Differential prognosis and utilization patterns among clinical subgroups of hospitalized geriatric patients.

While screening elderly inpatients on acute Veterans Administration (VA) hospital wards for a special geriatric program, we prospectively classified all patients age 65 and over, who had been hospitalized at least a week, into five clinical subgroups using specific diagnostic, prognostic, and functional criteria. These five subgroups were "geriatric evaluation unit (GEU) candidate", "severely demented", "medical", "terminal", and "independent". Medical record data from the initial admission and a full year of follow-up were collected from random samples of each subgroup and of nonscreened patients who had been hospitalized for less than a week. Analysis revealed that each subgroup had a distinctive pattern of survival, living location, and use of institutional services during the follow-up period. For one major subgroup ("GEU candidate"), a specific intervention (the GEU) has proved very effective in reducing mortality, increasing patient functioning, improving placement, and decreasing use of institutional services. Moreover, there are specific treatment and intervention strategies appropriate for each of the other subgroups (e.g., hospital-based home care, hospice, respite, and day treatment programs), although these services are not universally available nor clearly proved effective. The process of identifying patient subgroups illustrated in this study may be useful in needs assessment, in planning new intervention programs for frail elderly patients, and for identifying appropriate patients for these programs.