@article{ref1,
title="Introducing the ImPACT-5: an empirically derived multivariate validity composite",
journal="Journal of head trauma rehabilitation",
year="2020",
author="Erdodi, Laszlo and Korcsog, Kassandra and Considine, Ciaran and Casey, Joseph and Scoboria, Alan and Abeare, Christopher",
volume="ePub",
number="ePub",
pages="ePub-ePub",
abstract="OBJECTIVE: To create novel Immediate Post-Concussion and Cognitive Testing (ImPACT)-based embedded validity indicators (EVIs) and to compare the classification accuracy to 4 existing EVIImPACT.   METHOD: The ImPACT was administered to 82 male varsity football players during preseason baseline cognitive testing. The classification accuracy of existing EVIImPACT was compared with a newly developed index (ImPACT-5A and B). The ImPACT-5A represents the number of cutoffs failed on the 5 ImPACT composite scores at a liberal cutoff (0.85 specificity); ImPACT-5B is the sum of failures on conservative cutoffs (≥0.90 specificity).   RESULTS: ImPACT-5A ≥1 was sensitive (0.81), but not specific (0.49) to invalid performance, consistent with EVIImPACT developed by independent researchers (0.68 sensitivity at 0.73-0.75 specificity). Conversely, ImPACT-5B ≥3 was highly specific (0.98), but insensitive (0.22), similar to Default EVIImPACT (0.04 sensitivity at 1.00 specificity). ImPACT-5A ≥3 or ImPACT-5B ≥2 met forensic standards of specificity (0.91-0.93) at 0.33 to 0.37 sensitivity. Also, the ImPACT-5s had the strongest linear relationship with clinically meaningful levels of invalid performance of existing EVIImPACT.   CONCLUSIONS: The ImPACT-5s were superior to the standard EVIImPACT and comparable to existing aftermarket EVIImPACT, with the flexibility to optimize the detection model for either sensitivity or specificity. The wide range of ImPACT-5 cutoffs allows for a more nuanced clinical interpretation.<p /> <p>Language: en</p>",
language="en",
issn="0885-9701",
doi="10.1097/HTR.0000000000000576",
url="http://dx.doi.org/10.1097/HTR.0000000000000576"
}