Association for Public Policy Analysis & Management

This paper identifies a paradox that drives misunderstandings of the impact of Alzheimer’s on African American and white populations. Although research on the diagnosis of Alzheimer’s and other forms of dementia indicates a much higher incidence among African Americans than whites, evidence on racial disparities in mortality rates from autopsy records indicates the opposite. This paper tests two competing hypotheses for the divergence in the racial gaps in the recorded diagnoses of Alzheimer’s disease and related disorders vs official mortality statistics. The first hypothesis is that there is a pattern of misdiagnosis of African American patients with cognitive deficits perhaps due to lower quality medical care or lack of access to expensive modern screening technologies that would minimize misdiagnoses. The second hypothesis is that there is a statistical anomaly arising from the fact that relatively few deaths result in autopsies, the definitive assessment of the presence of Alzheimer’s disease. Using linked data on diagnosis rates and autopsies in Minnesota we can distinguish between the pure selection bias effects vs the misdiagnosis effects.