Age corrections and dementia classification accuracy.

Area of Science:

• Neuropsychology
• Biostatistics
• Gerontology

Background:

• Diagnostic classification accuracy in neuropsychological testing can be influenced by demographic factors such as age and education.
• Standardization of test scores using demographic corrections aims to reduce potential biases.
• Age is a known risk factor for dementia, complicating age-based demographic corrections.

Purpose of the Study:

• To investigate whether demographic corrections universally improve diagnostic classification accuracy.
• To evaluate the impact of age-based demographic corrections on dementia risk simulation and classification.
• To compare the sensitivity and specificity of raw versus demographically corrected test scores.

Main Methods:

• Simulated dementia risk based on demographic variables, including age.
• Analysis of classification accuracy using both raw and demographically corrected test scores.
• Comparison of score sensitivity and specificity at specific percentiles (16th and 9th) in simulated and clinical data.

Main Results:

• Demographic corrections did not universally enhance diagnostic classification accuracy.
• Simulated increased dementia risk based on age reduced classification accuracy for corrected scores compared to raw scores.
• In clinical data with weak age-dementia association, corrected and raw scores showed similar overall accuracy.
• Raw scores generally exhibited higher sensitivity, while demographically corrected scores demonstrated superior specificity.

Conclusions:

• Demographic corrections, particularly for age, may not always improve diagnostic accuracy and can potentially hinder it.
• The utility of demographic corrections depends on the specific demographic variable and its association with the diagnosis.
• While corrected scores may offer improved specificity, raw scores might be more sensitive for detecting cognitive impairment.