Key points

Even in trauma, signs and symptoms take time to develop and the diagnosis of “no injury” is more difficult to make than the positive.

Introduction

Emergency department observation units (EDOUs) have been used to manage patients with injuries over the past several decades in an effort to reduce hospital overcrowding and improve throughput. In 2012, the first prospective study evaluating the use of EDOUs for all types of trauma was performed. Ten percent required inpatient conversion from the EDOU, and there were no deaths, intubations, loss of vital signs, or other adverse outcomes for patients placed in the EDOU under the trauma protocol. Overall, the EDOU was shown to be a safe, cost-effective alternative to routine inpatient admission for the short-term management of injured patients.

Introduction

Emergency department observation units (EDOUs) have been used to manage patients with injuries over the past several decades in an effort to reduce hospital overcrowding and improve throughput. In 2012, the first prospective study evaluating the use of EDOUs for all types of trauma was performed. Ten percent required inpatient conversion from the EDOU, and there were no deaths, intubations, loss of vital signs, or other adverse outcomes for patients placed in the EDOU under the trauma protocol. Overall, the EDOU was shown to be a safe, cost-effective alternative to routine inpatient admission for the short-term management of injured patients.

Mild traumatic brain injury

Head injury is one of the most commonly encountered types of trauma in the ED, comprising 20% to 30% of all traumas. In 2010, 2.2 million ED visits in the United States were due to a traumatic brain injury (TBI). However, only 10% to 15% of these patients have severe head injuries that require hospitalization.

The severity of TBI is categorized according to the GCS. Approximately 80% of brain injuries are considered mild traumatic brain injury (mTBI) (GCS 13–15). The optimal evaluation and treatment strategy of mTBIs remains controversial. Some patients with an initially normal GCS will have an abnormal CT scan. Furthermore, some patients with normal neurologic examinations and CT scans will subsequently deteriorate. Mendelow and colleagues reported that mTBI patients occupied 1959 of 5288 total bed days and incurred 37% of total cost for all TBI patients, although the incidence of subsequent deterioration was low. In a separate retrospective analysis by Jones and colleagues, if observation was not performed, 35 additional beds would be made available for inpatients each year, but 30 patients with neurologic deterioration after mTBI would be erroneously discharged.

The ability to predict which alert, responsive ED patients with mTBI require monitoring for subsequent deterioration is limited. In one study, observation of patients with loss of consciousness or amnesia who had no evidence of impaired consciousness, focal neurologic deficit, seizure, vomiting, severe headache, or skull fracture led to no negative outcomes and no missed injuries. The incidence of significant neurologic deterioration after minor head injury is less well known and estimated at 0.59% to 3.9%. Unpredictable neurologic deterioration may occur regardless of GCS score, so focused and frequent reassessments in observation are necessary to detect any change that suggests the development or expansion of an intracranial hematoma or edema.

The predictive value of head CT in identifying patients at risk for neurologic deterioration has been unclear. A retrospective review concluded that in patients who presented with a GCS of 15, if the patient had a normal mental status and a normal neurologic examination in the ED, the chance of the patient developing a serious complication from mTBI was exceedingly small. However, in the presence of an abnormal mental status or focal neurologic deficit, even if no operative lesion is present on CT, the patient should be observed in the hospital. This recommendation was based on the finding that 3 of the 137 patients developed operative hematomas and an additional 3 had significant deterioration while in the hospital under observation.

It has also been noted that relying on neurologic signs at the time of arrival at the ED and observation may not be adequate in all settings. Stein and Ross reported 18% of patients with mTBI and a nonfocal neurologic examination had abnormalities on CT scan, which increased 3-fold as GCS decreased from 15 to 13. Five percent of the total study population eventually needed surgery as the result of their mTBI. Similar to other investigators, they were unable to determine who would experience delayed deterioration from intracranial lesions not yet clinically evident.

Two large investigations attempted to derive criteria to predict the need for CT in ED patients with mTBI. The investigators of the New Orleans Criteria (NOC) study attempted to develop and validate a set of 100% sensitive clinical criteria for identifying patients with mTBI presenting with a GCS of 15 who should undergo CT scanning. The primary endpoint of the NOC was the detection of any acutely positive finding on CT scan.

The following year, a second clinical decision tool, the Canadian CT Head Rule (CCHR), was derived to guide decision making regarding the use of CT in patients with mTBI. Patients were enrolled if they experienced mTBI and presented with a GCS of 13 to 15. The primary goal of the CCHR was predicting the need for neurosurgical intervention, whereas the secondary outcome was predicting clinically important brain injury on CT. The CCHR was not calibrated to detect any of the following in a neurologically intact patient: solitary contusion less than 5 mm diameter, localized subarachnoid hemorrhage less than 1 mm thick, isolated pneumocephaly or closed depressed skull fracture not extending through the inner table. The investigators concluded they had developed a highly sensitive tool for the use of CT in patients with mTBI who were at high risk for the primary or secondary outcomes. However, TBI that does not require immediate neurosurgical intervention may still lead to significant clinical problems, and patients with these injuries may benefit from being observed for the development of these conditions.

Subsequent validation studies of the CCHR and NOC have been attempted. Sensitivities ranged from 79% to 100% for the CCHR and 86% to 100% for NOC, although confidence intervals were wide when perfect sensitivity was reported and specificity was low. In addition, there is a reported risk of subsequent deterioration in those with initially normal CTs, or those with positive CTs with normal neurologic examinations that these instruments are unable to address. Missing any acute intracranial injuries may not be acceptable to some practitioners, regardless of their clinical significance. The benefits of using observation in such patients include the opportunity to provide symptomatic support as well as monitoring neurologic status and coordinating care with specialists for expedited evaluation and follow-up.

Anticoagulated patients with mTBI and normal initial CT scan are at an increased risk for delayed intracranial hemorrhage (ICH). The rate of delayed traumatic ICH in anticoagulated patients with mTBI and a normal CT scan was 0.6% in one prospective evaluation. In a separate review of 77 anticoagulated patients with mTBI, neurologic deterioration was most likely to occur at 8 to 18 hours after injury, whereas other small series demonstrated most patients deteriorated within 24 hours of injury.

Observation can also be used to manage patients with acute intracranial findings on CT that do not require immediate neurosurgical intervention. In a series of 1078 patients with mTBI, 110 patients with small intracranial bleeds were managed in an OU. No patients experienced delayed deterioration or needed neurosurgical intervention, and there were no readmissions. This concept was supported in a subsequent evaluation comparing observation to routine admission to manage patients with mTBI and acute intracranial hematoma on head CT. The hospital length of stay (LOS) was shortened by 2 days when patients were managed in an OU instead of routine admission. In a separate study of 1185 patients with mTBI, all patients with isolated traumatic subarachnoid hemorrhage managed in an OU did well. Some traumatic ICHs are at low risk for decompensating and should be considered safe for EDOU management to decrease resource utilization. More work is needed to identify these types of conditions.

Small traumatic pneumothoraces

A pneumothorax (PTX) is one of the most common complications of chest trauma. With more widespread use of CT in the evaluation of trauma patients, small PTXs not seen on the supine anteroposterior chest radiograph are being diagnosed with CT at an increasing rate, most of which would otherwise have gone unnoticed and presumably untreated. This type of PTX is called an occult pneumothorax (OPTX). Management of the OPTX is controversial, and it is not clear if this condition should be managed with tube thoracostomy, or if it can be managed with observation alone. Proponents of tube thoracostomy argue for the definitive procedure over the risk of progression of the PTX or development of a tension pneumothorax (TPTX). However, tube thoracostomy carries its own risks, such as lung injury, vascular damage, malpositioning, or infection in up to 20% of those receiving the procedure.

Several prospective, randomized evaluations have been conducted to address the management of this condition. Enderson and colleagues compared the risks and benefits of tube thoracostomy versus observation for 40 patients diagnosed with OPTX and included some patients receiving positive-pressure ventilation (PPV). The overall incidence of OPTX in their study population was 5.6%. LOS was shorter in the tube thoracostomy cohort, although not by a statistically significant duration. Eight patients in the observation group developed a need for tube thoracostomy while on PPV, with 3 of these developing a TPTX. The investigators concluded that patients with OPTX who require PPV are at risk for progression of their OPTX and development of TPTX.

These initial studies were followed by a separate investigation of 39 patients with OPTX to compare management with tube thoracostomy versus observation alone in patients with OPTX, regardless of the size of OPTX or requirement for PPV. The overall incidence of OPTX in this study was 5.9%. The median LOS was shorter for the observation cohort than the routine thoracostomy group (5 days vs 8 days). There was no statistically significant difference in rate of OPTX progression or OPTX size between the groups. Similarly, the overall likelihood of progression of OPTX for those patients on PPV was not statistically significant. There was no difference between the groups in overall complication rate, and most patients treated by observation that received PPV did not have progression of their OPTX or require tube thoracostomy placement.

These 2 prior trials yielded differing results, with Enderson and colleagues reporting that 53% of observed patients had progression of their OPTX, including 3 with TPTX, while Brasel and colleagues found that only 22% experienced progression of OPTX, both including patients with OPTX requiring PPV. A third prospective, randomized study of 24 patients with OPTX was performed in an effort to evaluate the management of patients with OPTX requiring PPV with tube thoracostomy or observation alone. Four of 13 patients in the observation cohort required tube thoracostomy; however, no patient required an emergent chest tube. The investigators concluded that observation may be at least as safe and effective as tube thoracostomy for managing OPTX with respect to mortality, progression of PTX, and complication.

Recently, Moore and colleagues conducted a prospective study over 2 years at 16 US trauma centers to determine which factors were associated with failed observation (defined as the need for tube thoracostomy while being observed) in patients with OPTX. Five hundred sixty-nine patients with OPTX were enrolled. Six percent of patients who were observed failed observation. No patients in the failed observation group developed a TPTX or an adverse event related to delayed thoracostomy. LOS was reduced by almost half in those who were successfully observed versus those who failed observation (7.6 vs 15 days), resembling findings from a separate study by Wilson and colleagues when OPTX were managed with observation alone versus tube thoracostomy.

Recently, an interim analysis of the American Association for the Surgery of Trauma workgroup was conducted to report the outcomes in an ongoing multicenter randomized controlled trial comparing tube thoracostomy with observation alone for the management of OPTX. There was no difference in the primary outcome noted between the 2 groups. Only one (2%) observed patient developed a TPTX, which occurred during sustained PPV and resolved with tube thoracostomy. There were no differences in mortality or LOS between the 2 groups.

Currently, the Eastern Association for the Surgery of Trauma (EAST) Practice Management Guidelines supports the position that OPTXs in a stable patient can be managed with observation alone. Most patients with OPTX will not have progression of the OPTX. However, the guidelines concluded that more information was needed in determining with accuracy which OPTXs are at risk of progression. Other nonclinical factors may impact decision making, including the level of monitoring while being observed, the availability of a clinician to perform an invasive intervention without delay should OPTX progress or TPTX develop, and patient preference regarding risks and benefits of tube thoracostomy versus observation alone for OPTX.

Traumatic orthopedic conditions

Data are limited regarding the use of EDOUs for isolated orthopedics trauma. Injuries well suited for EDOUs include nonoperative fractures requiring pain control and coordination of postdischarge services, injuries requiring bracing, and extremity fractures requiring monitoring for compartment syndrome (CS). The vast majority of patients placed in the EDOU with nonoperative orthopedic trauma can be discharged home on the index visit without a need for inpatient admission.

It has been demonstrated that older adults in the ED with orthopedic trauma can be managed with pathways incorporating resources available to most EDOUs. The EDOU provides an ideal setting for care coordination and discharge planning and an opportunity for assessment by a geriatrician, pharmacist, physical therapist, social worker, and care manager. These patients receive focused interventions in mobility, gait training, and functional status improvements. Therapists in the EDOU setting are skilled in managing patients with acute pain with the focus of safe disposition planning and also serve as an educational resource to patients and families in preventing future injuries.