The role of biopsychosocial factors in classifying pain intensity across various chronic pain conditions.

Chronic pain is a multifaceted condition shaped by biological, psychological, and social factors, which challenges its diagnosis and treatment. While prior research largely focused on sensory profiles to distinguish pain mechanisms, this study adopted a biopsychosocial perspective with the key addition of experimental pain paradigms to identify factors explaining variance in chronic pain severity across diverse etiologies. The dataset comprised 101 individuals with chronic pain and 63 pain-free controls. Participants were classified into three groups (no pain, mild-to-moderate pain, and severe pain) using three models: a sensory-profile-based model including quantitative sensory testing (QST model), a biopsychosocial model incorporating QST alongside psychological and social variables (biopsychosocial model), and a model using the same biopsychosocial data but excluding QST measures (noQST model). The QST model achieved an accuracy of 0.51 and an F1-score of 0.48. In contrast, the biopsychosocial model performed best, with both accuracy and F1-score of 0.71, while the noQST model reached 0.60 for both metrics. Key predictors in the best-performing model included quality of life, loss of sensation, depressive mood, pain catastrophizing, anxiety, fatigue, and general health. Overall, biopsychosocial factors enhanced chronic pain severity classification beyond sensory profiles alone across heterogeneous pain etiologies.