Fifteen years later: Enhancing the classification accuracy of the performance validity module of the Advanced Clinical Solutions

• Published in: Applied neuropsychology. Adult pp. 1 - 13
• Main Author: Erdodi, Laszlo A
• Format: Journal Article
• Language: English
• Published: United States 23.09.2024
• Subjects: performance validity; multivariate models; Erdodi Index; Advanced Clinical Solutions
• ISSN: 2327-9095; 2327-9109; 2327-9109
• DOI: 10.1080/23279095.2024.2406313

The study was designed to evaluate the performance validity module of Advanced Clinical Solutions (ACS) against external criterion measures and compare two alternative aggregation methods for its five components. The ACS was evaluated against psychometrically defined criterion groups in a sample of 93 outpatients with TBI. In addition to the default method, the component performance validity tests (PVTs) were either dichotomized along a single cutoff (VI-ACS) or recoded to capture various (EI-ACS). The standard ACS model correctly classified 75-83% of the sample. The alternative aggregation methods produced superior overall correct classification: 80-91% (VI-ACS) and 86-91% (EI-ACS). Mild TBI was associated with higher failure rates than moderate/severe TBI. Failing just one of the five ACS components resulted in a 3- to 8-fold increase in the likelihood of failing criterion PVTs. Results support the use of the standard PVT module for ACS: it is an effective measure of performance validity that is robust to moderate-to-severe TBI. Post-publication research on individual ACS components and methodological advances in PVT research provide an opportunity to enhance the overall classification accuracy of the ACS model. Passing stringent multivariate PVT cutoffs does not indicate valid performance.