Pathway-Based Reduction of Repeat Head Computed Tomography for Patients With Complicated Mild Traumatic Brain Injury: Implementation and Outcomes

Abstract

BACKGROUND

Routine follow-up head imaging in complicated mild traumatic brain injury (cmTBI) patients has not been shown to alter treatment, improve outcomes, or identify patients in need of neurosurgical intervention. We developed a follow-up head computed tomography (CT) triage algorithm for cmTBI patients to decrease the number of routine follow-up head CT scans obtained in this population.

OBJECTIVE

To report our experience with protocol implications and patient outcome.

METHODS

Data on all cmTBI patients presenting from July 1, 2018 to June 31, 2019, to our level 1, tertiary, academic medical center were collected prospectively and analyzed retrospectively. Descriptive analysis was performed.

RESULTS

Of the 178 patients enrolled, 52 (29%) received a follow-up head CT. A total of 27 patients (15%) were scanned because of initial presentation and triaged to the group to receive a routine follow-up head CT. A total of 151 patients (85%) were triaged to the group without routine follow-up head CT scan. Protocol adherence was 89% with 17 violations.

CONCLUSION

Utilizing this protocol, we were able to safely decrease the use of routine follow-up head CT scans in cmTBI patients by 71% without any missed injuries or delayed surgery. Adoption of the protocol was high among all services managing TBI patients.

Graphical Abstract

Graphical Abstract.

ABBREVIATIONS

ACS

acute care surgery

BIG

Brain Injury Guidelines

cmTBI

complicated mild traumatic brain injury

CT

computed tomography

DOAC

direct oral anticoagulant

ED

emergency department

EDH

epidural hematoma

GCS

Glasgow Coma Scale

ICH

intracranial hemorrhage

SDH

subdural hematoma

TBI

traumatic brain injury

tSAH

traumatic subarachnoid hemorrhage

Routine follow-up head imaging in complicated mild traumatic brain injury (cmTBI) patients has not been shown to alter treatment, improve outcomes, or identify patients in need of neurosurgical intervention.^1-3 Despite the abundance of prospective and retrospective studies, many academic and private institutions still obtain a routine follow-up head computed tomography (CT) scan in all patients with cmTBI.

About 2.8 million people sustain a traumatic brain injury (TBI) annually and about two-thirds of the injuries are classified as mild.^4,5 The term cmTBI refers to patients with a head CT scan positive for an intracranial traumatic finding with a nonfocal neurological examination and a Glasgow Coma Scale (GCS) score between 13 and 15. Of these, less than 10% of patients seeking medical evaluation have a positive CT scan and less than 1% need a surgical intervention based on this initial head CT scan. The probability of needing operative intervention is extremely low if the initial CT scan was none operative. This study focuses on patients who have cmTBIs deemed nonsurgical after their initial CT scan.

It is important to understand the natural history of traumatic intracranial hemorrhages (ICHs). Traumatic ICHs are associated with a progression rate from 17% to 20% in patients with or without anticoagulation.^2,6 But it is notable that this radiographic progression rate does not result in need for neurosurgical intervention.³

Based on our work and previous work by others, we developed a follow-up head CT triage algorithm for cmTBI patients. The objective of this study was to report our experience of the change in frequency of routine follow-up head CTs, our experience with protocol implications, and 30-d outcomes.

METHODS

Protocol Development

A cmTBI protocol to triage routing follow-up head CT scans (Figure 1) was developed using a multidisciplinary approach at our level 1 academic medical center. Representatives of the neurosurgery, emergency medicine, and acute care surgery (ACS) departments all were involved. The protocol resembled our existing TBI observation protocol, which allows for patients with cmTBI to be observed in the emergency department (ED) for a 6-h period, and subsequently discharged if their neurological exam and symptoms remained stable.

FIGURE 1.

cmTBI protocol used to triage need for follow-up head CT scan. CT = computed tomography; SDH = subdural hematoma; EDH = presence of epidural hematoma.

Educational sessions were held with ACS, emergency medicine, and neurosurgery faculty, residents, and advanced care providers prior to initiating the protocol.

Study Design

We implemented the protocol in July 2018. All patients who came through the ED with evidence of cmTBI on initial imaging were triaged according to our new cmTBI protocol. We defined cmTBI as a GCS score of 13 to 15 with a head CT scan positive for an intracranial, traumatic finding, such as traumatic subarachnoid hemorrhage (tSAH), contusion, subdural hematoma (SDH), or epidural hematoma (EDH). Patients were excluded if they required neurosurgical intervention based on initial presentation.^7,8

Routine follow-up head CT scans were deemed indicated only in patients with an acute SDH > 1 cm, EDH, history of taking a direct oral anticoagulant (DOAC) within 18 h of presentation, who were on warfarin or any antiplatelet agent other than aspirin, or who had a worsening neurological examination during the observation period. For all other patients, including those who were on aspirin only, routine repeat head imaging was not recommended.

All patients were seen and evaluated by the neurosurgery service. After their evaluation in the ED, patients were either discharged, admitted to the ED observation unit, or admitted to the hospital. All admitted patients were re-evaluated by the neurosurgery team prior to discharge from ED.

Variables

Following implementation of the new protocol, data on all cmTBI patients presenting from July 1, 2018 to June 31, 2019, were collected prospectively and analyzed retrospectively. Violations of the protocol, the reason for the violation, and the violating service were recorded. Outcome variables collected were discharge disposition, number of head CT scans avoided, need of operative intervention, and 30-d follow-up. We also collected data on patient characteristics such as GCS, age, gender, anticoagulation status, mechanism of injury, and head CT imaging findings.

Statistical Methods

We compared 3 groups of patients (protocol concordant no routine follow-up CT, protocol concordant routine follow-up CT, and protocol violation). We used SAS statistical software (SAS Institute, Cary, North Carolina). Statistical tests included analysis of variance for age, Kruskal-Wallis for GCS, extensions of Fisher exact test for contusion and EDH, and Pearson chi-square for all other comparisons. Data were extracted from an Institutional Review Board database with a waiver for consent as the data was deidentified.

RESULTS

Descriptive Data

A total of 178 patients presented to the emergency room with a cmTBI that was deemed nonsurgical based on the initial presentation and CT scan. The ages ranged from 18 to 98 yr old, with a mean age of 68.4 yr (standard deviation = 18.5), 53% were women and 47% were men. The majority of injury mechanisms were falls (88%). The median GCS was 15. Of patients studied, 46% (n = 81) were taking at least 1 antiplatelet or oral anticoagulant medication. Nine (5%) were on more than 1 of these medications. For details, see Table 1.

TABLE 1.

Summary of Patient Characteristics and Injury Mechanism

Patient characteristics	Number (%)
Age (mean)	18-98 yr (68.4 yr)
Gender (%)
Male (%)	84 (47.2%)
Female (%)	94 (52.8%)
Mechanism (%)
Fall	156 (88%)
Motor vehicle collision	13 (7.3%)
Sports-related accident	4 (2.2%)
Need for surgical intervention	0 (0%)
GCSa (median)	15
At least on antiplatelet or oral anticoagulation medication, including ASA	81 (45.5%)
Received follow-up head CT	52 (29.2%)

^aGCS = Glasgow Coma Scale; ASA = acetylsalicylic acid (aspirin) therapy; CT = computed tomography.

Outcome Data

Of the 178 patients, 29% (n = 52) received a follow-up head CT. A total of 15% (n = 27) of patients were scanned because of initial presentation (see Figure 2) and triaged to the group to receive a routine follow-up head CT. An additional 10% (n = 17) were scanned in violation of protocol, and 4% (n = 8) received a follow-up CT scan because of neurological exam change perceived by the treating physician. All of the noted neurological changes resolved during observation and did not trigger any intervention. None of the patients who had routine follow-up head CT scans or triggered follow-up CT scans had a change in management or required neurosurgical intervention.

FIGURE 2.

This flow diagram shows the triage of patients according to the cmTBI protocol, number of head CT scans, and violations. CT = computed tomography; SDH = subdural hematoma SDH; EDH = presence of epidural hematoma.

Of all cmTBI patients who were deemed nonsurgical based on the initial head CT, 85% (n = 151) could be triaged to the group without a routine follow-up head CT. With implementation of the protocol 71% (n = 126) of repeat head CT scans were avoided. Of the 151 patients who were triaged to the group without a routine follow-up head CT scan, there was an 89% (n = 134) protocol adherence. There were 17 violations of the protocol. All of these violations are unindicated repeat head CT scans. The violations were split nearly equally between neurosurgery 41% (n = 7) and emergency medicine 53% (n = 9). There was 1 violation by the ACS service. In total 48% (n = 85) of 178 patients were admitted. The remainder 52% (n = 93) were discharged from the ED. Three clinical services admitted the majority of the patients when indicated. Neurosurgery admitted 27% (n = 23); ACS admitted 35% (n = 30); and internal medicine admitted 38% (n = 32).

When comparing patients who did not have a repeat head CT scan to patients who did (either as indicated or by protocol violation), there was no statistically significant difference in age (P = .85), gender (P = .4), SDH (P = .74), tSAH (P = .69), contusion (P = . 06), or GCS score (P = .16). There was a statistically significant difference between the groups in patients who had EDH (P = .002) as expected by the protocol and anticoagulation (P < .001). See Figure 3 below.

FIGURE 3.

Patient characteristics compared between patients who were triaged to no follow-up head CT scan, triaged to have a follow-up head CT scan, and who received a head CT scan because of a violation of the protocol. SDH = subdural hematoma; tSAH = traumatic subarachnoid hemorrhage; EDH = epidural hematoma; GCS = Glasgow Coma Scale; anticoag = use of anticoagulant except aspirin; CT = computed tomography; f/u = follow-up.

Of the 178 patients included, 11% (n = 19) returned to the ED within 30 d. Eight of these visits were related to the initial TBI (see Table 2). One patient who returned needed an operation for a chronic SDH; this patient was from the routine follow-up head CT group.

TABLE 2.

Overview of Patient Who Returned to the ED Within 30 D of Discharge for Reasons Related to TBI

Patient number	Reason for return	CTa group	ED imagine findings	Disposition
1	Headache	No f/u	No change	Discharged from ED
2	Altered metal status	No f/u	No change	Discharged from ED
3	Symptoms changed	No f/u	No change	Discharged from ED
4	Symptoms changed	Routine f/u head	Aging SDH	Burr hole evacuation for chronic SDH
5	Headache	No f/u	No change	Discharged from ED
6	Headache	No f/u	Edema evolution	Discharged from ED
7	Headache	No f/u	Radiographic progression	Admission for observation no OR
8	New symptoms	No f/u	new ICH	Declined intervention as patient was DNR/DNI

Radiographic worsening refers to the head CT scan obtained when the patient re-presented to the ED. With CT group, we are referring to the group the patient was triaged to during their initial presentation.

^aCT = computed tomography; ED = emergency department; f/u = follow-up; b/o = because of; SDH = subdural hematoma; OR = operating room; ICH = intracranial hemorrhage; DNR/DNI = do not resuscitate/do not intubate.

DISCUSSION

Key Results

In an effort to decrease the number of routine follow-up head CT scans for patients with cmTBI, our institution implemented the cmTBI protocol based on medical history and findings from the neurological examination and initial head CT scan. Utilizing this protocol, we could avoid 71% (n = 126) unnecessary CT scans within the first year.

Adherence to the institutional protocol was high, especially since this was a substantial paradigm shift in cmTBI management. Of the 151 patients who were triaged to the group without follow-up head CT, 89% of decisions adhered to protocol. There were no missed injuries or delays in neurosurgical intervention. None of the patients in this group underwent delayed surgery. Protocol violations were distributed equally between neurosurgery and other services. Violations happened in the early phase of the new protocol implementation. This demonstrated that it takes some time for providers to adapt to the new protocol. The data show that the majority of our providers felt that it is safe to forgo routine follow-up CT scans in patients without neurological changes.

A total of 4% (n = 8) of patients had a follow-up head CT because of neurological change. None of the scans resulted in a treatment change or neurosurgical intervention during the patients’ admissions. In addition, 11% of patients returned to the ED within 30 d after their initial presentation. Seven patients re-presented with TBI-related complaints (see Table 2). There was 1 patient who needed surgery for a now-chronic SDH. It is important to note that this patient was triaged into the routine follow-up head CT group because of the size of the SDH. This is a good example of how performing a routine follow-up head CT scan does not prevent patients from developing a need for surgery later, and may contribute to a false sense of security.

As our protocol was written, we continue to triage patients with an EDH or on anticoagulant therapy, especially those on DOACs, to receive routine follow-up head CT scans. Prior research has shown that EDH has a higher risk of expansion and acute deterioration with need for surgical intervention.⁸ The only statistically significant differences found in this study between those having a routine follow-up head CT scan and the group receiving no routine follow-up imaging were presence of an EDH and anticoagulation status.

The prevalence of antithrombotic medication in elderly patients who present with cmTBI can range from 30% to 50%. In our group, 46% of patients were taking at least 1 antithrombic agent, including ASA (Table 3). A validation study in a level III trauma center showed that anticoagulant or antiplatelet use was not associated with worsened outcomes in cmTBI patients. The Brain Injury Guidelines (BIG) criteria were updated by Martin et al⁹ to allow aspirin therapy to be included in the routine follow-up head CT category. This change did not affect outcomes adversely. We allowed patients on ASA (29.12%) to be triaged to the group without a routine follow-up head CT. This led to 85% of patients being triaged to the no-routine-follow-up CT group compared to 56% if we had not allowed ASA in the routine follow-up head CT group.

TABLE 3.

Overview of Anticoagulation Status in Patient Population

Antithrombotic medications	Number (%)
At least on antiplatelet or oral anticoagulation medication including ASAa	81 (45.5%)
ASA alone	52 (29.1%)
Oral anticoagulation (coumadin and NOACS)	30 (16.8%)
Antiplatelet other than ASA	8 (4.5%)
>1 antiplatelet or oral anticoagulation medication	9 (5.0%)

^aASA = acetylsalicylic acid (aspirin) therapy; NOACS = novel oral anticoagulants.

Limitations

We are aware that the event rate in cmTBI is low and that our sample size of 178 might not be large enough to capture such an event. We recognize that we are working at a level 1 trauma center with continuous in-house neurosurgery coverage should a patient deteriorate and need neurosurgical intervention. Still, this study should be seen as a stepping stone to a large multicenter study.

Interpretation

One of the most commonly cited reasons for routine follow-up head CT is to make sure that the patient is safe, but safe from what? We know that 20% of intracranial findings are worse on the follow-up head CT scan without resulting in treatment changes or surgery.² A stable routine follow-up head CT scan does not protect the patient's condition from deteriorating later. Our protocol advocates forgoing the use of routine follow-up head CT, but still observing the patients for 6 h and monitoring for neurological decline. The monitoring does not need to be done by neurosurgery. Other services such as emergency medicine or ACS can safely monitor the patient, as previous studies have shown.^10,11 The BIG developed by Joseph et al^12,13 represent one of the first protocols to decrease routine follow-up head CT scans in TBI patients. After implementation, there was a 29% decrease in routine follow-up head CTs, a 20% decrease in neurosurgery consultations, and no change in mortality. The neurosurgical intervention rate was unchanged and there was not a higher readmission rate after the protocol was implemented.¹⁴

A routine follow-up head CT scan does not result in treatment changes 99% of the time.³ In this pilot study, none of the patients had surgical intervention based on triage to routine follow-up head CT. We understand that with an event rate of 1% or less, our patient sample might be too small to capture such patients. Still, the practice of substituting clinical vigilance and proper neurological examination with a routine follow-up head CT scanning is misplaced.

Another important point in questioning the validity of the routine follow-up head CT is the evolution of imaging technology. Our ability to diagnose traumatic brain hemorrhages has increased tremendously, resulting in the detection of trivial and clinically insignificant findings on CT scans and overdiagnosis of cmTBI, over-triage, and unnecessary repeat CT scans. Over the past decade, there has been an 80% increase in the TBI-related use of health care in patients with mild TBI. With our ability to diagnose the smallest, mostly insignificant or trivial ICH after trauma, combined with the 80% increase in mild TBI patients presenting to the ED and an aging population, we need to examine our protocols and indications for routine follow-up head CT in cmTBI. As medicine evolves, the care we are providing needs to evolve as well.

Although our study is not a cost-effectiveness study, 1 head CT examination costs on average $1162 at our institution. A cost savings of $146 412 was achieved with this study at our institution alone in 1 yr. The average national head CT scan costs are much higher ($3275), which could result in even higher savings. Our study proves that not all cmTBIs require follow-up head CTs and that many do not even require admission; instead, they can be observed safely in the ED and then discharged home. This may have dramatic effects on health care costs by eliminating unnecessary hospital admissions and CT scans.

Generalizability

Our study protocol can be implemented easily in any ED. The triage criteria are readily available and allow for the majority of cmTBI patients to be triaged without routine follow-up head CT. Also, allowing patients on ASA to be included in the group forgoing routine follow-up head increased our generalizability.

This being said, academic medical centers, which should lead this change in triage, tend to be the slowest in adapting to changes in the status quo. The additional work burden that arises is absorbed mostly by the residents, and rarely by the faculty dictating care. Academic medical centers also have a responsibility to educate the next generation of physicians to feel comfortable caring for cmTBI patients.

CONCLUSION

A cmTBI protocol based on medical history, findings from the neurological examination, and initial CT scan of the head can be successfully implemented in a multi-disciplinary fashion in collaboration with neurosurgery, ACS, and emergency medicine to help decrease the number of routine follow-up head CT scans performed in patients with cmTBI. Utilizing this protocol, we were able to safely decrease the use of routine follow-up head CT scans in cmTBI patients by 71% without any missed injuries or delayed surgery. Adoption of the protocol was high among all services managing TBI patients.

Funding

This study did not receive any funding or financial support.

Disclosures

The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article. The content is solely the responsibility of the authors and does not necessarily represent the official views of Harvard Catalyst, Harvard University, and its affiliated academic health-care centers, or the National Institutes of Health.