Differential Validity in the Written, Oral, and Electronic SDMT in Multiple Sclerosis

Background: Cognitive decline in multiple sclerosis (MS) is commonly assessed using the Symbol Digit Modalities Test (SDMT). However, the written (wSDMT), oral (oSDMT), and electronic (eSDMT) versions may display differing psychometric properties, despite often being used interchangeably. Objective: The objective of the study is to compare the wSDMT, oSDMT, and eSDMT, including inter-test reliability, and to assess demographic and motor disability influences on test performance. Methods: In this within-subjects study, 85 participants with MS completed all three SDMT versions. Statistical analyses included assessment of intraclass correlation coefficients (ICCs) and sequential multiple regression modeling. Results: The three SDMT versions demonstrated good inter-test reliability (ICC = 0.77) and similar mean scores (range: 53.5-56.2). Scores on all three test versions declined with age (p < 0.05), but the decline was significantly steeper in the eSDMT. Education correlated with eSDMT but not wSDMT or oSDMT scores. Paced Auditory Serial Addition Task (PASAT) scores predicted performance across all versions, while 9HPT times predicted only wSDMT and eSDMT scores (p < 0.01). Conclusion: The wSDMT and eSDMT show signs of motor disability influence, while the eSDMT displays the greatest sensitivity to the influences of age and education. This differential construct validity necessitates SDMT version-specific normative data and motor-adjustments for accurate cognitive assessment in MS.