Age and outcome in head injured

patients
Objective: To investigate the relationships between age and outcome in head
injured patients.
Methods: A prospective study of all patients with a head injury admitted to the
regional centre at Newcastle General Hospital, UK between 1987 and 2000
(excluding 1989 and 1991) has been carried out. Data collected included cause of
injury, age and gender of patient, GCS at admission, presence of haematoma,
whether surgery was performed and outcome at discharge using GOS.
Results: During the 12 years of the study a total of 7154 patients were
admitted: 28% were aged under 10, 20% 10-19, 15% 20-29, 10% 30-39 and
26% 40 or over. The age span was from 8 days to 99 years (median 21 years)
and three quarters were male. GCS was 15 for 54%, 17% had a mild injury (GCS
13 & 14), 9% a moderate injury (GCS 11 & 12) and 19% a severe injury (GCS
8). One fifth had a haematoma of which 11% were subdural, 6% extradural and
4% ICH. Over two thirds (70%) made a good recovery by the time of discharge
or transfer from the regional centre while 7% died or were vegetative. The main
causes of head injury were a fall (49%) road traffic accident (24%) and assault
(16%). During the period of the study there was little change in the age profile of
patients admitted in each year. Older patients were more likely to be female and
to have a subdural or intracerebral haematoma that required surgery. Their GCS
on admission tended to be lower than younger patients and the cause of their
head injury was more likely to be from a fall. Age, initial GCS and presence of a
haematoma were significant factors in relation to outcome.
Conclusions: The profile of head injury admissions over the 14 years of this
study has remained constant in relation to age, injury severity and mortality. The
unfavourable outcome for older patients remains even after adjusting for their
lower admission GCS and subdural or intracerebral haematomas.

Publication
Gregson BA, Stobbart L, Ogungbo B, Chambers IR, Mendelow AD. Relationships
between age and outcome in head injured patients in Newcastle. Euroacademia
Multidisciplinaria Neurotraumatologica in Graz, Austria, 21-23 May, 2003.

Elderly patients with severe head

injury
Advancing age is known to be a determinant of outcome in head injury. We have
sought to discover whether there has been any change in the outcome of elderly
patients with severe head injury in Newcastle, where these patients have
continued to be treated with maximum intervention.
A review of prospectively collected data from the Newcastle Head Injury Database
for the period 1990 to 2000 was carried out. All patients aged 70 years and above
who had sustained a severe head injury (Glasgow Coma Score of 8 or less from
the outset) were included. The Glasgow Outcome Score (GOS) was determined at
6 months. Seventy-one patients were identified. Fifty-seven (80%) died and 2
(3%) were in a vegetative state, 11 (16%) had severe disability, 1 (1%) had
moderate disability and no patients made a good recovery. The natural history of
this condition remains unchanged and due consideration should be given to this
when evaluating interventions for elderly patients with a severe head injury.

Publication
Ushewokunze S, Nannapaneni R, Gregson BA, Stobbart L, Chambers IR,
Mendelow AD. Elderly patients with severe head injury in coma from the outset has anything changed? British Journal of Neurosurgery,2004; 18(6): 604 - 607.

Paediatric head injured patients

Introduction: Severe head injury in childhood continues to be associated with
considerable mortality and morbidity. Early surgical decompression may be
beneficial but the ICP and CPP levels at which this might be indicated have yet to
be defined and may vary with age. The objective of this study was to examine the
relationship between thresholds of mean ICP & CPP over the first 6 hours and age
in relation to outcome in paediatric head injury patients.
Methods: A total of 209 head injured children admitted to five UK hospitals were
studied. Patients aged 2 to 16 years were included if they had a minimum of six
hours of invasive monitoring. Mean values of ICP and CPP over this period were
calculated and compared in those with independent (good recovery and moderate
disability) and poor outcomes (severe disability, and death) for different age
groups.

Results: There were 133 boys and 76 girls; 92 made a good recovery, 56 had
moderate disability (148 independent outcome), 30 remained severely disabled
and 31 died (61 poor outcome). There was a significant difference between those
with independent and poor outcomes in relation to ICP (p<0.001) and CPP
(p<0.001). Patients were divided into three groups according to age. The
sensitivity of ICP and CPP in predicting outcome were similar for all groups but
the specificity differed between groups. At a CPP of 50mmHg the specificity varied
between the groups (2 to 6 years 0.47, 7 to 10 years 0.28 and 11 to 16 years
0.10) as it did at an ICP of 25mmHg (2 to 6 years 0.53, 7 to 10 years 0.44 and
11 to 16 years 0.38).
Discussion: Younger children may be able to tolerate lower perfusion pressures
and still have an independent outcome. This is important in the identification of
patients that might benefit from new treatments e.g. surgical decompression.

Publication
Chambers IR, Jones PA, Minns RA, Stobbart L, Mendelow AD, Tasker RC,
Kirkham F. Which paediatric head injured patients might benefit from
decompression? Thresholds of ICP and CPP in the first six hours. Oral
presentation at the 12th International symposium on Intracranial Pressure and
Brain Monitoring. Hong Kong 16-26th August 2004.

Monitoring of head injury

Aims: (1) To establish the feasibility of myotatic reflex measurement in patients
with head injury. (2) To test the hypothesis that cerebral dysfunction after head
injury causes myotatic reflex abnormalities through disordered descending
control. These objectives arise from a proposal to use reflex measurements in
monitoring patients with head injury.
Methods: The phasic stretch reflex of biceps brachii was elicited by a servopositioned tendon hammer. Antagonist inhibition was evoked by vibration to the
triceps. Using surface EMG, the amplitude of the unconditioned biceps reflex and
percentage antagonist inhibition were measured. After standardisation in 16
normal adult subjects, the technique was applied to 36 patients with head injury
across the range of severity. Objective (1) was addressed by attempting a
measurement on each patient without therapeutic paralysis; three patients were
also measured under partial paralysis. Objective (2) was addressed by preceding
each of the 36 unparalysed measurements with an assessment of cerebral
function using the Glasgow coma scale (GCS); rank correlation was employed to
test a null hypothesis that GCS and reflex indices are unrelated.
3

Results: In normal subjects, unconditioned reflex amplitude exhibited a positive

skew requiring logarithmic transformation. Antagonist inhibition had a prolonged
time course suggesting presynaptic mechanisms; subsequent measurements
were standardised at 80 ms conditioning test internal (index termed "TI 80").
Measurements were obtained on all patients, even under therapeutic paralysis
(objective (1)). The unconditioned reflex was absent in most patients with GCS
less than 5; otherwise it varied little across the patient group. TI80 fell
progressively with lower GCS, although patients' individual GCS could not be
inferred from single measurements. Both reflex indices correlated with GCS
(p<0.01), thereby dismissing the null hypothesis (objective (2)).
Conclusion: Cerebral dysfunction in head injury is reflected in myotatic reflex
abnormalities which can be measured at the bedside. With greater reproducibility,
reflex measurements may assist monitoring of patients with head injury.

Publication
Cozens JA, Miller S, Chambers IR, Mendelow AD. Monitoring of head injury by
myotatic reflex evaluation. J Neurosurg Psychiatry 2000;68:581-588

Secondary insults after brain injury

Introduction: Traumatic brain injury, from accidental and non-accidental (NAHI)
(shaken infant syndrome) causes, is the commonest cause of death and long
term morbidity in children. If surviving to reach hospital, it is generally accepted
that preventing secondary brain damage is likely to be more efficacious than
pharmacological intervention. Therefore the detection and correction of secondary
insults remains paramount, and may improve the neurological outcome of
patients, but has never been adequately studied in children. Definitions of 'insults'
in children are not yet established, so this on-going study aims to determine the
total burden of physiological 'derangement' encountered during the patients stay
in Intensive Care, and investigate the contribution that makes to global outcome
in survivors, and neuropathological findings in non-survivors.
Methods: A 2-year prospective study where physiological data, downloaded
every minute from bedside monitors in ICU, is summarised using predetermined
thresholds of each physiological variable by age. A Kings Outcome Score of
Childhood Head Injury (KOSCHI) at 6 and 12 months post-injury is obtained by
questionnaire.
Results: 26 children with severe or moderate HI have been recruited to date as
follows: 15 males & 11 females; Age range 4 months - 14.5 years. Causes include
10 pedestrians, 5 falls, 4 RTAs, 3 struck on head and 2 NAHI; Paediatric Trauma
4

Score range 2-10; 1 child had fixed and dilated pupils on admission; 4/26 (15%)
have died, and at 6 months 1,3 & 3 have been classified as KOSCHI 3,4 and 5
respectively (11 not yet due). 17/26 had ICP monitoring, the duration of which
has ranged from 474 to 17,133 minutes (median 7968, IQ range 3174, 12188).
No child has had SvO2 monitored continuously.
Conclusions: Age-specific derangements have been found in all variables, and to
date no child has been totally 'derangement-free'. As more patients are recruited,
outcome and neuropathology will be analysed in relation to derangement, and it
is anticipated that age-specific treatment thresholds may be identified.

Publication
Jones PA, Lo TYM, Chambers IR, Wilson G, Mendelow AD, Forsyth R, Fulton B,
Andrews PJD, Minns RA. Edinburgh-Newcastle Study: secondary insults after
accidental and non-accidental brain injury in children, and relationship to outcome
- work in progress. European Journal of Neurosurgery 2002; 144: A19.

Traumatic brain injury

Objectives: This study evaluates the geo-social factors that contribute to the
incidence and outcome of the severely head injured patient by the first direct
comparison of head injury outcome between a western centre and that of a
developing country.
Method: A retrospective analysis of head injuries admitted to two centres (Doha
and Newcastle) over a five year period (January 1997-December 2001) was
carried out comparing demographic details, aetiology, severity of injury and
management outcome. The mode of participation of victims was identified and
the geo-social factors that contributed to the aetiology and severity of head injury
were evaluated in detail. Data analysis was with SPSS version 11.
Results: There was a significant male preponderance in Doha (12m:lf) compared
with Newcastle (2.7m:lf). The peak age group in Newcastle for all head injuries
was 0-10yrs, whereas in Doha the peak age was 20-30yrs. Road traffic accident
accounted for nearly 70% of all head injuries in Doha whereas falls accounted for
nearly 50% of all head injuries in Newcastle. Overall the good outcome was 74%
in Doha and 81% in Newcastle (NS). The proportion of patients presenting in
coma and mortality (26% ) in Doha was more than twice that in Newcastle. A
recent change in the law in Qatar allowing female drivers has not significantly
increased the incidence of head injury.

Conclusion: The combination of powerful cars and good roads with a young
population has resulted in an excessive mortality in Doha compared with
Newcastle. Recent changes in the law in Qatar have not significantly altered the
incidence of head injury related to road traffic accidents. The high male
preponderance in Qatar reflects the relatively protected role women enjoy in this
society.

Publication
Igbaseimokumo U, Bashir EF, Taha Z, Raza A, Mohammed A, Mendelow D,
Chambers I, Gregson B, Treadwell L. The influence of social and cultural factors
on the incidence and severity of traumatic brain injury: a comparative study of
traumatic brain injury in Doha, Qatar and Newcastle upon Tyne, UK. J
Neurotrauma 2002; 19: 1312.

Comparison of the Canadian CT Head Rule to

Physician Judgment
Ian Stiell, James Worthington, Jonathan Dreyer, Katherine Vandemheen,
Catherine Clement, Valerie De Maio, Michael Schull, Mark Reardon, Laurie
Morrison, Douglas McKnight, Iain MacPhail, Gary Greenberg, Mary
Eisenhauer, Daniel Cass, Robert Brison and George Wells
University of Ottawa, Ottawa, Ontario, Canada

ABSTRACT
Objectives: To compare the predictive accuracy of emergency physicians' clinical
judgment to the Canadian CT Head Rule, a recently developed and highly
sensitive clinical decision rule for the use of CT in patients with minor head injury.
Methods: This prospective cohort study was conducted as a component of the
Canadian C-Spine/CT Head (CCC) Study in 10 Canadian EDs and involved adults
with loss of consciousness, amnesia, or confusion and a GCS score of 13-15.
Physicians completed a 22-item assessment form prior to CT scan. The outcome
standards were `need for neurological intervention' and `clinically important brain
injury' on CT. Physicians also estimated, based upon clinical judgment alone, the
probability, from 0% to 100%, of brain injury and of neurological intervention.
Analyses included comparison of areas under the ROC curve with 95% CIs.
Results: Among 1,416 patients enrolled over 18 months, the mean age was 38.1
(range 16-96), 66.6% were male, 97 (6.9%) had a clinically important brain
injury, and 11 (0.8%) underwent neurological intervention. Comparing physician
judgment to the Canadian CT Head Rule for predicting brain injury on CT, the
respective areas under the ROC curve were 0.77 (95% CI 0.72-0.83) vs 0.87
6

(0.85-0.89) (p < 0.05); sensitivities were 91.8% vs 97.2% (p < 0.05); and
specificities were 36.5% vs 49.0% (p < 0.001). Comparing physician judgment to
the Canadian CT Head Rule for predicting neurological intervention, the respective
areas under the ROC curve were 0.75 (0.58-0.92) vs 0.96 (0.95-0.97) (p <
0.01); sensitivities were 72.7% vs 100% (p < 0.01); and specificities were 72.4%
vs 80.9% (p < 0.001). Conclusions: Compared to the Canadian CT Head Rule,
physicians fared poorly in predicting the presence of important brain injury on CT
or the need for neurological intervention among minor head injury patients.

045 Use and Yield of Investigations for Alert Patients with

Possible Subarachnoid Hemorrhage.
Perry JJ, Stiell IG, Wells GA, Spacek A. Division of Emergency Medicine, University of Ottawa.
Ottawa, ON.
OBJECTIVES: There is little evidence to guide investigation to rule out subarachnoid hemorrhage
(SAH) in alert ED patients with acute headache. This study evaluated the current use and yield of
CT and LP for SAH in ED patients with possible SAH. METHODS: This health records review
enrolled patients presenting to a tertiary care university teaching hospital ED if they were >15
years, alert with GCS 15, had no new neurological deficits, and had a complaint of headache,
syncope or possible SAH. Outcome measures included: CT use and yield, LP use and yield and
length of stay prior to discharge/referral to neurosurgery. Exclusion criteria included: maximal
intensity in >1 hour, referrals with SAH, recurrent headaches, head trauma, pain for >14 days,
focal neurological deficits, papilledema or decreased level of consciousness. Analysis included:
descriptive statistics including 95% CI and ANOVA for length of stay. RESULTS: The 891 patients
seen over a 10-month period had these characteristics: mean age 41.9, 66.4% female, 1.1%
SAH, 2.6% admitted, 42.2% vomiting, 40.8% transient loss of consciousness, 35.1% CT, and
6.8% LP. Only 9 (2.6%) CT and 2 (2.4%) LPs were positive for SAH. There were no missed cases
with CT. There was 1 positive LP without a prior CT and another with a positive CT. There were
144 (11.5%) patients who underwent a normal CT without subsequent LP. The mean length-ofstay for patients without SAH was as follows: 4.0 hours (3.8-4.1) without testing, 5.0 hours (4.75.4) with only CT, 7.1 hours (6.3-7.9) with LP (p = 0.001). CONCLUSIONS: This study
demonstrated that patients who underwent testing spent much more time in the ED. CT and LP
had very low yield suggesting the need for a clinical decision rule to guide physicians in the
investigation of acute headache to rule out SAH.

046 Is There Still a Role for Physical Examination of Patients

with Minor Head Injury?
Stiell IG, Clement C, Cass D, Rowe BH, Eisenhauer M, Schull M, McKnight RD, MacPhail I, Brison
R, Reardon M, Greenberg G, Battram E, for the CCC Study Group. Division of Emergency Medicine,
University of Ottawa, Ottawa, ON.
OBJECTIVES: In an era of ubiquitous CT scanning, is there still a role for physical examination of
minor head injury patients? This study evaluated the accuracy and reliability of common physical
signs. METHODS: This prospective cohort study was conducted in 10 tertiary care EDs and
involved adults with loss of consciousness, amnesia, or confusion and a GCS score of 13-15. MDs
performed a standardized exam for: pupillary size and reaction, lateralizing motor weakness,
suspected open skull fracture, basal skull fracture signs, and unreliability due to suspected
ethanol/drug intoxication. Where feasible, 2nd physicians performed interobserver assessments.
Patients had CT to determine the outcome, clinically important brain injury. Analyses included
univariate association, kappa, sensitivity, specificity, adjusted odds ratio by stepwise logistic

regression. RESULTS: The 3,121 patients enrolled over 36 months had these characteristics:
mean age 38.7, GCS score - 15 79.8%, important brain injury 8.1%, required neurological
intervention 1.4%. The Table shows % of brain injury (N = 254) and non-injury (N = 2867)
patients with the findings, p-value, kappa, sensitivity, specificity, adjusted odds ratio.

Injury

No
injury

P-Value

Kappa

Sens

Spec

O.R.

Abnormal
pupils

2.8%

1.1%

0.05

0.66

3%

99%

NS

Lateralizing
weakness

0.8%

1.1%

NS

0.66

1%

99%

NS

Suspected
open skull #

11.8%

2.6%

0.001

0.85

12%

97%

3.2

Signs of basal
skull #

30.4%

4.7%

0.001

0.76

30%

95%

5.0

Unreliable 2nd
Etoh/Drug

18.1%

11.8%

0.01

0.54

18%

88%

NS

Assessment

CONCLUSIONS: Assessment of pupils and motor weakness is not useful in minor head injury
patients. The most reliable and accurate physical findings for identifying the risk for brain injury
are 'suspected open skull fracture' and 'signs of basal skull fracture' and these should be
incorporated into decision rules for the management of minor head injury.

047 What can be Gained from the General Physical

Examination of Alert Patients with Potential Cervical Spine
Injury?
Stiell IG, Clement C, Worthington JR, Schull M, Eisenhauer M, MacPhail I, Cass D, Rowe BH,
Battram E, Bandiera G, Brison R, McKnight RD, for the CCC Study Group. Division of Emergency
Medicine, University of Ottawa. Ottawa, ON.
OBJECTIVES: In the assessment of potential C-spine injury patients, what can be gained from the
general physical examination? This study evaluated the accuracy and reliability of common
physical signs. METHODS: This prospective cohort study was conducted in 10 tertiary care EDs
and involved alert (GCS 15) and stable adult trauma patients. MDs performed a standardized
exam for: patient position, distracting painful injuries, unreliability due to ethanol/drug
intoxication, visible facial injury, visible head injury, motor deficit in extremities, sensory deficit in
extremities. 2nd physicians performed interobserver assessments in 150 cases. Patients
underwent radiography to determine the outcome, clinically important C-spine injury. Analyses
included univariate association, kappa, sensitivity, specificity, odds ratio by multivariate logistic
regression. RESULTS: The 8,924 patients enrolled over 36 months had mean age 36.8 and 1.7%
had important C-spine injury. The Table shows % of C-spine injury (N = 151) and non-injury (N =
8773) patients with the findings, p-value, kappa, sensitivity, specificity, adjusted odds ratio.

Physical
findings

Injury

No
injury

P-Value

Kappa

Sens

Spec

O.R.

6%

39%

0.001

0.74

6%

62%

NS

Distracting
injuries

15%

8%

0.001

0.41

15%

92%

NS

Unreliable
2nd
Etoh/Drug

8%

4%

0.05

0.22

8%

96%

NS

Facial injury

43%

19%

0.001

0.75

43%

81%

1.8

Head injury

48%

20%

0.001

0.76

48%

80%

1.6

Motor deficit

5%

1%

0.001

0.93

5%

99%

NS

Sensory
deficit

6%

2%

0.001

0.60

6%

98%

NS

Upright
position

CONCLUSIONS: The findings 'distracting painful injuries' and 'unreliable due to etoh/drug' show
poor interobserver agreement. The other 5 findings are reliable but only 'facial injury' and 'head
injury' independently predict a higher risk of C-spine injury in alert trauma patients.

049 Closed Reduction of Distal Radius Fractures in the

Emergency Room: Factors Associated with Orthopedic
Intervention.
Skoretz TG, Eisenhauer M, Amir H. London Health Sciences Centre, University of Western Ontario.
London, ON.
OBJECTIVE: To describe patient and radiographic characteristics associated with Orthopedic
intervention in patients that received closed reduction of a distal radius fracture in the Emergency
Room. INTRODUCTION: Closed reduction of a distal radius fracture is a time consuming and
laborious procedure. Despite closed reduction in the ER, many patients will go on to further
manipulation and/or operative repair by the Orthopedic Surgery service. Discovering factors that
predict future Orthopedic intervention may be helpful in the management of this common injury.
METHODS: The Emergency Room logbook and billing records were reviewed for all adult patients
that underwent a closed reduction of a distal radius fracture by the Emergency Medicine service at
London Health Sciences Centre-South Street, a tertiary care centre, in 2000. Hospital charts and
presenting wrist radiographs were reviewed. Multivariate analysis was used to calculate predictors
of Orthopedic intervention (repeat closed reduction or operative repair). In addition, multiple
reviewers assessed initial radiographs to calculate inter-rater reliability in extracting 5 different xray variables. RESULTS: A total of 71 distal radius fractures underwent closed reduction by the
Emergency Room medical staff at LHSC-South Street. Seventy charts were available for review.
Initial radiographs of the injured wrist were available in 56 patients. Analyses showed that patient
age (OR = 0.952) and initial dorsal angulation (OR = 0.964) were associated with Orthopedic
intervention. Analyses of inter-rater reliability demonstrated fair to excellent reliability in

extraction of radiographic variables. CONCLUSION: In the sample of patients reviewed, patient

age and initial dorsal angulation were predictive of future Orthopedic intervention. In addition,
inter-rater reliability in extracting x-ray data was considered good overall, but variability existed.

051 Does Urine Screening for Drugs of Abuse Change the

Management of ED Patients?
Eisen JS, Sivilotti MLA, Collier C. Department of Emergency Medicine, Queen's University.
Kingston, ON.
BACKGROUND: It is estimated that substance use is a frequent factor in Emergency Department
(ED) visits. Qualitative urine testing for drugs of abuse (U-DOA) is frequently ordered, but is
limited in its ability to establish the identity, timing or dose of substances used. Although previous
retrospective studies have demonstrated these limitations, their study design cannot be used to
determine whether U-DOA provides useful information to the ED physician when making patient
care decisions. Objective: To isolate and measure the impact of U-DOA on ED patient care.
METHODS: All U-DOA ordered in adult patients seen in 2 teaching EDs were eligible; screens that
were ordered for victims of trauma or sexual assault were excluded. Prior to reporting the test
results to the ED, ordering physicians were phoned by the investigators and queried about their
patient care plans before, and then immediately after the results were disclosed. This design
isolated the impact of the U-DOA screen on ED patient care decisions. Any changes in plan
reported by the physician were compared to a pre-determined set of changes that were
considered to be clinically important. RESULTS: To date, 81 U-DOA have been enrolled during a
period with approximately 42,000 ED visits. One ED physician reported a change in plan (CT head
deferred), but this change was not considered significant according to pre-specified criteria. UDOA thus led to a clinically important change in management in 0/81 cases (95% CI 0-3.7%).
CONCLUSIONS: U-DOA is rarely helpful in guiding patient care decisions in the ED. These results
call into question the need for the test in the ED setting.

052 Economic Evaluation of the Potential Impact of the

Canadian C-Spine Rule.
Coyle D, Stiell IG, Wells GA, Clement C, for the CCC Study Group. Division of Emergency Medicine,
University of Ottawa, Ottawa, ON.
INTRODUCTION: The Canadian C-Spine Rule (CCR) is designed to improve the efficiency of ED
management of potential cervical spine injury patients. This economic analysis estimated the
potential cost savings to the Canadian health care system with widespread use of the CCR.
METHODS: This economic analysis used a probabilistic-based decision analytic model comparing
current clinical practice to that assuming 100% uptake of the CCR. Costs savings were assessed
from a Canadian health care system perspective. The sensitivity and specificity of the rule was
estimated by combining data from the derivation (N = 8,924) and validation (N = 7,017) studies.
For our base analysis, current radiography rates were estimated to be 71.6%. Sensitivity analyses
assumed rates of 90% and 100%. Cost data were obtained from provincial health care fee
schedules, hospital cost accounting systems and, if required, the literature. The probabilistic
model employed Monte Carlo simulation that was based on 3,000 replications. We estimated the
expected values for potential cost savings and reduction in radiography rates. Results are in
Canadian dollars. RESULTS: In our base analysis, the expected value of cost savings with the CCR
was $8.54 (95% credibility interval $5.61-$11.91) per alert stable trauma patient. The rule is
forecasted to lead to an absolute reduction in radiography of 12.8% (95% CI 8.9-16.9) compared
to current rates of 71.6%. The total annual cost savings, assuming a Canadian adult trauma
patient population of either 185,000 or 400,000 patients, would be $1.6 million and $3.4 million
respectively. Assuming radiography rates of 90% and 100%, the expected annual cost savings
were $8.8 million and $11.5 million respectively, based on 400,000 patients per year.
CONCLUSIONS: Widespread use of the CCR is expected to lead to cost savings as low as $1.6

10

million per year or as high as $11.5 million. Future studies should evaluate the potential economic
impact of the CCR in other countries.

053 Transient Ischemic Attacks in the Emergency Department:

Description and Outcome.
Rowe BH, Yiannakoulias N, Bullard M, Spooner CH, Holroyd BR, Svenson L, Rosychuk R,
Schopflocher D. University of Alberta, Edmonton, AB.
OBJECTIVES: Defining the short-term prognosis and risk factors for stroke after transient ischemic
attacks (TIA) may provide guidance in determining which patients need rapid ED evaluation.
Recent US data suggest the 90-day risk of recurrent TIA (13%) or stroke (11%) are high. This
study examines ED presentations of TIA and links data to health care use in the subsequent year.
METHODS: All patients presenting to Alberta EDs were eligible for inclusion. Data were derived
from a sample of ED patients treated in 17 health regions over 1 year (98/99) with a diagnostic
code of TIA (434.x). Data were extracted from the Ambulatory Care Classification System (ACCS)
database, consisting of computerized abstracts coded similarly across regions. Diagnostic
categories were coded by medical record nosologists using ICD-9 codes for the primary discharge
diagnosis. Descriptive statistics and crude presentation rates are reported. RESULTS: Overall, 1.49
million ED visits were recorded in the year; the number of patients with a diagnosis of TIA was
2543 (1.7/1000 ED visits). Overall, 75% of patients were over the age of 65 years; males and
females were equally represented. Limited daily or seasonal variation exists; Monday (15.6%) and
December (10.7%) numbers were highest. Most patients are discharged (1810; 71%); admission
(770; 28%) is higher than the ED average and few leave prior to seeing a physician (<1%). The
rate of TIA varies between regions, with the average of 7.5/1000 population. Representation to
the ED is common (833; 37%); however, representation with a TIA (236; 10%) or stroke 24 (1%)
in the subsequent year was less frequently observed. CONCLUSIONS: These results indicate that
TIA is a relatively common presentation to the ED, but that development of recurrent TIA or
stroke was lower than expected based on recent US figures. Further evaluation of TIA patients in
the ED is urgently required to understand the observed variation and determine predictors of
recurrence.

054 Defining 'Irrelevant' CT Findings in Blunt Head Injury

Patients.
Atzema C, Mower WR, Hoffman JR, Holmes JF, Killain AJ, Greenwood SD, Shen A. University of
California. Los Angeles, CA, USA.
INTRODUCTION: Researchers developing a clinical decision instrument (CDI) for the use of
computed tomography (CT) in patients with minor head injury (MHI) must classify certain injuries
seen on CT as 'clinically unimportant'. This is necessary to identify which patients actually needed
a head CT. This study aims to evaluate the importance of various minor CT findings, based on
need for neurosurgical intervention and Glasgow Outcome Scale (GOS) scores. METHODS: NEXUS
II is a prospective observational study involving patients at 18 sites who received an emergent
head CT scan between April 1999 and December 2000. In this substudy of NEXUS II we
prospectively defined a number of CT findings generally considered clinically unimportant, and
identified patients at 6 sites for whom the official radiology report included 1 of these findings.
Two trained independent abstractors reviewed patient charts to determine presence of
neurosurgical intervention or a poor outcome (GOS 3-5). RESULTS: Prevalence of minor CT
findings was 1.86% (n 156) among the first 8374 trauma patients in the NEXUS II cohort. Eightytwo patients at the 6 sites met the inclusion criteria, and 11 (13%) patients fared poorly and/or
had neurosurgical intervention. Adequate follow-up information was available on 10 of these
patients, all of whom had an abnormal GCS at the time of the CT scan. Five of the 7 neurosurgical
patients had abnormal coagulation studies. CONCLUSIONS: Clinically unimportant findings are
diagnosed in less than 2% of head trauma patients undergoing CT scanning. While an important
minority of these patients do have neurosurgery or a poor outcome, this appears to be extremely

11

rare in patients who do not present with an abnormal mental status and/or a coagulopathy. With
certain qualifiers, clinicians and developers of a CDI for the use of head CT in MHI patients may
safely classify minor CT findings as insignificant.

055 Attitudes and Judgement of Emergency Physicians in the

Management of Patients with Acute Headache.
Perry JJ, Stiell IG, Wells GA, Mortensen M, Lesiuk H, Wallace G, Sivilotti M, Kapur A. Division of
Emergency Medicine, University of Ottawa, Ottawa, ON.
OBJECTIVE: Currently there is little objective evidence guiding physicians in the investigation
acute headache to rule out subarachnoid hemorrhage (SAH). This study assessed emergency
physicians (EP) in: 1) their attitudes to not ordering head CT before performing LP, this
demonstrates if EPs are willing to use Schull's model, which suggests a LP directly for patients
with normal neurological examination, and 2) the accuracy of their judgment for predicting SAH
without a decision rule. METHODS: This 1-year prospective cohort study was conducted in 2
tertiary care university EDs. Data was collected for consecutive alert patients with an acute
headache and a normal neurological examination. Prior to investigation, MDs recorded the pre-test
probability for SAH to the closest decile and comfort in 'performing an LP without obtaining a CT
with a 3-point Likert scale. The criterion outcome was SAH as diagnosed by SAH on CT,
xanthochromia in the cerebral spinal fluid (CSF), or the presence of red blood cells in the final
tube of CSF with positive cerebral angiography. Descriptive statistics and a receiver operating
characteristic (ROC) curve with a 95% CI were generated. RESULTS: The 222 enrolled patients
had a mean age 43.4 years, 58.1% were female, 9.0% had SAH, 83.2% underwent CT, and
41.9% underwent LP. EPs reported being 'uncomfortable' in performing an LP without first
ordering a CT in 54.6% of cases. They were 'very comfortable' in performing an LP without CT in
7.7% of the cases. The area under the ROC curve for the pre-test probability of SAH was 0.85
(95% CI, 0.76, 0.93). There were 3 positive cases with a pre-test probability of <5% and 4
patients with SAH with a pre-test probability of 10%. The remainder had a pre-test probability of
over 20%. CONCLUSIONS: EPs were uncomfortable in performing an LP without obtaining a CT.
Physicians showed only fair accuracy in predicting SAH and a clinical decision rule may improve
the accuracy and efficiency of SAH diagnosis.

056 The Influence of the Emergency Medicine Clinical

Teachers' National Certification Level on their Evaluation by
Residents.
Steiner IP, Yoon PW, Kelly KD, Donoff MG, Diner BM, Mackey DS, Rowe BH. University of Alberta.
Edmonton, AB.
INTRODUCTION: The evaluation of clinical teaching faculty is necessary and using accepted tools
it can be valid and reliable. Currently there are no data from the medical education literature
concerning specific faculty-related factors relating to teaching performance. This study examines
the influence of EM certification status of clinical faculty on the teaching performance scores
provided by residents. METHODS: A retrospective analysis of data accumulated between 07/199407/2000 on 1st, 2nd and 3rd year Family Medicine residents' evaluations of EM clinical teaching
faculty at the University of Alberta was conducted. Resident and teaching faculty related variables
were entered anonymously using the ED Scale (Acad Emerg Med 2000;7[9]:1015-21).
Credentialing and demographic information on EM teaching faculty was supplemented by data
obtained through a 9-question survey; public information resources provided data on some
teachers. Descriptive and ANOVA analyses are presented. RESULTS: 562 Family Medicine
residents completed EM clinical rotations during the study period and 777/831 (94%) had
voluntarily returned anonymous completed evaluation forms. 705/831 (85%) had valid data. 115
clinical teaching faculty members had been evaluated in 4 dispositional domains: Didactic
teaching, Clinical teaching, Approachability and Helpfulness for a total of 12,816 individual

12

evaluations. Complete information on 109/115 (95%) EM teaching faculty members was obtained.
Univariate analysis on the scores by EM certification level indicated that EM specialists rated
statistically highest in the didactic teaching category (p < 0.001), whereas EM certified family
physicians rated higher in the approachable and helpful categories (p < 0.001). Physicians without
any national certification were rated lowest on all 4 categories. Subgroup analysis revealed
superior teaching performance by formally trained teachers in certain evaluation domains.
CONCLUSIONS: EM national certification is a positive moderator of teaching performance. The
statistically significant differences found between specialist and family physician teaching groups
probably have no practical implications.

057 Medical Undergraduate Curriculum International Health

Enrichment Project.
Felix J, Bullard M, Hoyano D, Sowa B, Laing L, Baydala L, Fanning A. University of Alberta.
Edmonton, AB.
INTRODUCTION: Many recent applicants to Emergency Medicine training programs cite
international health (IH) opportunities as 1 reason for their specialty selection. A CAEP
International Emergency Medicine Committee has recently been formed. This study attempted to
identify the extent of international health (IH) content in a typical medical curriculum, in concert
with a proposal to enhance it. METHODS: A manual search of the first 2 years of undergraduate
course materials at the University of Alberta was completed by trained research assistants. IH
content was coded as: any material consistent with the previously established IH Core Content. IH
mentions were not weighted and any written IH content in handouts was included. In parallel, the
content of each educational block was surveyed to identify topic areas best suited to IH
enrichment. These results were reported to each block Coordinator. Finally 2 of the authors met
face-to-face with the major Block Coordinators to discuss implementation. RESULTS: In 2001,
Year 1 and 2 contained 11 distinct educational blocks. The number of IH mentions ranged from 2
(Cardiology) to 69 (Infection/Immunity/Inflammation) with a median of 22 (IQR: 11, 29). All
(100%) Block Coordinators supported the IH initiative; however, 80% requested assistance in
developing teaching materials and concepts. Our presentation to the Undergraduate Curriculum
Committee led to the incorporation of IH enrichments into the curriculum beginning in September
2001. CONCLUSIONS: Prior to this survey, IH was not a large component of most medical student
block training at this University. There are presently 7 hours of dedicated IH time in the first 2
years of the curriculum with several enrichment areas being developed. Optimizing the amount
and type of IH undergraduate enrichment is an ongoing project. Educators in other universities
need to repeat this study, and motivated EM educators need to participate in areas of IH teaching.

075 Emergency Department Treatment of Stable Acute Paroxysmal Atrial Fibrillation.

Kapur AK, Stiell IG, Wells GA, Brison RJ, Mortensen M. Division of Emergency Medicine, University
of Ottawa. Ottawa, ON.
OBJECTIVES: The optimal management of acute paroxysmal atrial fibrillation (PAF), a common ED
presenting complaint, remains undetermined. This study's purpose was to compare immediate
and short-term outcomes of aggressive (AGG) and conservative (CON) ED treatment of clinically
stable PAF. METHODS: This 6-month prospective cohort study, conducted at 3 university-affiliated
hospital EDs, enrolled all adult patients with <48 hours of clinically stable PAF. CON patients
received no treatment or only rate control agents. AGG patients had pharmacologic and/or
electrical cardioversion attempted. Patients were telephoned at 4 weeks to determine PAF
recurrence, complications, and quality of life using the SF-36 scale. Proportions of the AGG and
CON groups in sinus rhythm at ED discharge and at follow-up, as well as complications in the ED
and at 4 weeks, were compared using chi-square. Quality of life was compared using t-test.

13

RESULTS: We enrolled 169 patients, 32 in the CON group and 137 in the AGG group. The CON
group was slightly older (mean 70.3 vs 61.9 yrs, p = 0.001) and more had coronary artery
disease (78.9% vs 42.6%, p < 0.01). More AGG patients cardioverted to sinus rhythm in the ED
(82.5% vs 34.4%, p < 0.001) and fewer were admitted (8.0% vs 37.5%, p < 0.001). Fifteen
(8.9%) patients, all AGG, had complications in the ED; 2 (1.2%) required admission. 97.0% of
patients were followed up. More AGG patients cardioverted and stayed in sinus rhythm for 4
weeks (52.3% vs 30.0%, p = 0.03). AGG patients had higher physical summary scores on the SF36 at 4 weeks (47.1 vs 41.2, p = 0.01). No thromboembolisms occurred by four-week follow-up.
CONCLUSIONS: This is the first study to prospectively follow PAF patients treated in the ED.
Aggressive treatment for PAF is as safe as conservative and more successful for restoring sinus
rhythm in the ED and should be considered the optimal ED management of PAF.
076 A Descriptive Review of a Canadian Chest Pain Evaluation Unit.
Martin D, Sinclair D. Dalhousie University. Halifax, NS.
INTRODUCTION: Many North American centres have addressed the problem of evaluating nondiagnostic chest pain by developing dedicated units for rapid assessment and risk-stratification.
The QEII Health Sciences Centre opened its Chest Pain Evaluation Unit (CPEU) in 1999. In order to
present some initial data on the operation of such a unit in the Canadian setting, we conducted a
review of its operation over a 6 month period. METHODS: The sample consisted of all 136 patients
observed in the CPEU from April to October, 2000. Data were collected retrospectively from an
admissions log and patient charts. RESULTS: There was 21% utilization of theoretically available
patient-hours in the CPEU. Mean length-of-stay in the CPEU was 13.6 hours. 68.2% of patients
underwent exercise stress testing (EST), 2.3% underwent coronary angiography, and 0.8%
underwent perfusion scintigraphy, while 28.0% underwent no other investigation while in the unit.
Of those undergoing EST, 16.7% had positive tests, 44.4% had negative tests, and 38.9% had
non-diagnostic tests. 5% of patients with negative EST results were admitted to the Cardiology
service. 33.3% of patients with positive EST results were discharged home. Of all CPEU patients,
75.7% were discharged home, and 24.3% were admitted to the Cardiology service. No deaths
occurred in the CPEU. Rationale for admission was based on a positive EST result in 31.2%, EKG
and cardiac marker evidence in 18.8%, isolated cardiac marker evidence in 18.8%, and isolated
EKG evidence in 12.5%, while 15.9% of admissions involved subjective rationale. 9.8% of patients
observed in the CPEU ultimately received an acute coronary diagnosis. CONCLUSIONS: These
results identify a number of topics requiring further investigation. The effect of Canadian CPEUs on
admission rate and length-of-stay, and the role of specific technologies in the admit/discharge
decision are among those areas in which important questions remain.
077 Institutional Variation in the Emergency Department Management of Paroxysmal
Atrial Fibrillation: A Comparison of Two Canadian Centres.
Ip J, Cadieu T, McKnight D, Abu-Laban RB, Zed JP. Vancouver General Hospital. Vancouver, BC.
INTRODUCTION: Paroxysmal atrial fibrillation (PAF) is the most common ED dysrhythmia. A recent
paper described the ED management of 289 PAF patient encounters at Ottawa Civic Hospital
(OCH) (Ann Emerg Med 1999:04). Our suspicion was the ED management and disposition of PAF
at Vancouver General Hospital (VGH) varies significantly from OCH. The purpose of this study was
to evaluate this hypothesis. METHODS: PAF patients presenting between Jan/01/1999 and
Jun/01/2000 were retrospectively identified from the VGH ED database and their records
reviewed. Inclusion/exclusion criteria identical to the OCH study were employed. Institutional
variance was evaluated using appropriate comparative two-tailed statistics. RESULTS: 88 patient
encounters of stable PAF were identified: 74 (84%) were treated in the ED and 14 (16%) received
no ED interventions. Demographic and baseline characteristics were similar between VGH and
OCH patients, with the exception of previous PAF (42% vs 72% respectively). Twenty-six
encounters (30%) were treated only with rate controlling medications. The majority of
cardioversion attempts were chemical (43/88: 49%), 53% with prior rate control. Of those in
whom chemical cardioversion was attempted, 16 (37%) went on to electrical cardioversion. Only 5

14

encounters (6%) were treated with primary electrical cardioversion. Comparison between VGH
and OCH respectively, showed no statistically significant difference in overall and primary electrical
cardioversion proportions (24% vs 28% and 5.7% vs 4.5%), or success rate (91% vs 89%), but
significant variation in chemical cardioversion proportion (49% vs 62%, p = 0.025); chemical
cardioversion success rate (26% vs 50%, p = 0.004); mean ED length of stay (5.9 vs 5.0 hr, p =
0.040); consultation proportion (59% vs 13%, p < 0.001); admission proportion (34% vs 3%, p
< 0.001); and ED return within 7 days proportion (7% vs 14%, p = 0.006). CONCLUSIONS: There
is significant variation in the ED management of PAF between VGH and OCH. We suspect this
finding is reflective of a general wide variability in the ED management of PAF. Development of a
practice guideline may improve management of PAF and resource utilization.
078 A Survey of Emergency Physicians' Attitudes Towards Primary Electrical
Cardioversion for Stable Paroxysmal Atrial Fibrillation of Less Than 48 Hours Duration.
Ip J, Sandhu M, McKnight D, Abu-Laban R, Zed PU, Pharm D. Vancouver General Hospital.
Vancouver, BC.
INTRODUCTION: Recent studies of the emergency department management of stable paroxysmal
atrial fibrillation (sPAF) at Vancouver General Hospital and Ottawa Civic Hospital demonstrate that
electrical cardioversion (EC) for sPAF of <48 h duration (sPAF <48 h) is safe and effective
(conversion proportion 89-91%). Approximately 5% of patients in these studies underwent
primary EC. Primary EC of sPAF <48 h is neither common nor well studied; however, there are
reasons to believe it may be preferable to primary chemical cardioversion. Further research would
be useful to define the role of primary EC for sPAF <48 h. The purpose of this study was to
determine emergency physicians' (EPs) current usage of and attitudes towards primary EC for
treatment of sPAF <48 h, and their hypothetical willingness to participate in a future clinical trial
of this modality. METHODS: A 12-question survey was distributed to all board-certified EPs and
emergency residents in 4 British Columbia university-affiliated hospitals between Aug/01/2001
and Oct/15/2001. Reminder follow-ups were utilized to encourage responses and anonymity was
maintained. RESULTS: Seventy-six percent (51/67) of surveys were completed. Seventy-eight
percent of respondents (range by institution: 67%-92%) use EC for sPAF <48 h and 67% felt it
was safe and effective. No respondents felt EC was unethical or dangerous and 75% deemed EC
as safe or safer than chemical cardioversion. Approximately 70% of respondents felt primary EC
could improve patient comfort and/or expedite ED disposition. Forty-three percent felt EC was
easy to use, could reduce consultation frequency, and prevent confusion in choice of chemical
cardioversion agents. Ninety percent of respondents (range by institution: 83%-93%) indicated a
willingness to participate in a clinical trial on primary EC for treatment of sPAF <48 h (46/51, 95%
CI: 79%-97%). CONCLUSIONS: The results of this study indicate that a clinical trial of primary EC
for the treatment of sPAF <48 h would be supported by most emergency physicians and appears
warranted.
079 Development of a Tool for Predicting Length of Stay (LOS) for the Emergency
Department Clientele.
Afilalo M, Unger B, Colacone A, Gigure C, Boivin JF, Vandal A, Lger R, Stiell I, Xue X. Sir
Mortimer B. Davis-Jewish General Hospital, McGill University. Montreal, QC.
OBJECTIVE: To develop a tool that will quantify the predicted length of stay (LOS) of ED patients.
The "LOURDEUR TOOL" will be based on patients' intrinsic characteristics (PICs) and not factors
related to the organization or functioning of EDs. METHODS: An in depth review of the literature
and numerous discussions with emergency physicians (EPs) permitted the development of a
conceptual model of factors which affect ED LOS. This model was subsequently used in an expert
consultative process with other EPs and nurses from across Quebec. The goal of the consultative
process, in the form of focus groups, was to produce a list of PICs, measurable early on arrival to
the ED, that could potentially be associated with LOS. The list produced was the source for the
development of a questionnaire focusing on the PICs. The next phase of the study included a
prospective sampling of visits (n = 2841) in 6 EDs (Quebec n = 5; Ontario n = 1). Using a sample

15

size of 2146 patients and 110 variables from the questionnaire, a multivariable logistic regression
analysis and mixed linear modeling methods were employed to identify the most important PICs
associated to the LOS. RESULTS: Through a backward and stepwise model selection, the following
variables were found to have an impact on LOS: Age, reason for ED visit, number of hospital
admissions in the last 3 years, triage code, perception of severity of illness, autonomy, mode of
transport, presence of endocrine or memory problems, ED referral, having a family physician and
employment status. CONCLUSIONS: The "LOURDEUR TOOL" permits the estimation in LOS that is
based on the PICs. It will bring new insights on ED congestion and will enable comparisons both
within and between EDs irrespective of their functioning. It can also be used to evaluate the
impact of the various health system transformations on specific patient populations and thus
adjustments can be made more efficaciously.
080 Recent Increases in Left Without Being Seen in the Emergency Department.
Bullard M, Rowe BH, Yiannakoulias N, Spooner CA, Holroyd B, Craig W, Klassen T, Johnson D,
Rosychuk R, Svenson L, Schopflocher D. University of Alberta, Edmonton, AB.
OBJECTIVES: Patients who leave emergency departments (EDs) without being seen (LWBS)
constitute have the potential for increased morbidity and dissatisfaction. This study examines
LWBS trends over a three-year period. METHODS: All patients presenting to provincial EDs were
eligible for inclusion. Data were derived from a sample of ED patients treated in 17 health regions
over 3 years (98/99, 99/00, 00/01) with a disposition code of LWBS or a disposition code of left
against medical advice in conjunction with a refusal of service (V642). Data were extracted from
the Ambulatory Care Classification System (ACCS) database, computerized abstracts coded
similarly by medical record nosologists across all regions. Descriptive statistics and crude
presentation rates are reported. RESULTS: Overall, approximately 1.5 million ED visits were
recorded per year across the province. The number of patients LWBS has risen every year; 98/99:
21,195, 99/00: 25,865; and 00/01: 32,375. Young children (ages <5; 14%) and adults (ages:
20-29; 23%) represent the largest percentage of cases overall. The elderly (>64 years) represent
<5% of the overall LWBS sample. Wide seasonal variation (34%) was observed and December
rates were highest (9.7%). The rate of LWBS is increasing in all areas of the province, but rates
increased most in rural EDs over the time period (59% increase). For the 2 major urban centres,
rates per 1000 ED visits were higher in Calgary than in Edmonton for all 3 years; increases over
time were greater in Calgary (39%) than Edmonton (25%). CONCLUSIONS: Despite the most
rapid population growth of any province in Canada, in-patient capacity was not increased over the
study period. These results indicate that LWBS cases across a large population are increasing
steadily and can be considered a proxy marker for ED overcrowding. Further detailed evaluation of
LWBS should identify other reasons for premature departure.
081 Adventure and Adversity: Injury Patterns in an Extreme Sport.
Denny CJ, Schull MJ. Division of Emergency Medicine, University of Toronto. Toronto, ON.
INTRODUCTION: Adventure racing is a wilderness multisport endurance activity. These events
challenge teams with days of continuous travel through environmental extremes of temperature
and terrain. Despite increasing popularity, there is a paucity of literature examining patterns and
predictors of injury in this sport. Our purpose was to estimate the prevalence of adverse incidents
in adventure racing. METHODS: Prospective, cross-sectional survey of all athletes at the Canadian
Adventure Racing Championships. During a three-day race in September 2001, 15 teams of 4
athletes trekked, mountain biked, and canoed a 234 kilometre course in northern Ontario. Trained
interviewers administered a questionnaire to consenting team captains at the finish line, or at
basecamp if teams did not finish. An incident was defined as any injury or illness of sufficient
severity to impede team progress. Analyses included descriptive statistics with 95% confidence
intervals, and logistic regression to determine the association of adverse incidents with age,
gender, team, and level of adventure racing experience. RESULTS: All 15 team captains agreed to
participate. Of 60 athletes, 44 (73%) were male, with an a mean age of 31.5 years. Nine teams
(60%; 95% CI 32-84) failed to finish the race; 7 due to an incident (47%; 95% CI 21-73). Of the

16

60 athletes, 29 (48%; 95% CI 35-62) suffered an incident. Fifteen incidents occurred while
trekking, 10 while biking and 4 while canoeing. The most common adverse incident was
musculoskeletal injury (52%; 95% CI 32-71). There were no deaths and only 1 incident required
hospital care. In multivariate analysis, least experienced athletes (<1 year of adventure racing)
were more likely to have an adverse incident (OR 7.6 p = .02). CONCLUSIONS: Incidents affect
nearly half of adventure racers. Less than half of teams finish the race. Injury prevention
initiatives may be more effective if focused on least experienced athletes.
082 Treatment Strategies for Early Presenting Acetaminophen Overdose - A Survey of
Medical Directors of Poison Centres in North America and Europe.
Kozer E, McGuigan M. Division of Clinical Pharmacology and Toxicology, The Hospital for Sick
Children, and the Ontario Regional Poison Control Centre. Toronto, ON.
BACKGROUND: Acetaminophen is frequently used in self-poisoning in Western countries. Although
treatment with N-acetylcysteine (NAC) reduces liver injury, no consensus exists on the preferred
management of acetaminophen toxicity. OBJECTIVES: To describe the approach taken by
toxicologists in North America and Europe toward the management of acetaminophen toxicity.
METHODS: Medical directors of poison centres in the United States, Canada, and Europe were
surveyed by means of a questionnaire presenting 2 clinical scenarios of acetaminophen overdose:
a healthy adolescent with no risk factors who had an acute ingestion of acetaminophen, and an
adult with both acute ingestion and possible risk factors. For each case several questions about
the management of these patients were asked. RESULTS: Questionnaires were sent to medical
directors of 76 poison centres in North America and 48 in Europe, with response rates of 62% and
44% respectively. Forty percent of responders suggested using charcoal 4 hours after ingestion of
a potential toxic dose of acetaminophen, and 90% recommended treatment with NAC when levels
were above 150 mg/mL but below 200 mg/mL 4 hours after ingestion. Duration of treatment with
oral NAC ranged from 24 to 96 hours; 38 responders suggested a duration of 72 hours. Of 49
centres recommending oral NAC, 18 (36.7%) said they might consider treatment for less than 72
hours. Eleven of 29 (37.9%) responders suggested treatment with intravenous NAC for more than
20 hours as their usual protocol or a protocol for specific circumstances. CONCLUSIONS: Our
study showed large variability in the management of acetaminophen overdose. Variations in
treatment protocols should be addressed in clinical trials to optimize the treatment for this
common problem.
083 Shiftwork and Emergency Medical Practice: Systematic Narrative Review.
Frank JR, Ovens H. University of Toronto Division of Emergency Medicine, Toronto, ON.
INTRODUCTION: Shiftwork is an essential component of the demanding 24/7 practice of
emergency medicine. Unfortunately, shiftwork schedules are also known to have numerous
negative effects on shiftworkers. To our knowledge, no systematic narrative overview of the
shiftwork literature exists. METHODS: We sought to answer the question, "What are the effects of
shiftwork on emergency medical practice?" We conducted a systematic literature search using
multiple databases, including Ovid Medline (1966-2000), Psyc Info (1984-2000), and Emergency
Medical Abstracts (1995-2001) using a defined search strategy. We also searched the Web sites of
the American College of Emergency Physicians (www.acep.org), the American Academy of
Emergency Medicine (www.aaem.org), and the Canadian Association of Emergency Physicians
(www.caep.ca) for documents containing "shiftwork". We searched the internet for shiftwork
information using the Google (www.google.com) meta-engine. We also searched the University of
Toronto electronic library resources site for relevant journals and references (www.utoronto.ca).
Bibliographies were hand-searched for further references. Finally, we consulted experts in the
fields of chronobiology and emergency physician wellness. RESULTS: Thirty-two initial references
met all of our database inclusion criteria and 15 Web sites were incorporated. Additional sources
added a further 65 relevant references. Shiftwork has negative effects on sleep, performance,
mental, social, and physical health. Strategies for ameliorating these effects exist in 5 categories.
CONCLUSIONS: Shiftwork is essential to EM practice, but has numerous negative effects on EM

17

physicians. Our review identified 5 strategies for minimizing the impact of shiftwork on EM
practice.
084 First Aid Kit Availability and Content Among Trekkers in Nepal.
Fedder S, Abu-Laban RB, Fefer J. Department of Emergency Medicine, Langley Memorial Hospital.
Langley, BC.
OBJECTIVES: Adventure travel has increased the incidence of medical problems in isolated areas.
We sought to determine the preparedness of trekkers in Nepal as manifested by whether they
carried a first aid kit and, if carried, by the kits contents. Our primary objectives were to
determine the proportions of trekkers who: (1) carried first aid kits; (2) carried antibiotics from
each of 3 a priori-defined categories; and (3) carried prophylactic medications for altitude illness.
METHODS: A convenience sample of trekkers who attended free daily information lectures on
altitude illness from 01/October/1998 to 05/December/1998 in Manang, a village midway along a
3 week trek (altitude 3540 m, maximum trek altitude 5416 m) were invited to participate.
Subjects were asked a series of standardized questions during a brief interview by 1 of 2
researchers with multilingual abilities. RESULTS: 121 trekkers were enrolled, the majority of
whom were from Europe (44%), the USA (23%), Australia/New Zealand (20%) or Canada (7%).
The mean age of those studied was 32 years and 63% were male. 97% of subjects carried a first
aid kit (117/121, 95% CI: 91.8%-99.1%), 73% carried at least 1 antibiotic (88/121, 95% CI:
63.9%-80.4%), and 40% carried prophylactic medication for altitude illness (48/121, 95% CI:
30.9%-50.0%). Thirty percent of subjects carried an antibiotic agent from 1 category only; 33%
from 2 categories; and 10% from all 3 categories. Logistic regression models, fit for secondary
purposes, indicated that carrying a complete or near-complete selection of antibiotic categories
and carrying prophylactic medications for altitude illness were both independent of trekker age,
sex, region of origin, days trekking and total days travelling. CONCLUSIONS: Although the
majority of trekkers in the Nepalese Himalayan carry first aid kits, a significant proportion of these
kits lack agents from important antibiotic categories and/or lack prophylactic medications for high
altitude illness. Measures to improve the preparedness of trekkers for medical problems appear
warranted.
085 A Surveillance of Soccer Injury in Canadian Children: A Five Year Canadian
Hospitals Injury Report and Prevention Program (CHIRPP) Perspective.
Shore BJ, Joubert GI. Department of Pediatrics - Emergency Medicine, Children's Hospital of
Western Ontario, London, ON.
INTRODUCTION: In 2000, the Canadian Soccer Association reported that there are 644,028
children under the age of 19 playing organized soccer. The objective of this study was to examine
the relative frequency of soccer injury in children using the CHIRPP. METHODS: A retrospective
study was conducted using the CHIRPP database, incorporating data from 10 pediatric hospitals
and 6 general hospitals since 1990. Soccer injury reports between September 1, 1994 and August
31, 1999 were analyzed. Age ranged from >1 to <19 years of age. Over the five-year period
injuries were analyzed to describe age, gender, context of the injury, body part injured, and
severity of injury. RESULTS: Total data pool consisted of 10,647 records. The greatest number of
inuries was in the 10-14 age group (n = 6281, 62% male, 38% female). The rank order of injuries
were sprains (31.5%), fractures (29.4%) and superficial lacerations (25.5%). Significantly more
injuries resulted from non-competitive (64%) versus (36%) for competitive play (p > 0.01).
97.3% of all injuries were minor. Using hospital admissions as an indicator for injury severity, only
2.7% required admission. Fractures (73%) and head injuries (11%) were the 2 most common
diagnoses requiring hospital admission. Males had an overall higher admission rate (Odds Ratio =
1.37). Male competitive play resulted in higher rates of severe head injuries (Odds Ratio = 7.47)
compared to male non-competitive play (Odds Ratio = 1.21). CONCLUSIONS: Using the CHIRPP
surveillance tool, soccer injuries in Canadian children occur at a greater rate in non competitive
compared to competitive play. The majority of soccer injuries in children are minor in severity.

18

Males are at an increased risk for soccer injuries in general, in particular for those requiring
hospitalizations, and especially head injuries.
086 Patterns of Injury of Canadian Children in Non-Competitive Soccer: A Five Year
Canadian Hospitals Injury and Report Prevention Program (CHIRPP) Perspective.
Shore BJ, Joubert GI. Pediatric Emergency Medicine, Children's Hospital of Western Ontario,
London, ON.
INTRODUCTION: The Canadian Soccer Association reports that in 2000 there were 644,028
children under the age of 19 playing competitive (C) soccer. The objective was to examine the
relative frequency of injury in non-competitive (NC) soccer play using CHIRPP. METHODS: A
retrospective study was conducted using the CHIRPP database, incorporating data from 10
pediatric hospitals and 6 general hospitals. Soccer injury reports between September 1, 1994 and
August 31, 1999 were analyzed. Age ranged from >1 to <19 years of age. Over the five-year
period injuries were analyzed to describe age, gender, the context of the injury, the mechanism of
injury, and the severity of injury. RESULTS: Analysis was done on 8,424 completed records. A
significantly larger proportion of soccer injuries were as a result of NC 64% (n = 5361) play
versus 36% for C play (p > 0.01). Males were twice as likely to be injured in NC group (67.4%)
versus females (32.6%). The 10-14 age group had the greatest number of injuries (3084).
Contact accounted for 84% of all NC injuries. Majority of injuries were minor (96.3%). Using
hospital admission as an indicator of injury severity, only 3.3% required admission. Fractures
(71.9%) and head injuries (10.7%) were the 2 most common diagnoses requiring hospital
admission. Males had an overall higher admission rate (OR = 1.16) and more frequent severe
head injuries (OR = 1.21). CONCLUSIONS: Using the CHIRPP surveillance tool, soccer injuries in
Canadian children occur at a greater rate in non-competitive play. This data shows that the
majority of soccer injuries in children are minor in severity and associated with contact. Males are
at an increased risk for soccer injuries in general, and in particular for those requiring
hospitalizations, especially head injuries.
087 Practice Variation Among Pediatric Emergentologists and Pediatric Orthopaedic
Surgeons in the Management of Wrist Buckle Fractures.
Plint A, Clifford T, Perry J, Bulloch B, Nguyen BH, Miller K, Pusic M, Joubert G, Lalani A, Ali S.
Division of Emergency Medicine, University of Ottawa. Ottawa, ON.
OBJECTIVES: Buckle fractures are the most common wrist fractures in children and frequent
cause of ED visits but there is few studies regarding their management. The purpose of this study
was to examine practice patterns and attitudes of pediatric emergency physicians (EP) and
pediatric orthopedic surgeons. METHODS: A standardized survey assessing management of wrist
buckle fractures and attitudes for immobilization was mailed to all pediatric orthopedic surgeons
and EPs at 9 children hospitals. A modified Dillman's method was used for follow-up. RESULTS:
82% of physicians surveyed responded (31/39 orthopedic surgeons and 79/96 EPs). 63% of EPs
and 68% of orthopedic surgeons believed wrist buckle fractures need to be immobilized (p =
0.28). There was variation among orthopedic surgeons on the length of immobilization
recommended, 71% recommended 2 to 3 weeks and 10% treated only until pain free. EPs
showed great diversity on length of immobilization needed (until pain free [17%], 2 to 3 weeks
[35%], and 1-2 weeks [13%]). 52% of orthopedic surgeons preferred a below elbow cast, 30%
preferred a combination of splint and cast (30%), and 10% preferred a splint. EPs "usually or
always" used a cast (60%) or splint (31%). Among physicians who believed all fractures should be
immobilized, pain control was the most frequently cited reason (95% orthopedic surgeons, 90%
EPs, p NS). Orthopedic surgeons were more concerned about refracture than EPs (76% vs 55%, p
= 0.10). The remaining physicians did not believe all buckle fractures needed immobilization, cited
buckle fractures are stable (67% orthopedic surgeons, 79% EPs, p = 0.46) and have a low risk of
refracture (33% orthopedic surgeons, 67% EPs, p = 0.09). CONCLUSIONS: Although many
physicians believe wrist buckle fractures need immobilization, a significant number disagree.

19

There is variation in the type and length of immobilization used. Given this practice variation, the
optimal management of wrist buckle fractures needs further study.
088 Croup Presentations to the Emergency Department: Description and Outcome.
Rowe BH, Yiannakoulias N, Johnson D, Klassen TP, Bullard M, Spooner CH, Holroyd BR, Svenson L,
Rosychuk R, Schopflocher D. University of Alberta, Edmonton, AB.
OBJECTIVES: This study examines ED presentations of croup and subsequent visits for the same
problem within the year using a large administrative database. METHODS: All patients <20 years
of age presenting to Alberta EDs were eligible for inclusion. Data were derived from a sample of
ED patients treated in 17 health regions over 1 year (1998/99) with a diagnostic code of croup
(464.4). Data were extracted from the Ambulatory Care Classification System (ACCS) database,
consisting of computerized abstracts coded similarly across regions. Diagnostic categories were
coded by medical record nosologists using ICD-9 codes for the primary discharge diagnosis.
Descriptive statistics and crude presentation rates are reported. RESULTS: During the year, there
were 4706 unique croup-related visits to the emergency department by 3933 individuals under 20
years of age. These visits made up roughly 0.3% of the 1.5 million total visits to the emergency
department. Overall, 2702 (66%) of patients were between 1-4 years of age; males presented
more commonly than females. Impressive daily and seasonal variation exists; weekends (35%)
and December-February (38%) numbers were highest. Most visits resulted in discharges from the
emergency department (4209; 90%). There were 464 admissions (9.9%), including 10 to critical
care areas. The 2 urban health regions had lower or significantly lower than average rates of
croup presentation. Repeat visits to the ED for croup were not uncommon; 16.5% of the cases
made at least 1 additional visit to the ED for croup within a year of the first visit. CONCLUSIONS:
These results indicate that croup is a relatively common presentation to the ED. Repeat
presentations and variation in rates of presentation suggest that further evaluation of croup
patients is required to determine the treatment variation for this ED problem.
089 Patients with Community-Acquired Pneumonia Discharged from the Emergency
Department According to a Clinical Practice Guideline - A 2 Year Observational Study.
Campbell SG, Patrick W, Varley-Doyle S, Els M, Murray D, Urquhart D, Maxwell D, Hawass A,
McIvor RA, Hernandez P, McParland C, Haase D. Dalhousie University Department of Emergency
Medicine. Halifax, NS.
INTRODUCTION: Clinical practice guidelines (CPG) decrease admission rates for CAP, although the
safety of decreased admissions in a non-study setting remains unclear. According to the CPG at
our institution, patients with a pneumonia severity score (PSS) of <90, (Fine groups I to III) and
who met each of 4 additional discharge criteria, are discharged, with referral for telephone followup in 24-48 hours. OBJECTIVES: Primary objective: To assess the safety of discharging patients
with CAP according to a CPG based on a pneumonia severity scoring system. Secondary objective:
To assess the utility of a 24-48 hour follow-up call. METHODS: A retrospective chart audit of all
patients identified in the ED database as having been discharged with a diagnosis of pneumonia
during the period 3 Jan 1999-3 Jan 2001. Readmission or death rates within 2 weeks of the
emergency visit were evaluated, using data from all local hospitals and from the provincial
coroner. RESULTS: 867 patients were identified. The average age was 55.5 years. (range 16-98),
and the mean PSI score was 69.2 (range 6-187). 26 (3%) were readmitted within 2 weeks, 15
(1.7%) died within 2 weeks. Of 148 (17.1%) patients referred for follow-up, average age (58.8
vs. 54.9) and PSI scores (67.2 vs. 69.6) were similar to patients not referred. Referred patients
were twice as likely to be readmitted within 2 weeks (4.7% vs. 2.5%). There was no difference in
deaths within 2 weeks between the groups (1.4% vs. 1.8%). CONCLUSION: The use of a CPG to
guide the discharge of patients from the ED appears to be safe. Patient follow up referrals are
infrequently made, and the chance a referral does not appear to be linked to the PSI or age of the
patient, but does appear to be associated with a higher subsequent admission rate.

20

090 Concordance between Radiologist Reports and Emergency Physician Diagnosis of

Community-Acquired Pneumonia in Patients Discharged from an Emergency
Department.
Campbell SG, Patrick W, Varley-Doyle S, Els M, Murray D, Urquhart D, Maxwell D, Hawass A,
McIvor RA, Hernandez P, McParland C, Haase D. Dalhousie University Department of Emergency
Medicine. Halifax, NS.
INTRODUCTION: Chest x-ray (CXR) has long been considered the 'gold standard' for the diagnosis
of community-acquired pneumonia (CAP), however CXR may miss up to 30% of pneumonias seen
on chest CT. Recent suggestions that acute bronchitis not be treated with antibiotics have
highlighted the importance of differentiating the 2 conditions. Radiologists rarely have the benefit
of direct patient contact when deciding on the clinical relevance of seemingly unimportant x-ray
features. Although numerous studies (using CXR as the standard) have described 'miss' rates in
CAP, very little work has been done on cases where the treating physician 'overcalls' the diagnosis.
OBJECTIVES: To evaluate the level of concordance between radiologist reports, (received after
discharge of patients), with the diagnosis of CAP in patients discharged from an emergency
department. METHODS: Three investigators conducted a retrospective chart audit of all patients
identified in the ED database as having been discharged with a diagnosis of 'pneumonia' or
'possible pneumonia' during the period 3 Jan 1999-3 Jan 2001. Emergency physician (EP) and
radiology report (RR) diagnoses were categorized as 'pneumonia', 'possible pneumonia', 'nonpneumonia' and 'normal', and reports for each patient were compared. RESULTS: 867 patients
were identified for audit. Of these, x-rays were performed in 844 (97.3%). RR were not found in
31 cases (3.67%). Of 669 EP diagnoses of 'pneumonia', 304 (37.4%) RR's were in agreement,
although in 82 (10%), the RR diagnosis was of 'possible pneumonia) of 813 EP diagnoses of
'pneumonia' or 'possible pneumonia', 426 (52.4%) of RR's were in agreement. 214 (26.3%) of
RR's in the combined group were of diagnoses other than pneumonia, while 173 (21.3%) were
read as 'normal'. CONCLUSION: EP's and radiologists frequently disagree on whether a patient has
pneumonia or not. Perhaps it is time too revisit the 'gold standard' status of plain chest x-ray.
091 Maintaining Normocapnia Prevents Cerebral Vasoconstriction during Oxygen
Therapy.
Tesler J, Rucker J, Volgyesi G, Fedorko L, Fisher J. Department of Anesthesia, University Health
Network. Toronto, ON.
INTRODUCTION: O2 treatment is accompanied by cerebral vasoconstriction which offsets, or even
reduces, O2 delivery to the brain. Hyperoxia also causes hypocapnia which constricts CO2responsive vascular beds. We hypothesized that preventing hypocapnia during O2 treatment
would prevent oxygen-induced cerebovascular vasoconstriction. METHODS: We exposed 5 normal
subjects to >97% O2 for 3 consecutive 20 minute test periods. Normocapnia was maintained only
during the second test period but subjects were unaware when normocapnia was maintained. We
monitored tidal volume, respiratory rate, and middle cerebral artery blood velocity (MCABV) as an
index of cerebral blood flow. RESULTS: On the initial exposure to hyperoxia, minute ventilation
increased by 21% (P < 0.05), end-tidal PCO2 decreased by 3.7 mm Hg (p < 0.01, paired t test)
and MCABV decreased by 11.5% (p < 0.02, paired t test). During the second test period when
normocapnia was maintained, minute ventilation increased by 77% and MCABV remained at
control values. During the third test period, responses were not significantly different from those
during the first test period. CONCLUSIONS: Maintaining normocapnia prevents the fall in cerebral
blood flow associated with O2 inhalation. Maintaining isocapnia during O2 treatment should
improve O2 delivery to organs with CO2 responsive vascular beds, such as the brain, heart and
kidney.
092 Factors Associated with Activation of the Pediatric Trauma Team for Severely
Injured Children.
Au BL, Shephard AL, Brennan-Barnes M, Osmond MH. McMaster University. Hamilton, ON.

21

INTRODUCTION: Activation of the pediatric trauma team (PTT) in our tertiary-care pediatric centre
is based on specific criteria (physiologic, anatomic, and mechanism). However, there are instances
in which the PTT is not activated for severely injured children. OBJECTIVES: The primary objective
of the study was to determine factors associated with activation of the PTT for severely injured
patients. The secondary objective was to determine whether care by the PTT would decrease
length of stay in the emergency department (ED). METHODS: All patients seen from July 4, 2000
to June 1, 2001 with an Injury Severity Score (ISS) >11 were included. Data were collected from
a trauma registry database. Data collected included: age, gender, ISS, mechanism of injury, need
for surgery, length of stay in the ED, and final disposition (ward vs PICU). RESULTS: 69 patients
with an ISS >11 were seen during the study period. The PTT was activated for 20 patients, and
not activated (NTT) for 49. There were no significant differences between the PTT and NTT groups
(PTT vs NTT) in: mean age (years) (9.35 5.21 vs 9.16 5.19; p = 0.893), proportion male
(13/20 vs 34/49; p = 0.466), mean ISS (23.10 10.99 vs 17.88 4.64; p = 0.052), or fall as
mechanism of injury (3/20 vs 14/49; p = 0.358). The PTT was more likely to be activated for MVA
mechanism (15/20 vs 13/49; p = 0.0003). Proportion of patients admitted to the PICU was
significantly greater in the PTT group (14/20 vs 21/49; p = 0.037). Proportion of patients going to
surgery was similar in both groups (3/20 vs 3/49; p = 0.346). ED length of stay (min) was
significantly lower in the PTT group (177.70 74.36 vs 255.96 203.84; p = 0.026).
CONCLUSIONS: Severely injured patients managed by the pediatric trauma team had shorter
lengths of stay in the ED, were more likely have MVA mechanism, and were more likely to be
admitted to the PICU.
093 Major Injury Associated with All-Terrain Vehicle use in Nova Scotia: A Five Year
Review.
Sibley AK, Tallon JM. Dalhousie University. Halifax, NS.
BACKGROUND: All-terrain vehicle (ATV) riding is a popular recreational sport with approximately
1.5 million users in Canada. Despite legislation to lower ATV injury rates, ATV related incidents are
still a major cause of trauma and death. This paper reviews the epidemiology of major injury
associated with ATV use in Nova Scotia. METHODS: Using the Nova Scotia Provincial Trauma
Registry, all adult (age >15) trauma (ISS 12) related to ATV incidents over a 5 year period were
evaluated. Data were analyzed for demographic variables, temporal statistics, alcohol use, helmet
use, injury characteristics and as well injury outcome variables including Injury Severity Score
(ISS), Length of Stay (LOS), Glasgow Coma Score and discharge status. RESULTS: 25 patients
met the inclusion criteria. The majority of trauma was incurred by males (92.0%) and by persons
between the ages of 15-34 (64.0%), average age 34.4. 71.4% of all trauma occurred between
13:00 hr and 19:00 hr, 52.0% occurred on the weekend and 40.0% of all injuries occurred during
the spring season. Injuries to the central nervous system comprised 39.1% of all major injuries.
The average ISS was 22.1 and the average LOS 21.6 days. Alcohol was involved in up to 56.0% of
all incidents and only 5 patients (20.0%) were known to be wearing a helmet at the time of injury.
INTERPRETATION: ATV related incidents are a continuing source of major injury. This paper
describes the epidemiology of ATV related trauma presenting to the sole tertiary care referral
centre in 1 province. Information gained from this study should be used to influence ATV public
education programs.
in both improved patient care and significant cost-savings.

095 Does CT at a Primary Hospital Delay the Transfer of

Trauma Patients to a Tertiary Centre?
Onzuka J, Worster A. McMaster University. Hamilton, ON.
INTRODUCTION: We feel that delays in transfer of patients to the level 1 trauma centre are due to
imaging procedures done at the primary hospital, namely the CT scan. For this reason, we set out
to identify whether doing CT scans in the primary hospital would delay the transfer of trauma

22

patients to a level 1 trauma centre and whether this affected mortality rate. METHODS: We
undertook a retrospective chart review of all patients that were transferred to the Hamilton
General Hospital (HGH), which services 2.2 million people and 24 hospitals over an area of 13,434
km2, for management of traumatic injuries from primary hospitals in the period including April 1,
1999 to March 4, 2001. Assessments were made to whether doing a CT scan at these primary
hospitals delayed the transfer of patients to the HGH and to assess whether this contributed to a
higher mortality rate. RESULTS: Patients were transferred to the Hamilton General Hospital 85
mins. (95% CI - 65-108) (p < 0.00001) faster if they had not received a CT scan at the primary
hopital (power = 1.00, alpha = 0.05 and n = 72). The 2 groups of patients (those that had CT and
those that did not) were matched for ISS, age, gender and mode of transport to the Hamilton
General Hospital. Analysis of the mortality data comparing the group who received CTs at the
primary hospital vs the group that did not, revealed an OR = 0.87 (95% CI 0.37-2.05)
CONCLUSIONS: Our data clearly identifies a statistically significant delay in the transport of
multisystem trauma patients to a level 1 trauma centre if CT scans were performed on patients in
a primary hospital. At this point, however, we have not significantly correlated this with an
increase in mortality since the OR = 0.87 (95% CI 0.37-2.05) for having a CT at a primary
hospital.
096 Using Electronic Clinical Practice Guidelines in Emergency Medicine.
Meurer DP, Rowe BH, Bullard MJ, Holroyd BR. Emergency Medicine Research Group, Division of
Emergency Medicine, University of Alberta. Edmonton, AB.
OBJECTIVES: Previous efforts to incorporate clinical practice guideliens (CPG) into practice have
met with failure, especially in the emergency department (ED). This study examines the use of an
innovative CPG project as well as the characteristics of resource use recorded by a computerbased health information system. METHODS: The EM CPG is a single sign-on, intranet, desktop
application for emergency department clinical decision making. This product consists of decision
tools, in- and out-patient order sets, patient information and important links; all EM physician staff
at 1 major teaching hospital affiliated with the University of Alberta had access. The data for the
usage characteristics were derived anonymously from user logs and spanned the first 8 months
this resource was available. Whenever possible, clinicians completed brief questionnaires using a
7-point Likert scale at the conclusion of their encounter. RESULTS: 24 (96%) of 26 EM physicians
accessed the site and there were 322 recorded uses to the CPG program over the study period.
The "helpfulness" (median = 6.0; IQR: 5, 7) and "ease of use" (median = 6.0; IQR: 5, 7) was
rated as "high" by 130 users. Also, "increasing confidence with treatment" (median = 5.0; IQR: 4,
6) and "improving quality of care" (median = 5; IQR: 4, 6) received "moderately high" ratings.
Most (80%) EM physicians used 1 product during an interaction. Frequently used resources were
community acquired pneumonia decision rules (n = 55), swollen limb assessment sets with Well's
criteria (n = 94), the IV out-patient treatment order form (n = 25), and the head injury patient
information form (n = 27); specific clinic consult forms were also popular (n = 32). For 3 of the
most commonly used resources, forms with pre-formatted "no" responses were used 66% of the
time. Use of the CPG resources increased 43% in the second 4-month period of the study.
CONCLUSIONS: An intranet CPG dedicated to the management of common emergency
department problems has been well received by most staff and rated very highly for ease of use
and helpfulness. Further implementation and evaluation of interventions designed to improve the
use of EBM resources, such as CPGs, appear warranted.
097 Using Clinical Practice Guidelines in Emergency Medicine.
Rowe BH, Meurer DP, Bullard M, Holroyd BR. Division of Emergency Medicine, University of
Alberta, Edmonton, AB.
OBJECTIVES: Previous efforts to incorporate clinical practice guidelines (CPG) into practice have
met with failure, especially in the emergency department (ED). This study examines the use of an
innovative CPG project as well as the characteristics of resource use recorded by a computerized
system. METHODS: The EM CPG is an intranet-based desktop application for ED clinical decision

23

making. This product consists of decision tools, order sets, patient information and important
links; access was provided to all 26 EM physicians at 1 teaching hospital. The data for the usage
characteristics were derived anonymously from user logs and spanned the first 8 months of use.
Whenever possible, clinicians completed brief questionnaires using a 7-point Likert scale at the
conclusion of their encounter. RESULTS: 24 (96%) of staff physicians accessed the site and there
were 322 recorded uses to the CPG program over the study period. The overall helpfulness
(median = 6.0; IQR: 5, 7) and ease of use (median = 6.0; IQR: 5, 7) were rated as high by 130
users. Most (80%) EM physicians used only 1 product during an interaction. The most highly used
resources were community acquired pneumonia decision rules (n = 55), swollen limb assessment
sets (n = 94), IV out-patient treatment order form (n = 25), and the head injury patient
information form (n = 27); specific clinic consult forms were also popular (n = 15). For 3 of the
most commonly used resources, defaulted forms (with pre-formatted The overall "helpfulness"
(median = 6.0; IQR: 5, 7) and "ease of use" (median = 6.0; IQR: 5, 7) was rated as "high" by
130 users. Most (80%) EM physicians used only 1 product during an interaction. The most highly
used resources were community acquired pneumonia decision rules (n = 55), swollen limb
assessment sets with Well's criteria (n = 94), the IV out-patient treatment order form (n = 25),
and the head injury patient information form (n = 27); specific clinic consult forms were also
popular (n = 15). For 3 of the most commonly used resources, defaulted forms (with preformatted 'no' responses) were used 66% of the time. Use of the CPG resources increased 43% in
the second 4-month period of the study. CONCLUSIONS: An intranet CPG dedicated to the
management of common emergency department problems has been well-received by most staff
and rated very highly for ease of use and helpfulness. Further implementation and evaluation of
interventions designed to improve the use of EBM resources, such as CPGs, appear warranted.
098 Cellulitis in the ED: Factors Associated with Treatment Failure.
Murray HE, Stiell IG, Wells GA. Kingston General Hospital, Kingston, ON, and the Ottawa Hospital,
Civic Campus, Ottawa, ON.
OBJECTIVE: This preliminary study identified both the expected rate of treatment failure and the
historical features and clinical characteristics that are associated with treatment failure in ED
patients with cellulitis. METHODS: This prospective observational cohort study was performed in a
tertiary care centre with ~50,000 annual visits. Adult patients with cellulitis had a standardized
physician assessment performed prior to the initiation of treatment. The primary outcomes were
clinical response or treatment failure, which was defined as any 1 of the following poor outcomes:
I&D of abscess, change in antibiotics (not due to allergy/intolerance) specialist consultation or
admission to hospital. Comparison of the means and proportions between the 2 groups was
performed with univariate associations, using parametric or non-parametric tests where
appropriate. RESULTS: 80 patients with 78 episodes of infection were entered. The patients were
60% male, mean age 49 (SD 19) with 76 (95%) extremity cellulitis and 11 (14%) abscess with
cellulitis. 14 episodes (17.5%) were classified as treatment failures. This can be further broken
down into an oral antibiotic failure rate (6.8%) and an ED-based IV antibiotic failure rate (26.1%).
Patients with treatment failure were older (mean age 59 vs. 46, p = 0.02) and more likely to have
already taken oral antibiotics (50% vs. 17%, p = 0.01). Patients with olecranon bursitis were also
more likely to fail treatment (29% vs. 9%, p = 0.05). CONCLUSIONS: The treatment of cellulitis
with daily ED-based IV antibiotics is a relatively new phenomenon. A clinical trial of this practice is
needed to determine which patients require IV therapy or admission. Patients with previous
(failed) oral therapy and those with olecranon bursitis are more likely to fail ED treatment for
cellulitis and should not be randomized in a clinical trial of oral vs. ED based IV antibiotics.
099 Cellulitis in the ED: Factors Affecting Treatment Decisions.
Murray HE, Stiell IG, Wells GA. Kingston General Hospital, Queen's University. Kingston, ON.
OBJECTIVES: The correct ED treatment of cellulitis is not clear. This study examined the historical
and clinical characteristics that determine the severity of a cellulitis episode. METHODS: This was
a prospective cohort study from a tertiary care centre with ~50,000 annual visits. Adult patients

24

with cellulitis had a standardized MD assessment prior to initiating treatment. Relevant historical
features and objective measurements including infection size were recorded on the data form. The
primary outcome was a treatment-based severity classification: those treated with ED-based IV
antibiotics were considered 'severe'and those with oral antibiotics 'mild.' Means and proportions
were compared between the 2 groups with univariate associations (using parametric or nonparametric tests where appropriate). ROC curves were constructed for significant continuous data.
RESULTS: The 64 study patients had a mean age of 45 years, 61 (95%) had extremity infections
and 8 (12.5%) had abscesses with cellulitis. 27 episodes were 'mild' and 37 'severe.' Patients with
severe cellulitis were more likely to report a previous history of cellulitis (32.4% vs. 7.4%, p =
0.02), fever (31.4% vs. 11.1%, p = 0.05) or systemic symptoms (38.9% vs. 3.7%, p < 0.01).
There were no differences in demographics or the presence of co-morbidities. The size of infection
was larger in severe infections (637.7 cm2 vs. 219.9 cm2, p < 0.01). The area under an ROC curve
of size vs. severity was 0.78 (95% CI 0.67, 0.90). There was no size cut point with 100%
sensitivity for severe infections. CONCLUSIONS:This is the first prospective study to evaluate the
characteristics determining cellulitis severity. Patients with previous cellulitis, larger size of
infection and systemic symptoms were more likely to be treated with IV antibiotics. However, the
absence of a clear division between the groups allows ethical randomization of patients with all
size infections into a proposed clinical trial comparing oral vs. IV antibiotics in cellulitis.
100 A Survey of Influenza Vaccination Rates Amongst Emergency Department
Personnel.
Saluja IS, Theakston K. London Health Sciences Centre, Emergency Department. London, ON.
INTRODUCTION: During the influenza season of 1999-2000, emergency department (ED) health
care workers at UWO teaching hospitals were surveyed to investigate their influenza vaccination
rates, motivating factors and attitudes toward vaccination. METHODS: An anonymous 28-item
survey was distributed to emergency physicians and residents, nurses, respiratory therapists
(RTs), and other allied healthcare workers. Statistical analysis was done using SPSS v.10.
RESULTS: 343 surveys were returned for an overall response rate of 81%. The respondents were
75% female, 87% nonsmokers, with a mean age of 38. The overall vaccination rate was 37%. The
RTs had the highest vaccination rate of 46%, the allied healthcare workers the lowest at 27%, and
the physician's rate was 35%. Logistic regression analysis revealed that respondents with a
chronic medical condition were almost twice as likely to receive vaccination (OR 1.96, p = 0.018).
With regards to perceptions and attitudes, 28% felt adverse affects were common, 51% felt
vaccination was effective, 52% would support a program to improve vaccination rates, and 41%
would support mandatory vaccination. Only 27% felt that patients are at an increased risk of
getting influenza from ED staff, but 58% perceive that ED staff are at an increased risk of getting
ill from patients. CONCLUSIONS: While there is a perception of increased risk of influenza
transmission in the ED, the immunization rate amongst ED personnel was only 37%, and the
majority (59%) did not support mandatory immunization. When controlled for baseline
characteristics, the only significant motivator to get vaccinated that was identified was the
presence of a chronic medical condition. There is good evidence that influenza immunization of the
elderly and nursing home workers decreases mortality, however more work needs to be done
regarding the efficacy of ED personnel influenza vaccination.
101 Pneumonia Presentations in the Emergency Department: Description and Outcome.
Spooner CH, Rowe BH, Yiannakoulias N, Bullard M, Holroyd B, Craig W, Klassen T, Johnson D,
Svenson L, Rosychuk R, Schopflocher D. Division of Emergency Medicine, University of Alberta.
Edmonton, AB.

OBJECTIVES: Pneumonia is a common condition that presents to the emergency department (ED) but
the epidemiology of this problem is understudied. This study examines all ED pneumonia visits within
a large, standardized health care region for 1 fiscal year. METHODS: All patients presenting to Alberta
EDs were eligible for inclusion. Data were derived from a sample of ED patients treated in 17 health
regions over 1 year (98/99) with a diagnostic code of pneumonia (486.x; but not influenza). Data were
25

extracted from the Ambulatory Care Classification System (ACCS) database, consisting of
computerized abstracts coded similarly across regions. Diagnostic categories were coded by medical
record nosologists using ICD-9 codes for the primary discharge diagnosis. Descriptive statistics and
crude presentation rates are reported. RESULTS: Overall, 1.49 million ED visits were recorded in the
year; the number of patients with a diagnosis of pneumonia was 17,162 (1.2% ED visits). Overall, 70%
were under the age of 65 years with a peak at 1-4 yrs (15.2%); male / female representation 52%/48%.
Limited daily variation existed; Saturday-Monday (~15.5%), Thursday (12.8%). However, seasonal
variation was noted: December-February (11.5-13.5%) numbers were highest, June-September lowest
(6.2-6.5%). Most patients were discharged (63.8%); however, admission (5924; 34.5%) was higher
than the ED average (9%). Few patients left prior to seeing a physician (9, <1%). The rate of
pneumonia varied between regions, with an average of 3.8/1000 population across the province; urban
areas had the lowest rate of presentation. More than 1 presentation for pneumonia was recorded for
16% of visits (2 or more visits: 10.3%). CONCLUSIONS: These results indicate that pneumonia is a
relatively common presentation to the ED, and admission rates are high. Further evaluation of
pneumonia patients in the ED is required to understand the observed variation and to evaluate

S-100b protein levels as a predictor for long-term disability after

head injury
Report by:

John-Paul Lomas - Medical Student

Joel Desmond - RCS Research
Search checked by:
Fellow
Institution:
Manchester Royal Infirmary
Date submitted:
18th June 2003
Last modified:
25th July 2003
Status:
Blue (submitted but not checked)
Three part question
In [patients with a head injury] do [levels of S-100B protein] predict [long-term
disability]?
Clinical scenario
A 17 year old male presents to the emergency department after a road traffic
accident. His GCS was 8 on arrival but an immediate CT scan showed no focal
abnormality. His GCS returned to 14 after 4 hours. You are talking to his mother
who is reassured that he does not need urgent neurosurgery, but she asks
whether he will suffer any long term consequences from this injury. You tell her
that it is difficult to predict, but you have recently head that S-100 protein
measurement is available in your hospital for research purposes. You wonder
whether S-100 could help predict his long term prognosis.
Search strategy
Medline 1966-06/03 using the OVID interface.
[(exp S100 Proteins/ OR s100.mp OR s-100.mp) AND (exp Brain Injuries/ OR
brain injury.mp OR exp Craniocerebral trauma/ OR head inj$.mp)]
26

Search outcome
Out of 138 papers, 12 were found to be relevant. 2 relevant papers decribed the
same patient population. The remaining 11 papers are shown in the table.
Relevant paper(s)
Author,
date and
country

Patient
group

Study
type
(level of
evidence)

Outcomes

Key results

Study
weaknesses

Patients with
GOS 3-5 S-100
level mean
30 patients
Glasgow Outcome 1.2mcg SD 1.8
with a severe
Scale on discharge
head injury
(Mean day 19 in
Patients with
(GCS<=9) and
Nonsevere group and GOS 1-2
11 with minor
independent
mean day 1.3 in (unfavourable)
Rothoerl et head injury
gold
Diagnostic minor head injury S-100 level
al,
(GCS 13-15)
standard
Cohort
group)
mean 4.9mcg/l
1998,
admitted to a
Small,
study (4)
SD 5.3
Germany
neurosurgical
selected
unit
cohort of
P=0.0025
S-100 levels
patients
25 of 27
measured
Elevated S-100
mean 2.5 hrs
Detectable level of levels were
after injury
S-100 (>0.5mcg/l) found in the
minor head
injury group
For s-100 level
82 patients
Glasgow outcome
of >2.5mcg/l, No
after severe
score at 6 months
Raabe A et
unfavourable confidence
head injury
Diagnostic
al,
outcome was intervals
(GCS<=8)
cohort
Unfavourable
1999,
predicted with presented
s-100 taken at study (2b) outcome defined
Germany
Sensitivity
Nonadmission and
as severe disability
44%
consecutive
every 24 hours
or vegetative state
Specificity 97%
44 patients
For S-100 level
after severe
Tables 2, 3
Glasgow outcome of >2mcg/l,
Woertgen et head injury
and 4 are
Diagnostic score calculated at PCS symptoms
al,
(GCS score
incorrect,
cohort
mean 11 months predicted with
1999,
<=8)
with erratum
study (3b) after trauma (GOS Sensitivity
Germany
S-100 taken 1printed in a
1-3 unfavourable) 95%
6 hrs after
later edition
Specificity 70%
injury
Ingebrigtsen50 patients
Diagnostic Neuropsychological11/36 patients Very small
et al,
with minor
Cohort
testing at 3
had S-100
study with
27

1999,
Sweden

months (for
attention,
psychomotor
speed, trailmaking test,
memory, digit
span) In 36
patients

head injury
and LOC (GCS
13-15)
referred to
Neurosurgery
study (3b)
dept after CT
scan
S-100 taken
hourly up to
MRI and CT scan
12 hours
findings within
48hrs

>0.2mcg/l
There were
non significant
trends to
no sample
reduced
size
impairment in
estimates
the S-100
Non
negative group
consecutive
4 of 5 patients
Only 36 of
with brain
50 patients
contusion had
followed up
S-100
at 3 months
>0.4mcg/l

Sensitivity
80%
(p=0.035)
Patients with a
Rivermead
positive S-100
postconcussion
had mean RPQ
182 patients
symptoms
Ingebrigsten
6.0 vs 4.0 in S-No
from 3 centres
questionnaire
et al
100 negative sensitivities
with GCS 13score (RPQ)
2000
group p=0.07 or
15 and brief
Diagnostic
Scandinavia
Detectable S- specificities
Loss of
Cohort
(3 centres
100 predicted given for
Consciousness.Study (2b)
Sweden,
intracranial
prediction of
Intracranial
Denmark,
pathology
long term
S-100 taken
Pathology on CT
Norway)
with:
disability
on admission
scan at <24 hours
Sensitivity
90%,
Specificity 65%
Mussack T 80 patients
Diagnostic S-100 in Minor
Patients
No gold
et al,
presenting
study (4) Head Trauma pts discharged
standard
2000,
with a history
<=6hrs 0.29 outcome
Germany
of minor head
+/- 0.11 ng/ml measures
trauma (GCS
Non
13-15)
Patients
consecutive
Also 10pts
discharged >= Results not
with severe
24hrs 0.70 +/- clearly
head injury
0.19 ng/ml
presented
(GCS<8)
Non
S-100 taken at
Patients
significant
0h, 6h and
subsequently findings
24hrs post
admitted to
between
admission
ICU 5.03 +/- groups
28

50 patients
GCS 13-15
after normal
CT scan

3.18 ng/ml
Patients with
Low number
5.26 +/Severe head Injury
of patients
1.56ng/ml
GCS<8
At 2 weeks, S100 of
>0.14mcg/l
predicted
positive
outcome:
69 patients
Inclusion
Sensitivity
admitted to a
criteria for
Intracranial
69%
Herrmann neurosurgical
patients
pathology on CT Specificity 90%
et al,
unit (mostly
Diagnostic
unclear
scan at 2 weeks
2001,
GCS >13)
study (3b)
Only 29
and 6 months, or At 6 month, SGermany
S-100 taken at
patients
focal neurology
100 of
1, 2 and 3
followed up
>0.14mcg/l
days
to 6 months
predicted
positive
outcome:
Sensitivity
65%
Specificity 89%
Patients with
20 patients
GOS 1-3 S-100 Data not
with severe
mean level
clearly
head injury
Chatfield DA
Glasgow outcome 2.46 +/presented
(GCS<=8)
Diagnostic
et al,
score at 6 months 0.32mcg/l
Small study
admitted to
cohort
2002,
after trauma (GOS Patients with No cut off
neurosurgical study (4)
UK
1-3 unfavourable) GOS 3-5 S-100 points or
unit
mean level 0.6 ROC curves
s-100 on
+/-0.1mcg
calculated
admission
P<0.05
Townend WJ 148 adult head Diagnostic Extended Glasgow SWide
et al,
injury patients study (2b) outcome score at 100>0.32mcg/lconfidence
2002,
(GCS 4-15) in
1 month
predicted
intervals
UK
4 hospitals.
severe
Non
Most had a
disability (15 consecutive
minor head
patients with Wide
injury
GOSE<5):
definition of
S-100 levels
Sensitivity
head injury
taken within 6
93% (63%(including no
hours of head
100%)
LOC)
injury
Specificity 72% 80% follow
(54%- 79%) up rate
NPV 99%
29

Spinella et
al,
2003,
USA

27 children
Pediatric Cerebral
(<18yrs) with
performance
traumatic
Diagnostic
category score
brain injury
cohort
(PCPC) assessed
S-100 taken study (3b)
at discharge and 6
within 12
months
hours

172
Savola O & consecutive
Hillbom M, patients with
2003,
mild head
Finland
injury (GCS
13-15)

Post concussional
symptoms defined
Diagnostic by Rivermead
cohort
Post-Concussion
study (2b) Symptoms
Questionnaire at
2-6 weeks

(93%-100%)
For s-100 level
Very small
of >2.0mcg/l,
study
unfavourable
Confidence
outcome was
intervals not
predicted with
given
Sensitivity
Non
86%
consecutive
Specificity 95%
For s-100 level
of >0.50mcg/l, No
PCS symptoms confidence
predicted with intervals or
Sensitivity
sample size
27%
calculations
Specificity 93%

Comment(s)
All studies were under 200 patients in size and most were under 100 patients.
The studies find sensitivities from 27%-95% and specificities from 70% to 97%.
The reasons for this great variation in findings may in large part be due to the
small sample sizes. The specificities seem to perform better than the sensitivities
and thus the finding of a high S-100 may indicate that your patient is at high risk
of long term disability.
The cut-points for a significant S-100 level differ between studies also and are
generally much higher when applied to patients after a severe head injury. Most
studies agree that S-100 levels must be taken within 6 hours of head injury.
Clinical bottom line
A high S-100 level is a marker of poorer long term outcome post minor and major
head injury.

30