We report data on the validity and responsiveness (i.e. sensitivity to change) of assessment instruments including Goal Attainment Scaling (GAS), at a single site in a multicentre trial of the experimental therapeutic agent linopirdine. Fifteen people (11 women) were evaluated. GAS yielded a mean 3.7 goals per patient (range 2-6). The mean gain in the GAS scores, 2.7 +/- 16.4, was compared to changes in the Alzheimer's Disease Assessment Scale-Cognitive Section, the Global Deterioration Scale, Clinical Global Impression and the Mini-Mental State Exam. GAS had the largest relative efficiency (0.47) when compared to the standard. GAS also had the largest effect size (0.61). The data suggest that an individualized approach may have merit as an outcome measure and as a means to better understanding treatment effects. Qualitative analysis revealed consistent goal setting in self-care, behaviour, cognition and leisure, suggesting that these areas should routinely be evaluated.