Profiles of Cognitive Functioning in a Population-Based Sample of Centenarians Using Factor Mixture Analysis

Abstract

Background/Study Context: The goal of the study was to identify and characterize latent profiles (clusters) of cognitive functioning in centenarians and the psychometric properties of cognitive measures within them. Methods: Data were collected from cross-sectional, population-based sample of 244 centenarians (aged 98 to 108, 15.8% men, 20.5% African American, 38.0% community-dwelling) from 44 counties in northern Georgia participating in the Georgia Centenarian Study (2001–2008). Measures included the Mini-Mental State Examination (MMSE), Severe Impairment Battery (SIB), Wechsler Adult Intelligence Scale-III Similarities subtest (WAIS), Hand Tapping, Behavioral Dyscontrol Scale (BDS), Controlled Oral Word Association Test (COWAT), and Fuld Object Memory Evaluation (FOME). The Global Deterioration Rating Scale (GDRS) was used to independently evaluate criterion-related validity for distinguishing cognitively normal and impaired groups. Relevant covariates included directly assessed functional status for basic and instrumental activities of daily living (DAFS), race, gender, educational attainment, Geriatric Depression Scale Short Form (GDS), and vision and hearing problems. Results: Results suggest two distinct classes of cognitive performance in this centenarian sample. Approximately one third of the centenarians show a pattern of markedly lower cognitive performance on most measures. Group membership is independently well predicted (area under the curve [AUC] = .83) by GDRS scores (sensitivity 67.7%, specificity 82.4%). Membership in the lower cognitive performance group was more likely for individuals who were older, African Americans, had more depressive symptoms, lower plasma folate, carriers of the apolipoprotein E (APOE) ε4 allele, facility residents, and individuals who died in the 2 years following interview. Conclusions: In a population expected to have high prevalence of dementia, latent subtypes can be distinguished via factor mixture analysis that provide normative values for cognitive functioning. The present study allows estimates for normative cognitive performance in this age group.