Prioritizing determinants of cognitive function in healthy middle-aged and older adults: insights from a machine learning regression approach in the Canadian longitudinal study on aging

Sarah Singh; Shiran Zhong; Kem Rogers; Vladimir Hachinski; Stephanie Frisbee

doi:10.3389/fpubh.2023.1290064

Prioritizing determinants of cognitive function in healthy middle-aged and older adults: insights from a machine learning regression approach in the Canadian longitudinal study on aging

Front Public Health. 2023 Dec 22:11:1290064. doi: 10.3389/fpubh.2023.1290064. eCollection 2023.

Authors

Sarah Singh¹, Shiran Zhong², Kem Rogers³, Vladimir Hachinski^{1

4}, Stephanie Frisbee⁵

Affiliations

¹ Robarts Research Institute, University of Western Ontario, London, ON, Canada.
² Department of Geography, University of Western Ontario, London, ON, Canada.
³ Department of Anatomy and Cell Biology, Schulich School of Medicine and Dentistry, University of Western Ontario, London, ON, Canada.
⁴ Department of Clinical Neurological Sciences, and Epidemiology and Biostatistics, Schulich School of Medicine and Dentistry, University of Western Ontario, London, ON, Canada.
⁵ Department of Pathology and Laboratory Medicine, and Epidemiology and Biostatistics, Schulich School of Medicine and Dentistry, University of Western Ontario, London, ON, Canada.

Abstract

Introduction: The preservation of healthy cognitive function is a crucial step toward reducing the growing burden of cognitive decline and impairment. Our study aims to identify the characteristics of an individual that play the greatest roles in determining healthy cognitive function in mid to late life.

Methods: Data on the characteristics of an individual that influence their health, also known as determinants of health, were extracted from the baseline cohort of the Canadian Longitudinal Study of Aging (2015). Cognitive function was a normalized latent construct score summarizing eight cognitive tests administered as a neuropsychological battery by CLSA staff. A higher cognitive function score indicated better functioning. A penalized regression model was used to select and order determinants based on their strength of association with cognitive function. Forty determinants (40) were entered into the model including demographic and socioeconomic factors, lifestyle and health behaviors, clinical measures, chronic diseases, mental health status, social support and the living environment.

Results: The study sample consisted mainly of White, married, men and women aged 45-64 years residing in urban Canada. Mean overall cognitive function score for the study sample was 99.5, with scores ranging from 36.6 to 169.2 (lowest to highest cognitive function). Thirty-five (35) determinants were retained in the final model as significantly associated with healthy cognitive functioning. The determinants demonstrating the strongest associations with healthy cognitive function, were race, immigrant status, nutritional risk, community belongingness, and satisfaction with life. The determinants demonstrating the weakest associations with healthy cognitive function, were physical activity, greenness and neighborhood deprivation.

Conclusion: Greater prioritization and integration of demographic and socioeconomic factors and lifestyle and health behaviors, such greater access to healthy foods and enhancing aid programs for low-income and immigrant families, into future health interventions and policies can produce the greatest gains in preserving healthy cognitive function in mid to late life.

Keywords: CLSA; cognitive function; dementia prevention; determinants of health; machine learning.

Publication types

Research Support, Non-U.S. Gov't

MeSH terms

Aged
Aging*
Canada
Cognition*
Female
Health Status
Humans
Longitudinal Studies
Machine Learning
Male
Middle Aged

Grants and funding

The author(s) declare financial support was received for the research, authorship, and/or publication of this article. Funding for this research was provided by the Weston Family Foundation through the Weston Brain Institute (grant TR202092). Funding bodies did not have a role in design of the study and collection, analysis, and interpretation of data or in writing the manuscript.