Cluster Analysis of Patients with Traumatic Physical Injuries Seeking Outpatient Behavioral Health Services in a Level-1 Trauma Hospital

Abstract Introduction: The use of clustering analysis within healthcare has evidenced promise in grouping complex medical cases, with increasing attention being given in grouping mental health symptoms and outcomes. Limited information is available on clustering for mental health symptoms in patients who have experienced a recent traumatic physical injury. Similarly, it is unknown if different patient subtypes have diverse social risk factors or if treatment response may differ. Methodology: Patients who experienced a physical traumatic injury (N=67) were referred for outpatient behavioral health services through a Level-1 trauma hospital. K-means clustering was used to examine potential groupings based on trauma symptoms, depression scores, adverse childhood experiences, and physical, mental, and general health quality of life. Results: Three clusters emerged based on symptom severity across measures. Behavioral health treatment at 6-weeks resulted in significant decreases in trauma symptoms for the highest t(12)= 2.61, p = 0.01, d = 0.72 and the mid-range severity groups t(7) = 2.12, p = 0.03, d = 0.75. Significant decreases in depression symptoms were noted for the severe symptom group t(12)= 3.44, p= 0.002, d= 0.95, but not the mid-range group. Conclusion: Groupings based on symptom severity corresponded with social risk and needs such as food, clothing, housing, medical care, transportation, social activities, witnessing violence, and history of being violent. Patient clusters emerged based on symptom severity. Patients with the highest symptom severity also had the highest social risks and needs. Implications suggest that resources can be allocated to address other needs that may exacerbate symptom presentation.