Prospective prediction of PTSD diagnosis in a nationally representative sample using machine learning

Area of Science:

• Psychiatry
• Psychology
• Data Science

Background:

• Numerous pre-traumatic, peri-traumatic, and post-traumatic factors influence PTSD risk.
• Interactions among these risk factors and their contribution to PTSD development are not fully understood.
• Existing research often relies on univariate analyses, limiting the understanding of complex risk factor dynamics.

Purpose of the Study:

• To examine the impact of psychological, ecological, and sociodemographic factors on PTSD onset.
• To utilize machine learning to determine the relative contributions of various risk factors.
• To develop a comprehensive risk assessment profile for PTSD following adverse events.

Main Methods:

• Utilized data from the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC) Waves 1 and 2.
• Combined individual-level risk factors with concurrent community-level data.
• Employed machine learning for feature selection and classification to predict new onset of PTSD.

Main Results:

• Machine learning models achieved 89.7% to 95.6% accuracy in predicting new PTSD onset.
• Prior diagnoses of Borderline Personality Disorder, Major Depressive Disorder, or Anxiety Disorder were strong predictors.
• Distal risk factors, like prior psychiatric diagnoses, showed greater relative risk than proximal factors.

Conclusions:

• Machine learning effectively integrates numerous PTSD risk factors, accounting for complex interactions and collinearity.
• Findings support targeted mobilization of emergency mental health resources.
• The study informs the creation of more accurate PTSD risk assessment profiles.