Discovering Clusters of Support Utilization in the Canadian Community Health Survey–Mental Health

Mental illness is one of the most pressing medical challenges facing society. Thus, identifying gaps in mental-health support-seeking is crucial for public health. This exploratory study aims to reveal gaps and patterns in mental-healthcare support-utilization by employing unsupervised machine learning in the Canadian Community Health Survey–Mental Health that measures support-seeking for mental-health issues from 24,788 Canadians. Of the clustering methods compared (K-means, hierarchical agglomerative, and Fuzzy C-means), Fuzzy C-means clustering yielded the best model fit for the data and revealed four clusters: No Support , Social Support , Professional Support , and Mixed Support evaluated based on existing theory. Findings reveal differential effects by all variables, except for the variable concerning whether a respondent was white or a visible minority.