BACKGROUND: Due to the prevalence and impact of cognitive impairment on quality of life, regular screening for changes in cognition is the recommended standard care for PwMS. The shift from PASAT3 to SDMT as the recommended screening test poses challenges, especially for long-term analyses of patient cohorts, necessitating the development of an algorithm converting PASAT3 into SDMT scores. METHODS: Concurrent SDMT and PASAT3 scores of 459 PwMS and 147 HC from four clinical centers were analyzed retrospectively together with the demographic variables age, sex and education. Twelve linear (mixed) models were built and the dataset was stratified, split and bootstrapped for model training, evaluation and validation. The best performing model was refit to the full dataset to derive a practical formula estimating SDMT from PASAT3 scores. RESULTS: Demographic variables influence SDMT scores differently than PASAT3 scores. The final model includes a random effect for clinical center and estimates main effects of PASAT3 score (ptextless0.001), patient/control group (ptextless0.001), age (ptextless0.05) and education (ptextless0.01). Furthermore, significant interactions are found between PASAT3 score and age (ptextless0.001), and between sex and patient/control group (ptextless0.05). Predicted SDMT scores correlate significantly with observed SDMT scores (R = 0.77, ptextless0.001) and are well calibrated across the distribution of the data. CONCLUSION: A conversion of PASAT3 into SDMT scores can be performed based on our formula, accounting for demographic and test-performance interactions. Our approach not only addresses a methodological challenge by aligning past and present cognitive assessments, but also provides deeper insights into the demographic influences on cognitive performance in PwMS.