PREDICTING SHORT-TERM SURVIVAL FOR PATIENTS WITH ADVANCED ALZHEIMERS-DISEASE

Objective: The purpose of this study was to develop a statistical model for predicting short term survival in patients with dementia of the Alzheimer type (DAT). Design: A prospective cohort study. Setting: Three 25-bed intermediate medical care units using a structured approach to patient care management including palliative care options and patients from a second, traditional long-term care setting. Participants: Of 104 patients with advanced DAT monitored for 34 months, 68 patients (97% white male) who had at least one fever episode were included in the model development phase. Data from 71 additional DAT patients with at least one fever episode were used to test the statistical model. Main Outcome Measures: Six-month survival following a fever episode. Results: Older age and higher severity of DAT at the time of the fever episode, palliative care, and hospital admission for long-term care within 6 months prior to the fever were found to be positively associated with likelihood of mortality within 6 months of the fever onset. Adjusted odds ratios for each of these variables were statistically significant. The model performed well in subsequent testing on an independent sample of patients. Conclusion: Results provide a formula which can be used to predict likelihood of dying within 6 months following onset of a fever in DAT patients. This statistical prediction is recommended for use in combination with clinical judgment to certify DAT patients for Medicare hospice coverage.