RESUMO
INTRODUCTION: The accuracy of commonly used screening tests for Alzheimer's disease (AD) has not been directly compared to those that could be more appropriate for lower schooling. OBJECTIVE: To compare the diagnostic accuracy of usual screening tests for AD with instruments that might be more appropriate for lower schooling among older adults with low or no literacy. METHODS: The study included a clinical sample of 117 elderly outpatients from a Geriatric Clinic classified as literate controls (n = 39), illiterate controls (n = 30), literate AD (n = 30) and illiterate AD (n = 18). The tests were compared as follows: Black and White versus Colored Figure Memory Test; Clock Drawing Test versus Clock Reading Test; Verbal Fluency (VF) animal versus grocery category; CERAD Constructional Praxis versus Stick Design Test. RESULTS: The means of literate and illiterate controls did not differ in the Black and White Figure Memory Test (immediate recall), Colored Figure Memory Test (delayed recall), Clock Reading Test and VF animals and grocery categories. The means of the clinical groups (controls versus AD), in the 2 schooling levels, differed significantly in most of the tests, except for the CERAD Constructive Praxis and the Stick Design Test. Diagnostic accuracy was not significantly different between the compared tests. CONCLUSION: Commonly used screening tests for AD were as accurate as those expected to overcome the education bias in a sample of older adults with lower or no education.