2004, Editrice Kurtis

Transcription

1 Aging Clinical and Experimental Research Acute geriatric intervention increases the number of patients able to live at home. A prospective randomized study Ingvild Saltvedt 1,2, Turi Saltnes 2, Ellen-Sofie Opdahl Mo 1, Peter Fayers 2,3, Stein Kaasa 2,4, and Olav Sletvold 1,5 1 Section of Geriatrics, Department of Internal Medicine, University Hospital of Trondheim, Trondheim, Norway, 2 Unit of Applied Clinical Research, Norwegian University of Science and Technology (NTNU), Trondheim, Norway, 3 Department of Public Health, University of Aberdeen, Scotland, 4 Palliative Medicine Unit, Department of Oncology and Radiotherapy, University Hospital of Trondheim, Trondheim, Norway, 5 Department of Clinical Neurosciences, Faculty of Medicine, NTNU, Trondheim, Norway ABSTRACT. Background and aims: In a previous publication, we showed that treatment of acutely sick, frail elderly patients in a Geriatric Evaluation and Management Unit (GEMU) reduced mortality considerably when compared with the general Medical Wards (MW). The aim of this presentation was to study the impact of treatment in a GEMU on health care utilization. Methods: Acutely sick, frail patients, 75 years or older, who had been admitted as emergencies to the Department of Internal Medicine were randomized either to treatment in the GEMU (n=127) or to continued treatment in the MW (n=127). While usual treatment was given in the MW, the GEMU emphasized interdisciplinary and comprehensive assessment of all relevant disorders, early mobilization/rehabilitation, and discharge planning. After discharge from hospital, no specific follow-up was offered to any of the groups. Results: Of all subjects, 101 (80%) GEMU and 79 (64%) MW patients were still living in their own homes at three months (p=0.005); at six months the number was 91 (72%) and 74 (60%) (p=0.04) respectively. Median length of index stay was 19 days in the GEMU and 13 days in the MW group (p<0.001). After the initial stay, there were no statistically significant differences in admissions to or time spent in institutions. Conclusions: The results indicate the overall positive treatment effect of acutely sick, frail elderly in a GEMU, i.e. patients treated in the GEMU had increased possibilities of living in their own homes, an effect that was mainly related to considerably reduced mortality in the GEMU group. (Aging Clin Exp Res 2004; 16: ) INTRODUCTION Demographic trends in all Western societies consistently show increasing numbers of elderly people in need of health care facilities. At the same time, progress in medical science is creating increased expectations and demands, leading to considerable pressure on health care systems. A central issue is of course that medical treatment should be appropriate. Although patients and society may be more concerned about the quality of survival than about death, it has been shown that old people may have at least as much as the young to gain from active diagnostic work-up and invasive treatments in terms of preventing immediate death (1). Previous research has shown that elderly patients in hospital experience underdiagnosis and misdiagnosis, increased risk of iatrogenic conditions, and functional decline (2-5). A meta-analysis published in 1993 showed that treatment in Geriatric Evaluation and Management Units (GEMUs) increased the chance of living in one s own home and improved survival (6). Later, three randomized studies on the effectiveness of GEMUs from acute care settings were published, but failed to replicate these findings (7-9), even though in the study of Landefeld et al. discharges to long-term care institutions were reduced (7) and in that of Asplund et al. the length of the initial hospital stay was shortened (8). Two former rehabilitation studies have shown reduced time spent in institutions and reduced mortality for GEMU patients as compared with usual care (10, 11). In a recent randomized rehabilitation study, patients were assigned to receive care either in a GEMU or usual inpatient care after their condition had been stabilized (12). The mean number of days in long-term care was slightly lower for the GEMU patients and, al- Key words: Frail elderly, geriatric evaluation and management unit (GEMU), health care utilization, hospitalization, randomized clinical trial. Correspondence: I. Saltvedt, MD, Section of Geriatrics, University Hospital of Trondheim, 7006 Trondheim, Norway. Ingvild.Saltvedt@medisin.ntnu.no Received July 8, 2003; accepted in revised form May 14, Aging Clin Exp Res, Vol. 16, No. 4

2 Acute geriatric intervention though there was a significant reduction in functional decline, there was no difference in survival. Many hospital wards are overcrowded by frail elderly patients otherwise ready for discharge, who have to stay in hospital awaiting transfer to extended care facilities (13). These so-called bed-blockers may have substantial consequences for the daily organization of medical units. However, discharge of frail elderly patients depends upon several factors such as patients function, family members ability to care for the patient, access to and quality of general practitioners, and extent and quality of home care nursing. Additionally, many of these patients will need acute specialized care in order to prevent repetitive re-hospitalization, and close collaboration between the hospital and primary health care system is necessary. A GEMU was established in 1994 at the Department of Internal Medicine (DIM) at the University Hospital of Trondheim, Norway. As the Department had already performed a randomized study showing the beneficial effects of establishing a stroke unit (14), it was decided to evaluate the GEMU through a randomized study comparing treatment in the GEMU with treatment in the general medical wards (MW). In a previous publication, it was shown that mortality in the GEMU group was reduced by 50% during six months of follow-up (15). The present analyses were undertaken to explore the use of health care resources in the given population. When planning the study, we had the following expectations for the intervention patients: 1) an increased number of patients would be able to live at home; 2) the number of readmissions to and time in hospital and nursing homes during a six-month period of follow-up would be reduced; 3) the length of the initial stay would be longer in the intervention group, due to more time-consuming assessment and intervention, when compared with standard practice at the DIM. After reduced mortality in the GEMU group was found, it was debated that, if life were prolonged in those patients who had the highest morbidity, it was likely that overall health care utilization would be higher in the GEMU group. Hence, it would be a positive finding if no differences in health care utilization could be found in the two groups. METHODS Hospital setting Patients were recruited from the DIM at the University Hospital of Trondheim, Norway. This hospital serves both as the regional hospital for Central Norway, and as the local hospital for about 200,000 inhabitants of the city of Trondheim and neighboring municipalities. The DIM consisted of nine different sections with a total of 190 beds, and more than 90% of all its admissions were emergencies. In April 1994, a nine-bed GEMU was established as a section within the DIM. Six months later the present study started. Patients were included between October 31, 1994 and November 13, Patient selection and randomization Frail patients 75 years or older admitted acutely to the DIM (local hospital patients) were screened for enrolment. Patients should not be too healthy or too sick to benefit from treatment in a GEMU. To target frail patients, at least one of the Winograd targeting criteria had to be met (acute impairment of single activity of daily living (ADL), imbalance/dizziness, impaired mobility, chronic disability, prolonged bedrest, falls, confusion, depression, mild/moderate dementia, weight loss, malnutrition, vision or hearing impairment, urinary incontinence, social/family problems, polypharmacy) (16). Patients were excluded if they had terminal illnesses, including known cancer with metastases, severe dementia, or lived in nursing homes, as were also those who had been fully independent and seemed to recover quickly from the acute illness precipitating the emergency admission. Further, patients were not included if discharge was planned within two days. Suitable patients were screened when there was a free bed in the GEMU. Randomization was executed after patients had given their informed consent. Permuted block randomization with unknown and varied block size was used. During the study period, 254 patients were recruited and randomly allocated, 127 to the GEMU and 127 to continued treatment in the general medical wards (MW) where they were already staying. Patients allocated to the GEMU were transferred on the day of inclusion. More details on study design and methods have been described in an earlier publication (15). Patient care The staff in the GEMU included one geriatrician, one resident, two occupational therapists, and one physiotherapist. The number of nurses was comparable with that in the MW. All professionals evaluated each patient, and short- and long-term goals were set through regular interdisciplinary meetings. Comprehensive assessment and treatment of all relevant illnesses, prevention of complications, and early mobilization were essential. Rehabilitation was initiated in the GEMU, but if long-lasting rehabilitation was indicated, patients were referred to a rehabilitation institution. Discharge planning started as early as possible, in collaboration with the patient, family caregivers, and community nurses. When necessary, nurses or occupational therapists arranged visits to patients homes to evaluate their ability to manage to live at home, and if technical arrangements, aids and assistance from the home services was required. The staff in the GEMU was also used for (time-consuming) study-re- Aging Clin Exp Res, Vol. 16, No

3 I. Saltvedt, T. Saltnes, E-S. Opdahl Mo, et al. lated assessments in both groups during hospital stays and also during follow-up. Enrolled patients assigned to the MW were treated according to the routines of the DIM. Physiotherapy and occupational therapy were given when prescribed by the doctor, each therapist serving several wards. When considered appropriate by the hospital staff, arrangements for discharge were discussed with the community nurses. The health care system in Norway is public. In the municipality of Trondheim, at the time of this study, community nurses made the final decisions on the type and amount of service, including nursing home placement, offered to patients in both groups. Their decisions were based upon overall evaluation of patients medical and social situation, as well as the current availability of resources in the municipality, which had experienced a shortage of nursing home beds for years. After discharge from hospital, general practitioners were responsible for medical treatment in both groups. Baseline and outcome variables Sociodemographic characteristics were obtained through standard interviews with patients and caregivers. Information about diagnoses, hospital stays and number of readmissions to hospital was obtained from hospital records. Records of the Municipality of Trondheim gave information on nursing home admissions and duration of stays. Rehabilitation stays were registered partly from rehabilitation institutions records and partly from the National Registry. Official death certificates provided details on deaths. Sample size and statistical analysis Sample size estimations were based upon expected mortality reduction. It was estimated that one-year mortality would be 30% in the MW and 15% in the GEMU group, with α=0.05 and power 80%. This would require 113 patients in each group, and it was decided to include patients for one year or until this number of patients was reached (15). The chi-squared test was used for comparison of all categorical data. The Mann-Whitney U-test was used for comparison of age and time in institutions. Kaplan Meier plots and the log rank test were used for survival analyses and to compare time to readmission in hospital and time to nursing home placement. In order to estimate the magnitude of the treatment effect, Cox s proportional hazard model was used to estimate the hazard ratio (HR) of living at home. In this analysis individuals who were placed in permanent nursing homes or who died were regarded as experiencing a negative event in contrast to those still alive in their own homes. A p-value of less than 0.05 was considered statistically significant. All analyses were carried out by SPSS software version Ethics Participation in the trial was voluntary and according to the Helsinki Declaration. Written informed consent was obtained from all patients except those who were not able to write, when oral consent was accepted. If the patient was cognitively impaired, relatives also gave their written consent. The Regional Ethical Committee approved the protocol. RESULTS During the study period, 1426 patients aged 75 or more from the municipality of Trondheim were admitted to the Department of Internal Medicine. Of these, 254 (18%) were allocated to the study, with an identical number of patients in the two groups (Fig. 1). Five patients refused to participate, 1167 patients were not included either because they were not suitable for the study or because there was no free bed in the GEMU. One GEMU and two MW patients withdrew their consent after the index stay and were excluded from later analysis. Nursing home data were missing for ten GEMU and 16 MW patients; data on living location were missing for two MW patients (Fig. 1). Analysis Follow-up Allocation Admissions Patients 75 years from Trondheim admitted as emergencies, n=1426 Randomized, n=254 Allocated to GEMU, n=127 Lost to follow-up, n=1 (withdrew consent) Time in nursing homes, n=116 (missing n=10) Living location, n=126 All other analyses, n=126 Refused participation, n=5. Not suitable or no free bed in GEMU, n=1167 Allocated to MW, n=127 Lost to follow-up, n=2 (withdrew consent) Time in nursing homes, n=109 (missing n=16) Living location, n=123 (missing n=2) All other analyses, n=125 Figure 1 - Flow of study. GEMU: Geriatric Evaluation and Management Unit, MW: general Medical Wards. 302 Aging Clin Exp Res, Vol. 16, No. 4

4 Acute geriatric intervention Table 1 - Sociodemographic characteristics at baseline. GEMU (n=127) MW (n=127) Age, mean ±SD 82±5 82±5 Female 81 (64%) 84 (66%) Widowed/living alone 93 (73%) 85 (67%) Residence at time of inclusion Private home 115 (91%) 110 (87%) Sheltered housing 12 (9%) 17 (13%) Days in hospital before inclusion, 3.8 (±3.7) 4.6 (±4.3) mean (±SD) Previous diagnoses * Heart disease 46 (36%) 58 (46%) Infectious disease 30 (24%) 21 (17%) Gastrointestinal disorder 27 (21%) 22 (17%) Cerebrovascular disease 24 (19%) 17 (13%) Endocrine disease 20 (16%) 16 (13%) Airway disease 18 (14%) 9 (7%) Cancer 15 (12%) 12 (9%) Other 41 (32%) 44 (35%) No. of Winograd targeting criteria, 4 (3-5) 4 (3-5) median (iqr ) GEMU: Geriatric Evaluation and Management Unit. MW: general Medical Wards. *Diagnoses at earlier admissions to University Hospital of Trondheim. acute impairment of single activity of daily living (ADL), imbalance/dizziness, impaired mobility, chronic disability, prolonged bedrest, falls, confusion, depression, mild/moderate dementia, weight loss, malnutrition, vision or hearing impairment, urinary incontinence, social/family problems, polypharmacy (16). iqr= interquartile range. None of the differences were statistically significant. The GEMU and MW groups were comparable with respect to all demographic and clinical characteristics at baseline (Table 1). The mean number of days spent in hospital before entering the study was 3.8 (SD 3.7) in the GEMU and 4.6 (SD 4.3) in the MW, p=0.09. There were no differences in number of patients admitted to or time spent in hospital one year before inclusion in the study. At discharge, 38% of the GEMU patients and 7% of the MW patients had psychiatric diagnoses (mainly dementia, depression and delirium), p<0.001 (Table 2). There were no statistically significant differences in the distribution of the other diagnoses. Patients in the GEMU group had a median of three diagnoses at discharge, while in the MW group the median was two, p< As reported previously (15), the mortality in the GEMU and MW groups during the index stay was 8 (6%) and 17 (13%) (p=0.002), at three months 15 (12%) and 34 (27%) (p=0.004), and at six months 20 (16%) and 37 (29%) (p=0.02) respectively. After the index stay, a comparable number of patients in the GEMU and MW groups were discharged to their homes, 87 (73%) and 80 (73%); to rehabilitation institutions, 13 (11%) and 11 (10%); and to nursing homes, 15 (13%) and 15 (14%). Four (3%) patients in each group were transferred to another hospital department. Of those still alive, 10 (9%) GEMU and 11 (12%) MW patients were living in nursing homes at three months, and 15 (14%) and 12 (14%) at six months. The numbers living in their own homes were 101 (91%) in the GEMU group and 79 (88%) in the MW group at three months, and 91 (86%) and 74 (86%) at six months. Of all patients recruited in the study, 101 (80%) GEMU and 79 (64%) MW patients were living at home (p=0.005) at three months, and 91 (72%) and 74 (60%) (p=0.04) at six months. The hazard ratio (HR) of living at home versus living in nursing homes or having died, was 2.1 (95% CI ) after three months, and 1.6 (95% CI ) after six months, showing that the GEMU patients had significantly higher chances of living in their own homes. The partitioned survival curves (Fig. 2) show the proportion of observed time spent in nursing homes and hospital by the two groups. The proportion of observed time spent in hospital during the first weeks was higher in the GEMU group, because the index stay was longer (Fig. 3). The total length of the index stay had a median of 19 days (interquartile range 13 to 30) in the GEMU group and 13 days (interquartile range 7 to 18) in the MW group, p< Patients were not enrolled if discharge was planned within two days. Despite this criterion, 18 (14%) MW patients were discharged during this time period. After the index stay, the proportion of observed time spent in hospital and nursing homes had no statistically significant difference in the two groups (Fig. 2). Forty-one (35%) of the GEMU and 33 (31%) of the MW patients surviving the index stay were not readmitted to any kind of institution during follow-up (p=0.5). Each GEMU patient spent a mean of 35.4 (SD 48.5) days in institutions after the initial hospital stay, as compared with 30.9 (SD 47.3) days in the MW group (p=0.63). Table 2 - Diagnoses at discharge from index stay. GEMU MW (n=127) (n=127) n % n % Heart disease 72 (57) 69 (54) Infectious disease 20 (16) 25 (20) Gastrointestinal disease 22 (17) 15 (12) Cerebrovascular disease 38 (30) 28 (22) Endocrine disease 24 (19) 21 (17) Airway disease 17 (13) 12 (9) Cancer 12 (9) 10 (8) Psychiatric disorders* 48 (38) 9 (7) Other 51 (40) 68 (54) GEMU: Geriatric Evaluation and Management Unit., MW: general Medical Wards. *p< None of the other differences were statistically significant. Aging Clin Exp Res, Vol. 16, No

5 I. Saltvedt, T. Saltnes, E-S. Opdahl Mo, et al. Percentage Percentage GEMU MW In hospital In hospital At home At home Observation time (months) In nursing home In nursing home Figure 2 - Partitioned survival curves. Proportion of observed time that survivors spent in hospital, nursing homes and in own homes during six-month follow-up. GEMU: Geriatric Evaluation and Management Unit, MW: general Medical Wards. Forty-six (39%) of the GEMU and 42 (39%) of the MW patients surviving the index stay were readmitted to hospital during the six-month follow-up, and their mean number of days spent in hospital after the index stay was 18.9 (SD 17.9) and 15.2 (SD 16.8) respectively (p=0.26). There was no difference in time to the first readmission in the two groups. Within six months, 40 (35%) of the GEMU and 30 (28%) of the MW patients surviving the index stay had been in nursing homes (p=0.26). These GEMU and MW patients had spent a mean of 67.4 (SD 50.0) and 70.6 (SD 60.7) days respectively in nursing homes during the six-month follow-up (p=0.81). Twenty-two (19%) and 17 (16%) of these GEMU and MW patients respectively had been in nursing homes permanently. There was no difference in time from inclusion in the study to permanent nursing home placement, and there was no difference in time spent in rehabilitation institutions. The number of patients still living at home who received assistance from the community nurses was 58 (57%) in the GEMU and 43 (54%) in MW group at three months and 57 (63%) and 44 (59%) respectively at six months. DISCUSSION In contrast to other randomized trials from acute care settings (7-9, 17, 18), in this study an increased number of patients were still able to live in their own homes, an effect that was mainly related to the considerably reduced mortality in the GEMU group (Fig. 2). After the index stay, which was longer in the GEMU group, there were neither statistically significant differences in admissions to or days in hospital, nursing homes and rehabilitation institutions, nor in the percentage of patients who received assistance from the community nurses. The treatment of acutely sick frail elderly patients in the GEMU significantly reduced mortality, a result which has been discussed in a previous publication (15). The consequence of improved survival among frail elderly patients may theoretically be an increased number of patients in need of long-term care facilities and hospitalization, if life was prolonged in those who were the sickest and with the poorest function. However, our data demonstrate that the comprehensive one-time intervention performed in a GEMU by a multidisciplinary geriatric team increased survival with no extra need of home care services, rehabilitation, readmissions to hospital or nursing home stays, and this should be regarded as a positive finding. The index stay in the GEMU group was prolonged mainly due to a delay shortly after inclusion in the study (Fig. 3). There may be several explanations for this. Transfer of patients from another ward, the interdisciplinary approach, comprehensive medical examination, and the fact that the staff in the GEMU performed study-related assessments in both groups may all have prolonged the stay in the GEMU group. The short run-in period of six months after the GEMU was established, from scratch until the start of the trial, may also have contributed to an unnecessarily long index stay. However, Percentage of patients discharged Time from inclusion (days) GEMU MW Figure 3 - Length of index stay. GEMU: Geriatric Evaluation and Management Unit, MW: general Medical Wards. 304 Aging Clin Exp Res, Vol. 16, No. 4

6 Acute geriatric intervention continuous evaluation of the routines in the GEMU aimed at improving efficacy, and subsequently the length of stay has since been shortened considerably. The study design did not allow further analyses as to which elements in the intervention had impact on the outcomes. However, it is noteworthy that, although patients planned to be discharged within two days were excluded, 18 MW patients were discharged within this period (Fig. 3), showing that while discharge planning was emphasized in the GEMU group, this was not the case in the MW. Diagnoses given at discharge were usually confined to those relevant for that specific hospital stay, which means that the real number of diagnoses per patient are likely to be higher than reported. However, it was shown that patients in the GEMU group had more diagnoses at discharge, including an increased number of psychiatric diagnoses, than those in the MW group (Table 2). This is explained by the comprehensive assessment of all relevant disorders in the GEMU, not as a real difference between the two groups. In addition, internists are generally less engaged in geriatric conditions like cognitive decline and depression. Our intervention was based upon a conservative in-patient hospital approach, and the achievement of improved time spent at home is very encouraging. However, while designing the study, an important overall aim of the GEMU intervention was a general reduction of health care utilization by each patient, which was not achieved. Other randomized studies focusing on improved collaboration between hospital and community health care system have shown the following results: reduced length of hospital stay (18), reduced number of patients awaiting transfer to nursing homes (19), reduced number of unplanned readmissions (20-23), postponed permanent nursing home placement (18, 24), and increased number of days spent at home (18, 23). In hospitals coordinating geriatric and discharge planning services, a reduced percentage of beds occupied by patients awaiting long-term care placement has been shown (13). By combining GEMU treatment and post-discharge intervention, the meta-analysis of Stuck et al. showed increased ability to live at home (6). In the light of results from the present study and these findings, it seems that to reduce health care utilization, future treatment models should combine optimal in-hospital treatment with systematic follow-up of patients by the GEMU team, in collaboration with the primary health care system (25, 26). CONCLUSIONS This is the first study to demonstrate that treatment of acutely sick frail elderly patients in a GEMU can significantly improve survival (15) and increase the number of patients able to live in their own homes, yet without statistically significant extra demands on health care utilization after discharge from hospital. These results strongly support the introduction of GEMU facilities and a general change in the management of frail elderly patients in hospitals. ACKNOWLEDGEMENTS The authors thank the members of the staff in the Geriatric Evaluation and Management Unit for their enthusiasm and willingness during the study. The study was supported by the Norwegian Ministry of Health and Social Affairs and the Research Council of Norway. REFERENCES 1. Grimley EJ. 21st Century: Review: ageing and medicine. J Intern Med 2000; 247: Sager MA, Franke T, Inouye SK, et al. Functional outcomes of acute medical illness and hospitalization in older persons. Arch Intern Med 1996; 156: Steel K, Gertman PM, Crescenzi C, Anderson J. Iatrogenic illness on a general medical service at a university hospital. N Engl J Med 1981; 304: Gillick MR, Serrell NA, Gillick LS. Adverse consequences of hospitalization in the elderly. Soc Sci Med 1982; 16: Rubenstein LZ, Josephson K, Wieland GD, et al. Geriatric assessment on a subacute hospital ward. Clin Geriatr Med 1987; 3: Stuck AE, Siu AL, Wieland GD, Adams J, Rubenstein LZ. Comprehensive geriatric assessment: a meta-analysis of controlled trials. Lancet 1993; 342: 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: Asplund K, Gustafson Y, Jacobsson C, et al. Geriatric-based versus general wards for older acute medical patients: a randomized comparison of outcomes and use of resources. J Am Geriatr Soc 2000; 48: Counsell SR, Holder CM, Liebenauer LL, 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: 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: 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: Cohen HJ, Feussner JR, Weinberger M, et al. A controlled trial of inpatient and outpatient geriatric evaluation and management. N Engl J Med 2002; 346: Brymer CD, Kohm CA, Naglie G, et al. Do geriatric programs decrease long-term use of acute care beds? J Am Geriatr Soc 1995; 43: Indredavik B, Bakke F, Solberg R, Rokseth R, Haaheim LL, Holme I. Benefit of a stroke unit: a randomized controlled trial. Stroke 1991; 22: Saltvedt I, Mo ES, Fayers P, Kaasa S, Sletvold O. Reduced mortality in treating acutely sick, frail older patients in a geriatric evaluation and management unit. A prospective randomized trial. J Am Geriatr Soc 2002; 50: Aging Clin Exp Res, Vol. 16, No

7 I. Saltvedt, T. Saltnes, E-S. Opdahl Mo, et al. 16. Winograd CH, Gerety MB, Chung M, Goldstein MK, Dominguez F. Jr, Vallone R. Screening for frailty: criteria and predictors of outcomes. J Am Geriatr Soc 1991; 39: 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 NZ J Med 1991; 21: Nikolaus T, Specht-Leible N, Bach M, Oster P, Schlierf G. A randomized trial of comprehensive geriatric assessment and home intervention in the care of hospitalized patients. Age Ageing 1999; 28: Styrborn K. Early discharge planning for elderly patients in acute hospitals - an intervention study. Scand J Soc Med 1995; 23: Hansen FR, Poulsen H, Sorensen KH. A model of regular geriatric follow-up by home visits to selected patients discharged from a geriatric ward: a randomized controlled trial. Aging Clin Exp Res 1995; 7: Stewart S, Pearson S, Luke CG, Horowitz JD. Effects of homebased intervention on unplanned readmissions and out-of-hospital deaths. J Am Geriatr Soc 1998; 46: Naylor MD, Brooten D, Campbell R, et al. Comprehensive discharge planning and home follow-up of hospitalized elders: a randomized clinical trial. JAMA 1999; 281: Melin AL, Bygren LO. Efficacy of the rehabilitation of elderly primary health care patients after short-stay hospital treatment. Med Care 1992; 30: Hansen FR, Spedtsberg K, Schroll M. Geriatric follow-up by home visits after discharge from hospital: a randomized controlled trial. Age Ageing 1992; 21: Young J, Philp I. Future directions for geriatric medicine. Geriatricians must move with their patients into the community. BMJ 2000; 320: Boult C, Boult LB, Morishita L, Dowd B, Kane RL, Urdangarin CF. A randomized clinical trial of outpatient geriatric evaluation and management. J Am Geriatr Soc 2001; 49: Aging Clin Exp Res, Vol. 16, No. 4