ED utilization of computed tomography in a poisoned population

The objective of this prospective, analytic study was to identify predictors and describe the demographic and clinical correlates of head computed tomography (CT) evaluation in patients with poisoning or drug overdose and altered mental status. Forty-three patients that were evaluated by head CT and 109 that were not evaluated by head CT were entered into the study at a poison control center. None of the 43 scanned patients had any acute findings on head CT. A logistic regression model yielded 4 predictors that were statistically associated with the ordering of a head CT scan: Glasgow Coma Scale (GCS) ≤8 (odds ratio [OR]: 2.3; 95% confidence interval [CI] 1.03-5.7); age ≥41 years (OR 5.3; 95% CI 2.2-13); use of drugs or abuse by history (OR 2.8; 95% CI 1.04-7.6); and witnessed seizure activity (OR 4.8; 95% CI 1.3-17.9). We also tested 2 additional models to identify predictors of hospital admission, 1 with and 1 without CT scan included as a covariate. In the first model, only GCS ≤8 was a significant predictor of admission (OR 10.7; 95% CI 2.4-47.2). When the use of head CT was added to the second model, it also emerged as an independent explanatory predictor of admission (OR 4.8; CI 95% 1.2-20.4) in addition to GCS (OR 10.1; 95% CI 2.2-45.4). In this pilot study, patients presenting to the emergency department (ED) with suspected poisoning or drug overdose were found to have a low likelihood of abnormal findings on head CT scan. Those among whom CT scans were obtained were more likely to be admitted to the hospital from the ED, despite negative findings.