Machine Learning Models for Differentiating Alcohol and Substance Use Disorders Using Hemogram Parameters
A retrospective study explored whether routine blood count parameters analyzed with machine learning could help distinguish between alcohol use disorder, substance use disorder, and healthy controls in 228 participants. Random Forest models achieved 81.6% accuracy in classification, identifying monocyte count, basophil count, and RDW-CV as significant differentiating parameters, though sensitivity varied across disorder types.