeasures did not completely preclude our ability to use

raditional meta-analytic techniques around certain outcomeeasures. However, we believed that meta-analysis was not

ppropriate to the aims of this study.

ESULTSThe database search yielded a total of 352 citations, from which

4 articles were selected for further examination according to anbstract review. Of these 34, 6 did not meet the inclusion criteria,nd an additional 8 articles were excluded for not reportingdequate data on outcomes. Overall, 20 articles describing 18tudies were included in the final analysis (Figure 1). These articlesepresented 7 randomized controlled trials, 8 nonrandomizedlinical trials, and 3 observational studies or program descriptions,ith a mean sample size of 890 patients (range 12 to 3,977). Theey descriptive features and findings from these studies were

igure 1. Criteria used to define the final studies includedn this review.

bstracted and are summarized in the Table.

Volume , . : June

Table. ED geriatric assessment and management interventions.

Author, Year,Reference, andCountry

Design and SampleSize Study Population Exclusions Intervention Duration

Outcome Measures(Figure 3) and

Results Comments

Core Characteristic Component (Figure 2)Adherence

C1 C2 C3 C4 C5 C6 C7 C8

Basic and Conforti,2005,14

Australia*

RCT, N�224 High risk, age�65 y

Medicallyunstable,NH, unableto speakEnglish.

Nurse-led CGA based in ED,liaising with home caregiversand health care providers,organizing postdischargereferrals for out-of-hospitalassessment and supportservices for those goinghome from the ED anddocuments suggestions forassessment and referral forthose admitted to thehospital

One-time 1,†

3,†

19†

Poor overallcompliance byinpatient teamswith nurse’srecommendationswere observed

Yes Yes No Yes Yes No No Yes

Brazil et al, 1998,24

CanadaPD/OS, N�123 Age �65 y Not specified Nurse CM conducts assessment

in ED and coordinatesservices to transportmedically stable patientshome and support them withthe necessary professionaland support services for upto 5 days

5 days, withreassessmentwithin 72 hours ofinitial assessment

2,‡

14,‡

27,‡

28‡

Yes Yes No Yes Yes Yes Yes Yes

Brookoff andMinniti-Hill,1994,25 UnitedStates

PD/OS, N�670 Adult patients NH, hospitalized Nurse DC planning coordinatorsin the ED liaise with homecaregivers and health careproviders, coordinate therapid deployment of homecare services to facilitatedischarges home and avoidhospital admissions, andorganize postdischargereferrals for out-of-hospitalassessment

One-time 2,‡

9‡

A net financialgain wasachievedbecause netbillings to third-party payers forhome careservicesexceededinterventioncosts

Yes Yes No Yes Yes Yes No Yes

Brooks and Ertl,2000,26 UnitedStates

PP, N�12 High risk, age�65 y, with�6 ED visitsin thepreceding year

Not specified Social worker conducts homevisits in consultation with MDand organizes referrals forout-of-hospital assessmentand support services

12 months 3,‡

4,‡

6,‡

11,‡

12,‡

14‡

Decrease in EDand inpatientcosts but not inOP costsbecauseincreasednumber of OPvisits

Yes No Yes Yes No Yes Yes Yes

Caplan et al,2004,27

Australia

RCT, N�739 Age �75;dischargedhome

NH, hospitalized Nurse-led CGA performed athome within 24 hours ofindex ED visit. After adiscussion with PCP, thenurse formulates care plan,initiates urgent interventionsand referrals, and reviewspatient at weeklyinterdisciplinary meeting, atwhich further interventions orreferrals could be ordered forup to 28 days in total.

28 days, with ongoingfollow-up afterinitial homeassessment

4,†§

5,†§

6,†§

8,†

13,†

19,†

20†

Lower rate of EDrevisitationsobserved up to18 months

Yes Yes Yes Yes Yes Yes Yes Yes

Corbett et al,2005,28

Australia

PP, N�3,977 Adult patientspotentially atrisk

Not specified Multidisciplinary carecoordination team providescase management tofacilitate discharge andensure necessary communitysupports and linkages are inplace

Initial assessmentand follow-up asnecessary

1,†§

4,†

14,†§

23,†§

27†§

Yes Yes No Yes Yes Yes Yes Yes

Sinhaetal

The

Geriatric

Em

ergencyM

anagementN

ursingM

odel

Volum

e,

.

:June

Annals

ofE

mergency

Medicine

675

Table. Continued.

Author, Year,Reference, andCountry

Design and SampleSize Study Population Exclusions Intervention Duration

Outcome Measures(Figure 3) and

Results Comments

Core Characteristic Component (Figure 2)Adherence

C1 C2 C3 C4 C5 C6 C7 C8

Freeman, 1994,29

CanadaPD, N�874 Age �60 y Not specified Nurse case manager conducts

assessment and coordinatesservices to transport patientshome and support them withthe necessary professionaland support services for upto 5 days

5 days, withreassessmentwithin 72 hours ofinitial assessment

2, 7,‡

14,‡

27,‡

28‡

Yes Yes No Yes Yes Yes Yes Yes

Gagnon et al,199922 Canada*

RCT, N�427 Age �70 y,functionalimpairmentwith risk ofadmission

CI, NH,hospitalized

Nurse CM in consultation withMD for patients dischargedfrom the ED during theprevious 12 months;minimum of a monthlytelephone call and home visitevery 6 weeks for 10 months

10 months 3,†

4,†§

6,†

14,†

19,†

23†

A slight increasein the EDrevisitation ratewas observedin theinterventiongroup

Yes Yes Yes Yes No Yes No Yes

Guttman et al,2004,30 Canada

PP, N�1724 Age �75 y CI, NH,hospitalized

Nurse DC planning coordinatorsin the ED liaise with homecaregivers and health careproviders, coordinate therapid deployment of homecare services to facilitatedischarges home and avoidhospital admissions, andorganize postdischargereferrals for out-of-hospitalassessment

Telephone follow-up24 hours post-DCand RN availablefor telephone for 1week post–indexED visit

4,†§

6,†‡

15,†§

16,†§

17,†

23†§

The 17% reductionin thesubsequenthospitaladmission rate,because of lownumbers forcomparison, didnot showstatisticalsignificance

Yes Yes No Yes Yes Yes Yes Yes

Hegney et al,2006,31

Australia

PP, N�2139 High risk, age�70 y

NH, unable toconsent, anddialysis,chemotheraphy,palliative orpsychiatricpatients

Nurse DC planning coordinatorsin the ED risk screenpatients, assess eligibility forand organize referrals forcommunity support servicesand further assessment, andshare a copy of riskscreening assessment withpatients’ PCP

One-time 3,†‡ 4,

†§6

†§Privacy restrictions

made thestatisticalverification ofthe LOSreductions notpossible

Yes Yes Yes Yes Yes Yes No Yes

McCusker et al,2001,40 2003,33

2003,34 Canada

RCT, N�388(2001,40

200333),N�345(200334)

High risk, age�65 y

NH, hospitalized Nurses conduct geriatricassessment in ED, liaisingwith ED and geriatric staff asneeded, and organizepostdischarge referrals toaccess further communityassessment and supportservices. PCPs were routinelynotified about the visit andthe results of theassessment. Limited follow-up was provided after DC toensure that appointmentsand services were provided.

One-time with limitedtelephone follow-up

8,†

13,†§

14,†

19,†§

21,†

24,†

25,†

(2001)40,†

12†§

(2003)33,†

4,†§

9,†§

10,†§

11,†§

(2003)34†

Greater cost-effectivenessshown despiteincreased ratesof referral toPCPs and homecare services,and a slightincrease in EDrevisitation

Yes Yes Yes Yes Yes Yes Yes Yes

Miller et al, 1996,23

United StatesCohort, N�770 Age �65 y Acutely ill, �1-

hour stayNurse-led CGA based in ED on

patients identified with casefinding, withrecommendations madedirectly to patients, familymembers, and emergencyphysician. Helps to suggestpostdischarge referrals forout-of-hospital assessmentand support services forthose admitted or goinghome from ED and follow-upwith admitted patients withidentified active geriatricissues.

One-time withtelephone follow-upin 7 days if DC orsooner in person ifadmitted

4†‡

(P�.06), 8,†

10,†

14,†

18†‡

(P�.07), 19,†

23†

Poor complianceby patients andfamilies withtherecommendationsof the RN and aminimal uptakein new dentalor socialservices wasobserved.

Yes Yes Yes Yes Yes Yes Yes Yes

The

Geriatric

Em

ergencyM

anagementN

ursingM

odelSinha

etal

676A

nnalsof

Em

ergencyM

edicineV

olume

,

.

:

June

Table. Continued.

Author, Year,Reference, andCountry

Design and SampleSize Study Population Exclusions Intervention Duration

Outcome Measures(Figure 3) and

Results Comments

Core Characteristic Component (Figure 2)Adherence

C1 C2 C3 C4 C5 C6 C7 C8

Mion et al, 2003,35

United StatesRCT, N�650 High risk, age

�65 yNH, hospitalized Advanced practice nurse–led

CGA-based DC; planning in EDand help in organizingpostdischarge referrals forcommunity support services.Summary of assessment andDC plan is sent to PCP.Telephone follow-up isprovided as necessary untilcommunity agency personnelare in contact with patient.

One-time, limitedtelephone follow-up7–10 days later

4,†

6,†

8,†§

12,†

14,†§

22†

The interventionwas moreeffective forhigher-riskelders.

Yes Yes Yes Yes Yes Yes Yes Yes

Moss et al, 2002,32

AustraliaPP, N�2532 High risk, age

�65 yNot specified Nurse or allied health care

coordinators performcomprehensive DC riskassessments for all high-riskpatients, leading to theorganization of referrals forfurther assessment or supportservices. Consultation withhome caregivers and PCPsand patient and familyeducations is integral to thisprocess.

One-time, telephonefollow-up within 7days to those notreferred to acommunity provider

1,†§

4†

(P�.28),14,

†‡24,

†‡

27,†‡

28†‡

A downward butnonsignificanttrend in EDrevisitation wasobserved.

Yes Yes Yes Yes Yes Yes Yes Yes

Roberts et al,2007,36

Australia

PP, N�228 Age �65 y, highrisk for EDrevisitation andhospitaladmission

Not specified Multidisciplinary casemanagement team based inboth community and EDsettings performscomprehensive assessmentswith geriatrician and PCPsupport as required, coupledwith intensive short-term careplanning and communitysupport for up to and beyond12 weeks when necessary

Up to and beyond 12weeks whennecessary

4,†§

6,†§

7,†§

14,†‡

24,†‡

27†‡

Yes Yes Yes Yes Yes Yes Yes Yes

Runciman et al,1996,37 UnitedKingdom

RCT, N�414 Age �75 y NH, hospitalized Post-ED discharge home visit byhealth visitor (nurse) forstandardized CGA andarrangement of communityservices in consultation withPCP

One-time 4,†

6,†

14,†‡

19†§

Significantly lessIADLdependence at1 month but nodifference inADL dependenceobserved

Yes Yes Yes Yes No Yes Yes Yes

Warburton, 2005,38

CanadaQuasi experimental,

N�277High risk, age

�75 yNot specified Nurse-led screening and referral

program in the ED foridentified at-risk older adultsfor whom care plans aredeveloped to provide targetedand coordinated interventions

One-time 3,‡

4,‡

6‡

Cost-benefitanalysis showscost-effectiveness ofmodel

Yes Yes Yes Yes Yes Yes No Yes

Weir et al, 1998,39

Canada*RCT, N�77 Adults with

identified careneeds

Not specified ED-based home carecoordinators arrange promptdelivery of in-home servicesfor up to 10 days after anindex ED visit for dischargedadults with an indentifiedchange in care needs

Initially and ongoing forup to 10 days

12,†

14,†

22,†

24,†

26†

Sample size andpower too smallto detectsignificantdifferences

Yes No No No Yes No Yes Yes

RCT, Randomized controlled trial; NH, nursing home; CGA, comprehensive geriatric assessment; PD, program description; OS, observational study; CM, case manager; DC, discharge; PP, pre- and postinterventiondesign; MD, physician; OP, outpatient; PCP, primary care physician; RN, registered nurse; CI, cognitive impairment; NIH, National Institutes of Health; LOS, length of stay; ADL, activities of daily living; IADL, in-strumental activities of daily living.Sample sizes: Reported figures represent numbers used in the analysis of outcomes.*Indicates the study result was negative.†Results were statistically verified.‡Significant result.§Statistically significant result.

Sinhaetal

The

Geriatric

Em

ergencyM

anagementN

ursingM

odel

Volum

e,

.

:June

Annals

ofE

mergency

Medicine

677

tT(p((i(bi5Esi

tsoclssc

dttr(a2wiweb

fisadcadcdorib

The Geriatric Emergency Management Nursing Model Sinha et al

The qualitative analysis identified 8 distinct modelcharacteristic components. These components are listed with an

Figure 2. Core characteristic components common to ED-based geriatric case management models.

accompanying short descriptive explanation in Figure 2. i

678 Annals of Emergency Medicine

The determined adherence of each described intervention tohe 8 identified characteristic components is summarized in theable. Every intervention used evidence-based practice models

component 1) and established evaluation and monitoringrocesses (component 8) to report their outcomes. Seventeen94%) interventions incorporated focused geriatric assessmentscomponent 4). Sixteen (88%) interventions relied on nursingnvolvement or leadership to deliver the interventioncomponent 2) and developed interprofessional and capacity-uilding work practices (component 6), whereas 15 (83%)nitiated care and disposition planning in the ED (component) and 13 (72%) provided some form of more immediate post-D discharge follow-up with patients (component 7). High-risk

creening (component 3) was used in only 11 (61%) of thenterventions.

Fifteen studies achieved at least 1 positive effect, whereas 9 ofhe 13 that used statistical analyses also demonstrated atatistically significant positive effect in at least 1 measuredutcome. Of the 15 studies with positive results, 6 had all 8haracteristic components and 9 were found to be lacking ateast 1 component, of which the failure to use high-riskcreening processes (characteristic 3) was the reason in 5 of thesetudies. Two studies with positive results lacked 2 characteristicomponents and none lacked more than 2 components.

Three studies with negative results were reported thatemonstrated no positive effects according to any outcomeested.14,22,39 One lacked 2, one lacked 3, and one lacked 4 ofhe 8 model components. Two of the studies with negativeesults did not use high-risk screening processes14,39

characteristic 3), 2 did not actively develop interprofessionalnd capacity-building work practices14,39 (characteristic 6), anddid not provide more immediate post-ED discharge follow-upith their patients14,22 (characteristic 7). None of the

nterventions were found to cause measurable harm comparedith usual care, although 2 of the 13 studies that examined

ffects on ED revisitation rates actually demonstrated a smallut statistically significant increase in revisitations.22,34

Satisfaction levels were a commonly used outcomes measureor individual patients and their care providers. Thesenterventions demonstrated an ability to garner high patientatisfaction levels,22,26,28-30,32,34-37,40 although Gagnon et al22

nd McCusker et al40 were unable to demonstrate anyifferences in effect compared with usual care. High homearegiver satisfaction32,36,40 was also demonstrated, althoughgain McCusker et al40 were unable to demonstrate any effectifferences compared with usual care. High hospital staff andommunity care provider satisfaction28,29,32,34,36 was alsoemonstrated. Only 2 of the 6 studies examining functionalutcomes noted an ability to achieve significantly favorableesults.37,40 Finally, only 2 of the 6 studies that examined theirntervention’s effect on improving a patient’s perceived well-eing or quality of life were able to demonstrate an overall

mprovement.28,30

Volume , . : June

ctbmsptitiamiitcu

Sinha et al The Geriatric Emergency Management Nursing Model

At a systems level, of the 13 studies examining revisitationrates to EDs, 7 demonstrated reductions in the early post-EDdischarge period23,26,30-32,36,38 and 1 up to 18 months after anindex ED visit.27 Five of the 6 studies that examined the abilityof the intervention to immediately obviate inpatient admissionsdemonstrated success,25,28,29,32,34 whereas 6 of the 7 studies thatexamined their ability to reduce subsequent nonelective hospitaladmissions demonstrated this as well.26,27,30,31,36,38 Decreases inlengths of inpatient stays were achieved by 3 of the 5interventions examining this.26,31,38 The ability to reducesubsequent nursing home admissions was demonstrated in only1 of the 4 studies examining this outcome.40

Figure 3 identifies the 28 outcome variables that the 18studies collectively measured, representing the obviousheterogeneity in trial design, interventions, and outcomemeasures that were of interest to the researchers. The mainvariables measured included ED revisitations in 13 (72%),patient satisfaction in 12 (67%), subsequent hospital admissionsin 7 (39%), and functional decline in 6 (33%). Length ofinpatient stays and ED and hospital care provider satisfactionwere measured in 5 studies; 4 examined costs related to healthand social care services utilization, nursing home admissions,home caregiver satisfaction, and the patients’ perceived sense ofwell-being or quality of life; and 3 examined the satisfaction ofprimary and community care service providers, patient healthstatus, and health and social care utilization. Only studiesperformed in single-payer health systems such as Canada,Australia, and the US Veterans Affairs network measuredhospital admission rates or the number of obviated hospitaladmissions. According to our qualitative analysis of the 28outcome variables used, consensus emerged from discussionsamong the authors, identifying 13 of these outcome measuresspecifically denoted in Figure 3 that appeared to measure themost significant and relevant patient and systems core outcomesof interest in relation to this model. Although the 13 areincluded in those mentioned earlier, we also thought that theoutcome measure examining patient adherence to follow-upappointments that was examined in only 1 study should warranttreatment as a core measure as well.

LIMITATIONSAs with any study, ours has several limitations. To the extent

that we performed a systematic review, publication bias and thepotential for incomplete identification of relevant studies arealways considerations. We conducted our search with multipledatabases, used bibliographic searches of the reviewed articles,and finally surveyed the authors of the reviewed articles toidentify additional studies to minimize this possibility. Theheterogeneity in trial design, interventions, and outcomemeasures did not completely preclude our ability to usetraditional meta-analytic techniques around certain outcomemeasures. However, given the goals of this study, meta-analysiswould not have been appropriate.

Using the data from our systematic review, we performed a

qualitative analysis and drew inferences about model a

Volume , . : June

omponents associated with effectiveness. We acknowledge thathe analysis of studies by positive versus negative results cane highly tentative because small studies with large effectsight have had negative results, whereas large studies with

mall effects might have had positive ones. Nevertheless,revious studies have used similar methodologies to informhe development of other clinical models and health servicesnterventions, with useful results.19 Furthermore, we invitedhe authors of the included studies to verify ournterpretations of their work to improve the quality of ournalyses. Evaluating and developing complexulticomponent health and social interventions such as these

n randomized clinical trials is problematic, especially whent is difficult to recruit, standardize, blind, and randomize inhis setting.10,12,41 Berwick42 further argues that randomizedontrolled trials may be impoverished ways to learn andnderstand interventions when knowledge of mechanisms

Figure 3. Identified research outcome measures.

nd context will be integral to understanding outcomes.

Annals of Emergency Medicine 679

ErafaEimg

AgaFfvoimcndaiw

paaEfasetaeasbcmEbgsegb

ct

The Geriatric Emergency Management Nursing Model Sinha et al

DISCUSSIONThis systematic review and qualitative analytic approach

identifies components and outcome measures that are integral toeffective case management models for older patients in thecontext of an index ED visit. This study builds on the emergingdescriptive literature in this area and leverages the differences inthese models and their associated outcomes to support thefuture development of an evidence-based normative andeffective geriatric emergency management practice modeldesigned to address the special care needs of older patients andthereby improve their health and health service utilizationoutcomes.

We identified 8 described core characteristic componentsthat appear to be integral to the geriatric emergencymanagement model and drew inferences about how thepresence or absence of these core characteristic components mayhave influenced overall intervention effectiveness. Severalthemes emerged. Effective geriatric emergency managementinitiatives use validated risk stratification tools as a routineprelude to initiating an assessment and developing a care plan orreferral process in the ED with specialized clinicians. Teamcomposition and leadership strongly influenced modeleffectiveness, and nurses in particular appeared to be a criticalcomponent. We therefore hypothesize that nurses and midlevelclinicians, whose training and experience provide them withbroad-based skill sets that span both health and social caredomains, can best blend their understanding of emergency care,disease processes, utilization management, payer requirements,and available community resources43 when needed toimplement complex and often interrelated health and socialinterventions that a geriatric emergency management modelrequires. Although some found a social work–led interventionto be effective,26 others have found that without appropriatenursing support, social workers in general did not have thebroader skill set required to work as case managers withinthe ED.25

Our findings are buttressed by the results of models thatwere unable to demonstrate any significant effect.14,22,39 Wehypothesize that lack of effect may relate to a failure to targetindividuals with greatest need, focus on the development ofinterprofessional and capacity-building work practices withother ED and hospital care providers, and the absence orlimited presence of defined follow-up processes allowing earlierrecommendations to be reviewed, revised, or reinforced.

Our data and experience in implementing such modelssuggest that collaborative working practices and capacity-building efforts can influence and enhance geriatric emergencymanagement initiatives and may also be their most integralcomponent. Similar to what has been observed in other studiesexamining the implementation of complex health and socialinterventions,44-46 collaborative working practices are critical inmodel implementation and rely on the interpersonal skill sets ofthe clinicians delivering those initiatives and their ability to earn

the trust and respect of their colleagues within and beyond the M

680 Annals of Emergency Medicine

D. These practices have been underappreciated in previousesearch, perhaps because their influence is underrecognizednd measuring their effects remains challenging. Usingormer ED nurses in geriatric emergency management rolesnd embedding geriatric emergency management nurses asD staff members are 2 ways to possibly facilitate

nterpersonal, interprofessional, and capacity-buildingeasures to help formalize this new role in a setting in which

eriatric care principles are uncommon.The model we propose is consistent with previous literature.

recent review observed that ED interventions providingeriatric nursing assessment together with community servicesnd primary care linkages seem to be most effective.10

urthermore, the use of validated risk-stratification tools hasacilitated effective assessment and referral strategies in higher-ersus lower-risk older ED attendees.35,40 Our systematic reviewf the literature demonstrates that the model components wedentify may be broadly integral to any geriatric case

anagement model because all of the 8 characteristicomponents we identified appeared to underlie other effectiveon-ED case management models for at-risk community-welling and hospitalized older patients in particular.47,48 Thesend other studies attributed the overall effectiveness of thesenterventions to their ability to target at-risk older adults whoould most likely benefit from them.

The geriatric emergency management practice model weropose offers an evidence-based, nursing-led, interprofessionalpproach and philosophy of care that aims to improve allspects of the emergency care that older patients receive. ThisD-based model also provides a framework that can inform

uture research and development of the model. We also proposeset of 13 core outcome measures that, if adopted in future

tudies of geriatric emergency management programs, cannhance future research efforts to evaluate, monitor, and refinehe model. Unlike earlier geriatric ED interventions, thispproach is not limited to discharge planning but alsomphasizes appropriate assessment and care planning activities,long with patient, home caregiver, community, and hospitaltaff care provider education and interprofessional capacityuilding. Furthermore, unlike in traditional geriatriconsultation teams, the integration of geriatric emergencyanagement practitioners with strong interpersonal skills as fullD staff members can facilitate capacity building within andeyond these settings. We hope the development of a standarderiatric emergency management nursing practice model willupport its broader dissemination and refinement efforts,specially because an increasing number of hospitals are usingeriatric nurses and nurse practitioners within their EDs toetter address the needs of older patients.

The authors acknowledge the following individuals forontributing to this article by providing its authors feedback onheir studies that were reviewed in this article: David Basic, MBBS,

PH, Ken Brazil, PhD, Gideon Caplan, MBBS, Helen Corbett,

Volume , . : June

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

Sinha et al The Geriatric Emergency Management Nursing Model

MaHSc, Michelle Freeman, MSN, Constance Schein, MSc, JaneMcCusker, MD, PhD, Douglas Miller, MD, Lorraine Mion, RN,PhD, Ro Roberts, MSW, Rebecca Warburton, PhD.

Supervising editor: Robert L. Wears, MD, MS

Author contributions: SKS, ESB, NF, and BL conceived thestudy and designed the review. SKS acquired the data. SKS,ESB, NF, and BL participated in the analysis and interpretationof the data. SKS and BL drafted the article, and all authorscontributed substantially to its critical review. BL supervisedthe study. SKS takes responsibility for the paper as a whole.

Funding and support: By Annals policy, all authors arerequired to disclose any and all commercial, financial, andother relationships in any way related to the subject of thisarticle as per ICMJE conflict of interest guidelines (seewww.icmje.org). Dr. Sinha was jointly supported as a fellowof the Erickson and Donald W. Reynolds Foundations.

Publication dates: Received for publication May 5, 2010.Revisions received November 15, 2010, and January 23,2011. Accepted for publication January 28, 2011.

Reprints not available from the authors.

Address for correspondence: Samir K. Sinha, MD, DPhil,FRCPC, Mount Sinai Hospital, Suite 475-600 University Ave,Toronto, Ontario, Canada, M5G 1X5; 416-586-4800 ext 7859,fax 416-586-5113; E-mail ssinha@mtsinai.on.ca.

REFERENCES1. Nawar E, Niska R, Xu J. National Hospital Ambulatory Medical

Care Survey: 2005 Emergency Department Summary. Hyattsville,MD: National Center for Health Statistics; 2007. No. 386.

2. Aminzadeh F, Dalziel WB. Older adults in the emergencydepartment: a systematic review of patterns of use, adverseoutcomes, and effectiveness of interventions. Ann Emerg Med.2002;39:238-247.

3. Keim S, Sanders A. Geriatric emergency department use andcare. In: Meldon S, Ma OJ, Woolard R, eds. Geriatric EmergencyMedicine. New York, NY: McGraw-Hill; 2004, pages 1-3.

4. Strange GR, Chen EH. Use of emergency departments by elderpatients: a five-year follow-up study. Acad Emerg Med. 1998;5:1157-1162.

5. McNamara RM, Rousseau E, Sanders AB. Geriatric emergencymedicine: a survey of practicing emergency physicians. Ann EmergMed. 1992;21:796-801.

6. Baum SA, Rubenstein LZ. Old people in the emergency room: age-related differences in emergency department use and care. J AmGeriatr Soc. 1987;35:398-404.

7. Sanders AB, Morley JE. The older person and the emergencydepartment. J Am Geriatr Soc. 1993;41:880-882.

8. Grief CL. Patterns of ED use and perceptions of the elderlyregarding their emergency care: a synthesis of recent research.J Emerg Nurs. 2003;29:122-126.

9. Hwang U, Morrison RS. The geriatric emergency department. J AmGeriatr Soc. 2007;55:1873-1876.

10. Hastings SN, Heflin MT. A systematic review of interventions toimprove outcomes for elders discharged from the emergencydepartment. Acad Emerg Med. 2005;12:978-986.

11. McCusker J, Verdon J. Do geriatric interventions reduceemergency department visits? a systematic review. J Gerontol A

Biol Sci Med Sci. 2006;61:53-62.

Volume , . : June

2. Fealy G, McCarron M, O’Neill D, et al. Effectiveness ofgerontologically informed nursing assessment and referralinterventions for older persons attending the emergencydepartment: systematic review. J Adv Nurs. 2009;65:934-935.

3. Hickman L, Newton P, Halcomb EJ, et al. Best practiceinterventions to improve the management of older people inacute care settings: a literature review. J Adv Nurs. 2007;60:113-126.

4. Basic D, Conforti DA. A prospective, randomised controlled trial ofan aged care nurse intervention within the emergencydepartment. Aust Health Rev. 2005;29:51-59.

5. Moher D, Liberati A, Tetzlaff J, et al; PRISMA Group. PreferredReporting Items for Systematic Reviews and Meta-Analyses: thePRISMA statement. PLoS Med. 2009;6:e1000097.

6. Ryan R, Hill S, Broclain D, et al. Cochrane Consumers andCommunication Review Group. Study Quality Guide. March 2007.

7. Stroup D, Berlin J, Morton A, et al. Meta-analysis of observationalstudies in epidemiology: a proposal for reporting. JAMA. 2000;283:2008-2012.

8. Huberman M, Miles A. Research utilisation: the state of the art,knowledge and policy. Knowledge Policy. 1994;7:13-33.

9. Hedrick SC, Inui TS. The effectiveness and cost of home care: aninformation synthesis. Health Serv Res. 1986;20(6 pt 2):851-880.

0. Mosteller F, Gilbert J, McPeek B. Reporting standards andresearch strategies for controlled trials. Control Clin Trials. 1980;1:37-58.

1. Validation Procedures Manual. Baltimore, MD: Policy ResearchInc; 1978.

2. Gagnon AJ, Schein C, McVey L, et al. Randomized controlled trialof nurse case management of frail older people. J Am GeriatrSoc. 1999;47:1118-1124.

3. Miller DK, Lewis LM, Nork MJ, et al. Controlled trial of a geriatriccase-finding and liaison service in an emergency department.J Am Geriatr Soc. 1996;44:513-520.

4. Brazil K, Bolton C, Ulrichsen D, et al. Substituting home care forhospitalization: the role of a quick response service for theelderly. J Community Health. 1998;23:29-43.

5. Brookoff D, Minniti-Hill M. Emergency department-based homecare. Ann Emerg Med. 1994;23:1101-1106.

6. Brooks MM, Ertl JM. Social work home visits: impact onrecidivism and health care costs. Continuum Soc Soc WorkLeadersh Health Care. 2000;20:3-9.

7. Caplan GA, Williams AJ, Daly B, et al. A randomized, controlledtrial of comprehensive geriatric assessment and multidisciplinaryintervention after discharge of elderly from the emergencydepartment—the DEED II study. J Am Geriatr Soc. 2004;52:1417-1423.

8. Corbett HM, Lim WK, Davis SJ, et al. Care coordination in theemergency department: improving outcomes for older patients.Aust Health Rev. 2005;29:43-50.

9. Freeman M. Quick response programs: effective management ofa population in crisis. Leadersh Health Serv. 1994;3:36-39.

0. Guttman A, Afilalo M, Guttman R, et al. An emergencydepartment-based nurse discharge coordinator for elder patients:does it make a difference? Acad Emerg Med. 2004;11:1318-1327.

1. Hegney D, Buikstra E, Chamberlain C, et al. Nurse dischargeplanning in the emergency department: a Toowoomba, Australia,study. J Clin Nurs. 2006;15:1033-1044.

2. Moss JE, Flower CL, Houghton LM, et al. A multidisciplinary carecoordination team improves emergency department dischargeplanning practice. Med J Aust. 2002;177:435-439.

3. McCusker J, Jacobs P, Dendukuri N, et al. Cost-effectiveness of a

brief two-stage emergency department intervention for high-risk

Annals of Emergency Medicine 681

4

4

4

4

4

4

4

4

The Geriatric Emergency Management Nursing Model Sinha et al

elders: results of a quasi-randomized controlled trial. Ann EmergMed. 2003;41:45-56.

34. McCusker J, Dendukuri N, Tousignant P, et al. Rapid two-stageemergency department intervention for seniors: impact oncontinuity of care. Acad Emerg Med. 2003;10:233-243.

35. Mion LC, Palmer RM, Meldon SW, et al. Case finding and referralmodel for emergency department elders: a randomized clinicaltrial. Ann Emerg Med. 2003;41:57-68.

36. Roberts RM, Dalton KL, Evans JV, et al. A service model of short-term case management for elderly people at risk of hospitaladmission. Aust Health Rev. 2007;31:173-183.

37. Runciman P, Currie CT, Nicol M, et al. Discharge of elderly peoplefrom an accident and emergency department: evaluation of healthvisitor follow-up. J Adv Nurs. 1996;24:711-718.

38. Warburton RN. Preliminary outcomes and cost-benefit analysis ofa community hospital emergency department screening andreferral program for patients aged 75 or more. Int J Health CareQual Assur Inc Leadersh Health Serv. 2005;18:474-484.

39. Weir R, Browne G, Byrne C, et al. The quick response initiative inthe emergency department: who benefits? Health Care ManagSci. 1999;2:137-148.

40. McCusker J, Verdon J, Tousignant P, et al. Rapid emergency

department intervention for older people reduces risk of

682 Annals of Emergency Medicine

functional decline: results of a multicenter randomized trial. J AmGeriatr Soc. 2001;49:1272-1281.

1. Lindsay B. Randomized controlled trials of socially complexnursing interventions: creating bias and unreliability? J Adv Nurs.2004;45:84-94.

2. Berwick DM. The science of improvement. JAMA. 2008;299:1182-1184.

3. Brewer BB, Jackson L. A case management model for theemergency department. J Emerg Nurs. 1997;23:618-621.

4. Sinha SK. The Sociology of Interprofessional Relations: A CaseStudy of English Care Trusts [dissertation]. Oxford, UnitedKingdom: University of Oxford; 2007.

5. Rubenstein LZ, Josephson KR, Wieland GD, et al. Effectivenessof a geriatric evaluation unit: a randomized clinical trial. N EnglJ Med. 1984;27:1664-1670.

6. Sommers LS, Marton KI, Barbaccia JC, et al. Physician, nurse,and social worker collaboration in primary care for chronically illseniors. Arch Intern Med. 2000;26:1825-1833.

7. Boult C, Boult LB, Morishita L, et al. A randomized clinical trial ofoutpatient geriatric evaluation and management. J Am GeriatrSoc. 2001;49:351-359.

8. Naylor MD, Brooten D, Campbell R, et al. Comprehensivedischarge planning and home follow-up of hospitalized elders: a

randomized clinical trial. JAMA. 1999;281:613-620.

Volume , . : June