In 1985 1.3 million people were residing in nursing homes with costs of institutional care exceeding the direct medical costs of home care by as much as $11,000 per patient-year. Yet the factors contributing to institutionalization of dementia patients are not clearly understood.' The present study will undertake four related objectives. The first is to determine the cumulative incidence of institutionalization in population- based cohorts of dementia patients and their matched controls. The second is to determine risk for institutionalization of dementia patients associated with a) demographic factors, b) co-morbid illness c) non- nursing home health' care utilization and d) factors specific to the dementing illness. The third Objective is to determine the frequency with which institutionalization is initiated by family members versus health care professionals. The fourth objective is to distinguish between the presence of deteriorated functional status and caregivers' abilities to tolerate these impairments. Each of these aims has relevance for public policy and care innovations for dementia patients. The present study will capitalize on the Mayo Clinic Alzheimer's Disease Patient Registry (ADPR) and emulate the existing scheme of the ADPR in having retrospective and prospective components. The retrospective sample is population-based and includes 662 dementia patients with onset of dementia occurring between 1975-1984 and 662 age- and sex-matched controls. The medical records of cases and controls in 1975-1979 and 1980-1984 retrospective cohorts will be abstracted to determine the date and place of nursing home placement. Because of the medical records-linkage system at the Mayo Clinic these records are complete for all Olmsted County, MN patients. Rates of nursing home placement within the dementia sample will be calculated. Hazards models for placement associated with various demographic factors and an index of co-morbid illness will be calculated. The prospective sample is community-based and currently includes 536 matched pairs of dementia patients and controls. Data regarding residence will be collected during ADPR longitudinal follow-up procedures. Participants will be seen in their residence at follow-up so that place of residence will be determined by direct observation. Disease specific variables including cognitive status, functional status and behavioral disturbance will be integrated with co- morbidity and demographic data to develop hazards models for institutionalization in the prospective sample. Understanding which features of dementing disease processes are most directly related to institutionalization is a necessary precursor to intervention strategies seeking to delay or prevent the need for institutionalized care.
