Objective: To define a sensitive and specific index, based on nutritional indicators, predicting mortality in nursing home patients.
Design: A prospective cohort study.
Setting: A nursing home.
Measurements: Anthropometric and laboratory nutritional indicators were assessed in 104 nursing home elderly residents (20 males, 84 females). Patients were aged 60 years and older and had relatively good somatic health and nutritional status. Mortality data were collected over an 18-month follow-up period. Discriminant analysis was used to compute an adimensional measure (index) that could predict mortality with the highest sensitivity and specificity.
Results: Mortality was 0.20/year. A nonlinear, "U-shaped" relationship of cholesterol levels with mortality was found. A prognostic index based on cholesterol, lymphocyte count, mid-arm circumference, hemoglobin, age, and gender was obtained. Higher values of this index were associated with progressively higher risks at 12 months and 18 months. Sensitivity and specificity calculated on 18-month mortality were 80% and 78.4%, respectively.
Conclusions: This index, based on simple measures, can be a useful tool in the evaluation of health status of the elderly living in nursing homes.