Relationship Between the Digital Assessment of Cognition and Paper and Pencil Neuropsychological Tests

The concordance between k-mean and paper/pencil classification was 89%. When a linear, stepwise regression analysis (DAC memory index = dependent variable; five paper/pencil indices = independent variables) was conducted, the paper/pencil episodic memory index entered first (beta=0.538, p <0.001) followed by the paper/pencil language index (beta=0.302, p <0.001). When a similar regression analysis for the DAC executive index was conducted, the paper/pencil working memory index entered first (beta=0.503, p <0.001), followed by the paper/pencil attention index (beta=0.288, p <0.001), followed by the paper/pencil memory index (beta=0.163, p <0.017). A multinomial logistic regression analysis (NC = reference group) found that the paper/pencil attention (Wald=6.15, p <0.013) and working memory indices (Wald=7.88, p <0.005) classified patients into the dMCI group; only the paper/pencil episodic memory index (Wald=15.08, p <0.001) classified patients into the aMCI group; both the paper/pencil working memory (Wald=11.02, p <0.001) and episodic memory indices (Wald=5.43, p <0.020) classified patients into the mixed MCI group; and the paper/pencil working memory (Wald=9.11, p <0.001), language (Wald=3.72, p <0.050) and episodic memory indices (Wald=9.49, p <0.001) classified patients into the dementia group CONCLUSION: These data provide evidence for both the construct and criterion validity of the DAC. The parsimony and ease associated with DAC administration/scoring suggests that the DAC can be deployed as a frontline assessment to flag emergent cognitive impairment.