Comparison of performance on the CERAD neuropsychological battery of Hispanic patients and cognitively normal controls at two sites.

Based on 88 non-demented controls and 140 AD patients enrolled at the University of Southern California, and 102 Spanish-speaking dementia patients at the University of Pennsylvania, we developed norms, stratified by disease severity, for the Spanish-language CERAD Neuropsychological Battery, and examined the impact of demographic characteristics. Adequacy of information and level of performance differed between sites, indicating hazards that occur in the absence of crosssite planning, including characteristics of 'retrofitted' data (e.g., missing information), small samples, additional countries of origin, and multiple dementing disorders. Nevertheless, controls generally performed better than even very mild AD patients, and performance decreased with increase in severity of disease. In the absence of larger, more representative samples, norms for Spanish-speaking patients and control subjects may be site-specific, and should be treated with caution. [ABSTRACT FROM AUTHOR]