American

Journal of
Emergency Medicine
(2005)arrest
23, 304teams
310
EM residents
leadership
for cardiac

www.elsevier.com/locate/ajem

Emergency medicine residents effectively direct

B
inhospital cardiac arrest teams
Bruce D. Adams MDa,c,*, Kathy Zeiler RN, MSNb,
Walter O. Jackson MDc, Brian Hughes MDc
a

Department of Emergency Medicine, Brooke Army Medical Center, San Antonio, TX 78234-6200,
USA
b
Department of Emergency Medicine, University Hospital, Augusta, GA 30901,
USA
c
Department of Emergency Medicine, Medical College of Georgia, Augusta, GA 30912,
USA
Abstract
Study Objective: We compared 2 models of physician leadership for inhospital cardiac arrest teams
(CATs): emergency medicine (EM) residents and staff hospitalist physicians.
Methods: A before-after study was conducted on all adult inhospital CAT activations over a 2-year
period. The primary outcome was return of spontaneous circulation (ROSC).
Results: There were 749 total code blues during the 2-year study period. Ninety-one were excluded by
protocol. EM residents directed 288 codes, hospitalists directed 248 codes, and other specialties
directed the remaining 62. There was no statistically significant difference in percent ROSC or
survival to hospital discharge. EM residents responded first for 59.2% of the codes compared with a
first response rate of 28% for hospitalists ( P b .05). Time to achieve ROSC was quicker in the EM
resident cohort. Conclusion: Our findings validate the use of a 24-hour EM resident staffing model for
CAT response to inhospital cardiac arrests.
Published by Elsevier Inc.

1. Introduction
bCode blueQ or inhospital cardiopulmonary resuscitation
(CPR) occurs frequently and unexpectedly [1,2]. Unfortunately, outcomes continue to be poor with sustained return
of spontaneous circulation (ROSC) of about 40% to 60%
[3] and survival to hospital discharge of 15% or less
B

There were no external sources of financial support or grants for

this study. This study was presented as a poster at the Council of
Emergency Medicine Residency Directors Best Practices Conference
(Washington, DC, February 2003) and at the American College of
Emergency Physicians Scientific Assembly Research Forum (Boston,
Mass, October 2003).
T Corresponding author. Department of Emergency Medicine, Brooke
Army Medical Center, San Antonio, TX 78234-6200, USA.
E-mail address: bruce.adams@amedd.army.mil (B.D. Adams).
0735-6757/$ see front matter. Published by Elsevier Inc.
doi:10.1016/j.ajem.2005.02.013

[1,3,4]. Multiple inhospital CPR studies previously have

looked at variables that are difficult to modify such as
presenting rhythm [1,5], sex [6], comorbidity, [7] hospital
type [2,8], ward location [1,2], bdo-not-resuscitateQ (DNR)
status [2], and circadian variation (day shift vs night shift)
[2]. Rapidly responding with a multidisciplinary team is one
of the few controllable elements that leads to improved
outcomes [4,9,10].
Attending physicians often are not immediately
available to supervise advanced cardiac life support
(ACLS) measures on their patients. Accordingly, many
hospitals maintain a cardiac arrest team (CAT), the
composition of which varies considerably (Table 1)
[2,9,11-15]. Whereas well-trained team leaders in ACLS
improve patient survival [16], ACLS knowledge and
protocol compliance can be surprisingly low

Table 1

Select models for CPR code team leadership

Model

Comments

Nurse or
paramedic

Smaller hospitals Less

expensive

Advantages

Disadvantages
Medicolegal
issues, less
expertise
Not all
physicians
at ease
with ACLS

bAny doctorQ First available

physician
leads code

Possibly
rapid
response

Attending
physician

May direct
by phone

EM staff

EM attending
responds
to ward
until relieved

Familiar with Not always

family, staff physically
present
Experienced Abandonment
responder
of ER
patients

Resident
coverage

IM, EM,
anesthesiology,
and cardiology
fellows, etc
New model,
focus of
current study

Hospitalists

24-h coverage Variable

supervision

Experienced, May not

often
be on 24-h
subspecialists coverage

Many factors are involved in designing staff coverage including salary

of the physicians, staff experience, acuity of the inpatients, staffing of
the ED, familiarity with nursing staff and equipment, and feasibility of
bany doctorQ being able to abandon clinical responsibilities to respond
to a code. IM indicates internal medicine.

among nurses [17], house staff [18], and attending physicians [19,20]. If experienced leadership of a CAT is vital,
which type of physician (or perhaps even nonphysician)
should run the team? Some hospitals use emergency
medicine (EM) resident physicians to direct their CAT.
However, this practice has not been previously evaluated,
and it is unknown whether EM resident leadership and staff
physician leadership of the CAT are equally effective.
Survival outcome differences recently have surfaced
between different specialists and their respective management of several clinical conditions [21-24]. Whether this
medical specialty effect extends to inhospital CPR is
currently unknown. Particularly relevant to this issue is
the newly emerging specialty of the hospitalist. Hospitalists
have been shown to improve admission rates, length of
stay, hospital costs, bedside teaching, and even some
clinical outcomes [15,25], but the particular role of
hospitalists in code blue is also not well described. The
purpose of our study was to determine which physicianstaffing model is best for the CAT: EM residents or staff
hospitalists?

2. Materials and methods

This was a before-after study design in a 550-bed
community teaching hospital in the southeastern United
States. The hospital annually admits 21,000 patients and
performs 5,000 interventional cardiac procedures. The

institutional review board reviewed and approved the

protocol. Participant physicians were blinded to the presence of the study to avoid the Hawthorne effect (except for
1 author who directed 15 code blues as an EM resident and
4 code blues as an EM staff). A code blue was defined as
respiratory distress requiring emergent airway management
or a cardiorespiratory arrest requiring CPR or
defibrillation. The subjects were all adult inhospital code
blue activations between June 1999 and July 2001
excluding those that
occurred in the ED, operating room, or cardiac catheterization suite. Also excluded were visitors to the hospital,
pediatric arrests, patients with a preexisting DNR status
(only if no resuscitation efforts were made), and false
alarms where there was neither cardiac nor respiratory
arrest.
A CAT composed of a pharmacist, chaplain, respiratory
therapist, nurses and physician team leader responded to all
code blue activations (Table 1). The specialty of the
physician responder was the independent variable for our
study. During the first year of the study, the physician
designated for responding to the code was an EM resident
who carried a dedicated pager during ED duties. On June 20,
2000, the local EM residency withdrew, and the community
hospital replaced the residents with 24-hour in-house staff
hospitalists. On November 1, 2001, EM residents were
rehired to cover code blue response between 11:00 pm and
7:00 am, with the hospitalists now covering the balance.
Hospitalists were defined for this study as staff internal
medicine specialists or subspecialists with the assigned
responsibility to supervise the CAT. The EM residents
were all currently trained in ACLS. The hospitalists were
not required by the hospital to have either ACLS or basic
life support certification. Notification of the CAT was by
pager and overhead public address system. Multiple
physicians not on the CAT often initially responded to the
arrest.
Data were collected for all patients on 2 supplemental
forms designed along the Utstein style [26]. The CAT nurse
recorded clinical information contemporaneously, and the
physician directing the code then reviewed and signed the
forms at the codes conclusion. The data were entered into
a spreadsheet for analysis [27]. We pulled and reviewed
229 of the full inpatient charts because of missing data
(208 records) or because of concerns regarding physician
leadership or ACLS compliance (21 records). We evaluated
statistical significance with the t test and by Mann-Whitney
rank sum test for continuous parameters. We used the
Fishers exact test for categorical data and calculated the
95% confidence intervals (CIs) for survival outcomes.
Because many factors can influence CPR survival, we also
performed a multivariate logistic regression analysis.
The primary outcome was immediate survival (defined
as sustained ROSC at termination of code or at transfer).
Secondary outcomes included survival to hospital discharge, code duration, and the time intervals from arrest
(t 0) to first defibrillatory shock, endotracheal intubation,
first epinephrine dose, ROSC, CAT physician arrival, and
patients primary attending physician arrival (not necessar-

306
EM residents leadership for cardiac arrest teams
Table 2 Specialty composition of each major group of
physicians
Specialty
ED residents (total = 288)
PGY
4 (EM-1)
3 (EM-2)
2 (EM-3)
Hospitalists (total = 248)
Cardiology
Family medicine
Gastroenterology
General internal medicine
Hematology-oncology
Infectious disease
Nephrology
Endocrinology
Pulmonary/critical care
Other HCPs (total = 122)
Nurse only
Anesthesiologist/CRNA
Cardiothoracic surgery
General surgery
Emergency medicine staff
Urgent care physician
Other physician
Unknown

No. of
physicians

No. of codes
directed

9
12
14

99
109
80

22
2
2
31
4
1
9
1
9

46
2
2
124
5
1
31
3
34

6
11
9
4
4

5
4
26
16
20
42
4

Table 4

Initial rhythm of code blue

Initial rhythm EM residents Hospitalist Other HCP All

VF/VT
Asystole
Bradycardia
Sinus
PEA
Other
Not recorded

27
59
40
54
35
10
63

12.0
26.2
17.8
24.0
15.6
4.4

30
50
37
24
30
15
62

16.1
26.9
19.9
12.9
16.1
8.1

15
22
19
9
21
5
31

16.5
24.2
20.9
9.9
23.1
5.5

72
131
96
87
86
30
156

14.3
26.1
19.1
17.3
17.1
6.0

There was no significant difference between VF/VT and non-VF/VT

rhythms between groups ( P = .401, by Fishers exact test). PEA
indicates pulseless electrical activity.

clearly in charge,Q as determined contemporaneously by the

CAT supervising nurse and retrospectively by the hospital
quality assurance director. Finally, we recorded whether
the physicians performed their own intubation and use of
DNR orders.

3. Results

PGY indicates postgraduate year; EM, level of EM residency training;

CRNA, certified registered nurse anesthetist.

ily the hospitalist). We also determined which physician

arrived first. We implicitly evaluated physician compliance
with ACLS guidelines and whether bthe physician was

Table 3

B.D. Adams et 306

al.

There were 749 code blues over the 24-month study

period. Ninety-one code blues were excluded by protocol
(catheterization laboratory = 6, EMS or ED = 71, hospital
visitors = 3, operating room = 8, and pediatric = 3). This
left
658 adult inhospital cardiac arrests among 582 unique
patients (there were 64 patients with multiple codes). The
distribution of CAT leader by specialty was EM residents
(n = 288), hospitalists (n = 248), and all other health care
providers (HCP) (n = 122). The EM residents directed

Baseline clinical characteristics

Parameter
Age
Average
Range
Gender
Male
Female
Not recorded
Race
Black
White
Other
Not recorded
Time of day of code
Day shift (7:00 am-7:00 pm)
Night shift (7:00 pm-7:00 am)
Type of arrest
Respiratory only
Cardiorespiratory
T P b .05 by Fishers exact test.

EM residents

Hospitalists

Other HCP

All

69.0
16-102

66.8
19-102

66.3
29-90

67.3
16-102

109 (47.2%)
122 (52.8%)
57

103 (48.1%)
111 (51.9%)
34

54 (52.9%)
48 (47.1%)
2

266 (48.6%)
281 (51.4%)
111

121 (54.8%)
99 (44.8%)
1 (0.5%)
67

94 (44.9%)
114 (54.5%)
1 (0.5%)
39

45 (45%)
55 (55%)
0
22

260 (49.1%)
268 (50.6%)
2 (0.4%)
128

89 (30.9%)T
199 (69.1%)T

161 (64.9%)T
87 (35.1%)T

66 (54.1%)
56 (45.9%)

316 (48.0%)
342 (52.0%)

57 (19.8%)
231 (80.2%)

37 (14.9%)
211 (85.1%)

21 (17.2%)
101 (82.8%)

115 (17.5%)
543 (82.5%)

236 codes in their capacity as ED residents and led 52 more

codes in their capacity as bmoonlightingQ overnight CAT

Table 6

Airway management of code blue

physicians. Altogether, 150 individual physicians from

16 different specialties directed the 658 code blues (Table 2).
The baseline characteristics between the different physician groups were similar except that residents more likely
directed codes during the night shift and hospitalists
directed significantly more day shift codes (Table 3). The
distribution of presenting dysrhythmia was similar between
physician groups (Table 4). Codes located on monitored
units occurred more often for the hospitalists (75.4%
monitored, 24.6% unmonitored setting) than the residents
(65.3% monitored, 34.7% unmonitored) ( P = .024). There
was considerable year-to-year spillover in this before-after
study, but most resident codes (78.5%) occurred in the first
year of the study, and most hospitalist codes (57.5%)
occurred in the second year, P b .0001.
For the primary outcome of ROSC, there was no
significant difference between the groups: EM residents =
61.1% (95% CI, 55.5%-66.7%), hospitalists = 58.5%
(95% CI, 52.3%-64.6%), all other HCP 65.6% (95% CI,
57.1%-74.0%), and all patients combined = 61.3% (95%
CI,
57.8-64.8%). For the outcome of survival to hospital
discharge, there was also no significant difference between
the groups: EM residents = 24.3 % (95% CI, 19.7%29.6%), hospitalists = 22.6% (95% CI, 17.8%-28.2%),
all other HCP 33.6% (95% CI, 25.8%-42.4%), and all
patients combined = 25.4% (95% CI, 22.2%-28.8%). No
survival difference was observed between all EM
physicians (EM residents + EM staff) versus non-EM
physicians, nor between the various medicine
subspecialties. Analysis excluding the 19 unblinded code
blues directed by a coauthor did not change the statistical
findings.
Multivariate logistic regression reconfirmed that physician specialty was not associated with immediate survival.
Codes occurring in monitored intensive care unit settings
(vs unmonitored ward settings) and respiratory arrests
(vs cardiorespiratory arrest) were associated with better

Table 5 Average duration and time intervals (minutes:

seconds) for each group of physicians
Parameter

EM
Hospitalists All others All
residents

Duration of code
23:43
t0
Compressions
2:34
First shock
3:23
First drug
4:18
Achieve airway
7:53
First doctor arrival 5:13
ROSC
4:51T
Primary attending
physician arrival

29:50T

25:53

25:29

25:05

2:09
3:22
4:17
8:54
4:12
11:25T

2:12
1:55
4:04
11:26
8:32
12:35T

3:24
2:37
4:15
9:00
5:45
9:52

17:34T

20:14T

23:35

T P b .001 by Mann-Whitney rank sum test.

Intubated by:

EM
Hospitalist Other HCP All
residents
n

Directing physician
Respiratory therapist
CRNA or
anesthesiologist
Already intubated
Never intubated
DNI/DNR status
Not recorded

26 5.6 29
109 37.8 74
0 0.0 3

6.9
29.8
0.8

16
42
2

7.9
34.4
1.1

71
225
5

54 12.2 65
26 6.7 26
13 3.6 13
60 17.2 38

16.7
8.0
4.3
13.3

31
10
3
18

16.8
6.5
2.1
12.9

150
62
29
116

survival. Age, sex, duration of code, and time of day (shift)

were not associated with survival.
The secondary outcome Utstein styledefined time
intervals are posted in Table 5. Time to achieve ROSC
was significantly quicker in the resident cohort. In the
430 index cases where the nurse supervisor identified the
sequence of arriving physicians, EM residents were the first
to arrive 255 times (59.3%) compared with hospitalists
arriving first to 120 code blues (27.9%) ( P b .05). The time
until first physician arrival was contemporaneously
recorded for only 49 codes. The average response times
were not statistically different between the physician
groups, with both the residents and hospitalists arriving
within 4 minutes in 67% of code blues. The attending
physicians arrival was recorded for 564 code blues. The
attending of record was bedside at the start of the code in
41 (7.3%) instances, eventually responded in 446 (79.1%)
instances, and did not respond at all in 77 (13.6%)
instances. The attending physician took significantly
longer than the other groups ( P b .0002) with an
average of 30 minutes to respond (range, 0-54 minutes).
Adequate team leadership and an appropriate ACLS
protocol were implicitly considered by the CAT nurse and
the hospital quality assurance director to have been
followed in all cases. They did identify
6 instances (all with ROSC) when neither the hospitalist
nor the EM resident responded.
Residents performed only 19.3% (vs hospitalists =
27.4% and all other HCP = 26.7%) ( P = .16) of all
potential endotracheal intubations, delegating the
remainder to
respiratory therapists (Table 6). There were 2 difficult
intubations, both requiring the EM resident to take over
from a respiratory therapist and perform rapid sequence
intubation. The hospitalists ordered 2 patients to be made
DNR during the code (at t = 4 and 15 minutes). An
additional 4 hospitalist patients were of bdo-not-intubateQ
(DNI) status and subsequently expired after only a
bchemical code.Q The EM residents ordered 3 patients to
be made DNR during the code (at t = 3, 14, and 18
minutes).
An additional 2 DNR patients were inadvertently resuscitated until the nursing staff discovered the preexisting
order. Three other EM resident patients had DNI orders and
died after only a limited resuscitation.

4. Limitations
There were several limitations to our study. Because we
were principally interested in the CAT process, we
included isolated respiratory failure which skewed our
overall survival rates higher than expected [28]. However,
we also included the 14 DNR/DNI patients that received
resuscita- tive efforts which would tend to balance out
survival rates. Many code sheets lacked complete
information, and as in most similar CPR studies, data
were
recorded
after
the event and not
contemporaneously. Whereas physician arrival sequence
was usually documented, arrival times were often not
recorded properly in the bheat of battle.Q In retrospect, an
Utstein-based point-of-care CPR data entry program may
have produced more precise data than did our handwritten
forms (which were often completed after a hectic
code). With a single keystroke, a nurse recorder could
accurately input important code events in real time into a
Tablet PC or personal digital assistant [27].
Other recognized limitations include a considerable yearto-year spillover in this before-after design. However, the
bias was likely negligible because we analyzed by the
independent variable of physician specialty and not by year
of code blue. We measured ROSC and survival to hospital
discharge outcomes, but not the ideal 6- to 12-month
survival. This common shortfall in community hospital
based CPR research [6,29] should not detract from our key
focus on the timeliness of response, leadership, and
immediate outcome of these codes. We performed implicit,
not explicit, chart review when deciding on ACLS leadership and compliance, but because rigid ACLS compliance
is not necessarily associated with improved outcome, this
approach was a reasonable beginning [30,31]. Although
previous investigators showed a worse survival for night
shift cardiac arrests [2,32], our data did not. The preponderance of night shift code blues in the EM resident group
might have obscured a true survival benefit, but the subsequent multivariate logistic analysis did not confirm this.

5. Discussion
The residents reestablished ROSC significantly faster
than the hospitalists did. However, we found no survival
outcome difference between physician groups. We were
90% powered (2-tailed, a = .05) to detect a 15% difference
in survival. Therefore, in this hospital setting, the EM
resident model of CAT leadership was at least as
successful as the staff hospitalist model.
We found that EM residents arrived first twice as often
and achieved ROSC twice as fast as the staff hospitalists.
This, despite being handicapped by both more unmonitored and more night-shift arrests, both factors that others
have demonstrated to be independently associated with
mortality [1,2,32]. This raises the interesting possibility
that something intrinsic about the EM residents
training,

clinical style, or just better responsiveness may

actually be producing an outcome difference. EM residents
are highly trained, motivated, and oriented toward a team
resuscitation of the critically ill patient [33,34]. Therefore,
it is not surprising that they performed at least as well as
staff hospitalists. Although the residents, unlike the hospitalists, were required to have ACLS training, the clinical
impact of this is not determinable by the current study
[12,16,18,19,30].
On the other side, the hospitalists could boast equal
survival outcomes despite longer ROSC intervals. Because
their patients were more likely to be in an intensive care
unit or on mechanical ventilation, presumably, these
sicker patients would reasonably require a longer CPR
interval before ROSC. Perhaps postresuscitative management such as ventilatory settings, transport, critical care
medicine infusions, and subsequent procedures could have
been superior in the hospitalist cohort, thereby offsetting
their prolonged ROSC delay. Future studies comparing
CAT physician leadership might focus more on postresuscitation care.
We expected that EM residents would be more comfortable with and therefore personally secure the airway, but
they delegated to the respiratory therapist 4 of every 5
intubations. It is not clear why the EM residents passed on
this valuable training opportunity, but technical difficulty
did not seem to be a factor as previously reported [35]. We
could not confirm that one specialty would be more likely
than the other to pronounce patients to be on DNR status
[36]. The duration of the code blue for both survivors and
nonsurvivors was similar between physician groups. We
also did not observe the interspecialty (specialist vs
specialist) nor the intraspecialty (resident vs attending
physician) that has been demonstrated for clinical settings
other than CPR [21-24]. Finally, by our implicit nurse
supervisor review, all physicians followed ACLS
guidelines and effectively took charge of the CAT.
For this large community hospital setting, unless the
primary attending physician was already at bedside at t 0,
there was generally a significant time lapse until their
arrival. In two thirds of our code blues, the residents and
hospitalists both arrived within that recognized crucial
4-minute interval [37,38]. This emphasizes the importance
of having in-house physicians when feasible.
Do we even need a physician on the CAT since early
defibrillation and basic life support by first responders
(nurses in the hospital setting) account for so much of the
survival benefit [14,39]? Our data (Table 5) indicate that
for many codes, the CAT achieved ROSC even before
physician arrival. Certainly, physician response time
impacts survival if defibrillation is delayed (albeit
inappropriately) until the actual arrival of the doctor [38].
But with our CAT model, skilled nurses and respiratory
therapists often performed protocol-driven ACLS
interventions such as defibrillation and endotracheal
intubation before physician arrival. This reemphasizes that
the bwhat Q and bwhenQ those interven-

tions are done (especially early defibrillation) are probably

more important than bwho Q does them.
Beyond a need for speed, the qualified physician
potentially brings much added value to the CAT. An
experienced clinician best manages the complicated airway,
decides DNR and medical futility [2] issues, and stabilizes
complicated nonventricular fibrillation/ventricular tachycardia (VF/VT) patients [5,40]. Furthermore, physician
interpretation of bedside tests such as capnometry and
ultrasonography may improve the CPR decision-making
process [41,42]. Finally, doctors can deviate from ACLS
algorithms when the clinical situation warrants [30]. For
example, if a patient with a suspected pulmonary embolism
suddenly arrests, the physician might decide to administer
life-saving thrombolytics [43]. The physician role on the
CAT will continue to evolve.
The importance of early therapy, especially
defibrillation, by nursing personnel for cardiac arrest
cannot be over- emphasized. But physicians certainly
serve a vital role during inhospital CPR. Whenever
feasible, community hospitals should appoint physician
leadership to the CAT. The EM resident model appears to
be an acceptable alternative to the hospitalist model for a
24-hour physician staffing of the CAT for in-hospital
cardiac arrest.

References
[1] Tunstall-Pedoe H, Bailey L, Chamberlain DA, Marsden AK, Ward
ME, Zideman DA. Survey of 3765 cardiopulmonary resuscitations
in British hospitals (the BRESUS Study): methods and overall
results. BMJ 1992;304:1347 - 51.
[2] Dumot JA, Burval DJ, Sprung J, et al. Outcome of adult
cardiopulmonary resuscitations at a tertiary referral center including results of
blimitedQ resuscitations. Arch Intern Med 2001;161:1751 - 8.
[3] McGrath RB. In-house cardiopulmonary resuscitation after a
quarter of a century. Ann Emerg Med 1987;16:1365 - 8.
[4] Parr M. In-hospital resuscitation: review and revise. Resuscitation
2001;50:13 - 4.
[5] Pepe PE, Levine RL, Fromm Jr RE, Curka PA, Clark PS, Zachariah
BS. Cardiac arrest presenting with rhythms other than ventricular
fibrillation: contribution of resuscitative efforts toward total survivorship. Crit Care Med 1993;21:1838 - 43.
[6] Ravakhah K, Khalafi K, Bathory T, Wang HC. Advanced cardiac life
support events in a community hospital and their outcome:
evaluation of actual arrests. Resuscitation 1998;36:95 - 9.
[7] Bedell SE, Delbanco TL, Cook EF, Epstein FH. Survival after
cardiopulmonary resuscitation in the hospital. N Engl J Med 1983;
309:569 - 76.
[8] Camp BN, Parish DC, Andrews RH. Effect of advanced cardiac life
support training on resuscitation efforts and survival in a rural
hospital. Ann Emerg Med 1997;29:529 - 33.
[9] Schade J. An evaluation framework for code 99. Qual Rev Bull 1983;
9:306 - 9.
[10] Lim YH, Anantharaman V. Hospital resuscitation of cardiac arrest
patients. Singapore Med J 1999;40:165 - 70.
[11] Whitcomb J, Brierton D, Kuhagen K, et al. Code teams and the
review of cardiac arrests. Qual Rev Bull 1990;16:156 - 62.
[12] Fields ML. The C.P.R. team in a medium-sized hospital. Am J Nurs
1966;66:87 - 90.

[13] Kapadia FN. Code 99 an international perspective. Chest 1999;

115:1483.
[14] Soar J, McKay U. A revised role for the hospital cardiac arrest team?
Resuscitation 1998;38:145 - 9.
[15] Diamond HS, Goldberg E, Janosky JE. The effect of full-time faculty
hospitalists on the efficiency of care at a community teaching
hospital. Ann Intern Med 1998;129:197 - 203.
[16] Lowenstein SR, Sabyan EM, Lassen CF, Kern DC. Benefits of
training physicians in advanced cardiac life support. Chest 1986;89:
512 - 6.
[17] Kaye W, Mancini ME. Retention of cardiopulmonary resuscitation
skills by physicians, registered nurses, and the general public. Crit
Care Med 1986;14:620 - 2.
[18] Skinner DV, Camm AJ, Miles S. Cardiopulmonary resuscitation skills
of preregistration house officers. BMJ 1985;290:1549 - 50.
[19] Kurrek MM, Devitt JH, Cohen M. Cardiac arrest in the OR: how are
our ACLS skills? Can J Anaesth 1998;45:130 - 2.
[20] Fielden JM, Bradbury NS. Observational study of defibrillation in
theatre. BMJ 1999;318:232 - 3.
[21] Bell CM, Redelmeier DA. Mortality among patients admitted to
hospitals on weekends as compared with weekdays. N Engl J Med
2001;345:663 - 8.
[22] Auerbach AD, Hamel MB, Davis RB, et al. Resource use and
survival of patients hospitalized with congestive heart failure:
differences in care by specialty of the attending physician.
SUPPORT Investigators. Study to Understand Prognoses and
Preferences for Outcomes and Risks of Treatments. Ann Intern Med
2000;132:191 - 200.
[23] Porter JM, Ursic C. Trauma attending in the resuscitation room: does
it affect outcome? Am Surg 2001;67:611 - 4.
[24] Chen J, Radford MJ, Wang Y, Krumholz HM. Care and outcomes
of elderly patients with acute myocardial infarction by physician
specialty: the effects of comorbidity and functional limitations. Am J
Med 2000;
108:460 - 9.
[25] Wachter RM, Goldman L. The emerging role of b hospitalists Q in the
American health care system. N Engl J Med 1996;335:514 - 7.
[26] Cummins RO, Chamberlain D, Hazinski MF, et al. Recommended
guidelines for reviewing, reporting, and conducting research on inhospital resuscitation: the in-hospital dUtstein style.T American Heart
Association. Circulation 1997;95:2213 - 39.
[27] Adams B, Whitlock W. Utstein Style spreadsheet and database
programs based on Microsoft Excel and Microsoft Access software for
CPR data management of in-hospital resuscitation. Resuscitation
2004;61:37 - 40.
[28] Peberdy M, Kaye W, Ornato J, Larkin G, Nadkarni V.
Cardiopulmo- nary resuscitation of adults in the hospital: a report of
14720 cardiac arrests from the National Registry of
Cardiopulmonary Resuscitation. Resuscitation 2003;58:297 - 308.
[29] Stratton SJ, Niemann JT. Reconsideration of proximate Utstein-style
end points. Crit Care Med 2002;30:S137 - 9.
[30] Callaham M. Epinephrine doses in cardiac arrest: is it time to
outgrow the orthodoxy of ACLS? Ann Emerg Med 1989;18:1011 2.
[31] Cline DM, Welch KJ, Cline LS, Brown CK. Physician compliance
with advanced cardiac life support guidelines. Ann Emerg Med
1995;25:
52 - 7.
[32] Muller JE. Circadian variation in cardiovascular events. Am J Hypertens 1999;12:35S - 42S.
[33] Hockberger RS, Binder LS, Graber MA, et al. The model of the
clinical practice of emergency medicine. Ann Emerg Med 2001;37:
745 - 70.
[34] Risser D, Simon R, Rice M, Salisbury M. A structured teamwork
system to reduce clinical errors. In: Spath P, editor. Error reduction
in health care: a systems approach to improving patient safety.
San Francisco7 Jossey-Bass; 2000. p. 235 - 78.
[35] Adams K, Scott R, Perkin RM, Langga L. Comparison of intubation
skills between interfacility transport team members. Pediatr Emerg
Care 2000;16:5 - 8.

[36] Kelly WF, Eliasson AH, Stocker DJ, Hnatiuk OW. Do specialists differ on do-not-resuscitate decisions? Chest 2002;121:
957 - 63.
[37] Bottiger BW, Grabner C, Bauer H, et al. Long term outcome after
out- of-hospital cardiac arrest with physician staffed emergency
medical services: the Utstein style applied to a midsized
urban/suburban area. Heart (British Cardiac Society) 1999;82:674 9.
[38] Colquhoun MC. Defibrillation by general practitioners. Resuscitation
2002;52:143 - 8.
[39] Spearpoint KG, McLean CP, Zideman DA. Early defibrillation and
the chain of survival in din-hospitalT adult cardiac arrest; minutes
count. Resuscitation 2000;44:165 - 9.

[40] Kaye W, Mancini ME. Use of the MegaCode to evaluate team leader
performance during advanced cardiac life support. Crit Care Med
1986;
14:99 - 104.
[41] Marik PE, Bankov A. Sublingual capnometry versus traditional
markers of tissue oxygenation in critically ill patients. Crit Care Med
2003;31:818 - 22.
[42] MacCarthy PWA, McCarthy G, Davies J. The use of transthoracic
echocardiography to guide thrombolytic therapy during cardiac arrest
due to massive pulmonary embolism. Emerg Med J 2002;178 - 9.
[43] Bottiger BW, Martin E. Thrombolytic therapy during cardiopulmonary resuscitation and the role of coagulation activation after cardiac
arrest. Curr Opin Crit Care 2001;7:176 - 83.