Related Experiment Video
Updated: May 22, 2025
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
Optimizing selectivity of the Cerebellar Cognitive Affective Syndrome Scale by use of correction formulas, and
Andreas Thieme1, Kerstin Rubarth2, Raquel van der Veen3
1Department of Neurology and Center for Translational Neuro- and Behavioral Sciences (C-TNBS), Essen University Hospital, University of Duisburg-Essen, Hufelandstraße 55, 45147, Essen, Germany. andreas.thieme@uk-essen.de.
Background:
Cerebellar disease may result in Cerebellar Cognitive Affective Syndrome (CCAS). The CCAS-Scale, designed to screen for CCAS, has been validated in English Hoche (Brain 141:248-270, 2018) and adapted to other languages.
Methods:
Here, the German CCAS-Scale Thieme (Neurol Res Pract 2:39, 2020) was validated in 209 patients with cerebellar disorders and 232 healthy controls. Correction formulas for the outcome parameters [failed test items (range: 1-10) and sum raw score (range: 0-120)] were developed, controlling for age, education, and sex effects. Diagnostic accuracy and reliability were assessed.
Results:
Correction formulas improved selectivity in controls, reducing false positives (failed items: 40%; sum score: 13% vs. original method Hoche (Brain 141:248-270, 2018): 67%), while maintaining moderate sensitivity (failed items: 69%; sum score: 48% vs. original method Hoche (Brain 141:248-270, 2018): 87%). Word fluency tests differentiated best between patients and controls, while other items did not. Internal consistency (α = 0.71) was acceptable. Removal of word fluency tests worsened it. Retest and interrater reliability were high [intraclass correlation coefficients (ICC): 0.77-0.95]. However, these ICCs yielded a large minimal detectable change (MDC; 2.2-2.4 failed items, 9.5-11.4 raw score points) in patients, limiting the use of the CCAS-Scale in follow-up examinations.
Conclusion:
The correction formulas improved diagnostic accuracy of the CCAS-Scale, particularly for the sum raw score. Therefore, we recommend using the corrected sum raw score for evaluation instead of the uncorrected number of failed items, proposed originally Hoche (Brain 141:248-270, 2018). Some test items, however, did not differentiate well between patients and controls and MDCs were large, highlighting the need for refined CCAS assessment instruments as progression or treatment outcomes.
