Objective: To determine which variables best predict prognosis--time to nursing home admission (NHA) and death--in patients with dementia.
Design: Survival analysis employing the Cox proportional hazards model with the use of risk variables pertaining to dementia severity and its rate of progression and caregiver functioning.
Setting: Patients and their caregivers participating in a controlled intervention study of training for caregivers in home management of dementia.
Participants: Patients with mild Diagnostic and Statistical Manual of Mental Disorders, Third Edition--defined dementia (N = 91; 68 patients with Alzheimer-type dementia, 20 with vascular dementia, and three with other types) and their caregivers.
Follow-up: All subjects had repeated assessments in the first year to determine rates of change and thereafter annually to determine the date of NHA and/or death. RISK VARIABLES: (1) Caregiver training; (2) dementia severity at index assessment; (3) caregiver stress, neuroticism, and socialization; (4) changes in patients and caregivers during the first 12 months; and (5) patient characteristics.
Results: By 5 years' follow-up, 76% of patients had entered a nursing home and 42% had died. Dementia severity and rate of deterioration ("how far" and "how fast") and caregiver psychological morbidity significantly influenced rates of NHA and death. Training of caregivers was significantly associated with delayed NHA and reduced mortality. Greater patient age, non-Alzheimer's dementia, and, unexpectedly, greater caregiver psychological morbidity were associated with shorter survival to death.
Conclusions: Both severity ("how far") and rate of deterioration ("how fast") influence time to NHA and death. Caregiver training may have important ameliorating effects on the prognosis of dementia.