The present postmarketing surveillance (PMS) study is the first large scale systematic and prospective clinical trial of pharmacotherapeutic intervention in advanced stages of dementia. Within a validation program this study aimed at demonstrating the sensitivity of the D-Scale of change (DS-C) for measuring ADL-function. Efficacy of treatment with the NMDA-antagonist memantine was investigated in 531 patients with advanced dementia employing a parallel group design that stratified patient cohorts by severity according to GDS stages (Reisberg, 1992). Efficacy was determined on two independent levels: by the assessment of the physicians' Clinical Global Impression of Change (CGI-C) at the end of a 6-week observation period, and by the assessment of change in elementary ADL-functions by the caregivers using the D-Scale-of-Change. With the D-Scale-of-Change the caregivers can assess a change in broad functional items, i.e. cognitive and motor functions and also elementary functions of daily life. The effect size of this improvement increased constantly during the observation period. Even in patients of GDS stage 7 an improvement could be measured. These results were also seen by the physicians, who recorded an overall clinical improvement in 75.5% of the patients after 6 weeks. Tolerability evaluations resulted in the ratings "very well" by 59.5% or "well" by 35.0% of the patients. No serious adverse drug reactions occurred. A correlation analysis demonstrated a high congruency of both assessments. Furthermore, the observed time course of these improvements paralleled with the time course of symptomatic benefit by effects of memantine that had been repeatedly demonstrated in randomised, double-blind, placebo-controlled studies in mild to moderate dementia. Together with the evaluation of the scale-properties of the D-Scale for assessment of severity the D-Scale and the D-Scale-of-Change can be regarded as validated.