Do All Patients With Mild Traumatic Brain Injury Require Admission?

 

Authors:  Zachary Il’Giovine, Damon Campbell, Melissa Whitmill, Ronald Markert, Jonathan M. Saxe

 

Abstract

Introduction: Strategies to curtail the increasing cost to hospitals while preserving the quality of care has come to the forefront of protocol design.  When hospitals appropriately shorten length of stay, unnecessary harm is avoided.  The purpose of this study was to determine if admission is necessary for a patient whose Glasgow Coma Scale (GCS) remains unchanged over 6 to 8 hours of observation.  Methods: We conducted a retrospective review of the trauma registry of our level one trauma center.  Data obtained included age, sex, Injury Severity Score (ISS), mechanism of injury, serial GCS, injury type, computed tomography(CT) results, hospital length of stay, and discharge location.  Results:  We identified 2009 patients with TBI admitted over a three-year period. Mild TBI (GCS 13-15) was identified in 1496 (75%) of these patients, and the most common mechanism of injury were falls (88%).  Patients age 65 and older were considered elderly, 933 of which were identified.  Of the 1390 patients in the mild category on both their first and second GCS, 76% showed no change on their repeat CT and 18% showed worsening repeat CT’s.  The mean length of stay for patients with a persistently mild GCS score was 5.20±6.56 days.  Of the 251 patients who showed worsening repeat CT despite consistent GCS scores, 233 (93%) were discharged without intervention, 5 (2%) had surgical intervention, and 5 (2%) died.  Conclusion:  This study supports the opinion that admission is not required with mild TBI and unchanged GCS over a 6 to 8 hour period, and that observation followed by discharge with home supervision is appropriate unless clinical examination suggests the need for admission or additional imaging.

 

Introduction

Mild traumatic brain injury (TBI), especially in elderly patients, has increased markedly over the last
decade.[1]  Concurrently, the affordability of healthcare has become increasingly uncertain.  Strategies to curtail cost while preserving quality has come to the forefront of protocol design.  When hospitals appropriately shorten length of stay, unnecessary risks and harms can be avoided. The current protocol at our hospital for a patient with TBI requires a preliminary computed tomography (CT) scan, admission to a critical care unit for observation with serial Glasgow Coma Scale (GCS) recordings, platelet transfusion for any patient on aspirin or clopidogrel, and a second CT scan in 8 to 12 hours.  Typically, patients who present with TBI at our institution are admitted for multiple days.

In a previous study, we found that repeat CT scanning is not necessary for every patient who presents with mild TBI, and instead patients can be observed for neurologic decline with serial clinical examinations and GCS recordings. Previous studies performed at other institutions have made similar conclusions.[2-3] Our belief was that by reducing the number of unnecessary CT scans, which is likely the reason many patients with mild TBI are admitted, the average length of stay could also be reduced.  This would significantly decrease the costs associated with longer hospital stays and abate unintended harm from unnecessary tests or procedures.

 

Methods

We used the trauma registry at our community-based, academically-affiliated hospital with a Level I Trauma Center to conduct a retrospective review of patients admitted for TBI from January 2009 through December 2011. Patients were categorized as mild (GCS 13-15) or severe/moderate (GCS 3-12) and assessed for serial GCS scores and repeat CT scanning.  Other data collected included age, sex, Injury Severity Score (ISS), hospital length of stay, mechanism of injury, and discharge location.

Means and standard deviations are reported for continuous variables, and counts and percents are reported for categorical variables.   The independent samples t test was used to compare the mild and severe GCS groups on continuous variables while the chi square test or Fisher’s Exact Test was used for group comparisons of categorical variables.  Inferences were made at the 0.05 level of significance with no corrections for multiple comparisons.  Analyses were conducted using IBM SPSS Statistics 21.0 (IBM, Armonk, NY).  The study was approved by our  Institutional Review Board. 

 

Results

During the three-year period (2009-2011) 2009 patients were admitted for TBI, 1897 (94.4%) of which had both a preliminary and follow-up GCS recorded.  Of the 1897 with at least two GCS scores, 1496 (79%) had mild TBI (GCS 13-15) and 401 (21%) had severe/moderate TBI (GCS 3-12) at presentation. The severe/moderate TBI group was younger than the mild TBI group (49.1±22.8 vs. 62.3±21.8, p < 0.001) and more likely to be male (69.8% vs. 59.2%, p < 0.001).  Of the patients admitted for TBI, 933 were aged 65 or older. 

As expected, the severe/moderate TBI group had a higher mean ISS (25.0±10.0 vs. 15.1±7.2, p < 0.001) and longer mean hospital length of stay (14.4±13.9 vs. 5.8±7.5,

p < 0.001).  The most frequent mechanism injury was falls: severe/moderate TBI group (38.7%) and mild TBI group (65.0%).  For the severe/moderate TBI group 28.3% died in the hospital while 28.8% were discharge to a rehabilitation facility, 20.5% to a nursing home, 18.4% home, and the remaining 4% elsewhere. For the mild TBI group most were discharge home (59.0%) while 24.8% were discharged to a nursing home and 8.0% to a rehabilitation facility.  In the mild TBI group 2.7% died in the hospital.     

Patients were also assessed for anticoagulation, specifically if they were taking aspirin, clopidogrel, warfarin, or Aggrenox (an aspirin/dipyridamole combination). 

Table 1 shows that the mild TBI group was more likely to be taking aspirin (32.3% vs. 14.2%, p < 0.001) and clopidogrel (12.5% vs. 5.7%, p <0.001) while the two groups did not differ in their use of warfarin or Aggrenox.

Among the total of 2009 TBI patients, 1308 (65.1%) had no change on their repeat CT, and 429 (21.4%) were worse.  A smaller number of patients either went to surgery (108, 5.4%) or died (50, 2.5%) before a repeat CT could be performed.  One hundred and fourteen (5.7%) did not have a repeat CT ordered for some other reason. 

In addition, of the 1897 patients with GCS recordings at presentation and subsequently throughout their hospital course, 1746 (92.0%) did not change from their initial mild or severe/moderate category (severe = 356 patients; mild = 1390 patients).  Thirty-seven initially severe/moderate TBI patients improved to the mild category (45 of 401 or 11.2%) while 106 initially mild TBI patients regressed to the severe category (106 of 1496 or 7.1%). Of the 106 patients with mild TBI at admission who regressed to severe TBI, 41 (38.7%) had no change on their repeat CT, 42 (39.6%) were worse, and 16 (15.1%) went to surgery before a repeat CT could be collected, and 1 (0.9%) died prior to repeat CT being collected.

Of the 1390 patients who remained in the mild category from presentation to discharge, 1063 (76.5%) showed no change on their repeat CT, 251 (18.1%) showed worsening repeat CT’s, and 38 (2.7%) required surgical intervention while 1 (0.1%) died before a repeat CT could be collected. Of the 251 (18%) of patients who showed worsening repeat CT’s despite a consistently mild GCS score, 233 (92.9%) were discharged to home (138 or 55%), a nursing home (71 or 28.3%, or a rehabilitation facility (24 or 9.6%).  Of the remaining 7.1%, 5 (2%) patients required surgical intervention despite a consistent GCS, and 5 (2%) patients died. Of the patients persistently in the mild category, the average length of stay was 5.20±6.56 days.

Finally, of the 356 patients in the severe/moderate TBI category throughout the course of their hospital stay, 132 (37.1%) had no change on their repeat CT, 118 (33.1%) were worse, and 42 (11.8%) went to surgery.  Sixty patients (16.9%) did not have a repeat CT.  Of the 45 patients with severe/moderate TBI at admission who improved to mild TBI, 33 (73.3%) had no change on their repeat CT, 5 (11.1%) were worse, and 5 (11.1%) went to surgery before a repeat CT scan was done.  None died and 2 (4.4%) did not have a repeat CT.

 

Discussion

It is common at most institutions for any patient who suffers a TBI to be admitted to a critical care unit so that serial repeat CT scans can be obtained.  It is reasonable, however, to question the benefit of admitting patients who suffer mild TBI in the absence of clinical indication.  Acquiring repeat imaging regardless of clinical picture is a risk-averse practice that results in extended admissions to accommodate needless testing. Washington and Grubb concluded that patients with mild TBI do not require ICU admission.[4] While their study  specified subarachnoid hemorrhage, small convexity contusion, small intraparenchymal hemorrhage, or small subdural hematoma, the conclusions made are applicable to our patients.  Intracranial injury type was not specified, but positive repeat scanning, a GCS of 13-15, and the absence of clinical decline were all present. This supports the opinion that admission is not required in mild TBI when GCS is unchanged over a 6 to 8 hour period.

 It was just over 10 years ago the Centers for Disease Control and Prevention called TBI the “silent epidemic”.[5]  Thompson et al added that “within this silent epidemic, there is a seemingly silent population: older adults”.[1]  The trend over past decade suggests that elderly patients with TBI are a distinct patient population that needs to be approached in a unique way.  It is also prudent to recognized that higher proportions of elderly patients are prescribed anticoagulation medication like warfarin, which predisposes to more frequent and more severe head injuries.[6]  It follows then that minor traumas which would be otherwise inconsequential can lead to TBI in anticoagulated elderly patients.  The high prevalence of anticoagulation in our sample, and the finding of no difference between mild and severe patients on warfarin may be explained by an exacerbation of minor trauma that would otherwise not result in TBI due to a hypocoagulable state. Likewise, trauma that would normally result in mild TBI without the presence of anticoagulation is made worse when the patient is medicated.

A concurrent study performed at our institution found that repeat CT scan alone did not contribute to patient care decision-making.  All but five patients who required intervention had an accompanying clinical decline as measured by GCS.  An additional five patients died despite GCS findings, but these were more likely associated with comorbid conditions or other injuries unrelated to the patients’ head trauma.  The nearly 93% of patients who did not require further intervention despite worsening imaging is consistent with Velmahos et al, who found that in no case was ordering a repeat head CT clinically helpful for patients with minimal head injury. Further, they reported that the 4% of patients in their study who did require treatment exhibited clinical deterioration that would have prompted repeat imaging anyway.[2]This, along with the findings of Washington and Grubb[4] suggest that for patients with mild TBI, clinical observation can be used in lieu of admission and repeat CT scanning.

Admitting a patient or extending their stay in the hospital to accommodate multiple CT scans can result in unnecessary transport, radiation exposure, and allocation of personnel.  While some resources (e.g., additional staffing or time spent in the CT scanner) can be difficult to quantify, the risks of transporting patients with TBI are recognized and can lead to adverse outcomes.[7] Additionally, extended hospital stays that include multiple repeat scans can expose patients significant levels of radiation. One head CT scan subjects patients to approximately 3-7 rads,[8] which is noteworthy over the course of multiple studies. 

Along with the obvious risks to the patient stated above, the financial burden of admission and longer stays is very tangible.  At our institution, the cost for a room in the advanced care unit for a single night is $3,739.[9]  For a room in the intensive care unit, the cost is closer to $4,305.[9]  Even if it was found that observation for 6 to 8 hours was inadequate, by decreasing the length of stay to just 1 day rather than the current average of 5.23 days, the total cost would be reduced by approximately $22-$26 million over a three year period.  More so, this is an underestimation as it does not take into account additional labs, scans, medications, oxygen, staffing, or other services required during an admission. It is also only for a single hospital.  If this were expanded to all healthcare institutions, the savings would be astronomical.  By revamping the admission protocol for these patients, spared resources could be directed towards other areas of a strained healthcare system.  With this in mind, it is our suggestion that the current institutional protocol be revised to allow for serial clinical examination during an observation period of 6 to 8 hours. Clinical decline could then be used to guide admission or the need for additional imaging.  We believe that 6 to 8 hours is suitable for observation, however further study is needed to determine the exact length of an adequate period. Furthermore, we suggest that these patients only be released under the supervision of family members or friends, with instruction to return to the Emergency Department if they appear to decline.

Our study had limitations.  By its design, a retrospective analysis is confined to the data available in the record.  Thus, any pertinent information that was either not recorded or occurred after discharge was unavailable.  While attempts were made to be as thorough as possible in collecting all of the interventions and outcomes for each patient, it is difficult to interpret the full extent to which imaging affected decision-making and admission in conjunction with the clinical picture.

 

Conclusion

When evaluating patients with mild traumatic brain injury, many hospitals include repeat CT scanning after 8 to12 hours of observation as part of their care protocol. A concurrent study we performed supports the notion that routine CT scanning in patients admitted with mild TBI and unchanged GCS does not appear to contribute to the care of these patients.  In an effort to avoid undue risk to patients with and misallocation of healthcare resources, we suggest an observation period of 6 to 8 hours followed by discharge with home supervision unless clinical examination suggest the need for admission or additional imaging in the setting of mild TBI.

 

References

[1] Thompson H, McCormick W, Kagan S. Traumatic Brain Injury in Older Adults:

Epidemiology, Outcomes, and Future Implications. J Am Geriatr Soc. 2006 October; 54(10): 1590-1595.

[2] Smith J, Chang E, et al. The Role of Early Follow-Up Computed Tomography Imaging in the Management of Traumatic Brain Injury Patients with Intracranial Hemorrhage. J. Trauma, 2007; 63:75-82

[3] Velmahos G, Gervasini A, et al. Routine Repeat Head CT for Minimal Head Injury is Unnecessary. J. Trauma, 2006; 60:494-501

[4] Washington C, Grubb R. Are routine repeat imaging and intensive care unit admission necessary in mild traumatic brain injury? J Neurosurg. 2012 Mar; 116(3):549-57

[5] Centers for Disease Control and Prevention. Report to Congress on Mild Traumatic Brain Injury in the United States: Steps to Prevent a Serious Public Health Problem; Atlanta, GA: 2001

[6] Lavoie A, Ratte S, Clas D, et al.  Preinjury warfarin use among elderly patients with closed head injuries in a trauma center. J Trauma, 2004 April; 56(4):802-7

[7] Insel J, Weissman C, Kemper M, et al. Cardiovascular Changes During Transport of Critically Ill and Postoperative Patients. Crit Care Med, 1986; 14:539-542

[8] Nickoloff E, Alderson P. "Radiation Exposures to Patients from CT: Reality, Public Perception, and Policy," (Commentary). American J. Roentgenology, 2001 Aug;Vol.177: 285-287.

[9] Miami Valley Hospital, Pricing, Billing & Insurance Guide (2013). Retrieved from: http://www.miamivalleyhospital.org/mvhdefault.aspx?id=12200. Accessed 3/3/2014

 

Contact

*Department of Surgery, Wright State University Boonshoft School of Medicine
One Wyoming Street, Suite 7819 WCHE
Dayton, OH 45409-2793 USA

 ** Corresponding Author:
Department of Surgery, Wright State University Boonshoft School of Medicine
One Wyoming Street, Suite 7819 WCHE
Dayton, OH 45409-2793 USA
Tele: 937. 208. 8322
Fax: 937. 208. 2105
Email: amnewsome@mvh.org

 

Editor’s note: The views and opinions expressed in the articles contained in the International NeuroTrauma Letter are those of the authors and contributors alone and do not necessarily reflect the views, policy or position of the International Brain Injury Association or all members of the INTL Editorial Board. The INTL is provided solely as an informational resource and the inclusion of any particular article does not establish or imply IBIA’s endorsement of its contents.