standard deviations of the scores of the first follow-up assessment of participants without dementia for the first three follow-ups (n = 573). The resulting composite score was validated against the CDR score by calculating the correlation between the composite score and CDR of the samples of the first (total n 5 1,099, with dementia n 5 235) and second (total n 5 683, with dementia n 5 190) followups and by calculating the correlation of the changes in both scores between the first and second follow-up. Participants without demenita were assigned a CDR score of 0.The Pearson correlation coefficient between the composite score and CDR was 0.88 at the first follow-up, 0.83 at the second follow-up, and 0.62 for in scores between the first and second follow-ups. All correlation coefficients were significant at Po.001. The composite score interquartile range (IQR) was ( À 0.66-0.32) for participants with a CDR of 0, ( À 2.58 to À 0.90) for a CDR of 1, ( À 5.50 to À 2.56) for a CDR of 2, and ( À 9.61 to À 5.23) for a CDR of 3.The results indicate high correlations between the Composite Cognitive and ADL Functioning Score and the CDR score at one point in time. Changes over time in the composite score also correlated highly with change in CDR. These observations support the validity of the composite score that was developed. The fact that the IQRs of the composite scores per CDR score overlapped little shows that the composite score could accurately separate CDR categories. This supports the clinical relevance and provides information for the clinical interpretation of the Composite Cognitive and ADL Functioning Score. Based on the validation by CDR scores in this study population, the Composite Cognitive and ADL Functioning Score is easy to use and interpret and clinically relevant, and because it reduces the need for clinical judgment of each individual required in comparable composite measures, it is highly useful for large population-based studies.