Objective: To compare prospectively the concordance between the diagnosis of dementia based on clinical criteria and using the DAT Inventory.
Design, setting, and participants: A prospective study of 81 consecutive patients referred to a Memory Clinic. Only patients for whom a definitive diagnosis of dementia was established after 8 to 20 months follow-up were retained in the study (n = 76).
Measurements: The sensitivity, specificity, positive and negative predictive values, and overall diagnostic accuracy of the DAT Inventory were calculated. Kappa values were also computed.
Results: Based on all patients (n = 76), sensitivity and specificity were 71% and 95%, respectively, with 98% positive prediction, 56% negative prediction, 78% overall accuracy, and kappa of 0.54. Of 21 cases not meeting NINCDS/ADRDA criteria for DAT, one patient with multi-infarct dementia was misclassified as DAT on the DAT Inventory. Of 55 DAT cases (NINCDS/ADRDA criteria), 16 patients, predominantly very mild or mixed cases, were classified as non-DAT on the DAT Inventory. When mixed, very mild, and borderline cases were excluded (remaining n = 54), DAT Inventory sensitivity increased to 94%, and specificity remained unchanged at 95%, with 97% positive and 91% negative prediction, 94% overall accuracy, and kappa of 0.88.
Conclusions: In general, scores above the designated cutoff point (> 14/20) on the DAT Inventory are consistent with a clinical diagnosis of DAT (NINCDS/ADRDA criteria). Concordance is best in cases of mild to moderate dementia (Clinical Dementia Rating 1-2). The Inventory is less discriminating as a differential diagnostic instrument in cases of very mild dementia, atypical presentations of DAT, or in cases of mixed pathology.