Derivation of embedded symptom validity indicators for the Brief Symptom Inventory-18.

Area of Science:

• Neuropsychology
• Psychometrics
• Traumatic Brain Injury (TBI) research

Background:

• Symptom validity assessment is crucial for evaluating psychological effects following traumatic brain injury (TBI).
• Many existing personality inventories include embedded symptom validity scales (SVTs).
• Research is exploring the development of SVTs within brief symptom and self-report measures.

Purpose of the Study:

• To develop and evaluate embedded symptom validity indicators using the Brief Symptom Inventory-18 (BSI-18).
• To identify specific BSI-18 items predictive of questionable validity (QV) in TBI patients.

Main Methods:

• Secondary analysis of TRACK-TBI data from 1,694 participants (1,515 TBI, 179 orthopedic injury).
• Development of two novel SVTs: BSI-V (using Bayesian model averaging) and BSI GSI (Global Severity Index cutoff).
• Receiver operating characteristic (ROC) analysis to compare the classification accuracy of BSI-V and BSI GSI against independent SVTs.

Main Results:

• Four BSI-18 items were identified as optimal predictors of questionable validity (QV).
• Both BSI-V and BSI GSI demonstrated excellent classification accuracy, with an area under the ROC curve (AUC) of 0.95.
• Established cutoffs for BSI-V (≥5) and BSI GSI (≥64 T) achieved 83% sensitivity and 90% specificity.

Conclusions:

• Preliminary evidence supports the utility of novel embedded SVTs (BSI-V and BSI GSI) within the BSI-18.
• Both developed SVTs show comparable and strong performance in identifying questionable validity.
• Further independent cross-validation is recommended to confirm classification accuracy and refine cutoffs for clinical practice.