Leveraging machine learning to improve trauma-related health outcomes

Severe and multiple traumas can lead to post-traumatic stress disorder (PTSD) which causes distress and difficulty in many areas of life. Military members, veterans, and public safety personnel such as police officers, firefighters, and rescue workers are often exposed to traumatic events on-the-job (seeing natural disasters, human violence, or death). One severe subtype of PTSD is PTSD with dissociation, where the person feels as if one’s body is not their own, or as if the outside world is not real. Currently, we have limited knowledge about PTSD with dissociation and whether other subtypes exist. Advanced computer models will be used to analyze data from over 800 patients (military members, veterans, public safety personnel, civilians) who received treatment for PTSD. Findings may uncover additional PTSD subtypes, refine what we know about dissociation, and predict what factors lead to better functioning and recovery. These models will then be used to predict which subtype of PTSD an individual has and identify which type of treatment will most likely lead to recovery.

Faculty Supervisor:

Margaret McKinnon

Student:

Anna Park

Partner:

Homewood Research Institute

Discipline:

Psychology

Sector:

University:

McMaster University

Program: