Reducing misclassification of mild cognitive impairment based on base rate information from the uniform data set

The current study aimed to define and validate the criteria for characterizing possible and probable cognitive deficits based on the psychometric approach using the Uniform data set Czech version (UDS-CZ 2.0) to reduce the rate of misdiagnosis. We computed the prevalence of low scores on the 14 subtests of UDS-CZ 2.0 in a normative sample of healthy older adults and validated criteria for possible and probable cognitive impairment on the sample of amnestic Mild Cognitive Impairment (MCI) patients. The misclassification rate of the validation sample using psychometrically derived criteria remained low: for classification as possible impairment, we found 66-76% correct classification in the clinical sample and only 2-8% false positives in the healthy control validation sample, similar results were obtained for probable cognitive impairment. Our findings offer a psychometric approach and a computational tool to minimize the misdiagnosis of mild cognitive impairment compared to traditional criteria for MCI.

Automated summary

The current study used a psychometric approach and the Uniform data set Czech version to define and validate criteria for characterizing possible and probable cognitive deficits. The findings showed that the psychometrically derived criteria had low misclassification rates and could accurately classify possible and probable cognitive impairment, reducing the misdiagnosis rates compared to traditional criteria for MCI.