Subscale Validation of the Neuropsychiatric Inventory Questionnaire: Comparison of Alzheimer's Disease Neuroimaging Initiative and National Alzheimer's Coordinating Center Cohorts

Abstract

Neuropsychiatric symptoms are prevalent in mild cognitive impairment (MCI) and Alzheimer disease (AD) and commonly measured using the Neuropsychiatric Inventory (NPI). Based on existing exploratory literature, we report preliminary validation of three NPI Questionnaire (NPI-Q-10) subscales that measure clinically meaningful symptom clusters. Cross-sectional results for three subscales (NPI-Q-4-Frontal, NPI-Q-4-Agitation/Aggression, NPI-Q-3-Mood) in amnestic MCI and AD dementia cases from the National Alzheimer's Coordinating Center (NACC) and Alzheimer's Disease Neuroimaging Initiative (ADNI) databases were analyzed using confirmatory unrotated principal component analysis. ADNI contributed 103 MCI, 90 MCI converters, and 112 AD dementia cases, whereas NACC contributed 1,042 MCI, 763 MCI converters, and 3,048 AD dementia cases. NACC had higher baseline mean age (75.7 versus 74.6), and more impaired mean scores (at month 24) on Mini-Mental State Exam (19.5 versus 22.4) and NPI-Q-10 (5.0 versus 4.3), and all NPI-Q subscales than ADNI. Medians were not different between cohorts for NPI-Q-4-Agitation/Aggression, and NPI-Q-3-Mood, however. Each item on all scales/subscales contributed variance in principal component analysis Pareto plots. All items in Factor (F) 1 for each scale/subscale projected in a positive direction on biplots (revealing coherence), whereas F2 and F3 items showed more spatial separation (revealing independence). There were remarkable similarities between cohorts for factor loadings and spatial patterns of item projections, although factor item identities varied somewhat, especially beyond F1. The similar pattern of results across two cohorts support validity of these subscales, which are worthy of further psychometric evaluation in MCI and AD patients and preliminary application in clinical settings.