Joseph B, Pandit V, Haider AA, Kulvatunyou N, Zangbar B, Tang A, Aziz H, Vercruysse G, O’Keeffe T, Freise RS, Rhee P. Improving Hospital Quality and Costs in Nonoperative Traumatic Brain InjuryThe Role of Acute Care Surgeons. JAMA Surg. 2015;150(9):866-872. doi:10.1001/jamasurg.2015.1134
The role of acute care surgeons is evolving; however, no guidelines exist for the selective treatment of patients with traumatic brain injury (TBI) exclusively by acute care surgeons. We implemented the Brain Injury Guidelines (BIG) for managing TBI at our institution on March 1, 2012.
To compare the outcomes in patients with TBI before and after implementation of the BIG protocol.
Design, Setting, and Participants
We conducted a 2-year analysis of our prospectively maintained database of all patients with TBI (findings of skull fracture and/or intracranial hemorrhage on an initial computed tomographic scan of the head) who presented to our level I trauma center. The pre-BIG group included patients with TBI from March 1, 2011, through February 29, 2012, and the post-BIG group included patients from July 1, 2012, through June 30, 2013.
Main Outcomes and Measures
The primary outcome measures were patients with repeated computed tomography of the head and neurosurgical consultations. Secondary outcome measures were findings of progression of intracranial hemorrhage on repeated computed tomographic scans, neurosurgical intervention, hospital admission, intensive care unit admission, hospital and intensive care unit length of stay, 30-day readmission rate, and hospital costs per patient.
A total of 796 patients (415 in the pre-BIG group and 381 in the post-BIG group) were included. There was a significant reduction (19.0%) in the rate of neurosurgical consultation (post-BIG group, 273 patients [71.7%]; pre-BIG group, 376 [90.6%]; P < .001), repeated computed tomography of the head (post-BIG group, 255 patients [66.9%]; pre-BIG group, 381 patients [91.8%]; P < .001), hospital (post-BIG group, 330 [86.6%]; pre-BIG group, 398 [95.9%]; P < .001) and intensive care unit admission (post-BIG group, 202 [53.0%]; pre-BIG group, 257 [61.9%]; P = .01), hospital length of stay (post-BIG group, 5.4 [4.5] days; pre-BIG group, 6.1 [4.8] days; P = .03), and hospital costs per patient ($4772 per patient; P = .03) with implementation of BIG. There was no difference in the in-hospital mortality rate (post-BIG group, 62 patients [16.3%]; pre-BIG group, 69 patients [16.6%]; P = .89), progression of intracranial hemorrhage on repeated scans (post-BIG group, 41 patients [10.8%]; pre-BIG group, 59 patients [14.2%]; P = .14), neurosurgical intervention (post-BIG group, 61 patients [16.0%]; pre-BIG group, 59 patients [14.2%]; P = .48), and 30-day readmission rate (post-BIG group, 31 patients [8.1%]; pre-BIG group, 37 patients [8.9%]; P = .69) after implementation of BIG.
Conclusions and Relevance
Implementation of BIG is safe and cost-effective. BIG defines the management of TBI without the need for neurosurgical consultation and unnecessary imaging. Establishing a national, multi-institutional study implementing the BIG protocol is warranted.
Traumatic brain injury (TBI) is a major medical and socioeconomic problem that affects 1.7 million people and results in more than 50 000 deaths annually.1 According to the US Centers for Disease Control and Prevention, the incidence of TBI-related emergency department visits and hospitalizations has increased by 20% from 2002 through 2006.1,2 Along with the loss of life and long-term disability, TBI costs up to 10% of the total health care budget, which accounts for nearly $30 billion annually.3 The initial diagnosis of TBI on a computed tomographic (CT) scan of the head is often followed by intensive care unit (ICU) admission, neurosurgical consultation, and repeated imaging. However, recent publications now support selective treatment of these patients.4- 7
Acute care surgeons treat patients with TBI in the acute phase of trauma and form an integral component of their nonoperative treatment. Ninety percent of patients with TBI are treated nonoperatively, and there is a paucity of well-defined guidelines that highlight the role of acute care surgeons in the nonoperative management of TBI. Moreover, with advancements in CT scanners, the number of patients presenting to emergency departments with miniscule intracranial bleeds is increasing, leading to a rising number of neurosurgical consultations. These consultations, however, rarely alter the course of treatment of patients who present with normal neurological examination findings and miniscule intracranial hemorrhage (ICH).8 In an effort to standardize the need for neurosurgical consultation and repeated CT scans of the head (RHCT) in patients with TBI, our institute formulated the Brain Injury Guidelines5 (BIG) based on a retrospective analysis of nearly 3000 patients with TBI who presented to our level I trauma center. These guidelines were based on patients’ medical history and findings from neurological examination and initial CT of the head. These guidelines, which were formulated in collaboration with our neurosurgical colleagues, addressed the role of the acute care surgeon in the management of TBI and had the potential to reduce the burden of nonoperative TBI on our neurosurgical colleagues. These guidelines were formally implemented on March 1, 2012; since then, we have been treating our patients with blunt TBI based on these guidelines. The aim of this study was to compare the outcomes of patients with TBI before and after implementation of the BIG protocol.
After approval and waiver of consent from the Institutional Review Board at the University of Arizona College of Medicine, we conducted a 2-year (pre- and post-BIG) analysis of our prospectively maintained database of all patients with TBI who presented to our level I trauma center. A waiver of patient consent was obtained because the BIG protocol was implemented as an institutional policy for management of TBI. The BIG protocol was implemented at our trauma center on March 1, 2012. The pre-BIG group included patients with TBI from March 1, 2011, through February 29, 2012, and the post-BIG group included patients from July 1, 2012, through June 30, 2013. The period from March 1, 2012, through June 30, 2012, was excluded from the analysis owing to excessively low compliance with the BIG protocol during the initial training period.
Patients with a blunt mechanism of trauma who had an ICH and/or a finding of skull fracture from an initial CT scan of the head were included in the study. We excluded patients who were transferred from other institutions, underwent emergency neurosurgical intervention, or were dead on arrival.
The following data points were prospectively recorded for each patient: patient demographics (age, sex, and mechanism of injury); vital parameters on presentation, including systolic blood pressure, heart rate, and temperature; Glasgow Coma Scale (GCS) score; neurological examination findings on presentation; intoxication (drug or alcohol); details regarding antiplatelet and anticoagulation therapy; intubation; loss of consciousness; findings from an initial CT scan of the head (type and size of ICH); reasons for and findings from an RHCT; neurosurgical consultation; neurosurgical intervention details; hospital and ICU length of stay; discharge disposition; GCS score at discharge; and in-hospital mortality rate. The 30-day readmission rate was assessed by reviewing electronic medical records for hospital admissions within 30 days of the initial discharge. The Injury Severity Score and head Abbreviated Injury Scale score were obtained from the trauma registry.
We developed BIG based on patients’ medical history (antiplatelet or anticoagulation therapy, loss of consciousness, and intoxication), findings from physical examination (focal neurological examination, pupillary examination, and GCS score on admission), and CT scan findings (size and location of ICH and type of skull fracture).5 Patients had to meet all the criteria for categorization into BIG 1 or BIG 2 (Table 1). Failure to meet even 1 criterion (in BIG 1 or BIG 2) categorized the patient into the BIG 3 category and altered the treatment plan for the patient based on the BIG 3 category.
Table 1 describes the 3 categories of brain injury guidelines. Patients who were categorized as BIG 1 (minor head injury) had normal findings on neurological examination, were not taking any antiplatelet or anticoagulation medications, and had minuscule findings on an initial CT scan of the head. We proposed a 6-hour period of observation in the emergency department for patients who were categorized as BIG 1 without the need for neurosurgical consultation or an RHCT scan. The BIG 2 category was composed of moderately injured patients with a nondisplaced skull fracture and/or a localized ICH of 5 to 7 mm. Patients who were categorized as BIG 3 had a severe head injury, and the optimal therapeutic plan for these patients consisted of hospitalization, neurosurgical consultation, and a follow-up RHCT scan. Patients who were categorized as BIG 3 had at least 1 of the following high-risk features: an abnormal neurological examination finding, intoxication, antiplatelet or anticoagulation medication use, concerning CT scan findings (displaced skull fractures, diffused subarachnoid hemorrhage, multiple types of bleeding, or an ICH ≥8 mm). Patients who could not be examined and those who were intubated were also categorized as BIG 3.
The compliance rate was defined as the percentage of patients in each BIG category for whom the predefined protocol for management (RHCT scan or neurosurgical consultation) was appropriately followed.
After implementation of BIG, patients with suspected TBI underwent an initial CT scan of the head; those with an intracranial injury were evaluated by the trauma surgeon.
Patients were categorized into 3 BIG categories based on patients’ medical history and findings from neurological examination and the initial CT scan of the head, as assessed by the on-call trauma team.
Patients in the BIG 1 category were observed for 6 hours without neurosurgical consultation and had no RHCT. We defined abnormal neurological examination findings as altered mental status, focal neurological deficits, and an abnormal finding from pupillary examination.
Patients in the BIG 2 category were observed for 24 hours without an RHCT scan or neurosurgical consultation.
Neurological examination was conducted every 2 hours. Patients who had findings of deterioration on clinical examination were upgraded to a higher category, warranting an RHCT scan and neurosurgical consultation.
Patients were stratified into 2 groups: pre- and post-BIG. The primary outcome measures were patients with RHCT scans and neurosurgical consultations. Secondary outcome measures were findings of progression of ICH on RHCT, neurosurgical intervention based on hospital admission, ICU admission, hospital and ICU length of stay, 30-day readmission rate, and hospital costs per patient.
A single investigator (B.J., a trauma surgeon) reviewed the CT scans after the attending radiologist’s reading for the type and size of ICH. Neurosurgical intervention was defined as craniotomy and/or craniectomy or intracranial pressure monitoring.
Data were reported as mean (SD) for continuous variables, median (range) for ordinal variables, and proportion for categorical variables. We used the Mann-Whitney test for nonparametric continuous variables and an unpaired 2-tailed t test for parametric continuous variables to explore for differences between the 2 groups (pre- and post-BIG). We used the χ2 test to identify differences in outcomes between the 2 groups for categorical variables. Ordinal variables were compared using the median test. For our study, we considered P < .05 statistically significant. All statistical analyses were conducted using SPSS, version 21 (IBM).
A total of 796 patients (415 in the pre-BIG group and 381 in the post-BIG group) with TBI were included in our study. The mean age was 39.5 (24.7) years, 528 patients (66.3%) were male, the median GCS score on admission was 13 (interquartile range, 8-15), and the median head Abbreviated Injury Scale score was 2 (interquartile range, 2-3). Overall, 567 patients (71.2%) had loss of consciousness and 165 (20.7%) had abnormal neurological findings on examination. The 2 groups were similar in demographics, preadmission antiplatelet and anticoagulant use (P = .23), blood pressure level on admission (P = .43), mechanisms of injury (P = .24), and head Abbreviated Injury Scale score (P = .74). Most patients were classified as BIG 3 (n = 451) in both groups. The 2 groups were similar in distribution of patients among the BIG 1 (P = .78), BIG 2 (P = .41), and BIG 3 (P = .65) categories. Table 2 highlights the demographics of the study population.
In terms of findings on the initial CT scan of the head, subdural hematoma (307 patients [38.6%]) followed by subarachnoid hemorrhage (273 patients [34.3%]) were the most common types of ICH in the study population. A total of 491 patients (61.7%) had a skull fracture, of which 290 (59.1%) had a displaced skull fracture. There was no difference in the type of ICH before and after implementation of BIG. Table 3 highlights the findings of the initial CT scan of the head in the study population.
The overall compliance rate with BIG at our center following their implementation was 85.5%. The compliance rate for BIG 1 was 91.6%; for BIG 2, 39.2%; and for BIG 3, 100%.
During the study, 649 patients received neurosurgical consultations and 636 patients received RHCT scans. There was a 19.0% reduction in neurosurgical consultations after implementation of BIG (post-BIG group, 273 patients [71.7%]; pre-BIG group, 376 patients [90.6%]; P < .001). Patients who were treated after implementation of BIG were less likely to get an RHCT scan compared with patients who were treated before implementation of the guidelines (post-BIG group, 255 patients [66.9%]; pre-BIG group, 381 patients [91.8%]; P < .001). The overall rate of neurosurgical intervention was 15.1% (n = 120). There was no difference in the rate of neurosurgical intervention before and after guideline implementation (pre-BIG group, 59 patients [14.2%]; post-BIG group, 61 patients [16.0%]; P = .48). Similarly, there was no difference in findings regarding the rate of progression on RHCT (pre-BIG group, 59 patients [14.2%]; post-BIG group, 41 patients [10.8%]; P = .14), neurosurgical intervention based on RHCT (pre-BIG group, 8 patients [1.9%]; post-BIG group, 12 patients [3.2%]; P = .27), or discharge GCS score (median [interquartile range] for post-BIG group, 15 [14-15]; pre-BIG group, 15 [14-15]; P = .88) before and after BIG implementation (Table 4).
After implementation of BIG, there was a 9.3% reduction in the hospitalization rate for TBI. Patients who were treated after implementation of BIG were less likely to require hospital (post-BIG group, 330 [86.6%]; pre-BIG group, 398 [95.9%]; P < .001) or ICU admission (post-BIG group, 202 [53.0%]; pre-BIG group, 257 [61.9%]; P = .01) and had shorter hospital lengths of stay (post-BIG group, 5.4 [4.5] days; pre-BIG group, 6.1 [4.8] days; P = .03) compared with patients who were treated before BIG implementation. Implementation of BIG was associated with lower mean hospital costs of $4772 per patient (P = .03) (Table 4).
On subanalysis of patients from the BIG 1 category, there was a significant reduction in RHCT scans (pre-BIG group, 59 [67.8%]; post-BIG group, 6 [7.2%]; P < .001), neurosurgical consultations (pre-BIG group, 76 [87.4%]; post-BIG group, 7 [8.4%]; P < .001), hospitalizations (pre-BIG group, 68 [78.1%]; post-BIG group, 42 [50.6%]; P < .001), and ICU admissions (pre-BIG group, 24 [27.5%]; post-BIG group, 6 [7.2%]; P = .001) following implementation of BIG (Table 5).
At most centers, any patient with a TBI is admitted to the hospital and ICU for repeated imaging and neurosurgical consultation. However, most cases of TBI are managed nonoperatively, and neurosurgical consultation or repeated imaging alone does not alter the course of management.9,10 Moreover, admitting every patient with TBI for repeated imaging and neurosurgical consultation incurs unnecessary radiation exposure, requires allocation of personnel, and puts a substantial burden on the neurosurgical services. Our center has been pushing to reserve the use of repeated imaging and consultation for only the patients with TBI who require it. However, before formulation of BIG, no practice guidelines were available for the management of nonoperative TBI by acute care surgeons. The implementation of BIG standardized the management protocols for nonoperative TBI, resulting in a major paradigm shift in the treatment of these patients exclusively by acute care surgeons at our center.
There are 2 main findings of our study. First, with the implementation of BIG, we were able to reduce unnecessary neurosurgical consultations, hospital admissions, and RHCT scans. Second, there was no difference in the outcomes of patients who were treated before and after the implementation of BIG. We believe that implementation of BIG is safe and cost-effective in standardizing the practice of nonoperative care for patients with TBI.
The BIG protocol was developed and implemented at our level I trauma center in collaboration with our neurosurgical colleagues. Our center has robust 24-hour on-call neurosurgical coverage, providing a safety net for the potential deviations from protocol. By following the BIG protocol, we were able to avoid neurosurgical consultation and repeated imaging for more than 30% of our patients with TBI. The greatest effect from implementation of BIG was observed in our BIG 1 category of patients, who were the minimally injured subgroup of patients with normal findings from neurological examination and no high-risk factors. In accordance with BIG, these patients were treated exclusively by acute care surgeons, with a 6-hour observation period in the emergency department only. This practice resulted in a reduction in neurosurgical consultations and RHCT to less than 10% of these patients. Many investigators, including us, have questioned the need for mandatory neurosurgical consultation for minuscule ICH in patients with TBI who can be examined, and our recommendations for these patients are consistent with those from other studies.7,11,12 However, none of the previous studies followed a strict protocol for such recommendations or determined their effects.
The role of RHCT scans in examinable patients has been questioned repeatedly.6,13- 15 In a prospective 3-year study at our center, we concluded that RHCT scans were unwarranted in people with normal findings from neurological examination.4,16 We implemented these findings in our BIG protocol by limiting RHCT scans only to patients in the BIG 3 category. This decision is based on the fact that an altered neurological examination finding or the inability to obtain a reliable neurological examination will place the patient in the BIG 3 category irrespective of their initial BIG category. Moreover, every CT scan exposes a patient to 3 to 7 rads of radiation.17 The implementation of this practice has helped us limit unnecessary radiation exposure in our patients.
Along with the above risks, the financial burden associated with current practices in TBI care is considerable. The cost of a hospital room is more than $3000 per day,18,19 which does not include other costs associated with hospital admission, including laboratory tests, personnel, imaging, and consultations. With the implementation of BIG, we reduced hospital and ICU admissions and hospital lengths of stay at our center. This practice translated to an average saving of more than $4000 per patient in our population.
The biggest concern with minimizing the use of RHCT or treating patients with TBI without neurosurgeons is the potential for adversely affecting patient outcomes. Although we did not perform a routine RHCT scan and did not consult neurosurgeons for most BIG 1 and BIG 2 patients, we did not find any difference in the in-hospital mortality rate, 30-day readmission rate, progression on RHCT, or discharge GCS scores before and after implementation of BIG.
There are multiple implications from the findings of our study. The BIG protocol helps limit the neurosurgical workforce, which is a scarce resource in the United States. In a national survey, more than 75% of emergency department directors reported a shortage of neurosurgeons.20 BIG can help reserve this limited resource for severely injured patients. Moreover, implementation of BIG nationwide can help reduce millions of dollars in hospital costs.21 Protocols for hospital admission in mild TBI should be redefined. Clinical deterioration during the 6-hour period of observation should be used as a guide to hospital admission, neurosurgical consultation, and further imaging in cases of mild TBI.22,23
The BIG protocol was implemented at a level I trauma center with a 24-hour on-call neurosurgical service. Therefore, we recommend that generalization of our findings should be done in this context. Our study is limited by the fact that we did not assess for long-term outcomes in our study patients, such as the Extended Glasgow Outcome Score or Functional Independent Measure. Variables not included in the BIG protocol are still an area of active debate, such as platelet transfusion in patients with brain injury who received anticoagulant therapy. We did not perform a robust cost analysis and only reviewed overall hospital costs. However, despite these limitations, we demonstrate effective implementation of BIG for the treatment of patients with TBI.
Implementation of BIG is safe and cost-effective. The BIG protocol standardizes the treatment of patients with TBI without the need for neurosurgical consultation and unnecessary imaging. It is an important contribution to the growing body of evidence demonstrating the extended utility of acute care service. Establishing a national, multi-institutional study implementing the BIG protocol is warranted.
Accepted for Publication: March 25, 2015.
Corresponding Author: Bellal Joseph, MD, Division of Trauma and Acute Care Surgery, Department of Surgery, University of Arizona Medical Center, 1501 N Campbell Ave, Room 5411, PO Box 245063, Tucson, AZ 85727 (firstname.lastname@example.org).
Published Online: June 24, 2015. doi:10.1001/jamasurg.2015.1134.
Author Contributions: Dr Joseph had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Joseph, Pandit, Haider, Zangbar, Aziz, Vercruysse, O’Keeffe, Freise, Rhee.
Acquisition, analysis, or interpretation of data: Joseph, Pandit, Kulvatunyou, Zangbar, Tang, Aziz, Vercruysse, O’Keeffe, Rhee.
Drafting of the manuscript: Joseph, Pandit, Haider, Zangbar, Tang, Aziz, Vercruysse, Rhee.
Critical revision of the manuscript for important intellectual content: Joseph, Kulvatunyou, Zangbar, Tang, O’Keeffe, Freise, Rhee.
Statistical analysis: Joseph, Pandit, Zangbar, Aziz, Rhee.
Administrative, technical, or material support: Joseph, O’Keeffe, Rhee.
Study supervision: Joseph, Haider, Tang, Vercruysse, O’Keeffe, Freise, Rhee.
Conflict of Interest Disclosures: None reported.
Previous Presentation: This study was presented as a poster at the 73rd Annual Meeting of the American Association for the Surgery of Trauma; September 11, 2014; Philadelphia, Pennsylvania.