Background: ED observation units are increasingly utilized for patients with acute traumatic injuries; however, there is limited evidence to identify appropriate patients for this level of care. This study was designed to identify a set of predictors of short stay trauma treatment that are available at the time of ED disposition.
Methods: This is a retrospective cohort study using data from the trauma registry of a level 1 trauma center from 2013 through 2017. We included adult patients who presented as trauma activations, required observation or admission, and survived through the time of discharge. Patients were excluded if missing length of stay data or mechanism of injury was burn. Logistic regression was used to identify predictors associated with short stay (defined as ≤ 48 hours from arrival to discharge) as the outcome of interest.
Results: Of the 5143 patient encounters included, 1783 (35%) were identified as meeting the short stay criteria. Significant predictors of short stay were mechanism of injury (blunt non-vehicular (OR 2.318, 95%CI 1.648-3.260) compared to penetrating), positive alcohol screen (OR 1.374, 95%CI 1.158-1.630), each year increasing age (OR 0.979, 95%CI 0.976-0.983), normal heart rate (OR 1.272, 95%CI 1.108-1.461), Glasgow coma scale (15 (OR 1.746, 95%CI 1.086-2.806) and 3-8 (OR 0.433, 95%CI 0.247-0.759) compared to 9-12), non-low sBP (OR 2.635, 95%CI 1.354-5.131), hypothermia (OR 0.315, 95%CI 0.107-0.926), and number of body regions injured (none (OR 10.691, 95%CI 7.714-14.815), one (OR 8.807, 95%CI 6.579-11.790), two (OR 5.984, 95%CI 4.480-7.993), three (OR 2.728, 95%CI 1.984-3.750) compared to multiple body regions). Comorbidities with significant associations included congestive heart failure (OR 0.540, 95%CI 0.339-0.860), chronic obstructive pulmonary disease (OR 0.640, 95%CI 0.460-0.890), dementia (OR 0.572, 95%CI 0.351-0.932), and psychiatric conditions (OR 0.626, 95%CI 0.460-0.851).
Conclusions: We have identified clinically relevant features associated with increased and decreased odds of short stay trauma treatment. These clinical features are available at the time of disposition and can inform the ED provider’s decisions regarding level and location of care for trauma patients. Further studies will seek to validate a clinical decision tool to identify trauma patients appropriate for ED observation.