Comparing embedded performance validity indicators within the WAIS-IV Letter-Number sequencing subtest to Reliable Digit Span among adults referred for evaluation of attention-deficit/hyperactivity disorder
John-Christopher A. FinleyVioleta J. RodriguezBrian M. CernyFini ChangJulia M. BrooksGabriel P. OvsiewDevin M. UlrichZachary J. ReschJason R. Soblea Department of Psychiatry and Behavioral Sciences,Northwestern University Feinberg School of Medicine,Chicago,IL,USAb Department of Psychiatry,University of Illinois Chicago College of Medicine,Chicago,IL,USAc Department of Psychology,University of Illinois,Urbana-Champaign,Champaign,IL,USAd Department of Psychology,Illinois Institute of Technology Chicago,IL,USAe Department of Psychology,University of Illinois Chicago College of Medicine,Chicago,IL,USAf Department of Neurology,University of Illinois Chicago College of Medicine,Chicago,IL,USA
DOI: https://doi.org/10.1080/13854046.2024.2315738
2024-02-15
The Clinical Neuropsychologist
Abstract:Objectives: This study investigated the Wechsler Adult Intelligence Scale-Fourth Edition Letter-Number Sequencing (LNS) subtest as an embedded performance validity indicator among adults undergoing an attention-deficit/hyperactivity disorder (ADHD) evaluation, and its potential incremental value over Reliable Digit Span (RDS). Method: This cross-sectional study comprised 543 adults who underwent neuropsychological evaluation for ADHD. Patients were divided into valid ( n = 480) and invalid ( n = 63) groups based on multiple criterion performance validity tests. Results: LNS total raw scores, age-corrected scaled scores, and age- and education-corrected T-scores demonstrated excellent classification accuracy (area under the curve of .84, .83, and .82, respectively). The optimal cutoff for LNS raw score (≤16), age-corrected scaled score (≤7), and age- and education-corrected T-score (≤36) yielded .51 sensitivity and .94 specificity. Slightly lower sensitivity (.40) and higher specificity (.98) was associated with a more conservative T-score cutoff of ≤33. Multivariate models incorporating both LNS and RDS improved classification accuracy (area under the curve of .86), and LNS scores explained a significant but modest proportion of variance in validity status above and beyond RDS. Chaining LNS T-score of ≤33 with RDS cutoff of ≤7 increased sensitivity to .69 while maintaining ≥.90 specificity. Conclusions: Findings provide preliminary evidence for the criterion and construct validity of LNS as an embedded validity indicator in ADHD evaluations. Practitioners are encouraged to use LNS T-score cutoff of ≤33 or ≤36 to assess the validity of obtained test data. Employing either of these LNS cutoffs with RDS may enhance the detection of invalid performance.
psychology, clinical,clinical neurology