For the NP measures, a multivariate analysis of variance (MANOVA) was carried out, with Group and Gender as main effects, and average Z-scores for the 9 domains as dependent variables. In addition, a Group × Gender ANOVA was performed for the average Z-score across all domains. Finally, for each subject, a global clinical impairment score was generated as follows: 1) Each domain Z-score was converted to a percentile score. Next, a clinical impairment score of 0, or no impairment, was assigned to domain Z-scores falling above the 15th percentile, a score of 1, or moderately impaired, was assigned to domain Z-scores falling at or below the 15th and above the 5th percentile, and a score of 2, or severely impaired, was assigned to domain Z-scores falling at or below the 5th percentile. The domain clinical impairment scores (0, 1, or 2) were then summed across domains to yield the GCIS, with greater GCIS scores indicating more severe impairment. The cutoff points for the clinical impairment scores were designed to make the GCIS sensitive to clinically relevant impairment. A nonparametric Mann-Whitney U test was performed in order to determine whether NAC and TNAD differed on the GCIS.
