Development and validation of an electronic Symbol-Digit Modalities Test for remote monitoring of people with multiple sclerosis

Michelangelo Dini,Giulia Gamberini,Marta Tacchini,Angela Boschetti,Alessandro Gradassi,Luca Chiveri,Mariaemma Rodegher,Giancarlo Comi,Letizia Leocani
DOI: https://doi.org/10.1111/ene.16454
2024-09-04
Abstract:Background: Computerized cognitive tests may extend the reach of cognitive screening and monitoring to those with mobility issues or living in remote areas. Moreover, it could enable frequent and autonomous remote cognitive assessments in people with multiple sclerosis (pwMS) on account of its reduced economic and organizational costs. This may further improve our understanding of longitudinal trends and significantly improve the standard of care for pwMS living in remote areas or with mobility limitations. We aimed to evaluate the psychometric properties of an electronic Symbol-Digit Modalities Test (eSDMT) designed to allow pwMS to perform a rapid cognitive assessment independently from home using their own PC/laptop. Methods: Sixty-two participants underwent a neuropsychological evaluation, and then performed the eSDMT in the clinic. Forty-two participants also repeated the eSDMT at home. We assessed concurrent validity (eSDMT vs. oral SDMT), test-retest reliability (in the clinic vs. at home), discriminant validity (pwMS with/without cognitive impairment), and other psychometric characteristics of the eSDMT (effect of age, sex, and education on test scores). Results: We observed good-to-excellent concurrent validity (r ≥ 0.84, all p < 0.0001) and test-retest reliability (intraclass correlation coefficients [ICCs]>0.87, all p < 0.0001). Discriminant validity was excellent (area under the curves [AUCs] >0.84, all p < 0.0001). eSDMT scores were only slightly influenced by demographic characteristics (all R2 < 0.200). Conclusions: We provided evidence which supports the use of our eSDMT as a feasible, valid, and reliable remote assessment of cognitive function in pwMS. Future studies will investigate long-term reliability and predictive power.
What problem does this paper attempt to address?