Background: : Although the 14-item Starkstein Apathy Scale (SAS) is recommended to screen for and measure the severity of apathetic symptoms in Parkinson disease (PD), the psychometric attributes of this scale have not yet been fully evaluated.
Objective: : The authors examined the reliability, factor structure, and discriminant validity of the SAS in 194 nondemented patients with early untreated PD.
Design: : Cross-sectional multicenter population-based study from Western and Southern Norway.
Measurements: : Standardized rating scales for parkinsonism and neuropsychiatric symptoms.
Results: : The SAS showed fair internal consistency (Cronbach's α = 0.69) and exploratory factor analysis identified two factors: the first factor (24.2% of the variance) represented cognitive-behavioral aspects of apathy (items 1, 2, and 4-8; Cronbach's α = 0.74) and the second factor (15.0% of the variance) a general apathy dimension (items 3 and 9-14; Cronbach's α = 0.52). The correlation between these two factors was low (Spearman's r = 0.19, N = 194, p = 0.008), indicating clinically distinct dimensions, but both factor scores were strongly related to the total SAS score (Spearman's r > 0.6, N = 194, p < 0.0005). Item 3 (insight or self-reflection) showed a negative item-total correlation, and removing this item raised the Cronbach's α of the second factor to 0.70, but did not substantially alter the other results. Both the total score and factor scores of SAS showed fair discriminant validity.
Conclusions: : Although the SAS showed fairly good psychometric properties and the exploratory factor analysis suggested a two-factor solution, the results with this PD sample indicate that item 3 is ambiguous and should be considered removed from the scale.