Background: Because primary prevention trials will require large samples and modest treatment effects are expected, the use of standard clinician-administered, clinic-based measures are unlikely to be feasible. There is a need for proxy-administered outcome measures. The goal of the Alzheimer's Disease Cooperative Study (ADCS) Prevention Instrument Project was to conduct a simulated Alzheimer disease prevention trial in 650 nondemented elderly (Ferris et al, 2006). This involved comparison of data acquisition from both home and clinic and the use of both informant-ratings and self-ratings. Important outcomes included clinical global impressions of change (CGIC) as indicators of clinically meaningful change. Such ratings provide verification that the effects of a medication as measured on rating scales are readily observable and clinically meaningful. One objective was to develop self-rated and study partner-rated CGICs optimized for nondemented elderly or people with very early Alzheimer disease. An important consideration was whether global assessments are specific and sensitive measures of change during a prevention trial.
Methods: A self-administered CGIC and a study partner-rated CGIC were developed to be used either in the clinic or at home. Using 3-month follow-up data, we determined its reliability and validity with 317 subject-partner pairs. We compared subject-ratings with partner-ratings, clinic-based with home-based ratings, and ratings based on severity as determined by the Clinical Dementia Rating scale.
Results: There were no differences between clinic and home ratings. Overall, 24% of subjects rated themselves, and 10% of study partners rated the subjects, as minimally to markedly improved. Subjects and partners agreed to within 1 point of their ratings 83% of the time on the 7-point scale. There were weak correlations, generally <0.20, with change scores of selected clinical rating scales.
Discussion: The CGICs behaved as expected, showing no overall change over 3 months, no difference between administrations at home compared with clinics, and concurrent validity. Some subjects tended to rate themselves better than their partners rated them. These analyses show the potential for using home-based CGICs which can be completed with minimal supervision and allow assessments of potential preventative interventions.