Printer Friendly

Risk of admission within 4 weeks of discharge of elderly patients from the emergency department--the DEED study.


Objective: to identify risk factors for admission for patients aged 75 years and older after discharge from the emergency department (DEED: discharge of elderly from emergency department).

Design: prospective evaluation of discharged elderly patients from the emergency department who were followed up after 4 weeks.

Setting: emergency department of a teaching hospital for 1 year.

Subjects: patients aged 75 and over discharged to their home or hostel. Main outcome measures: demographic data, indices of function and cognitive status and admission to hospital within 4 weeks.

Results: 468 patients aged 75 and over (mean age 81.7 years; range 75-98) were enrolled; 80 patients (17.1%) were admitted to hospital during the subsequent 4 weeks. Risk factors for admission included dependence in the following activities of daily living (relative risk; 95% confidence interval): bathing (2.41; 1.32-4.41); dressing (2.38; 1.22-4.63); stairs (1.60; 1.09-2.33); finance (1.66; 1.23-2.25); shopping (1.39; 1.12-1.73) and transport (1.61; 1.25-2.06), as well as use of a community nurse (1.88; 1.12-3.17). Logistic regression analysis found two variables to be significant in predicting admission or not: dependence in transport and use of a community nurse.

Conclusions: older patients are at increased risk of readmission within 4 weeks of being sent home from the emergency department. It is possible to identify high-risk patients by a questionnaire. This allows targeting of these patients for more intensive follow-up in an attempt to ameliorate further deteriorations in their health.

Keywords: activities of daily living, emergency department, frailty, old age


Patients over 75 more commonly present to and are also more often admitted to hospital from the emergency department (ED) than younger patients [1, 2]. Many of those who are not admitted to hospital, and are therefore discharged from the ED, will suffer a deterioration of their health or function, and approximately 20% will be admitted to hospital within the next month [3-5]. Many of these admissions may be preventable.

It is not clear from the published data why older patients have an increased risk of admission. This increased risk is not unique to the ED, but also occurs after discharge from surgical, medical and geriatric wards [6]. The effects of disease on frail, older people frequently lead to a cascade of illness and functional decline resulting in excess physiological morbidity and mortality [7]. This suggests that the increased risk of admission is due to frailty [8]. Frailty has never been precisely defined, but is used to describe patients, mainly over age 65, whom the burden of chronic disease leaves more vulnerable to environmental challenges. We hypothesized that standardized assessments of function, mental and social status would identify frail patients at risk of admission within the next month and designed the DEED (discharge of elderly from emergency department) study to test this hypothesis.


This study was conducted in the ED of a 361-bed teaching hospital attached to a university. The ED is an area trauma centre staffed by staff specialists and registrars in emergency medicine as well as by residents and interns on rotation. The population in the catchment area of the hospital includes 7.1% aged 75 years and over [9]. At our ED patients over 75 comprised 14.4% of all presentations, approximately double their representation in the population. This over-representation by a factor of two is in accord with other data [10].

We studied all patients aged 75 years and over who were discharged from the ED from 7 April 1994 to 6 April 1995. During the year of our study 3974 patients aged 75 and over presented to the ED, 2343 of these were admitted and 57 were dead on arrival. Patients were excluded if they met the following criteria: living out of area (121), living in a nursing home (246), eligible for the care of already existing intensive outreach services in orthogeriatrics, palliative care or psychiatry (56) or already in the study (211). These intensive outreach services already offer multidisciplinary interventions--medical, nursing and allied health--treating at home patients who may otherwise require admission. Only after the decision was made to discharge the patient was informed consent obtained. An additional group of patients declined consent (209) or were missed (263).

All patients who consented to enter the study answered a questionnaire covering living arrangements, background of the presentation to the ED, the Barthel index of activities of daily living (ADL) [11], a modified instrumental ADL index (IADL) [12] and the Short Portable Mental Status Questionnaire (MSQ) [13].

After 28 days patients were contacted by phone and visited at home, where detailed inquiry into the use of health services and the reason for that use during the study period was carried out. The study was approved by the hospital's ethics committee.

Risk of admission was calculated by dichotomizing individual's performance in each item of the ADL and IADL indices into fully independent or dependent to any degree. Statistical analyses were performed using Epi Info (USD Inc., Stone Mountain, GE, USA) for relative risks (RRs) and SPSS for Windows (SPSS Inc., Chicago, IL, USA) for logistic regression. For the logistic regression analysis, variables were entered into the model in a variety of combinations and analysed by automated model building to find the combination of variables with the best interpretability and parsimony, as recommended by the software.


The 468 patients enrolled in the study had a mean age of 81.7 years (range 75-98). Women outnumbered men 289 to 179 (61.8% female). Two-fifths of the subjects lived alone and three-fifths used no community services (Tables 1 and 2). One hundred and forty-four patients were referred by themselves (30.8%), 105 by their family (22.4%) and 113 by their general practitioner (24.1%); 59 were picked up by ambulances in public places (12.6%) and 47 (10.0%) came through other channels. The most common diagnoses causing presentation to the ED were 183 related to falls or collapses (39.1%; Table 3).
Table 1. Baseline characteristics of patients enrolled in
the study (n = 468)

                                       Value            %

Mean age (and range)                    81.7 (75-98)    -
Sex ratio, F:M                         289:179          62:38
No. of patients
  Living alone                         192              41.0
  Self-referred to ED                  144              30.8
  Receiving no community services(a)   304              65.0
Mean (and median) score
  Barthel (/20)                         19.0 (20)       -
  IADL (/12)                             9.0 (10)       -
  MSQ (/10)                              8.2 (9)        -

(a) Including private services (e.g. cleaner).

ED, emergency department: IADL, instrumental activities of daily living: MSQ, Short Portable Mental Status Questionnaire.
Table 2. Use of community services before entry into
the study (n = 468)

Service                     No. of patients(a)     %

Community nurse             47                    10.0
Community rehabilitation     5                     1.1
Home care                   69                    14.7
Meals on wheels             49                    10.5
Community options            6                     1.3
Private cleaner             24                     5.1
Other(b)                    38                     8.1
None                       304                    65.0

(a) Many subjects used more than one service.

(b) Includes 24 with private cleaners.
Table 3. Diagnostic groups

Diagnosis                             Frequency       %

Angina/congestive cardiac failure         46         9.8
Chest pain                                22         4.7
Fall [+ or -] soft tissue injury          82        17.5
Fall + fracture                           25         5.3
Collapse/syncope/dizziness                76        16.2
Respiratory                               33         7.1
Gastrointestinal                          57        12.2
Genitourinary                             15         3.2
Pain (back etc.)                          20         4.3
Other                                     90        19.2
Missing                                    2         0.4
Total                                    468       100.0

During the 4-week follow-up period 80 patients (17.1%) were admitted, of whom nine died. Three additional patients died at home. The principal diagnoses for the admissions are summarized in Table 4. Subjects had a significantly greater chance of being admitted to hospital during the follow-up period if they were dependent in any one of the following ADLs: bathing [RR, 2.41; 95% confidence interval (CD, 1.32-4.41; P= 0.0088), dressing (RR, 2.38; CI, 1.22-4.63; P= 0.020) or stairs (RR, 1.60; CI, 1.09-2.33; P= 0.030; Figure 1a) or the following IADLs: finance (RR, 1.66; CI, 1.23-2.25; P=0.0038), shopping (1.39; 1.12-1.73; P=0.011) or transport (RR, 1.61; CI, 1.25-2.06; 0.0014; Figure 1b). Patients with a total MSQ score of <9/10 had higher risk of admission (RR, 1.33; CI, 1.02-1.74; P= 0.050; Figure 1c).

Table 4. Reasons for admission within 1 month of discharge of elderly from the emergency department
Diagnostic group                 No.

Gastroenterology                  12
Cardiac                           11
Respiratory                        9
Orthopaedic/falls                  7
Neurology                          7
Urology/gynaecology                7
Vascular surgery/leg ulcers        5
Other                             11
Unknown (other hospital)          11
Total                             80


Use of support services and isolation appeared to be associated with increased risk of admission (Figure 1c), particularly use of a community nurse (RR, 1.88; CI, 1.12-3.17; P= 0.038). While there was a tendency for living alone (RR, 1.53; CI, 1.00-2.36; 0.052), and receiving meals on wheels (RR, 1.71; CI, 0.99-2.94; P=0.063) to be associated with increased risk of admission, they did not reach statistical significance.

There was no relationship between the presenting problem, sex of the patient or pattern of referral to the ED and admission within the next 4 weeks.

Thus, on univariate, or unadjusted, analysis eight variables were significantly associated with increased risk of admission. To adjust for confounding we performed a mulitivariate (logistic regression) analysis. The logistic regression analysis calculated the probability of being admitted and showed that there was a great deal of overlap between the variables from the ADL and IADL indices. This left two adjusted variables which were significantly associated with increased risk of admission in the equation: dependence in transport [adjusted odds ratio (OR), 2.03; CI, 1.09-3.77] and use of a community nurse (OR, 2.68; CI, 1.25-5.77; Table 5). The results of the logistic regression equation correlate with the unadjusted RRs calculated.

Patients independent for transport and not using a community nurse have a 10.0% rate of admission, whereas patients dependent in both have a 43.5% rate of admission within the next month (Figure 2). However, 40% of admissions were not dependent in either of these two variables.



We found that functional deficits, as measured by widely available scales for the assessment of elderly patients, and use of community services are predictive of readmission. Use of the logistic regression equation suggests that our two-variable equation can accurately predict the outcomes of 86% of patients aged 75 and over. It also has a relatively high positive predictive value of 84% but we could not recommend its use as a diagnostic instrument as it has not been validated. However, the multiple risk factors we have identified lead us to endorse the use of the simple screening instrument (ADL + IADL + MSQ) on all people over 75 presenting to the ED.

Previous studies of discharge of elderly patients from EDs did not attempt to identify factors predictive of admission, or any other deterioration in health status [2-5]. Our data demonstrate that a brief review of mental and physical function and use of community services identifies patients at risk of deterioration, as well as providing useful information for completing a care plan for that patient. It is important to assess patients over 75 individually and not simply refer them all to more intensive follow-up. The group as a whole showed a high level of independence. Most were living independently of all community services (Table 2) [we had already excluded the patients living in nursing homes]. The median scores were 20/20 on the Barthel index and 10/12 for the modified IADL, with less than 20% scoring below 19/20 on the Barthel index or below 6/12 on the IADL, confirming a high level of independent function. Therefore use of these instruments allows follow-up to be targeted at a high-risk group.

The high-risk group is conceptually identical with the group that has been called `the frail elderly'. Frailty has never been precisely defined, for example one geriatrics textbook defines frailty as an inherent vulnerability to challenge from the environment. A better definition of frailty would aid research and communication between geriatricians, and all who care for older patients. It has been suggested that no single marker, such as functional disability, is an accurate predictor of frailty because of inter-individual variation among older people. Some authorities define frail elderly people as dependent in ADLs--for example not independently mobile--and often in institutional care. This definition certainly identifies one end of the spectrum of frailty, but lacks sensitivity. Our data suggest that clusters of functional disabilities, especially more complex functional disabilities as measured by IADL, can define frailty for community-living elderly subjects, at least in the context of representations to the ED. We are not proposing a new definition of frailty, but refining the existing understanding.

The deficits of function associated with admission to hospital differ from those associated with admission to other care facilities. Incontinence of bladder and bowel will often trigger admission to nursing homes. Perhaps that is why the dependence in bladder, bowel and toileting were not associated with increased odds of admission to hospital: incontinent people tend to get admitted to nursing homes. However, dependence in bathing and dressing (assistance for which is easily, and commonly, supplied at home by community services) is suggestive of the sort of frailty which will allow people to stay at home. Similarly, assistance with financial affairs, shopping, transport and housework is often supplied by family, friends, commercial or community services when old people reach a degree of frailty. Thus persistence in one's own home is likely despite these disabilities. Use of the telephone stands out from the other IADLs in its clear lack of association with increased odds of admission, perhaps because it is a less complex task and can be accomplished by most very disabled elderly people.

The deterioration after discharge of elderly patients from EDs that leads to admission may be iatrogenic, due to the initial visit to the ED, although we consider this to be rare. The deterioration appears generally to be a continuation or exacerbation of an existing problem that was present at the first visit to the ED or, less commonly, a new problem. This phenomenon of more frequently returning elderly patients is not unique to the ED. Data from older patients readmitted after discharge from inpatient stays in hospital identified various reasons for readmission within the first month, including relapse of the original problem (51%), new problem (15%), carer problem (14%), terminal care (6%) medication problems (6%), complications of initial illness (5%) and problems with services (3%) [14]. Most returns are due to the original problem and this probably holds for all situations.

Hospital EDs face an increasing number of elderly patients, and not because elderly people are now presenting more frequently. The available longitudinal data show that the increase follows closely the ageing of the population [15]. Older people are not presenting for more frivolous complaints than younger patients. On the contrary, studies show that elderly patients in the ED are actually more seriously ill than younger patients [1, 16]. Therefore it is imperative that ED planning takes account of the ageing of the population.

Our data form a derivation set which has identified a small number of predictors of admission. We have begun a project to validate these data which, if successful, will provide EDs with a useful tool to identify older patients in need of follow-up after discharge.

Our data confirm that elderly patients discharged from the ED have a high rate of subsequent admission. We found that this is more common in frail elderly subjects, those with deficits of ADLs or IADLs. Prospective identification of this group, which is a small part of those over 75 years, will allow targeted strategies to be studied in an attempt to improve their care, and thereby their health and admission rate.

Key points

* Patients over 75 present to the emergency department at twice their proportion in the population.

* Within 4 weeks of being sent home from the emergency department, 17% will be admitted to hospital.

* Frail elderly people, as identified by dependency in instrumental activities of daily living and receipt of assistance from a community nurse, are at greater risk of admission.

* It is possible to identify frail elderly people in the community by dependence in instrumental activities of daily living.


We would like to thank Professor William Dunsmuir, Head of the Department of Statistics at the University of New South Wales, for his advice and assistance with statistical methods. We would also like to thank the staff of the Post Acute Care Services and the Emergency Department at Prince of Wales Hospital, without whose help this study would not have been possible.


[1.] Strange GR, Chen EH, Sanders AB. Use of emergency departments by elderly patients: projections from a multicenter data base. Ann Emerg Med 1992; 21: 819-24.

[2.] Richardson DB. Elderly patients in the emergency department: a prospective study of characteristics and outcome. Med J Aust 1992; 157: 234-9.

[3.] Rosenfeld T, Fahey P, Price M, Leeder S. The fate of elderly patients discharged from the accident and emergency department of a general teaching hospital. Community Health Stud 1990; 14: 365-72.

[4.] Currie CT, Lawson PM, Robertson CE, Jones A. Elderly patients discharged from an accident and emergency department--their dependency and support. Arch Emerg Med 1984; 1: 205-213.

[5.] Denman SJ, Ettinger WH, Zarkin BA, Coon PJ, Casani JA. Short term outcomes of elderly patients discharged from an emergency department. J Am Geriatr Soc. 1989; 37: 93743.

[6.] Tierney AJ, Worth A. Review: readmission of elderly patients to hospital. Age Ageing 1995; 24: 163-6.

[7.] Fretwell MD. Acute hospital care for frail older patients. In: Hazzard WR, Andres R, Bierman EL, Blass JP, eds. Principles of Geriatric Medicine and Gerontology, 2nd edition. New York: McGraw-Hill, 1990; 247-53.

[8.] Woodhouse KW, Wynne H, Baillie S, James OFW, Rawlins MD. Who are the frail elderly? Quart J Med 1988; 68: 505-6.

[9.] Australian Bureau of Statistics. Age and Sex Distribution of the Estimated Resident Population of Statistical Local Areas in New South Wales. Catalogue number 3209.1; 1991.

[10.] Richardson D. Emergency medicine for the elderly. Emerg Med 1995; 7: 73-4.

[11.] Wade DT. Stroke rehabilitation. In: Evans JG, Williams TE eds. The Oxford Textbook of Geriatric Medicine. Oxford: Oxford University Press, 1992; 321-8.

[12.] Lawton MP, Moss M, Fulcomer M, Lenban MH. Research and service oriented multilevel assessment instrument. J Gerontol 1982; 37: 91-5.

[13.] Pfeiffer E. A short portable mental status questionnaire for the assessment of organic brain deficit in elderly patients. J Am Geriatr Soc 1975; 23: 433-41.

[14.] Williams EI, Fitton E Factors affecting early unplanned readmission of elderly patients to hospital. Br Med J 1988; 297: 784-7.

[15.] Dove F, Dave S. Elderly patients in the accident department and their problems. Br Med J 1986; 292: 807-9.

[16.] Stathers GM, Delpech V, Raftos JR. Factors influencing the presentation and care of elderly people in the emergency department. Med J Aust 1992; 156: 197-200.

Received 20 August 1997; accepted 31 October 1997


Post Acute Care Services, Dickinson Building, Prince of Wales Hospital, Randwick, NSW 2031, Australia (1) Department of Emergency Medicine and (2) Emergency Department, Prince of Wales Hospital, Randwick, NSW, Australia

Address correspondence to: G. A. Caplan. Fax: (+61) 2 9382 2477; E-mail
COPYRIGHT 1998 Oxford University Press
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1998 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Publication:Age and Ageing
Geographic Code:4EUUK
Date:Nov 1, 1998
Previous Article:Jaw reflexes in healthy old people.
Next Article:The effect of walking aids on exercise capacity and oxygenation in elderly patients with chronic obstructive pulmonary disease.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters