A Population Health Informatics and Predictive Analytics Approach to the Risk Assessment of Traumatic Brain Injury and Associated Mental Health Outcomes

Abstract

Traumatic brain injury (TBI) is a public health challenge causing considerable long-term disability and mortality. Despite its burden, our understanding of TBI is limited by critical gaps along the injury continuum. Specifically, the spatial and socioeconomic factors influencing TBI occurrence remain underexplored. In the post-injury phase, research on mental health service use among individuals with TBI in Canada is limited, leaving unanswered questions about how these services are utilized. Furthermore, few studies have investigated the risk of new-onset psychiatric disorders (NPDs) and long-term depression in patients with TBI.

The goal of this dissertation, composed of five complementary studies, was to address these knowledge gaps by providing an integrated understanding of the TBI continuum, from injury risk to associated mental health outcomes. Leveraging large-scale administrative and clinical data, this dissertation employed an interdisciplinary approach, utilizing a range of analytical methods to examine TBI from population to patient levels. In the pre-injury phase, Studies 1 and 2 revealed that TBI risk is spatially clustered and specific dimensions of neighborhood deprivation are associated with varying injury mechanisms. In the post-injury phase, Study 3 showed that individuals with TBI are 60% more likely to use mental health services, with disparities in utilization linked to age, sex, and education. Building on the increased rates of mental health service use, Study 4 identified distinct clinical phenotypes of TBI, each associated with varying risks of NPDs. Finally, Study 5 demonstrated that depression trajectories vary widely over a 10-year period, and that incorporating individual-level variability improves the prediction of depression scores. Together, these findings highlight the heterogeneity of TBI, from geographically clustered injury risk influenced by socioeconomic factors to variable psychiatric outcomes over time.

This research advances theoretical understanding of injury etiology and post-injury outcomes by integrating diverse perspectives across the TBI continuum. The findings inform public health prevention strategies targeting spatial and socioeconomic risk factors. The clinical implications reinforce the need for risk stratification and early intervention for post-injury mental health outcomes. Methodologically, this work demonstrates the value of applying a population informatics and predictive analytic approach to large-scale health data for comprehensive TBI research.