Introduction

Top Abstract Introduction Subjects and methods Results Discussion References

Improving the ability of health care systems to respond to the demands of older people is among the greatest challenges of our time. Most elderly people, even with considerable disability, prefer to stay at home,¹ and hospitals are shortening lengths of stay. Community care has therefore acquired greater relevance.^2-6 Responsibility among various health professionals for care management of older people living in the community, however, remains poorly defined, and patients falling between primary and secondary health care and social services are at risk of being forgotten.⁷ A possible solution may be the integration of medical and social services in a continuum of care with case management programmes. ^{7 8} None the less, the cost effectiveness of this approach remains untested.

We conducted a randomised trial to evaluate the impact of such an integrated programme on admissions to institutions, use and costs of health services, and functional decline among frail elderly people living in the community.

	Subjects and methods

Top Abstract Introduction Subjects and methods Results Discussion References

During the early 1990s, to comply with the national health plan, the health agency of Rovereto, a town in northern Italy of nearly 35 000, created a broad array of health services for older people. These ranged from a hospital geriatric evaluation unit to a skilled nursing facility and a home health agency. However, no coordination of these components nor integration with social services in the municipality was considered.

Subjects
In 1995 we identified all people aged 65 and over who were recipients of home health services or home assistance programmes (n=224). Usually, patients were receiving these services because of multiple geriatric conditions (for example, dementia, immobility, incontinence, and stroke deficits), but the evaluation preceding care planning was not based on a comprehensive geriatric assessment. Among the total number 24 declined to participate: six were not interested in the project; nine had been advised against it by relatives; and nine had been advised against it by their general practitioner. The remaining subjects were randomly stratified by age and sex according to a computer generated list. One hundred subjects (control group) received primary and community care with the conventional and fragmented organisation of servicesthat is, general practitioner's regular ambulatory and home visits, nursing and social services, home aids, and meals on wheels. Another hundred subjects (intervention group) received case management and care planning by the community geriatric evaluation unit and general practitioners. All the services considered necessary were provided in an integrated fashion after a formal agreement between the municipality and the local health agency. Twenty one of 24 general practitioners agreed to participate in the trial and to be involved in care planning, meetings, and emergency situations. Informed consent was obtained from all patients. The study was approved and monitored by the steering committee of the National Research Council's aging project and the local state authority (Provincia Autonoma of Trento).

View larger version (24K): [in this window] [in a new window]   Fig 1.   Integrated network of services designed for intervention

Outcomes and expenditure
Main outcomes included admission to an institution, use and related costs of health services, and physical and cognitive function. Information on outcomes such as admission and use of health services was collected every 2 months by a research assistant unaware of patients' assignments. In the event of admission to hospital or a nursing home patients remained in the study. Vital status was obtained from general practitioners and confirmed by the National Death Registry.

Statistical analysis
Quantitative parameters are presented as means (SD). The impact of intervention on functional outcomes was evaluated by analysis of covariance with follow up measures adjusted for baseline values. The impact of the intervention on time to admission to hospital or a nursing home was tested by comparing the survival curves obtained with the Kaplan-Meier method. Differences between curves were evaluated with the log rank test. A P<0.05 level was chosen for significance. Statistical analysis was performed with SAS and SPSS software.

	Results

Top Abstract Introduction Subjects and methods Results Discussion References

There were no significant differences at baseline in the intervention and control groups across several functional and clinical variables (table 1). No difference in 1 year mortality was observed; 12 subjects in the intervention group died compared with 13 in the control group (hazard ratio 0.99; 95% confidence interval 0.89 to 1.09). Figure 2 shows changes in functional outcomes. In the control group all functional indices deteriorated (activities of daily living 13.0%; instrumental activities of daily living 6.9%; mental status 9.4%; depression 11.8%). In the intervention group less consistent changes were observed (5.1%; unchanged; 3.8%; 4.0%, respectively). Differences between intervention and control groups were all significant (table 2). Also the adjusted mean number of medications was reduced in the intervention group (4.7 (0.2) v 5.4 (0.2); P<0.05).

The benefit obtained in the intervention group was achieved without increases in use of health services. Use of home support (intervention group 120 (20) v 154 (29) hours/patient/year), nursing care (13 (3) v 12 (3) hours/patient/year), and meals on wheels (54 (12) v 39 (10) meals/patient/year) was not significantly different, although the lack of differences might be due to insufficient sample size. Conversely, more home visits by general practitioners were needed in the control group (10.2 (1.1) v 13.1 (0.8); P=0.04).

View larger version (19K): [in this window] [in a new window]   Fig 2.   Percentage change in functional outcomes after 1 year of follow up

View larger version (17K): [in this window] [in a new window]   Fig 3.   Proportion of subjects admitted to hospital or nursing home in intervention and control groups

Thirty six subjects in the intervention group and 51 in the control group were admitted at least once to acute hospital (P<0.05). The hazard ratio was 0.74 (0.56 to 0.97), while that for visits to an emergency room was 0.64 (0.48 to 0.85) compared with control group (table 3). Furthermore, patients in the control group had a trend toward a higher rate of admission to nursing homes, although this was not significant. The cumulative number of days per year spent in either nursing home (1087 v 2121) or acute hospital (894 v 1376) was reduced by up to half in the intervention group.

Results of survival analyses on the basis of time to first admission to hospital or nursing home are shown in figure 3. Thirty eight subjects in the intervention group and 58 subjects in the control group were admitted at least once (hazard ratio 0.69; 0.53 to 0.91). Even when subjects in the intervention group entered hospital or nursing homes, they did so later (and less often) than control subjects (P<0.003).

Finally, we calculated total per capita health care costs over the follow up period. The intervention group accounted for 23% less than the control group. The overall saving, after addition of salaries of case managers, was estimated at around £1125 per person per year. Apart from reductions in community health services costs (£744 v £919; 19%), intervention group savings resulted mainly from substantial decreases in nursing home (£644 v £1244; 48%) and hospital expenses (£1763 v £2688; 34%).

	Discussion

Top Abstract Introduction Subjects and methods Results Discussion References

Our study shows that an integrated community care programme implemented by an interdisciplinary team including a general practitioner and a case manager reduced the risk of hospital admission and length of stay in either hospital or nursing home. Despite a similar use of supportive home care resources, subjects in the intervention group showed less physical and cognitive decline; total health care costs per capita were also reduced.

Some features of our study may explain these results. Firstly, we intensively trained case managers, providing them with case management skills and geriatric assessment technology.⁹ This determined the ability to design care plans and coordinate all available agencies, thus assuring integrated care.¹⁵

Secondly, in our model the community geriatric evaluation unit represented the gatekeeper to health services. This provided a unique community based setting for the referral of patients, regardless of specific needs. Importantly, the role of case managers was to support and integrate the activity of general practitioners who, by law, retained full responsibility of the patient. This is consistent with previous recommendations⁷ and has been suggested by the public policy committee of the American Geriatrics Society.¹⁶

Finally, the close collaboration between case managers, community geriatric evaluation unit, and general practitioners was critical to the success of the intervention; this may determine the effectiveness of any community based programmes.¹⁷

Although randomised, our study could not be performed with the rigid criteria of a clinical trial. Because of the nature of the intervention (that is, a change in the provision of care) all the professionals concerned were aware of the assignment of patients to either group. Also, as a consequence of the informed consent, patients and physicians were aware of the ongoing project. Case managers, however, performed the assessment simply as a part of their routine activities; both patients and professionals remained blind about the outcomes under study and the length of follow up. This greatly limited the risk of introducing a bias. Furthermore, differences in functional outcomes collected by case managers were consistent with the objective outcomes based on medical record review (that is, number of medications and admissions to nursing homes and hospital). In this respect, our study is no different from many others of elderly patients.¹⁸ Also, general practitioners who followed both control and intervention participants may have introduced a contamination bias, though the ratio of physicians to patients (1:5) rendered a directional bias unlikely. Moreover, the higher number of visits in the control group was inconsistent with less intensive treatment in these patients. We believe that a more critical consideration is that the benefits of an integrated care approach can be achieved only in circumstances where all the parties concerned are sufficiently motivated.