The WAIS Coding Subtest As an Embedded Performance Validity Measure in Cases of Traumatic Brain Injury
MetadataShow full metadata
Most recent estimates have indicated that around 69 million people worldwide sustain traumatic brain injury (TBI) every year. Often following TBI, a neuropsychological assessment is conducted in order to determine the existence and the extent of the cognitive impairment. A good effort is required from the examinee to obtain valid results on neuropsychological tests. However, extensive literature has demonstrated that not everybody performs to the best of their ability and that some may feign or exaggerate their cognitive deficits. As a result, performance validity testing (PVT) research is considered to be crucial and is currently one of the most dominant themes in the field of neuropsychology. The aim of the current study was to explore the efficacy of a number of embedded validity indicators derived from the WAIS Coding subtest in a mixed traumatic brain injury sample (n = 650), namely: presence of a Coding Error; the Number of Coding Errors; the Coding ACSS; a derived Coding Combination score (i.e. the sum of adding the Coding Error and Coding ACSS). The study also examined if these Coding embedded validity indicators are valid in the range of traumatic brain injury severities or if they are biased against those with more significant brain injuries. Results of logistic regression analyses revealed that all Coding embedded effort variables were significantly predictive of low effort. Furthermore, findings indicated that the injury severity affects performance on all Coding embedded validity indicators. Results showed an inverse relationship between presence of a Coding Error and injury severity such that individuals with more severe brain injuries were less likely to make Coding Errors than those affected by mild traumatic brain injuries. The Coding ACSS and the Coding Combination score were sensitive to TBI severity, such that those with more severe brain injuries obtained lower scores on the ACSS and higher scores on the Coding Combination score, indicating a more impaired performance. Based on these findings, a ROC analysis was employed to identify recommended cut-off scores for each severity group with an aim to minimise false positive errors. Furthermore, a positive predictive value (PPV) and negative predictive value (NPV) were reported for 20%, 30%, 40%, and 50% base rates. Overall, results demonstrated excellent specificity rates, but low and variable sensitivity rates. Employing a Coding Error cut-off score of > 0 for all severity groups resulted in 97% specificity and 38% sensitivity. Based on the analysis of the Coding ACSS, in cases of mTBI a cut-score of ≤ 5 resulted in 91% specificity and 56% sensitivity; for cases of Moderate TBI employing a cut-score of ≤ 4 resulted in 94% specificity and 33% sensitivity; in cases of Severe and Very Severe TBI employing a cut-score of ≤ 3 resulted in 94% specificity and 57% sensitivity. Results showed that the Coding Combination score achieved higher sensitivity than the Coding ACSS for the same level of specificity. Based on the findings in this study, in cases of mild TBI, a Coding Combination cut-score of > 14 resulted in 91% specificity and 63% sensitivity; in cases of Moderate TBI, employing a cut-score of > 15 resulted in 90% specificity and 40% sensitivity; in cases of Severe and Very Severe TBI, employing a cut-score of > 16 resulted in 93% specificity and 70% sensitivity. Overall, the findings in this study show that the Coding embedded validity indicators show good potential. However, due to low and variable sensitivity levels, they should be used in a combination with other PVTs in order to determine performance validity during neuropsychological evaluation and are likely to be used as complementary or confirmatory of other, more sensitive, standalone PVTs.