Gaining precision on the Alzheimer's Disease Assessment Scale-cognitive: a comparison of item response theory-based scores and total scores.

BACKGROUND The Alzheimer's Disease Assessment Scale-cognitive (ADAS-cog) is a commonly used measure for assessing cognitive dysfunction in patients with Alzheimer's disease (AD). The measure has 11 subscales, each of which captures an important aspect of cognitive dysfunction in AD. Traditional scoring of the ADAS-cog involves adding up the scores from the subscales without regarding their varying difficulty or their strength of relationship to AD-associated cognitive dysfunction. The present article analyzes problems associated with this approach and offers solutions for gaining measurement precision by modeling how the subscales function. METHODS We analyzed data collected at the Baylor College of Medicine Alzheimer's Disease and Memory Disorders Clinic from 1240 patients diagnosed with varying degrees of dementia. Item response theory was used to determine the relationship between total scores on the ADAS-cog and the underlying level of cognitive dysfunction reflected by the scores. RESULTS Results revealed that each total score corresponded to a spectrum of cognitive dysfunction, indicating that total scores were relatively imprecise indicators of underlying cognitive dysfunction. Furthermore, it was common for two individuals with the same total score to have significantly different degrees of cognitive dysfunction. CONCLUSIONS These findings suggest that item response theory scoring of the ADAS-cog may measure cognitive dysfunction more precisely than a total score method.