Objective: To evaluate the extent to which mortality data, which is often used to track secular trends for specific diseases, underestimates the prevalence of dementia.
Design: Retrospective analysis of existing data.
Setting: Department of Mental Health inpatient facilities in South Carolina.
Subjects: Inpatients at Department of Mental Health facilities who were listed in the South Carolina Statewide Alzheimer's Disease and Related Disorders Registry and who died between 1988 and 1990 (n = 450).
Main outcome measures: Sensitivity and specificity of dementia diagnoses on death certificates compared to medical record diagnoses for inpatients with a pre-mortem dementia diagnosis.
Results: Twenty-three percent of death certificates contained any dementia diagnosis (104/450). The sensitivity of death certificates varied by type of dementia (28 percent for Alzheimer's disease; 8 percent for multi-infarct dementia) as well as by race, sex, and age.
Conclusions: Mortality statistics substantially underestimate the prevalence of dementing illnesses and do not fully represent the public health burden of dementia.