Optimizing PTSD Diagnosis: A Mathematical Approach to Simplifying Complexity to Six Symptoms Without Compromising Diagnostic Accuracy

Diagnostic criteria for posttraumatic stress disorder (PTSD) in the Diagnostic and Statistical Manual of Mental Disorders (DSM) have evolved substantially since their introduction, now encompassing 20 symptoms across four clusters. This complexity creates substantial heterogeneity and challenges for both clinical practice and research. Although various simplification approaches have been proposed, none have utilized mathematical optimization to identify symptom combinations that maximize diagnostic reliability while reducing complexity. Using PTSD Checklist for DSM-5 assessment data from 52,609 veterans, we identified optimal six-symptom combinations for a probable PTSD diagnosis that minimized missed diagnoses compared to DSM-5-TR criteria. We compared two approaches: hierarchical, which required at least one symptom from each cluster in the four symptoms needed for diagnosis, and non-hierarchical, which required any four symptoms irrespective of cluster membership. Results revealed that the non-hierarchical approach achieved remarkable diagnostic accuracy with 99.30% sensitivity, missing only 0.70% of cases (N=353) and outperforming the hierarchical approach (94.01% sensitivity). These findings demonstrate that the assessment of probable PTSD can be effectively simplified to six symptoms while still maintaining reliability, thereby challenging the necessity of cluster-based criteria. Our analysis code is publicly available to facilitate validation across different populations, potentially informing future diagnostic systems that could streamline clinical assessment without compromising accuracy.