Page 1 of 1

Chapter 30. Geriatric Evaluation and Management Units for Hospitalized Patients

Joseph V. Agostini, M.D.
Dorothy I. Baker, Ph.D., RNCS
Sidney T. Bogardus, Jr., M.D.
Yale University Schools of Medicine and Public Health


Inpatient care of elders with multiple co-morbidities requires close attention to the special needs and geriatric syndromes that arise in this vulnerable population.1,2 One strategy to address the risks of hospitalization is to provide care by a multidisciplinary team in a dedicated geriatric unit, principally Geriatric Evaluation and Management (GEM) Units. This model of care, in many ways similar to coordinated stroke units,3 may improve mortality and other clinically relevant outcomes compared with outcomes achieved on a general medical ward. An alternative strategy, reviewed elsewhere in this Compendium, is the comprehensive geriatric consultation service, analogous to a typical medical consultation team (see Chapter 29).

In a GEM unit, a multidisciplinary team provides comprehensive geriatric assessment, detailed treatment plans, and attention to the rehabilitative needs of older patients. A typical team is composed of a geriatrician, clinical nurse specialist, social worker, and specialists from such fields as occupational and physical therapy, nutrition, pharmacy, audiology, and psychology. GEM units are typically separate hospital wards that have been redesigned to facilitate care of the geriatric patient. Multidisciplinary team rounds and patient-centered team conferences are hallmarks of care on these units, which, in contrast to geriatric consultation services, have direct control over the implementation of team recommendations.

Practice Description

In all the studies reviewed in this chapter, the GEM unit team included a physician experienced in geriatric medicine, skilled geriatric nurses and rehabilitation specialists. (The latter may not have been on site but were accessible). The teams completed multidimensional geriatric assessments, conducted interdisciplinary team rounds, and provided comprehensive discharge planning. Units were physically separate wards that were designed to facilitate geriatric care. Acute Care of Elders (ACE) Units 4 incorporate the GEM unit design with additional enhancements and admit patients with acute illnesses. ACE units often provide improved flooring and lighting, reorienting devices, and other environmental improvements such as common rooms for patient use. For example, in the ACE unit studied by Landefeld and colleagues5 the GEM unit concept was enhanced by the use of more nurse-initiated protocols and a greater number of environmental and design modifications.

Both styles of unit emphasize the early assessment of risk factors for iatrogenic complications and the prevention of functional decline.4,6 Studies of both are included in this chapter.

Prevalence and Severity of the Target Safety Problem

One-third of hospitalized patients are aged 65 years and older. In 1996, although comprising only 13% of the US population they accounted for 38% of the approximately 31 million discharges from non-government, acute care hospitals.7 Since all hospitalized older patients are potentially at risk, the target population is quite large. Appropriate selection of patients who are at risk for hospital-related complications and who are likely to receive benefit from this practice, however, would decrease this number.8

The target safety problems are preventable deaths and hospital-related functional decline in older persons. The number of deaths that could be prevented if the practice were to be widely implemented is unknown. On the other hand, since hospital-related functional decline occurs in 25 to 60% of older hospitalized patients, there is substantial opportunity to improve clinical outcomes.9 Other clinical problems explicitly studied in controlled trials include cognitive status and nursing home placement.

Opportunities for Impact

Data from the American Hospital Association (AHA)10 indicate that fewer than half of hospitals providing care for the elderly have geriatric acute care units or offer comprehensive geriatric assessment. (see footnote in Chapter 29). Researchers in the field believe the number is increasing. The Department of Veterans Affairs reports that in 1997 there were 110 active GEM units, although some concentrate solely on outpatient assessment.11 A recent national survey12 identified at least 15 active ACE units, with average daily censuses ranging from 5 to 25 patients. Depending on the screening and targeting criteria used to identify eligible patients, the potential for impact could be quite large and raises the question of the physical and manpower capacity of these units to meet the apparent demand.

Study Designs

A structured literature search identified a systematic review13 that included 6 studies (4 randomized controlled trials,14-17 one retrospective cohort study with historical controls,18 and one published abstract of a randomized controlled trial19). We also identified 2 randomized controlled trials of ACE units5,20 that were published later (see Table 30.1). All of the cited studies were single center in design. Five of the studies5,14,16,17,20 provided sufficient data to evaluate the baseline level of function of enrolled patients.

Study Outcomes

All-cause mortality was reported in each study. For this outcome, most patients were followed 6 or more months after hospitalization. Other clinically important outcomes measured in some studies were functional status,5,14,16,17,20 cognitive status,14,16,17 length of stay,5,14-16,18,20 and discharge rates to institutional settings.5,14-20

Evidence for Effectiveness of the Practice

In some studies mortality during hospitalization, at 3 months, or at 6 months was reduced in the intervention group but the differences failed to achieve statistical significance. A meta-analysis13 of 6 studies14-19 found the summary odds ratio for 6-month mortality was 0.65 (95% CI: 0.46-0.91), using both published and unpublished data from the included trials. Tests for heterogeneity across studies were reported with p<0.10 for the pooled analyses.

Cognitive function, as measured by the Kahn-Goldfarb Mental Status Questionnaire,21 showed no statistical improvement over the course of one study,17 nor did 2 other studies14,16 demonstrate improvement, using the Mini-Mental State Examination22 to assess mental status. Two trials of ACE units examined functional status, using the basic activities of daily living (ADL).23 Landefeld et al5 reported statistically significant improvements, while Counsell et al20 found benefit in a composite outcome of ADL improvement and nursing home placement, but not in discharge ADL levels alone. Two other studies also demonstrated statistically improved functional status in the six months after randomization14 and at 12 months follow-up.17

In individual studies, GEM units were associated with a higher likelihood of home residence, rather than in an institutional setting (skilled nursing facility or nursing home).5,14,17 The meta-analysis by Stuck et al calculated a combined odds ratio that revealed a statistically significant increase in patients discharged from GEM units who were living at home at 6 months (summary odds ratio 1.80, 95% CI: 1.28-2.53) and 12 months (summary odds ratio 1.68, 95% CI: 1.17-2.41) thereafter.13 ACE unit trials in a community hospital20 and a university hospital5 were also both successful in decreasing patient placement in institutional settings and would likely have strengthened the summary estimate if included. Extrapolating their study findings to the US population, Rubenstein and colleagues17 estimated that approximately 200,000 nursing home admissions per year could be avoided using their geriatric evaluation unit approach. Winograd noted that this multidisciplinary practice would potentially be more effective if target populations were better identified and enrolled using specific criteria.8

Potential for Harm

No data suggest that GEM units were associated with harm.

Costs and Implementation

Implementation of the practice requires construction or redesign of hospital ward(s) to create a suitable environment, training or recruiting experienced staff, establishing selection criteria to determine patient eligibility, and implementing a continuous evaluation process to assess clinical and non-clinical outcomes.

A working group has recommended including costs as an important outcome measure in future studies of GEM units.24 Applegate and colleagues reported in a later analysis25 of their randomized controlled trial14 that the increased costs associated with their intervention study were not balanced by savings in subsequent healthcare spending, but if charges were adjusted for days spent at home (versus in long-term care) the charges were similar. Lower direct costs were demonstrated by one intervention study,17 particularly after adjusting for differences in survival (mean institutional-care costs per year survived, $22,597 for intervention patients vs. $27,826 for control-group patients). The study by Landefeld et al5 reported mean hospital charges of $10,289 for intervention patients compared with $12,412 for control patients, with similar median charges (p=0.3). Additional costs of about $75,000 attributable to staffing and physical redesign of the unit resulted in a cost of about $230 per patient in the intervention group. Counsell et al,20 in a recent large randomized trial of an ACE intervention, reported no difference in hospital costs for patients in the intervention group compared with the usual care group ($5640 vs. $5754, respectively; p=0.68). Included in the costs was $28 per hospital day per intervention patient, representing costs of the geriatric nurse specialist, unit medical director, and unit renovations ($300,000).


Reasonable evidence supports the use of GEM units, despite varying findings across individual studies with respect to their effectiveness at preventing outcomes such as mortality. Nonetheless, mortality appears to be improved after pooling results of smaller trials. There is good evidence that this model of care decreases nursing home placements, which in itself is a noteworthy finding. Furthermore, the intervention costs may not be significantly higher in these specialized units. The generalizability of the practice requires further examination, and the need for multicenter studies, as advocated by a previous consensus group,26 has thus far not been undertaken.

Limitations of this practice compared with multidisciplinary geriatric consultation teams (Chapter 29) include limited bed availability in most units, decreased transferability of geriatric practices throughout a hospital, and a larger resource commitment compared with a hospital-wide consultation team. The advantages of the GEM and ACE unit model are direct control over implementation of clinical recommendations, the presence of dedicated geriatric nursing and rehabilitative staff associated with the unit, and the beneficial effects of a ward designed to address older patients' needs. In sum, the practice of a dedicated GEM or ACE unit carries much promise.

Table 30.1. Randomized Controlled Trials of Geriatric Evaluation and Management Units*

StudyStudy SettingStudy DesignAll-Cause Mortality and Other Outcomesa
Stuck, 1993136 studies (published 1983-1991) in the US, UK, Australia, and Canada, involving 1090 patients (meta-analysis of Refs. 14-19)Level 1A

6-month mortality: summary odds ratio 0.65 (95% CI: 0.46-0.91)

12-month mortality: summary odds ratio 0.77 (95% CI: 0.56-1.06)

Applegate, 199014155 patients in a university-affiliated community hospital, 1985-1987Level 1

6-month mortality: 10% vs. 21% (p=NS)

After 6 months: greatest difference p=0.08 by log-rank test

Improvement in ADLs: 3 of 8 ADLs better in intervention group (p<0.05)

Counsell, 2000201531 patients in a community hospital, 1994-1997Level 1

Inpatient mortality: 3% vs. 4% (p=0.30)

Length of stay: no significant difference

Long-term placement or decline in ADLs

At discharge: 34% vs. 40% (p=0.027)

At 12 months: percentages not reported (p=0.022)

Gilchrist, 199815222 women on an orthopedic-geriatric service in the UK, 1984-1986Level 1

Inpatient mortality: 4% vs. 10% (p=0.06)

6-month mortality: 14% vs. 18% (p>0.1)

Harris, 199116267 patients in an Australian hospital, 1985-1986Level 1

Inpatient mortality: estimated from Figure 1 in paper, 8% vs. 6% (p=NS)

12-month mortality: 23% vs. 29% (p=NS)

Landefeld, 19955651 patients in a university-affiliated hospital, 1990-1992Level 1

Inpatient mortality: 7% in both groups (p=NS)

3-month mortality: 14% vs. 13% (p=NS)

Improvement in ADLs at discharge: 34% vs. 24% (p=0.009)

Discharged to nursing home: 14% vs. 22% (p=0.01)

Powell, 199019203 patients in two Canadian teaching hospitals, year not statedLevel 1

Mortality: lower in intervention group; timing not stated (p not stated)

Transferred to long-term care: fewer in intervention group (p not stated)

Rubenstein, 198417123 patients in a VA hospital, 1980-1982Level 1

Inpatient mortality: 14.3% vs. 15.0% (p=NS)

12-month mortality: 23.8% vs. 48.3% (p<0.005)

12-month improvement in basic functional status: 48.4% vs. 25.4% (p<0.01)

12-month improvement in mental status: 35.6% vs. 22.4% (p=NS)

Teasdale, 198318124 patients in a VA hospital, 1981-1982Level 3

Inpatient mortality: 12% vs. 14% (p=NS)

6-month mortality: 28% vs. 35% (p=NS)

* ADL indicates activities of daily living; NS, not statistically significant.
a Comparisons are reported as intervention group vs. control group.


1. Creditor MC. Hazards of hospitalization of the elderly. Ann Intern Med 1993;118:219-223.

2. Hirsch CH, Sommers L, Olsen A, Mullen L, Winograd CH. The natural history of functional morbidity in hospitalized older patients. J Am Geriatr Soc 1990;38:1296-1303.

3. Stroke Unit Trialists Collaboration. How do stroke units improve patient outcomes? A collaborative systematic review of the randomized trials. Stroke 1997;28:2139-2144.

4. Palmer RM, Counsell S, Landefeld CS. Clinical intervention trials: the ACE unit. Clin Geriatr Med. 1998;14:831-849.

5. Landefeld CS, Palmer RM, Kresevic DM, Fortinsky RH, Kowal J. A randomized trial of care in a hospital medical unit especially designed to improve the functional outcomes of acutely ill older patients. N Engl J Med 1995;332:1338-1344.

6. Palmer RM, Landefeld CS, Kresevic D, Kowal J. A medical unit for the acute care of the elderly. J Am Geriatr Soc 1994;42:545-552.

7. Kramarow E, Lentzner E, Rooks R, Weeks J, Saydah S. Health and aging chartbook. Health, United States, 1999. Hyattsville, Maryland: National Center for Health Statistics, 1999.

8. Winograd CH. Targeting strategies: an overview of criteria and outcomes. J Am Geriatr Soc 1991;39S:25S-35S.

9. Palmer RM. Acute hospital care of the elderly: minimizing the risk of functional decline. Cleve Clin J Med 1995;62:117-128.

10. American Hospital Association. Health Statistics, 2001 edition. Chicago, Illinois: Health Forum, 2001.

11. Veterans Administration. VA programs for senior veterans. Geriatric evaluation and management program. Available at: Accessed May 14, 2001.

12. Raziano DB, Jayadevappa R, Valenzula D, Weiner M, Lavizzo-Mourey RJ. Do Acute Care for the Elderly (ACE) units work? Paper presented at: American Geriatrics Society Annual Meeting; May 10, 2001; Chicago, IL. 2001.

13. Stuck AE, Siu AL, Wieland GD, Adams J, Rubenstein LZ. Comprehensive geriatric assessment: a meta-analysis of controlled trials. Lancet 1993;342:1032-1036.

14. Applegate WB, Miller ST, Graney MJ, Elam JT, Burns R, Akins DE. A randomized, controlled trial of a geriatric assessment unit in a community rehabilitation hospital. N Engl J Med 1990;322:1572-1578.

15. Gilchrist WJ, Newman RJ, Hamblen DL, Williams BO. Prospective randomised study of an orthopaedic geriatric inpatient service. BMJ 1988;297:1116-1118.

16. Harris RD, Henschke PJ, Popplewell PY, et al. A randomised study of outcomes in a defined group of acutely ill elderly patients managed in a geriatric assessment unit or a general medical unit. Aust N Z J Med 1991;21:230-234.

17. Rubenstein LZ, Josephson KR, Wieland GD, English PA, Sayre JA, Kane RL. Effectiveness of a geriatric evaluation unit. A randomized clinical trial. N Engl J Med 1984;311:1664-1670.

18. Teasdale TA, Shuman L, Snow E, Luchi RJ. A comparison of placement outcomes of geriatric cohorts receiving care in a geriatric assessment unit and on general medicine floors. J Am Geriatr Soc 1983;31:529-534.

19. Powell C, Montgomery P. The age study: the admission of geriatric patients through emergency [abstract]. Age & Aging 1990;19(Suppl. 2):P21-P22.

20. Counsell SR, Holder CM, Liebenauer MA, et al. Effects of a multicomponent intervention on functional outcomes and process of care in hospitalized older patients: a randomized controlled trial of Acute Care for Elders (ACE) in a community hospital. J Am Geriatr Soc 2000;48:1572-1581.

21. Kahn RL, Goldfarb AI, Pollack M, Peck A. Brief objective measures for the determination of mental states in the aged. Am J Psychiatry 1960;117:326-328.

22. Folstein MF, Folstein SE, McHugh PR. "Mini-mental state": a practical method for grading the cognitive state of patients by the clinician. J Psychiatr Res 1975;12:189-198.

23. Katz S, Ford AB, Moskowitz RW, Jackson BA, Jaffe MW. Studies of illness in the aged. The index of ADL: a standardized measure of biological and psychosocial function. JAMA 1963;185:914-919.

24. Hedrick SC, Barrand N, Deyo R, et al. Working group recommendations: measuring outcomes of care in geriatric evaluation and management units. J Am Geriatr Soc 1991;39S:48S-52S.

25. Applegate WB, Graney MJ, Miller ST, Elam JT. Impact of a geriatric assessment unit on subsequent health care charges. Am J Public Health 1991;81:1302-1306.

26. Kramer A, Deyo R, Applegate W, Meehan S. Research strategies for geriatric evaluation and management: conference summary and recommendations. J Am Geriatr Soc 1991;39S:53S-57S.

Return to Contents
Proceed to Next Chapter

Page last reviewed July 2001
Internet Citation: Chapter 30. Geriatric Evaluation and Management Units for Hospitalized Patients. July 2001. Agency for Healthcare Research and Quality, Rockville, MD.