Development and validation of the Uniform Data Set (v3.0) executive function composite score (UDS3-EF)

Introduction Cognitive composite scores offer a means of precisely measuring executive functioning (EF). Methods We developed the Uniform Data Set v3.0 EF composite score (UDS3-EF) in 3507 controls from the National Alzheimer's Coordinating Center dataset using item-response theory and applied nonlinear and linear demographic adjustments. The UDS3-EF was validated with other neuropsychological tests and brain magnetic resonance imaging from independent research cohorts using linear models. Results Final model fit was good-to-excellent: comparative fit index = 0.99; root mean squared error of approximation = 0.057. UDS3-EF scores differed across validation cohorts (controls > mild cognitive impairment > Alzheimer's disease-dementia approximate to behavioral variant frontotemporal dementia; P < 0.001). The UDS3-EF correlated most strongly with other EF tests (beta s = 0.50 to 0.85, Ps < 0.001) and more with frontal, parietal, and temporal lobe gray matter volumes (beta s = 0.18 to 0.33, Ps <= 0.004) than occipital gray matter (beta = 0.12, P = 0.04). The total sample needed to detect a 40% reduction in UDS3-EF change (n = 286) was approximate to 40% of the next best measure (F-words; n = 714). Conclusions The UDS3-EF is well suited to quantify EF in research and clinical trials and offers psychometric and practical advantages over its component tests.