Management of Minor Traumatic Brain Injury in an ED Observation Unit

Matthew A. Wheatley, MD; Shikha Kapil, MD; Amanda Lewis, MSSc, PA-C; Jessica Walsh O'Sullivan, MD; Joshua Armentrout, MD; Tim P. Moran, PhD; Anwar Osborne, MD, MPM; Brooks L. Moore, MD; Bryan Morse, MD; Peter Rhee, MD; Faiz Ahmad, MD; Hany Atallah, MD


Western J Emerg Med. 2021;22(4):943-950. 

In This Article


During the study period 209 patients were placed on the mTBI protocol. We excluded 40 patients from this analysis because they did not have an acute TIH or were admitted as inpatients to the trauma service but boarding in the EDOU. The control group consisted of 53 patients. Demographic and clinical information for the intervention and comparison groups are summarized in Table 3.

The primary outcome is presented in Figure 1. Median LOS (IQR) for EDOU patients was 24.8 (IQR: 18.8 – 29.9) hours compared with a median LOS for the comparison group of 60.2 (IQR: 45.1 – 85.0) hours. This 35.4 (95% CI, 27.3 – 43.5) hour reduction was significant (P < .001). In the adjusted analyses, the intervention was associated with a 35.5 (95% CI, 27.2 – 43.8, P < .001) hour reduction is LOS. In the EDOU group 47 (27.8%) patients got a repeat HCT compared with 40 (75.5%) inpatients, and 106 (62.7%) had a neurosurgical consultation compared with 53 (100%) inpatients (Figure 2). Subdural hematoma was the most common type of hemorrhage. It was found in 60 (35.5%) of patients, and subarachnoid hemorrhage was found in 56 cases (33.1%). Eleven patients had multicompartment hemorrhage of various classifications.

Figure 1.

A) A box and whisker plot depicting length of stay as a function of intervention group. The solid lines within the boxes depict the median for each group and the diamonds within the boxes depict the means for each group. Note that the data are presented on a log10 scale. B) The results of the quantile regressions evaluating the association between the protocols and length of stay. The solid lines depict the difference between the intervention and control groups (eg, the median/50th percentile for the intervention group was approximately 35 hours shorter than for the control group; however, the 75th percentile was approximately 55 hours shorter for the intervention group than for the control group). Negative coefficients indicate that the intervention group had reduced lengths of stay relative to the control group. Shaded regions depict the 95% confidence intervals. The inset section of panel B highlights the change in cost between the 25th and 75th percentiles
LOS, length of stay.

Figure 2.

Graphic representation of difference in neurosurgical consultation and repeat head computed tomography between intervention (EDOU) and control (Inpatient) groups.
HCT, head computed tomography; EDOU, emergency department observation unit.

Twelve (7.1%) patients required hospital admission from the EDOU. Reasons for admission are explained in Table 4. Average inpatient LOS was 3.25 days. Only three patients required ICU care, and four were admitted to the internal medicine service. Ten of the admitted patients were able to be discharged home following their hospitalization. One patient was transferred to hospice, and one was discharged to rehab. None of the patients managed in the EDOU required neurosurgical intervention. There was only one patient death in the EDOU group. Based on review of clinical records, this was thought to be due to metabolic encephalopathy and not head injury.

Follow-up information was available on only 45 (26.6%) patients. Twelve patients reported mild symptoms of headache or dizziness. One patient had persistent headache three months later. No patients required readmission or neurosurgical intervention due to their head injuries. Two patients were called back to the ED due to CT over-reads. Neither of these visits resulted in an admission. Seven patients received outpatient imaging due to persistent symptoms, but no neurosurgical intervention was required for these patients.