A-227 A Story Told in Four Parts: Embedded Validity Indicators’ Effectiveness in Predicting Dot Counting Test Validity Classification

Abstract

Abstract Objective: Combining multiple embedded performance validity tests (PVTs) can produce similar classification accuracy to freestanding PVTs. However, there is a lack of research on the incremental predictive power of various combinations of PVTs. Thus, we compared eight embedded PVTs to assess those that best predict classification accuracy on the Dot Counting Test (DCT). Method: Sample included 225 patients (mean age=45.96; mean education=13.96; 56% female, 44% male; 35% White, 40% Black, 17% Hispanic, 5% Asian, 2% Other) undergoing neuropsychological evaluation that included multiple embedded PVTs, including: Brief Visuospatial Memory Test-Revised Recognition Discrimination (BVMT-R RD), Stroop Color and Word Test Word Reading T-Score (Stroop-T), Trail Making Test Part A T-Score (TMT-A), Rey Auditory Verbal Learning Test (RAVLT) Forced Choice, RAVLT Effort Score, Digit Span Age Corrected Scaled Score, Reliable Digit Span, and Letter Fluency T-Score. Patients were classified into valid/invalid groups based on four independent criterion PVTs. Results: A forward stepwise logistic regression was performed to predict DCT pass/fail using the aforementioned embedded PVTs as predictors. The model was achieved in three steps (p<.05); Step 1: BVMT-R RD (Classification Accuracy=87.6%; Nagelkerke R2 =.30), Step 2: BVMT-R RD + Stroop-T (Classification Accuracy=89.6%; Nagelkerke R2 =.44); Step 3: BVMT-R RD + Stroop-T + TMT-A (Classification Accuracy=90.1%; Nagelkerke R2 =.49). Conclusion: BVMT-R RD + Stroop-T + TMT-A reliably predicted the DCT pass/fail group. Thus, this combination of embedded PVTs may be reliable predictors of validity classification when time prohibits delivery of freestanding PVTs, such as the DCT.