Cross-validation of the Utility of Test of Memory Malingering (TOMM) Cut-offs in a Large Colombian Sample

This study sought to cross-validate the utility of Test of Memory Malingering (TOMM) cut-offs derived in a large Colombian sample with further exploration of the impact of age, education, and Mini-Mental Status Examination (MMSE) scores on TOMM trials. To address these aims, the study examines concordance rates of the TOMM scores by demographics and Hopkins Verbal Learning Test-Revised (HVLT-R) test performance. Additionally, item response theory analysis (IRT) focused on the interaction between demographic variables and the psychometric properties of the TOMM items. Data were collected from 1416 healthy controls (58.6% female; M age = 58.19; M education = 9 years) who completed the TOMM, MMSE, and HVLT-R as part of a comprehensive battery conducted in Spanish. Frequency analysis was used to assess for concordance rates of passing TOMMe10, Trial 1, and Trial 2 scores, further broken down by median split for MMSE, education, and age. Additionally, IRT is used to examine, in detail, the psychometric properties of the TOMM items and relationships to demographic variables. Validity classification differed across the TOMM trials, with 82.8% passing e10, 88.6% Trial 1, and 92.5% Trial 2. When dichotomized by a median split, passing rates across all TOMM trials were significantly different (p ≤ .001) for MMSE scores, education, and age with the largest discrepancy (~ 21%) observed on TOMMe10. IRT confirmed that all TOMM items are easy to answer correctly with Trial 2 items being appreciably easier (difficulty range = − 4.90 to − 1.52) than Trial 1 items (difficulty range = − 6.22 to − 3.50). Age, education, MMSE, and HVLT-R scores all were significantly related to latent trait scores on the TOMM (p