Geospatial and machine learning methods in traumatic brain injury prediction and prevention in Nova Scotia

Traumatic brain injury (TBI) is a leading cause of death and disability. One area of injury prevention research is to use geographic information systems to identify at-risk neighbourhoods for targeted intervention. However, few studies have been conducted to investigate the geographical distribution of TBI in Canada. The proposed project is based on the linked trauma registry data, emergency medical service data and Canadian census data, which contains about 4500 TBI patients from year 2002-2018. TBI incidence will be mapped by age, sex, and major mechanisms of injury (e.g., falls or motor vehicle collisions) over years to identify high risk geographic areas of TBI. Bayesian spatial regression models will be applied to investigate to what extent the spatial pattern of TBI risk is attributable to neighborhood-level factors. Further, machine learning methods will be used to predict in-hospital mortality and hospital length of stay in TBI patients for preventing future most severe outcomes among TBI patients.

Faculty Supervisor:

Cindy Feng;Syed Sibte Raza Abidi;David Clarke

Student:

Partner:

Super GeoAI Technology Inc.

Discipline:

Mathematics

Sector:

Information and cultural industries; Professional, scientific and technical services

University:

Dalhousie University

Program: