Validation and Pattern Discovery in the Canadian Community Health Survey - Mental Health (CCHS-MH) Support Utilization

• Author / Creator

  Southcott, Jordan

• Mental illness is one of the most pressing medical challenges facing society. Although identifying gaps in mental-health support utilization is important for public health, this topic has not been widely explored in the literature. The latest Canadian Community Health Survey - Mental Health Component on mental-health support utilization was conducted by the Canadian government and sampled 24,788 Canadians. It collected information on twelve mental-health support utilization items and nine sociodemographic items, namely province of residence, residence in a metropolitan/non-metropolitan area, age, sex, marital status, visible minority status, immigrant status, highest level of attained education, and household income. However, this instrument has not been validated yet. Hence, this research aims to 1) probe the structural validity and reliability of the CCHS-MH instrument using exploratory factor analysis (EFA) and confirmatory factor analyses (CFA); 2) use clustering unsupervised machine learning algorithms to find patterns of mental-health support utilization by grouping participants based on their support utilization; and 3) compare and contrast these patterns using chi-square analyses to examine group differences in demographic characteristics. Findings show that the reliability (i.e., internal consistency) of the measure was adequate (α = .79). There is agreement among the EFA, CFA, and clustering analyses in revealing a 4-factor optimal model fit and in the nature of the factors: No Support, Social Support, and Professional Support were always relevant. The fourth factor, Mixed Support, which combines professional and social support systems, seems to yield the best fit, as reflected by the CFA. The final model yields 4 factors underlying mental-health support utilization: No Support, Social Support, Professional Support, and Mixed Support. The findings also show that Fuzzy C-Means clustering outperform the other two clustering algorithms employed (K-Means and Hierarchical Agglomerative Clustering). Post-hoc analyses found significant differential patterns of utilization in every demographic variable, except for visible minority status. Theoretical implications include support for the validation and reliability of a 4-factor model of the CCHS-MH support utilization and for the effectiveness of Fuzzy C-Means Clustering in finding patterns underlying large quantities of psychological data. Practical implications include more evidence for established patterns of support utilization observed in both the Canadian and global context as well as campaigns to encourage communities to talk openly about mental health, reverse biases in the field, and emphasize mental health in medical training.

• Subjects / Keywords

  • Psychology
  • Machine Learning
  • Public Health
  • Unsupervised Clustering

• Graduation date

  Spring 2021

• Type of Item

  Thesis

• Degree

  Master of Education

• DOI

  https://doi.org/10.7939/r3-0fet-me36

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.