A-78: Updated Normative Data and Cognitive Indices for Use in Parkinson's Disease and Older Adults

Abstract

Objective: The Parkinson’s Progression Markers Initiative cognitive battery (PPMI-CB) is used to understand cognitive sequelae in Parkinson’s disease (PD). Despite its prominence, analytical approaches and application of norms is psychometrically lacking. Our aim is to apply robust norming (i.e., omitting those with concern and impairment at baseline and followup) and latent structure procedures to identify optimal and consistent standardized scores and indices for the PPMI-CB. Method/Data Selection: The PPMI is a longitudinal cohort study of individuals with PD, prodromal PD, and controls. Tests include Letter-Number Sequencing, Symbol Digit Modalities, Judgement of Line Orientation, Category fluency, Hopkins Verbal Learning Test, Trail Making Test, Letter fluency, and Boston Naming Test. We derived norms from unimpaired controls combined with prodromals (UCPs) using quantile smoothing and multiple fractional polynomials, then conducted factor analyses with measurement invariance. Results/Data Synthesis: Robust selection resulted in 1,266 UCPs, of which 48 were controls. Controls were younger (64vs.67, p<.05), more educated (17vs.16, p<.05; all had some college), and more racially diverse (83%vs.96% White; p<.05) than UCPs. Test scores did not differ among groups, save for BNT being lower in controls (55vs.57, p<.05), but normative modeling functioned better for UCPs. We then identified three-factor, hierarchical, invariant models for indices. Conclusions: In sum, many so-called controls in PPMI may have meaningful cognitive concerns or impairment, which we addressed by combining with conservative-robust sampling from at-risk for PD, but cognitively healthy, individuals. In addition to updated norms for several tests, we produced recommended indices for those using the full PPMI-CB, and provide methods to optimize cognitive analyses within PPMI.