Multilevel modeling of cognitive function at individual- and area-levels: A profile of middle-aged and older adults in Canada

Authors: Matthew Quick*, University of Waterloo
Topics: Spatial Analysis & Modeling, Quantitative Methods
Keywords: multilevel, disease mapping, spatial, cognitive function
Session Type: Paper
Day: 4/3/2019
Start / End Time: 2:35 PM / 4:15 PM
Room: Marshall South, Marriott, Mezzanine Level
Presentation File: No File Uploaded


Past research has shown that neighbourhood characteristics influence cognitive function, however, few studies account for the multilevel structure of individuals nested in neighbourhoods and examine both individual- and group-level risk factors. This study applies a Bayesian multilevel modelling approach to explore the geographical variation of cognitive function in a nationally representative sample of middle-aged and older Canadian adults. Data measuring executive function and memory, two dimensions of cognitive function, were retrieved from the Canadian Longitudinal Study on Aging. Each participant was nested within one forward sortation area. Three multilevel models with different assumptions regarding the relationships between individuals and their geographical contexts were compared. Accounting for individual risk factors, including age, gender, and education, the best fitting models show that forward sortation areas explained approximately five and three percent of the variation of executive function and memory, respectively. Maps of the area-level random effects parameters highlight locations with clustering of high (or low) memory and executive function scores. The advantages and disadvantages of using multilevel models to analyze sparse spatial health data are discussed.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login