Predicting PTSD development with early post-trauma assessments: a proof-of-concept for a concise tree-based...

Area of Science:

• Psychiatry
• Computational Neuroscience
• Clinical Psychology

Background:

• Traumatic events affect a majority globally, with significant risk of developing posttraumatic stress disorder (PTSD).
• Effective PTSD prevention and treatment strategies are critical due to its prevalence and impact.

Purpose of the Study:

• To develop a concise, tree-based adaptive test using the Classification and Regression Trees (CART) framework for predicting PTSD development.
• To assess the feasibility of using a minimal set of questions for early PTSD risk identification.

Main Methods:

• Adult trauma survivors (n=143) from a longitudinal neuroimaging study were analyzed.
• Data included demographic, trauma-related, and clinical symptom features collected within two weeks post-trauma.
• CART model performance was compared against Random Forest (RF) and Gradient Boosting (GB) machine learning algorithms.

Main Results:

• The CART model, using only three questions, achieved predictive performance comparable to RF and GB models.
• CART accuracy was 0.641 and AUC was 0.663, demonstrating high efficiency.
• The model's conciseness suggests practical utility for early PTSD detection.

Conclusions:

• The CART framework offers a streamlined and efficient method for predicting PTSD onset in trauma survivors.
• Further validation and refinement are needed to enhance predictive accuracy and establish broader clinical utility for early intervention.