Download Developing a Decision Instrument to Guide Computed Tomographic

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Transcript
The Journal of TRAUMA威 Injury, Infection, and Critical Care
Developing a Decision Instrument to Guide Computed
Tomographic Imaging of Blunt Head Injury Patients
William R. Mower, MD, PhD, Jerome R. Hoffman, MA, MD, Mel Herbert, MBBS, Allan B. Wolfson, MD,
Charles V. Pollack Jr., MA, MD, and Michael I. Zucker, MD, for the NEXUS II Investigators
Background: Computed tomographic
(CT) head scanning of blunt trauma patients is expensive, delays care, and necessitates radiation exposure, while detecting intracranial injuries in a minority of patients.
Clinical characteristics may be able reliably
identify patients who do not have intracranial injuries and consequently, do no require imaging.
Methods: Physicians assessed blunt
trauma patients undergoing imaging for
the presence or absence of specific criteria. Recursive partitioning was used to
identify criteria that predict intracranial
injuries with high sensitivity.
Results: Intracranial injuries were
found in 917 of 13,728 enrolled patients
(6.7%). Injuries were rare among patients
under age 65 who had no evidence of skull
fracture, scalp hematoma, neurologic def-
icit, abnormal alertness, abnormal behavior, coagulopathy, or persistent vomiting.
These characteristics would have identified 901 injury cases (sensitivity 98.3%
[CI: 97.2–99.0]), while classifying 1,752
patients (12.8%) as “low risk.”
Conclusions: Clinical characteristics
can reliably identify patients who are unlikely to have intracranial injuries and
who do not require CT imaging.
J Trauma. 2005;59:954–959.
U
nrecognized intracranial injury can produce permanent
brain damage, severe disability, and even death. Reports of “occult” injuries among patients with benign
clinical presentations have led many clinicians to obtain imaging on virtually all blunt head trauma patients. Consequently, nearly one million blunt trauma patients undergo
head CT imaging annually in the U.S.,1,2,3 but fewer than
60,000 (about 6%) prove to have significant intracranial
injuries.4,5,6
Recent work suggests that a clinical decision instrument
could be developed to identify reliably blunt head trauma
patients who have essentially no risk of significant brain
injury.7,8 The use of such an instrument could reduce CT
imaging and decrease radiographic charges, time in the ED,
and radiation exposure among blunt trauma patients.1,6,8 –11
In an attempt to derive such a decision instrument, we conducted this prospective study of blunt head trauma patients.
PATIENTS AND METHODS
Our methodology (NEXUS-II), a multi-center, prospective, observational study, has been described in detail
elsewhere.11 Briefly, NEXUS-II enrolled all blunt trauma
patients for whom head CT scanning was ordered by the
managing physician in 21 participating centers. Study participants represent a variety of facilities and were selected to
increase the study’s generalizability to the majority of trauma
victims evaluated in North American EDs.
Study Subjects
The study population consisted of all acute blunt head
trauma victims undergoing CT head imaging at participating
centers. Patients with a delayed presentation or without blunt
trauma were excluded. Clinicians made imaging decisions
based on their clinical judgment and not by study protocol.
Data Collection
Submitted for publication May 11, 2004.
Accepted for publication March 23, 2005.
Copyright © 2005 by Lippincott Williams & Wilkins, Inc.
From the UCLA Emergency Medicine Center (W.R.M., J.R.H.),
UCLA School of Medicine, Los Angeles, CA; Department of Emergency
Medicine (M.H.), KECK USC School of Medicine, LAC⫹USC Medical
Center; Department of Emergency Medicine (A.B.W.), University of Pittsburgh School of Medicine, Pittsburgh, PA; Department of Emergency Medicine (C.V.P.), Pennsylvania Hospital, Philadelphia, PA; Department of Radiology (M.I.Z.), UCLA School of Medicine, Los Angeles, CA.
This work was funded by Grant # RO1 HS09699, from the Agency for
Healthcare Research and Quality (AHRQ).
Address for reprints: William R. Mower, MD, PhD, UCLA Emergency
Medicine Center, 924 Westwood Blvd., Los Angeles, CA 90024; email:
[email protected]
DOI: 10.1097/01.ta.0000187813.79047.42
954
Patients were enrolled when a physician ordered CT
imaging. To maximize compliance, participating institutions
agreed to adopt a rigid protocol, described in detail
elsewhere,11 whereby CT scanning would not be performed
until study data had been collected and recorded.
Data Collected
Upon enrollment, clinicians collected limited demographic information (date and time of evaluation, age, sex,
and race), documented the presence or absence of each candidate variable (Table 1), and assessed each patient’s Glasgow Coma Scale (GCS) score. Candidate variables were
drawn from among those previously found to be reliably
assessable in the clinical environment.12
October 2005
CT Imaging for Blunt Head Injury Patients
Table 1 Physician Instructions, Guidelines and Medical Definitions
Each patient seen in the emergency department will receive medical care consistent with current practice standards (ATLS, ACLS, PALS,
etc). Study execution should not influence clinical decisions or care. The following definitions are for the study purposes only, and do not
represent recommendations for patient evaluation.
Eligible patients: Any victim with blunt head trauma is eligible for inclusion in this study. Every blunt trauma patient who undergoes
computed head tomography must be enrolled in the study.
It is not currently possible to reliably exclude intracranial injuries in blunt head trauma victims having any of the following: 1) post-traumatic
seizure, 2) loss of consciousness (particularly if longer than five minutes), 3) severe or progressive headache, 4) coagulopathy (whether
hereditary, drug induced or acquired), 5) abnormal behavior, 6) abnormal level of alertness, 7) signs of basilar or depressed skull fracture,
8) recurrent or forceful vomiting, 9) evidence of intoxication, 10) motor deficit, 11) gait abnormality, 12) cerebellar abnormality, 13) cranial
nerve abnormality, 14) inability to read and write name, 15) age 65 years or greater. Such patients are at risk for intracranial injuries and
should be managed appropriately.
Exclusions: Data forms should be completed on all patients undergoing computed head tomography. Patients undergoing tomography for
reasons other than blunt trauma may have additional indications for imaging studies. This would include patients being evaluated for
penetrating trauma, infections, cerebrovascular accidents, tumors, or any other atraumatic indications. Data will be collected on these
patients even though they will not be entered into the blunt head injury study.
Clinical variables: The following terms are defined for purposes of clarity and to ensure consistent data collection. They do not represent
recommendations for patient evaluation, and they should be considered subject to interpretation by individual physicians.
A post-traumatic seizure is any seizure following the traumatic event (witnessed either by examining physicians, or other reliable observer).
Loss of consciousness is based on the patients report of being knocked unconscious, or a witnesses report that the patient lost
consciousness, or did not response to external stimuli (including verbal stimuli, or physical stimuli such as prodding, shaking, pinching,
among others) for some interval following the event. Loss of consciousness longer than five minutes is based on a report by a reliable
witness (such as a paramedic, health care worker, or examining physician), that the patient lost consciousness for more than five minutes.
Severe or progressive headache is any head pain deemed by the patient to be severe or progressive in nature.
Coagulopathy is any impairment of normal blood clotting such as occurs in hemophilia, secondary to medications (coumadin, heparin,
aspirin, etc.), hepatic insufficiency and other conditions.
Abnormal behavior is any inappropriate action displayed by the victim. It includes such things as excessive agitation, inconsolability,
refusal to cooperate, lack of affective response to questions or events, and violent activity.
Abnormal level of alertness is evidenced by a variety of findings, including but not limited to a Glasgow coma score of 14 or less; delayed
or inappropriate response to external stimuli; excessive somnolence; disorientation to person, place, time or events; inability to remember
three objects at five minutes; perseverating speech; and other findings.
Significant skull fracture includes, but is not limited to, any signs of basilar skull fracture (periorbital or peri-auricular ecchymoses,
hemotympanum, and drainage of clear fluid from the ears or nose). Signs of depressed or diastatic skull fracture (a palpable step-off of
the skull, a stellate laceration from a point source, or any injury produced by an object striking a localized region of the skull [such as a
baseball bat, club, pool cue, golf-ball, baseball, pipe, etc.]).
High-risk vomiting is evidenced by recurrent, projectile or forceful emesis (either observed or by history) after trauma, or vomiting
associated with altered sensorium.
Evidence of intoxication includes: a) a history of intoxication or recent intoxicating ingestion is provided by a patient or observer; b) test of
bodily secretions (blood, urine, saliva, breath, etc) is positive for drugs or alcohol; c) patient has physical evidence suggesting intoxication
(odor of alcohol, slurred speech, ataxia, dysmetria or other cerebellar findings), or behavior consistent with intoxication and unexplained
by medical or psychiatric illness.
A motor deficit is a finding of abnormal weakness in any one or more of the four extremities, as determined by systematic testing of
muscle strength in all four limbs.
Gait abnormality is the inability to walk normally due to inadequate strength, loss of balance or ataxia; it is determined by systematic
testing of gait, including tandem and heel-to-toe walking, and Romberg testing.
Cerebellar abnormality is manifested by ataxia, dysmetria, dysdiadokinesis, or other impairment of cerebellar function; it is determined by
systematic testing of cerebellar function, including tests of ataxia, and finger-nose-finger, heel-to-shin, and rapid alternating movement
testing.
Cranial nerve abnormality is an abnormality of cranial nerves II-XII; it is determined by systematic testing of each of these cranial nerves.
The ability to read and write is determined by asking the patient to read the physician’s name from an identifying badge, or a written
piece of paper, and subsequent ability to write that same name.
Age 65 years or more is determined by available clinical history.
Neurologic deficit was considered to be present if a patient exhibited an abnormal Glasgow Coma Scale score, motor deficit, gait
abnormality, abnormal cerebellar function, or cranial nerve abnormality (patients were deemed to have a “neurologic deficit” if any
constituent variables was abnormal, and were considered “normal” only if all constituent variables were normal).
A composite variable, “neurologic deficit,” was constructed by combining the elements of the Glasgow Coma
Scale score with the presence or absence of motor deficit, gait
abnormality, abnormal cerebellar function, and cranial nerve
abnormality. Patients were considered to have a ‘neurologic
deficit‘ if any of constituent variables were abnormal, and
Volume 59 • Number 4
were considered “normal” only if all constituent variables
were normal.
Interpretation of Imaging Studies
All study patients underwent head CT scanning. Initial
imaging could be supplemented by other studies deemed
955
The Journal of TRAUMA威 Injury, Infection, and Critical Care
necessary. CT scans were interpreted by clinical radiologists
at each site. Copies of all final radiographic readings were
collected and abstracted to determine the presence or absence
of intracranial injuries. The diagnosis of intracranial injury
was based solely on final radiologic interpretations of all
imaging studies, including subsequent inpatient studies. Investigators determined final injury classification while
blinded to information about clinical variables (and thus to
the patient’s classification by possible decision instruments).
The definition of significant intracranial injury was
based on outcomes documented in previous clinical work,13
and consists of any injury that may require neurosurgical
intervention (craniotomy, intracranial pressure monitoring,
mechanical ventilation), or lead to rapid deterioration or significant long-term neurologic impairment.
the performance among patients with “minor” head injury
(defined as a cases presenting with GCS scores of 15).
Assessing Potential for Bias
Because we enrolled only patients who underwent CT
imaging and did not verify injury status among un-imaged
patients, our methodology is vulnerable to “verification” bias.
To assess the potential for verification bias, we conducted
three-month follow-up interviews with all blunt trauma patients presenting to our core facility who did not undergo CT
imaging. Each patient (or a surrogate family member) was
asked whether they had eventually undergone radiographic
imaging, been diagnosed with intracranial injuries, or ultimately required neurosurgical intervention.
Patient Consent
Intracranial Injuries Considered Significant:
Mass effect or sulcal effacement
Signs of herniation
Basal cistern compression or midline shift
Substantial epidural or subdural hematomas (greater the
1.0 cm in width, or causing mass effect)
● Substantial cerebral contusion (more than 1.0 cm in
diameter, or more than one site)
● Extensive subarachnoid hemorrhage
● Hemorrhage in the posterior fossa
● Intraventricular hemorrhage
● Bilateral hemorrhage of any type
● Depressed or diastatic skull fracture
● Pneumocephalus
● Diffuse cerebral edema
● Diffuse axonal injury
●
●
●
●
Data Analysis
Data from clinical evaluations were merged with head
CT scan results to form the final derivation data set. Cases
lacking radiographic reports were deleted from the final data
analysis.
Formulation of the Optimal Decision Instrument
We used ␹2 recursive partitioning to evaluate the importance of individual criteria, and to identify combinations of
criteria that could be used to form tentative decision instruments that predicted intracranial injuries with high sensitivity
and excluded intracranial injuries with high negative predictive value.14 We calculated the specificity of each tentative
instrument, and defined the instrument exhibiting the highest
specificity as the “optimal” decision instrument. By design,
this optimal instrument has very high sensitivity and NPV,
and the highest possible specificity. We employed exact
methods to determine the confidence intervals associated
with each operator characteristic (sensitivity, NPV and
specificity).15 Operator characteristics were calculated for the
performance of the instrument among all blunt head injury
victims, while separate calculations were made to determine
956
NEXUS-II was an observational study that did not mandate or direct any aspect of patient care and posed no risk to
patients. The study protocols, methodology and request for
waiver of informed consent were reviewed and approved by
the Federal Office for the Protection from Research Risks, as
well as by the institutional review board at each site. Waivers
of informed consent were granted to each participating institution.
Results
In total, we enrolled 13,728 patients in the study, including 917 patients diagnosed with clinically important intracranial injuries. This latter group included 330 patients who
presented with minor head injuries. The median age of the
blunt head trauma population was 37 years (inter-quartile
range [IQR]: 23 – 52); the median age of those with clinically
important intracranial injuries was 40 (IQR: 22 – 62) and the
median age for minor head injury victims with intracranial
injuries was 41 (IQR: 24 – 58).
The ethnic distribution of the overall population included
6,912 Whites (50%), 2,418 Blacks (18%), 1,985 Hispanics
(15%), 330 Asians (2%), 99 Middle Easterners (1%), 157
Native Americans (1%), and 1,827 others (13%). There were
8,988 males (66%) and 4,718 females (34%); gender was not
recorded in 22 patients (0.2%).
The 917 patients with clinically important intracranial
injuries included 528 Whites (58%), 118 Blacks (13%), 114
Hispanics (12%), 29 Asians (3%), 3 Middle Easterners
(0.3%), 5 Native Americans (0.5%), and 120 others (13%).
This population included 656 males (72%), 260 females
(28%), and 1 individual of unknown gender (0.1%).
Table 2 presents the prevalence of the individual criteria
in patients with and without intracranial injuries. Among all
patients, loss of consciousness, noted in 48%, was the most
frequently observed criterion, while only a small proportion
of subjects presented with evidence of significant skull fracture (4%). Among patients with intracranial injuries, neurologic deficit was the most frequently observed criterion
(noted in 65% of such patients), while seizure and coaguOctober 2005
CT Imaging for Blunt Head Injury Patients
Table 2 Prevalence of Individual Criteria among Blunt
Head Injury Patients
Seizure
Loss of consciousness
Prolonged loss of consciousness
Severe headache
Coagulopathy
Abnormal behavior
Abnormal level of alertness
Significant skull fracture
Persistent vomiting
Evidence of intoxication
Inability to read/write
Scalp hematoma
Extreme age
Neurological deficit
4%
48%
8%
17%
4%
22%
28%
4%
6%
24%
6%
33%
14%
28%
5%
63%
29%
12%
5%
43%
64%
21%
10%
19%
19%
47%
19%
65%
lopathy were the least frequently observed (present in 5%). In
contrast to patients without intracranial injuries, those with
injuries were much more likely to present with evidence of
significant skull fracture (21% versus 3%).
Recursive partitioning identified eight criteria that were
independently and highly associated with intracranial injuries. These include: 1) evidence of significant skull fracture,
2) scalp hematoma, 3) neurologic deficit, 4) altered level of
alertness, 5) abnormal behavior, 6) coagulopathy, 7) persistent vomiting, and 8) age 65 years or more. The remaining
variables failed to show a significant association with intracranial injuries under any partitioning scheme.
A decision instrument based on these eight criteria correctly identified 901 of the 917 patients with clinically important intracranial injuries (sensitivity 98.3 [CI: 97.2–
99.0]), while classifying 1,752 patients as “low risk” (negative predictive value [NPV] 99.1% [CI: 98.5–99.5]). Specificity was 13.7% [CI: 13.1–14.3]. Table 3 describes the patients and injuries that would have been missed by this
combination of criteria. Most of these “false negative” patients had relatively minor injuries, and while most were
admitted for observation, only one required immediate neurosurgical intervention (ICP monitoring).
Among patients with minor head injury, the decision
instrument correctly identified 314 of 330 with clinically
important intracranial injuries, while correctly classifying
1,752 as “low risk” (sensitivity 95.2% [C.I. 92.2–97.2], NPV
99.1% [C.I. 98.5–99.5], and specificity 17.3% [C.I. 16.5–
18.0].
Several other decision instruments were constructed,
which had slightly less useful test characteristics (Table 4). In
each case, these alternate combinations of clinical variables
could not achieve equivalent sensitivity without further sacrificing specificity.
We identified 2,397 patients who had sustained blunt
head trauma, but did not undergo emergent head imaging,
including 1,266 patients (52.8%) who agreed to participate in
the long-term outcome study. CT scanning was ultimately
performed in 27 of these patients (2.1%), magnetic resonance
imaging in 29 (2.3%), and skull radiography in 14 (1.1%). No
significant intracranial injuries were found in any of these
patients. There were no cases of missed intracranial injuries
among the follow-up cohort, no subsequent hospitalizations
or neurosurgical interventions to treat intracranial injuries,
and no deaths due to intracranial injuries.
DISCUSSION
Decision instruments are frequently used to screen patients for radiographic imaging. An ideal instrument recommends imaging for all patients who have a significant injury
(high sensitivity), guarantees that injury is absent in patients
not selected for imaging (high NPV), and limits imaging
among patients who do not have injury (high specificity). To
be clinically useful, however, such an instrument must also
Table 3 Intracranial Injuries Not Identified By an Optimal Combination of Criteria. Patient Injuries
Age (yrs)
Race/Sex
Injuries
64
53
White/male
White/male
51
48
43
41
40
28
26
23
23
23
22
19
8
7
White/female
Other/female
Black/male
White/female
White/male
White/male
White/male
White/male
White/male
Hispanic/male
White/male
White/male
Black/male
Other/female
Small subdural hematoma in the middle cranial fossa. Small amount of subarachnoid blood.
Hemorrhage in the left frontal base, probably a combination of epidural and parenchymal hemorrhages. Sliver of
subdural hemorrhage in the left anterior convexity.
High left medial parietal convexity parafalcine focus of contusion.
Left temporal cortical hemorrhages. Mild diffuse brain swelling.
Diffuse edematous changes within the cerebral hemispheres bilaterally slightly greater on the left than the right.
Subdural/subarachniod hemorrhage with dilated lateral and third ventricles and sulcal effacement.
Occipital skull fracture with small right occipital collection that is epidural or subdural blood.
Comminuted left temporal bone fracture with subjacent air.
Comminuted, depressed fracture, frontal bone, with subjacent right frontal lobe contusion and pneumocephalus.
Frontal sinus fracture with air and probable blood near the anterior left falx, just posterior to the frontal bone.
Small focus of hemorrhage in the right frontal lobe deep white matter
Right subdural hematoma.
Left subdural hematoma with parasagittal component.
Minimal pneumocephalus. Very small subarachnoid hemorrhage lining the posterior falx.
Diastasis of the left lamboidal suture which is wider than the right.
Right sided occipito-temporal hemorrhagic contusion.
Volume 59 • Number 4
957
The Journal of TRAUMA威 Injury, Infection, and Critical Care
Table 4 Operator Characteristics of Alternate Decision Rules Formed By Adding Additional Elements to the
Optimal Combination of Criteria
Additional Element(s)
Sensitivity
NPV
Specificity
PPV
Seizure
Loss of consciousness
Prolonged unconsciousness
Headache
Seizure, loss of consciousness, headache
Seizure, prolonged unconsciousness, headache
98.5 (97.5–99.2)
99.3 (98.6–99.8)
98.9 (98.0–99.5)
98.8 (98.0–99.5)
99.8 (99.2–100)
99.6 (98.9–99.9)
99.1 (98.5–99.5)
99.0 (97.8–99.6)
99.2 (98.5–99.6)
99.2 (98.5–99.6)
99.4 (97.9–99.9)
99.5 (98.7–99.9)
12.3 (11.8–12.9)
4.5 (4.2–4.9)
9.4 (8.9–9.9)
10.2 (9.7–10.8)
2.7 (2.4–3.0)
6.2 (5.8–6.6)
7.5 (7.0–8.0)
6.9 (6.5–7.4)
7.2 (6.8–7.7)
7.3 (6.9–7.8)
6.8 (6.4–7.3)
7.1 (6.6–7.5)
offer some advantage over clinical judgment in selecting
patients for imaging.
The decision instrument developed in this study includes
eight criteria, and may at first glance appear somewhat cumbersome. However, it has excellent construct validity, and
suggests that intracranial injuries are rare in patients under
age 65 who do not exhibit signs of significant trauma to the
calvarium (as manifested by evidence of skull fracture or
scalp hematoma), neurologic dysfunction (including focal
neurologic deficits, altered level of alertness, or abnormal
behavior), persistent vomiting, or coagulopathy. A decision
instrument based on these criteria would have identified
nearly all of the patients with intracranial injuries in this
study, and would have produced a modest decrease in CT
imaging. We were unable to identify any other set of variables that performed as well or better.
Our findings also suggests that clinicians are already
adept at identifying patients who require CT head imaging,
and it appears that any decision instrument is likely to provide
only marginal benefit over current clinical practice. Instruments exhibiting higher sensitivity inevitably suffer from
poorer specificity and are less efficient at limiting imaging,
while instruments that could significantly reduce imaging do
so at the cost of unacceptable rates of missed injury.
Many previous efforts to develop screening tools for
traumatic brain injury have had difficulty attaining the sensitivity required for clinical application.1,2,4 – 6 In contrast,
instruments recently developed by Haydel7 and Stiell8 exhibit
adequate sensitivity, but suffer in other ways that limit their
utility. First, both instruments apply only to patients who
have sustained loss of consciousness (while Stiell also enrolled patients with amnesia and disorientation, these criteria
were subsequently found to exclude patients from low-risk
classification), and neither provides guidance for patients
who have not lost consciousness. Second, these instruments
used surrogate markers, rather than validated clinical measures, to determine patient outcomes, limiting the clinical
relevance of the final rule. Furthermore, the rule developed
by Haydel mandates imaging for all patients with evidence of
trauma above the clavicles, including those with small contusions, abrasions, lacerations and minor facial injuries.7
While this criterion helps to ensure adequate sensitivity, it
requires imaging for many patients who would otherwise be
958
cleared on the basis of clinical judgment, and is likely to
result in increased imaging at many centers.8
Clinical decision instruments are unlikely to be used if
they are complex or cumbersome. The rule proposed by
Stiell8 is very complex, and it also requires close observation
of head trauma patients for two hours before they can be
judged low-risk. Such a period of observation is likely to
prove difficult in busy trauma centers; in addition, it might
lead to harm for that patient who fails to improve during this
period, and is found on subsequent CT to have a lesion
requiring neurosurgery.16 Consequently, we suspect many
clinicians would abandon this rule in favor of clinical judgment, or worse, rote imaging.
Several factors must be considered in assessing the benefits of a decision instrument over clinical judgment. Reduced imaging decreases radiographic charges, ED waiting
times, and exposure to potentially dangerous ionizing
radiation.11,17,18,19 Decision instruments may also offer other
less quantifiable benefits, including the ability to standardize
care and inform medicolegal practice. However, the use of a
decision instrument is not without cost. If the instrument in
this study is validated, it would be expected to miss approximately 1.7% of cases with “clinically important” intracranial
injuries. Virtually all such missed injuries, however, are relatively minor and should not result in serious immediate
consequence. Few of these patients require hospitalization or
neurosurgical intervention, but all require careful instructions
regarding signs and symptoms that could indicate the need for
additional evaluation. Such “head injury precautions” should
be expected to further reduce the possibility of a serious
adverse outcome among patients considered to be “low-risk,”
in whom an initial CT scan would not be done. Follow-up for
such patients is important even in the absence of this concern,
because post-traumatic sequelae occur in many head injury
patients, so the “safety net” of careful follow-up does not
represent an additional requirement mandated by use of this
instrument.
Since our study enrolled only individuals who underwent
imaging, this instrument would almost certainly exhibit
higher specificity when applied to the population of all blunt
head trauma patients. This is because clinicians are more
likely to omit imaging among patients who do not exhibit
worrisome clinical findings, such as our criteria. At the same
October 2005
CT Imaging for Blunt Head Injury Patients
time, because there were no missed injuries among un-imaged patients, the sensitivity of our decision instrument, if
applied to the entire blunt head trauma population, is likely to
be no worse than reported here.
It is important to recognize that our decision instrument
is intended to identify patients who are unlikely to have
significant injuries revealed by CT imaging. While most of
these patients can probably be safely discharged following
their evaluation, it is beyond the scope of this study to
determine their ultimate management. Physicians will need to
use clinical judgment in deciding subsequent care.
Our study has some important limitations. Although recursive partitioning is very efficient in identifying criteria to
produce an optimal rule among a derivation cohort, this does
not guarantee equivalent performance among a second cohort. Thus while the sheer size of our study is reassuring, and
provides the best information to guide imaging decisions to
date, our results should be validated in an independent cohort.
In addition, the criteria used in this instrument are relatively
subjective. Nevertheless, they can be consistently assessed by
different observers,12 and appear to identify reliably patients
with intracranial injuries, as suggested by the current study.
Finally, we were unable to derive any instrument with adequately high sensitivity for important intracranial injuries that
also exhibited specificity dramatically better than current
clinical practice.
ACKNOWLEDGMENTS
Acknowledgment: The authors wish to thank Guy Merchant, NEXUS
Project coordinator, for his outstanding contributions to the project, as well
as the house officers and attending physicians at each of the study sites,
without whose cooperation and hard work the study would not have been
possible. The following centers and investigators collaborated in this study.
Principal Investigator: W. Mower. Co-Investigators: J. Hoffman, M. Herbert.
Steering Committee: M. Herbert, J Hoffman, W. Mower, C. Pollack, A.
Wolfson, M. Zucker. Site Investigators: Boston University, Boston Medical
Center (Boston,MA): N. Rathlev and G. Barest; Case Western Reserve
University, Metrohealth System (Cleveland, OH): RK. Cydulka and B.
Karaman; Cooper Hospital, University Medical Center (Camden,NJ): C.
Terregino, A. Nyce and S. Ross; Emory University Medical Center (Atlanta,
GA): D. Lowery and S. Tigges; Hennepin County Medical Center (Minneapolis, MN): B. Mahoney and J. Hollerman; Louisana State University
Medical Center, Charity Hospital (New Orleans,LA): M. Haydel and E.
Blaudeau; Maricopa Medical Center (Phoenix,AZ): C. Pollack and M. Connell; Ohio State University Medical Center (Columbus, OH): DR. Martin and
C. Mueller; Stanford University Medical Center (Stanford,CA): SV. Mahadevan and G. Arabit; State University of New York at Stonybrook (Stonybrook,NY): P. Viccellio and S. Fuchs; University of Calgary, Foothills
Hospital (Calgary, Canada): G. Lazarenko and C. Fong; University of California, Davis, Medical Center (Sacramento,CA): J. Holmes and RA McFall;
University of California, Irvine (Irvine,CA): J. Krawczyk; University of
California, Los Angeles, Center for the Health Sciences (Los Angeles,CA):
J. Hoffman and M. Zucker; University of California San Diego Medical
Center Hillcrest,(San Diego,CA): D. Guss and D. Meltzer; University of
California, San Francisco at Fresno, University Medical Center (Fresno,CA):
G. Hendey and R. Lesperance; University of Cincinnati, University Hospital
(Cincinnati,OH): GJ. Fermann and HH. Hawkins; University of Missouri,
KS City, Truman Medical Center (Kansas City, MO)): E. Westdorp and S.
Go; University of Pennsylvania Medical Center (Philadelphia, PA): J Hol-
Volume 59 • Number 4
lander; University of Pittsburgh Medical Center (Pittsburgh, PA): A. Wolfson and J. Towers; Vanderbilt University Medical Center (Nashville, TN): V.
Norton.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
Borczuk P. Predictors of intracranial injury in patients with mild
head injury. Ann Emerg Med. 1995;25:731–736.
Madden C, Witzke DB, Sanders AB, Valente J, Fritz M. High-yield
selection criteria for cranial computed tomography after acute
trauma. Acad Emerg Med. 1995;2:248–253.
Subcommittee on Health for Families and the Uninsured, Committee
on Finance, U.S. Senate. Emergency Departments Unevenly Affected
by Growth and Change in Patient Use. Washington, D.C.: United
States General Accounting Office, 1993, pp 4.
Jeret JS, Mandell M, Anziska B, et al. Clinical predictors of
abnormality disclosed by computed tomography after mild head
trauma. Neurosurgery. 1993;32:9–16.
Schynoll W, Overton D, Krome R, et al. A prospective study to
identify high yield criteria associated with acute intracranial
computed tomography finding in head-injured patients. Am J Emerg
Med. 1993;11:321–326.
Miller EC, Derlet RW, Kinser D. Minor head trauma: Is computed
tomography always necessary? Ann Emerg Med. 1996;27:290–294.
Haydel MJ, Preston CA, Mills TJ, et al. Indications for computed
tomography in patients with minor head injury. N Engl J Med. 2000;
343:100–105.
Stiell IG, Greenberg GH, Worthington JR, et al. The Canadian CT
head rule for patients with minor head injury. Lancet. 2001;
357:1391–1396.
Levitt MA. Minor head injury. Ann Emerg Med. 1994;23:1143–1144.
Stiell IG, Wells GA, Vandemheen K, et al. Variation in ED use of
computed tomography for patients with minor head injury. Ann
Emerg Med. 1997;30:14–22.
Mower WR, Hoffman JR, Wolfson AB, Pollack CV, Herbert M,
Zucker MI. Selective computed tomography in blunt head injury:
Methodology of the National Emergency X-Radiography Utilization
Study II. Ann Emerg Med. 2002;40:505–514.
Hollander J, Go S, Lowery D, Wolfson AB, Pollack CV, Herbert M,
Hoffman JR, Mower WR. Inter-rater reliability of clinical findings
used in assessing the need for emergent computed head tomography.
Acad Emerg Med. 2003;10:830-835.
Atzema C, Mower WR, Hoffman JR, Holmes JF, Killian AJ, Oman
JA, Shen AH, Greenwood SD. Defining “Therapeutically
inconsequential” head CT findings in patients with blunt head
trauma. Ann Emerg Med. 2004:44;47-56.
Hair JF, Tatham RL, Anderson RE, Black W. Multivariate Data
Analysis, 5th Ed. Prentice-Hall, Englewood Cliffs, N.J., 1998.
Young KD, Lewis RJ. What is confidence? Part 2: detailed
definition and determination of confidence intervals. Ann Emerg
Med. 1997;30:311–318.
Seelig JM, Becker DP, Miller JD, Greenberg RP, Ward JD, Choi
SC. Traumatic acute subdural hematoma: major mortality reduction
in comatose patients treated within four hours. N Engl J Med. 1981;
304:1511–1518.
Brenner DJ, Elliston CD, Hall EJ, Berdon WE. Estimated risks of
radiation-induced fatal cancer from pediatric CT. Am J Roertgerol.
2002;176:289–296.
Hall P, Adami HO, Trichopoulos D, Pederson NL, Lagiou Pagona,
Ekbom, A, Ingvar M, Lundell M, Granath F. Effect of low doses of
ionizing radiation in infancy on cognitive function in adulthood:
Swedish population based cohort study. BMJ. 2004;328:19–23.
National Research Council, Committee on the biological effects of
ionizing radiations (BIER V). Health Effects of Exposure to Low
Levels of Ionizing Radiation. Washington, D.C.; National Academy
Press, 1990.
959