Agitation severity and emergency department outcomes in substance-related behavioral emergencies: a retrospective cohort study

BACKGROUND: Substance-related emergency department (ED) visits increasingly present with agitation and behavioral instability rather than discrete toxidromes. Early decisions regarding escalation and disposition may be influenced by observable behavioral severity in addition to suspected substance category and initial pharmacologic regimen. We evaluated predictors of pharmacologic escalation and psychiatric admission among adults requiring psychoactive medication for substance-related behavioral disturbance. METHODS: This retrospective cohort study was conducted in a tertiary ED between January 2024 and January 2025. Adult patients with clinician-judged substance-related presentations who received antipsychotics or benzodiazepines for immediate behavioral control were included. Agitation severity was assessed using the Overt Agitation Severity Scale (OASS) and categorized as mild (0), moderate (1), or severe (≥ 2). The primary outcome was pharmacologic escalation, defined as the addition of a second antipsychotic or benzodiazepine. The secondary outcome was psychiatric admission. Predictors were evaluated using multivariable logistic regression. Model discrimination was assessed using bootstrap internal validation (1,000 resamples). An exploratory receiver operating characteristic (ROC) analysis was performed to evaluate the discriminative performance of OASS. RESULTS: Ninety encounters were analyzed (median age 32 years; 86.7% male), of which 61.1% had severe agitation (OASS ≥ 2). Pharmacologic escalation occurred in 23.3% and psychiatric admission in 71.1% of cases. Both outcomes increased across OASS strata (p = 0.012 and p = 0.016, respectively). In multivariable analysis, severe agitation independently predicted psychiatric admission (adjusted OR 7.87, 95% CI 2.28-27.17), while increasing age was inversely associated (adjusted OR 0.91 per year, 95% CI 0.86-0.96). The optimism-corrected AUC for psychiatric admission was 0.76 (95% CI 0.66-0.85). In exploratory analysis, OASS ≥ 2 demonstrated a sensitivity of 81.8% and a negative likelihood ratio of 0.36. CONCLUSIONS: Agitation severity was independently associated with psychiatric admission and showed a graded association with pharmacologic escalation in unadjusted analyses across OASS strata. Structured behavioral assessment using OASS may support early risk stratification and disposition planning in substance-related agitation.