Health Services and Outcomes Research

Dispatcher-Assisted Cardiopulmonary Resuscitation

Time to Identify Cardiac Arrest and Deliver Chest
Compression Instructions
Miranda Lewis, BS; Benjamin A. Stubbs, MPH; Mickey S. Eisenberg, MD, PhD

Background—Dispatcher-assisted cardiopulmonary resuscitation (DA-CPR), in which 9-1-1 dispatchers provide CPR

instructions over the telephone, has been shown to nearly double the rate of bystander CPR. We sought to identify
factors that hampered the identification of cardiac arrest by 9-1-1 dispatchers and prevented or delayed the provision of
dispatcher-assisted CPR chest compressions.
Methods and Results—We reviewed dispatch recordings for 476 out-of-hospital cardiac arrests occurring between January
1, 2011, and December 31, 2011. We found that the dispatcher correctly identified cardiac arrest in 80% of reviewed cases
and 92% of cases in which they were able to assess patient consciousness and breathing. The median time to recognition
of the arrest was 75 seconds. Chest compressions following dispatcher-assisted CPR instructions occurred in 62% of
cases when the dispatcher had the opportunity to asses for consciousness and breathing and bystander CPR was not
already started. The median time to first dispatcher-assisted CPR chest compression was 176 seconds.
Conclusions—Dispatchers are able to accurately diagnose cardiac arrest over the telephone, but recognition is likely not
possible in all circumstances. In some cases, recognition of cardiac arrest may be improved through training in the
detection of agonal respirations. Delays in the delivery of dispatcher-assisted CPR chest compressions are common
and are attributable to a mixture of dispatcher behavior and factors beyond the control of the dispatcher. Performance
standards for the successful and quick recognition of cardiac arrest and delivery of first chest compressions
should be adopted as metrics against which emergency medical services systems can measure their performance.  
(Circulation. 2013;128:1522-1530.)
Key Words: cardiopulmonary resuscitation ◼ death, sudden ◼ heart arrest
Downloaded from http://ahajournals.org by on August 22, 2022

E arly provision of bystander cardiopulmonary resuscita-

tion (CPR) is an important factor in survival from out-
of-­hospital cardiac arrest.1 Although the benefit of bystander
CPR is well established, rates remain relatively low in most
communities, with roughly a third of patients receiving
bystander CPR.2,3 Dispatcher-assisted CPR (DA-CPR), in
which dispatchers provide timely CPR instructions to 9-1-1
callers over the telephone, has been shown to nearly double
the rate of bystander CPR4 yet many communities have not
implemented DA-CPR programs.
Editorial see p 1490
Clinical Perspective on p 1530
Provision of DA-CPR compressions as early as possible
after the start of the call is ideal.5 However, there are currently
no national performance standards for the time to recogni-
tion of arrest and the time to first DA-CPR chest compres-
sion. There is significant variation in the time to recognition
of cardiac arrest by dispatchers, with reports ranging from
50 seconds to 2 minutes 38 seconds.6–9 A recent evaluation
reports a mean time to first compression of 4 minutes after
call initiation.10
Understanding the factors that lead to nonrecognition
of cardiac arrest and the reasons for delay in the delivery
of chest compressions represents an important step toward
improving DA-CPR rates. The purpose of this study was to
identify and characterize the specific factors that led to non-
recognition of cardiac arrest by dispatchers. In cases when
cardiac arrest was recognized, we sought to identify the rea-
sons for and to quantify delays in the time from 9-1-1 call to
delivery of the first DA-CPR chest compression. The study
procedures were approved by the University of Washington
Human Subjects Division.
Description of the Study Community and
Dispatch Protocol
This study was undertaken in King County, Washington
(excluding Seattle). King County implemented a telephone

Continuing medical education (CME) credit is available for this article. Go to http://cme.ahajournals.org to take the quiz.
Received March 15, 2013; accepted July 30, 2013.
From the School of Medicine (M.L.) and Department of Medicine (M.S.E.), University of Washington, Seattle; and Public Health Seattle & King County,
Emergency Medical Services Division, Seattle, WA (B.A.S., M.S.E.).
Correspondence to Mickey S. Eisenberg, MD, PhD, Public Health Seattle & King County, Emergency Medical Services Division, 401 5th Ave, Ste 1200,
Seattle, WA 98104. E-mail gingy@u.washington.edu
© 2013 American Heart Association, Inc.
Circulation is available at http://circ.ahajournals.org DOI: 10.1161/CIRCULATIONAHA.113.002627

1522
 Lewis et al   Timeliness of Dispatcher-Assisted CPR   1523
CPR program in 1982.11 Dispatchers undergo 32 hours of
emergency medical training, with 6 hours dedicated to the
recognition of cardiac arrest and delivery of CPR instructions
over the telephone. Dispatchers receive 8 hours of continuing
education each year and undergo regular performance evalu-
ations to ensure that quality assurance standards are met. For
the purposes of this study, we use the term dispatcher to refer
to the person responsible for questioning and giving instruc-
tions to the 9-1-1 caller, although this term may not be used
in all emergency response systems. In the 2 communications
centers participating in this study, the dispatcher handles only
1 call at a time and stays on the line until the call is terminated.
Dispatchers are trained to approach every call with a high
index of suspicion for cardiac arrest and to assertively provide
CPR instructions in cases of suspected arrest. After the inci-
dent address is confirmed, protocol dictates that dispatchers
ask every caller 2 questions: “Is the patient conscious?” and
“Is the patient breathing normally?” If the caller answers no
to both questions, then the patient is presumed to be in cardiac
arrest and CPR instructions are provided. The protocol allows
dispatchers to use information spontaneously provided by
the caller, so it is not always necessary to ask both questions.
When CPR instructions are given, the rescuer is encouraged
to count out loud while giving chest compressions so that the
compression rate can be monitored and corrected.
Methods
This study is a retrospective cohort study of out-of-hospital cardiac
arrests occurring between January 1, 2011, and December 31, 2011.
Downloaded from http://ahajournals.org by on August 22, 2022
Cases of confirmed cardiac arrest in adults >17 years of age occurring
before the arrival of emergency medical services (EMS) personnel
were included. Arrests resulting from trauma, arrests that occurred in
a medical clinic or nursing home, and incidents that were handled by
a nonparticipating dispatch center were excluded.
For all cases meeting the case definition, we attempted to obtain
a copy of the communication center report that records the typed
comments and dispatch times for each event and the recording of the
9-1-1 call. Before reviewing these materials, we determined that we
would consider the dispatcher to have recognized the arrest if CPR
instructions were given or if the dispatcher made a comment during
the recording that indicated an arrest was suspected (eg, “We need to
do CPR,” or “Those are agonal respirations”). If review of the record-
ing did not allow us to determine whether the dispatcher recognized
the arrest, we reviewed the written report for references indicating
a cardiac arrest (“CPR needed,” dispatch codes specific to cardiac
arrest or text notations indicating that cardiac arrest dispatch proto-
cols were being implemented) and upgrading the call to a paramedic
response. If both criteria were met, we concluded that the dispatcher
recognized that the patient was in cardiac arrest.
The time to key events in the call sequence was determined, including
confirmation of incident address, query about the patient’s level of con-
sciousness or provision of this information by the caller, query about the
patient’s breathing status or provision of this information by the caller,
determination of the need for CPR, delivery of CPR instructions, and
delivery of the first chest compression following DA-CPR instruction.
For the purposes of this study, determination of the need for CPR was
said to have occurred when the dispatcher verbalized the need to perform
CPR or when CPR instructions began. Delivery of the first chest com-
pression following DA-CPR instruction was said to have occurred when
the caller or other rescuer began to count compressions out loud or when
the caller first verbally confirmed that compressions were being done.
We also reviewed recordings for the clarity of questions asked
by the dispatcher, the quality of information provided by the caller,
and the reasons for delays. For example, when assessing patient’s
breathing, the dispatcher may use questions such as “Is the patient
breathing normally?” or “Is the chest rising and falling?” The caller
may respond with a simple “yes” or “no” or may give an answer
necessitating further evaluation such as “kind of” or “I don’t think
so.” We documented the questions asked by the dispatcher to deter-
mine whether the patient was conscious and breathing, as well as the
caller’s responses to the dispatcher’s queries.
Reasons for delays in the recognition of cardiac arrest were also
documented. A delay was defined as a deviation from protocol that
resulted in a time lapse of ≥5 seconds. Each delay was recorded indi-
vidually, and the reason for the delay was identified. Impediments in
recognition of cardiac arrest by dispatchers can largely be attributed
to one of the following: dispatcher-related factors, caller-related fac-
tors, and call circumstance–related factors. Dispatcher-related factors
included inquiries about medical history when enough information
had been provided by the caller to determine that the patient was in
cardiac arrest, inquiries about the incident such as queries concerning
what the patient was doing before the event, redundant assessment of
breathing, redundant assessment of consciousness, and determination
of patient age and sex. Caller-related factors included emotional state,
interactions between the caller and other rescuers, indirect answers to
dispatcher questions or giving unnecessary information, leaving the
phone, and hesitancy to move the patient. Call circumstance–related
factors, which are largely nonmodifiable factors that could be attrib-
uted to neither the caller nor the dispatcher, included language barri-
ers, difficulty confirming the address, time spent moving the patient,
delays in gathering information when the caller was not in direct
contact with the patient, time spent unlocking the door, patient in a
position that did not allow chest compressions, and patient vomiting.
We used a similar approach to classify delays in the provision of
chest compressions when CPR instructions were provided by the dis-
patcher. Reasons for a delay were categorized into caller delays, dis-
patcher delays, and circumstance-related delays as described above.
Data were analyzed with SPSS version 20.0. Patient characteristics
and outcomes were described using means for continuous variables
and counts for categorical variables. Differences across groups were
assessed with the Student t test or the Pearson χ2 test as appropriate.
A multivariate logistic regression analysis was conducted to identify
patient and incident characteristics associated with the recognition
of cardiac arrest and the provision of DA-CPR chest compressions.
Dispatcher-caller interactions were explored descriptively, using
counts and percentages to search for trends. Nonparametric times to
key events were described using quartiles, with differences in dis-
tribution across subgroups explored with the Mann-Whitney U test.
Results
There were 901 out-of-hospital cardiac arrests during the
study period, 590 (65%) of which met the inclusion criteria
for this study (Figure 1). Of the cases meeting the inclusion
criteria, recordings were available and reviewed for 476 cases
(81%). There were 114 recordings that were unavailable for
review because they were not transmitted from the dispatch
center to be archived at the King County EMS Department or
the file that was archived was corrupted.
Table 1 shows patient characteristics for cases for which the
recording was reviewed compared with cases for which the
recording was not reviewed. Subjects in the group for which
recordings were not reviewed were more likely to be admitted
to the hospital (P=0.04) and to survive to hospital discharge
(P=0.03). There were no other statistically significant differ-
ences between the 2 groups.
Figure 1 shows that dispatchers recognized cardiac arrest in
80% of the cases reviewed (381 of 476). The dispatcher was not
able to asses consciousness and breathing in 13% of the cases
reviewed (n=60). Reasons included the following: the caller
was not at the scene with the patient (eg, a call from a medi-
cal alarm company, n=27); phone contact was lost and could
 1524  Circulation  October 1, 2013
not be reestablished (n=7); the patient was not in arrest during
Downloaded from http://ahajournals.org by on August 22, 2022
the call but apparently arrested before the arrival of EMS per-
sonnel (n=18); or caller factors such as emotional state (n=2),
inability to move the patient to assess breathing and conscious-
ness (n=4), or refusal to evaluate the patient (n=2). Therefore,
we determined that when the dispatcher had the opportunity to
assess consciousness and breathing, the dispatcher did not rec-
ognize a cardiac arrest in 8% of the cases reviewed (35/416).
We assessed patient and incident characteristics that may be
associated with recognition of the cardiac arrest or with provi-
sion of chest compressions following dispatcher instructions.
The results are shown in Table 2. In a univariate analysis, we
found that dispatchers were less likely to correctly diagnose
a cardiac arrest when the event was witnessed (P=0.009) and
that bystander CPR was less likely to be provided when the
dispatcher did not recognize the arrest (P<0.001). In a logistic
regression analysis with recognition of the arrest (yes/no) as
the outcome and with adjustment for age, sex, witnessed sta-
tus, time to EMS arrival, and arrest location, with all covari-
ates included in the final model, dispatchers were less likely
to correctly diagnose a witnessed arrest (odds ratio, 0.38; 95%
confidence interval, 0.17 – 0.82; P=0.014). The independent
variables included in the model were chosen a priori, and no
covariates other than witnessed status were significantly asso-
ciated with recognition of the arrest.
Table 2 also shows that older patients may be less likely to
receive DA-CPR chest compressions (P=0.06). In a logistic
regression analysis with bystander CPR as the outcome (CPR
with dispatcher instructions compared with no bystander CPR)
and with adjustment for age, sex, witnessed status, time to EMS
arrival, and arrest location, with all covariates included in the
Figure 1. Study enrollment and key end
points. Call circumstances included
patient position, patient vomiting, time
spent unlocking the door, time spent
moving the patient, and delays resulting
from the caller’s proximity to the patient.
Caller factors included the emotional
state of the caller, interactions between
the caller and other rescuers, indirect
answers to dispatcher questions, leaving
the phone, giving unnecessary information,
and hesitancy to move the patient.
DA-CPR indicates dispatcher-assisted
cardiopulmonary resuscitation; and EMS,
emergency medical services.
final model, the odds of receiving DA-CPR chest compressions
decreased ≈1% with each additional year of patient age (odds
ratio, 0.985; 95% confidence interval, 0.969–1.000; P=0.054).
The independent variables included in the model were chosen
a priori, and no covariates were significantly associated with
the provision of DA-CPR at the P=0.05 level.
Table 3 shows the caller response to questions about con-
sciousness and breathing normally for cases in which the car-
diac arrest was recognized compared with cases in which the
arrest was not recognized. For cases in which the arrest was
not recognized, the caller was more likely to provide uncer-
tain or contradictory information. For example, the patient
was reported to be conscious or the caller gave contradictory
information in 54.3% of the cases when cardiac arrest was not
recognized compared with 22.0% of cases when the arrest was
recognized. Similarly, patients were reported to be breathing
or contradictory information was given in 74.3% of nonrec-
ognized cases compared with 35.7% of recognized cases. The
accurate assessment of breathing may have been complicated
by the higher rate of agonal respirations in the nonrecognition
group (40.0% compared with 29.1% in the recognized group).
Table 3 also shows that, when call circumstances allowed, dis-
patchers followed the protocol and assessed for consciousness
and breathing in 99% of cases (412 of 416).
Figure 1 shows that dispatcher-assisted bystander chest
compressions were provided in 62% of cases (210 of 339)
when the dispatcher had the opportunity to asses for con-
sciousness and breathing and excluding cases in which
bystander CPR was already in progress (77 of 416, 18.5%).
Table 4 shows the reasons that bystander CPR was not
provided for the remaining 94 cases. In 40 cases (42.6%),
 Lewis et al   Timeliness of Dispatcher-Assisted CPR   1525

Table 1.  Characteristics of Eligible Cases*

Eligible for Study, Recording Eligible for Study, Recording
Characteristics Reviewed (n=476) Not Reviewed (n=114) P Value†
Age, mean (SD), y 63.7 (16.7) 63.8 (17.1) 0.94
Female, n (%) 167 (35.1) 49 (43.0) 0.12
Witnessed arrest, n (%) 234 (49.2) 55 (48.2) 0.86
Bystander CPR, n (%) 295 (62.0) NA‡ NA
 With dispatcher assistance, 210 (71.2) NA‡
n (% receiving bystander CPR)
 Without dispatcher assistance, 85 (28.8) NA‡
n (% receiving bystander CPR)
Minutes to EMS arrival, mean (SD), n 6.1 (3.0) [463] 6.3 (3.8) [107] 0.45
Initial rhythm shockable or VF/VT, n (%) 139 (29.2) 35 (30.7) 0.75
Location of arrest, n (%) 0.08
 Home 400 (84.0) 86 (75.4)
 Other residence 8 (1.7) 4 (3.5)
 Public indoors 25 (5.3) 8 (7.0)
 Public outdoors 43 (9.0) 16 (14.0)
Admitted to hospital, n (%) 205 (43.1) 62 (54.4) 0.03
Discharged alive from hospital, n (%) 102 (21.4) 36 (31.6) 0.02
CPC score 1–2, n (% of discharged alive 95 (93.1) 33 (91.7) 0.77
from hospital)
CPC indicates cerebral performance category; CPR, cardiopulmonary resuscitation; EMS, emergency medical services;
and VF/VT, ventricular fibrillation/ventricular tachycardia.
*Eligible cases were out-of-hospital cardiac arrests occurring between January 1, 2011, and December 31, 2011, in
which the 9-1-1 call was received by a participating agency. Patients <18 years of age, who arrested after the arrival of EMS
personnel, or who arrested in a medical facility or nursing home were excluded.
†Pearson χ2 for categorical variables or independent Student t test for comparison of means.
Downloaded from http://ahajournals.org by on August 22, 2022

‡Provision of bystander CPR was determined by review of the 911 recording without consulting the EMS incident report.

physical limitations prevented the caller from performing Potentially modifiable caller factors accounted for 22 of the
CPR or moving the patient. In 25 cases (26.6%), call circum- cases (23.4%) in which bystander CPR was not performed.
stances prevented the performance or instruction of DA-CPR. Finally, there were 7 cases (7.4%) in which there was a

Table 2.  Characteristics of Cases in Which Arrest Was Recognized Compared With Not Recognized and DA-CPR Chest
Compressions Were Given Compared With Not Given
Cardiac Arrest Cardiac Arrest DA-CPR Chest Bystander Chest
Recognized Not Recognized* P Compressions Compressions Not P
Characteristics (n=381) (n=35) Value Provided (n=210) Provided† (n=94) Value
Age, mean (SD), y 63.8 (16.4) 64.5 (19.1) 0.80 63.0 (17.3) 67.0 (15.9) 0.06
Female, n (%) 133 (34.9) 8 (23.5) 0.15 77 (36.7) 28 (29.8) 0.24
Witnessed arrest, n (%) 174 (45.7) 24 (68.6) 0.009 95 (45,2) 36 (38.3) 0.26
Bystander CPR, n (%) 287 (75.3) 3 (8.6) <0.001 210 (100.0) 0 (0.0) NA
  With dispatcher assistance, n (% receiving bystander CPR) 210 (73.2) 0 NA 210 (100.0) 0 (0.0)
 Without dispatcher assistance, n (% receiving bystander CPR) 77 (26.8) 3 (100.0) 0 (0.0) 0 (0.0)
Time to EMS arrival, mean (SD) [n] 5.9 (2.9) [370] 6.2 (3.1) [34] 0.69 5.7 (2.0) [203] 5.9 (2.7) [92] 0.6
Initial rhythm shockable or VF/VT, n (%) 113 (29.7) 12 (34.3) 0.57 62 (29.5) 22 (23.4) 0.27
Location of arrest, n (%) 0.22
 Home or other residence 333 (87.4) 28 (80.0) 190 (90.5) 89 (94.7) 0.22
 Public 48 (12.6) 7 (20.0) 20 (9.5) 5 (5.3)
Admitted to hospital, n (%) 160 (42.0) 18 (51.4) 0.28 78 (37.1) 35 (37.2) 0.99
Discharged alive from hospital, n (%) 82 (21.5) 9 (25.7) 0.57 38 (18.1) 16 (17.0) 0.82
CPC score 1–2, n (% of discharged alive from hospital) 77 (93.9) 8 (88.9) 0.57 35 (92.1) 15 (93.8) 0.83
CPC indicates cerebral performance category; DA-CPR, dispatcher-assisted cardiopulmonary resuscitation; EMS, emergency medical services; NA, not applicable;
and VF/VT, ventricular fibrillation/ventricular tachycardia.
*Excludes 60 cases in which the dispatcher did not have the opportunity to assess consciousness and breathing.
†Excludes 77 cases in which bystander CPR was provided without dispatcher instructions.
 1526  Circulation  October 1, 2013
Table 3.  Caller Responses for Cases in Which the Dispatcher Had the Opportunity to Assess Consciousness and Breathing*
Caller response to “Is the patient conscious?”
 “No” or patient reported to be unconscious before the question was asked
 “Yes” or uncertain/contradictory answer
 Question not asked and information not volunteered by caller
Caller response to “Is the patient breathing normally?”
 “No” or reported that patient was not breathing before the question was asked
 “Yes” or uncertain/contradictory answer
 Question not asked and information not volunteered by caller
Agonal respirations
 Heard by reviewer or described by the caller
*Excludes cases in which the patient was not in cardiac arrest during the call, the caller hung up before an assessment could be completed, and the caller refused
or was unable to assess the patient.
technical error in the recording, causing it to cut off before the
reason that bystander CPR was not provided could be verified.
Table 5 shows the time from the start of the recording to
key events. For all calls reviewed, the median time to recogni-
tion of cardiac arrest was 75 seconds (n=381). The median
time to the first DA-CPR chest compression was 176 seconds
(n=210). Table 5 also shows that it took longer for the dis-
patcher to gather information about consciousness and breath-
ing in cases when the arrest was not recognized (median, 69
seconds; n=31) compared with cases in which the arrest was
recognized (median, 43 seconds; n=360; P<0.001 for differ-
ence, Mann-Whitney U test).
Downloaded from http://ahajournals.org by on August 22, 2022
Table 6 shows the factors involved in a delay of ≥5 seconds
from the time of recognition of a cardiac arrest to the provision
of the first DA-CPR chest compression. Delays in the deliv-
ery of the first DA-CPR chest compression ≥5 seconds were
present in 92.9% of cases (n=354) in which cardiac arrest was
recognized. The longest delays were found in cases in which
there was a language barrier (median time, 59 seconds; n=13)
and in cases in which the caller left the phone (median time,
43 seconds; n=27). The most common delays were associated
Table 4.  Reasons Bystander CPR Was Not Provided in Cases
in Which the Dispatcher Recognized the Arrest (n=94)
n %
Physical limitations
 Rescuer unable to position the patient for CPR 40 42.6
Call circumstances
 EMS arrived before DA-CPR could be performed 15 16.0
 CPR was deemed futile according to the caller’s description 6 6.4
 Caller reported that the patient would not have wanted CPR 2 2.1
 Language barrier with caller prevented provision of CPR 2 2.1
instructions
Caller factors
 Caller left the phone and never returned 14 14.9
 Caller refused CPR instructions 4 4.3
 Caller’s emotional state prevented provision of CPR instructions 4 4.3
Technical error with recording
 Reason could not be assessed 7 7.4
DA-CPR indicates dispatcher-assisted cardiopulmonary resuscitation; and
EMS, emergency medical services.
Cardiac Arrest Recognized Cardiac Arrest Not Total
(n=381), n (%) Recognized (n=35) , n (%) (n=416) , n (%)
297 (78.0) 14 (40.0) 311 (74.8)
84 (22.0) 19 (54.3) 103 (24.8)
0 2 (5.7) 2 (0.5)
245 (4.3) 5 (14.3) 250 (60.1)
136 (35.7) 26 (74.3) 162 (38.9)
0 4 (11.4) 4 (1.0)
111 (29.1) 14 (40.0) 125 (30.0)
with moving the patient (n=168, 44.1% of cases in which the
arrest was recognized; median time, 32 seconds) and unneces-
sary questions about the age, sex, or chief complaint of the
patient (n=126, 33.1% of cases in which the arrest was recog-
nized; median time, 10 seconds).
Discussion
Nonrecognition of cardiac arrest by the layperson presents
a significant barrier to the delivery of bystander CPR. With
every minute that a patient in cardiac arrest goes without CPR,
the chances of survival decrease.12,13 DA-CPR programs have
been shown to increase the rate of bystander CPR.4,11 This
study was undertaken to better understand the factors asso-
ciated with the recognition of cardiac arrest by emergency
dispatchers and to describe the delays associated with recog-
nition of cardiac arrest and the provision of CPR instructions
over the telephone. We did not attempt to show a survival ben-
efit for dispatcher-assisted CPR, which has previously been
established compared with no bystander CPR.4
We found that dispatchers were able to recognize cardiac
arrest in 80.0% of the cases reviewed. Although we observed
a high rate of recognition, other published findings indicate
a great deal of variation in the detection of cardiac arrest by
dispatchers over the telephone, with reports ranging from 47%
to 97%.8,9,14–17 A systematic review of the literature suggests
that in most EMS systems, dispatchers are able to recognize
cardiac arrest with a sensitivity of ≈70%.18
Although 100% recognition of cardiac arrest would be ideal,
this is likely not possible given the nature of cardiac arrest
calls. As we observed in this study, call circumstances that
are not amenable to change negatively affect the ability of the
dispatcher to detect cardiac arrest. These circumstances such
as reports from callers who are not with the patient (eg, alarm
companies), loss of phone contact with the caller, and callers
who refuse or cannot assess the patient were found to contrib-
ute to nearly half of the cases of nonrecognition in this study.
The remaining cases in which cardiac arrest was not recognized
by dispatchers because of potentially modifiable factors such as
failure to recognize agonal respirations, conflicting information
provided by the caller, or early termination of the call by the
dispatcher represent an opportunity to improve cardiac arrest
recognition through training and quality improvement.
 Lewis et al   Timeliness of Dispatcher-Assisted CPR   1527

Table 5.  Time From the Start of the Call to Key Events for Cases That Had Recordings Available for Review
Median Time (25%–75%), s
Cardiac Arrest Not Recognized, Cardiac Arrest Not Recognized,
Dispatcher Had Opportunity to Dispatcher Did Not Have Opportunity
Cardiac Arrest Assess Consciousness to Assess Consciousness and All Calls Reviewed
Recognized (n=381) and Breathing (n=35) Breathing (n=60) (n=476)
Confirm address 23 (17–32, n=380) 21 (17–39, n=35) 25 (19–33, n=58) 23 (17–32, n=473)
Ask about consciousness or information 31 (21–46, n=340) 40 (30–61, n=33) 29 (21–41, n=32) 32 (22–47, n=405)
provided by caller
Ask about breathing or information 43 (29–66, n=360) 69 (44–116, n=31) 45 (34–77, n=30) 44 (31–70, n=421)
provided by caller
Recognition of cardiac arrest 75 (47–121, n=381) NA NA 75 (47–121, n=381)
Instruction to position patient/bare the 85 (58–141, n=319) NA NA 85 (58–141, n=319)
chest
Start of chest compression instructions 151 (114–218, n=237) NA NA 151 (114–218, n=237)
First DA-CPR chest compression 176 (141–242, n=210) NA NA 176 (141–242, n=210)
EMS arrival or end of recording 338 (261–422, n=381) 247 (163–343, n=35) 184 (115–335 n=60) 325 (241–413, n=476)
DA-CPR indicates dispatcher-assisted cardiopulmonary resuscitation; and EMS, emergency medical services.

The assessment of breathing presented the most significant
challenge to the recognition of cardiac arrest by dispatchers
in this study, possibly because of the presence of agonal res-
pirations. Callers may lack the experience necessary to assess
Table 6.  Delays in Provision of the First DA-CPR Chest
Compression in Cases in Which Cardiac Arrest Was
Recognized (n=381)
Median Time
Downloaded from http://ahajournals.org by on August 22, 2022
for and accurately describe breathing, especially in the high-
stress context of a cardiac arrest call. It is therefore impera-
tive that dispatchers not only assess for whether the patient is
breathing but also evaluate the quality of breathing.
We found that callers more often provided erroneous or
ambiguous information about breathing in cases when cardiac
arrest was unrecognized (Table 3). This may partially explain
the finding of our logistic regression analysis that dispatchers
were less likely to recognize a witnessed arrest (odds ratio,

(25%-75%), s Calls, n 0.38; 95% confidence interval, 0.17–0.82). When the caller
witnesses the collapse, he or she may be more likely to mistake
Call circumstance delays
agonal breathing for normal breathing. Although the quality of
 Language barrier 59 (11–148) 13 information provided by the caller is largely out of the control
 Caller moving the patient or baring the chest 32 (18–58) 168 of the dispatcher, dispatchers can be trained to approach each
 Caller not near patient 24 (13–43) 23 call with a high index of suspicion for cardiac arrest.
 Difficulty determining the address or location 18 (10–32) 76 Given the challenges inherent in recognition of cardiac
of the patient arrest over the telephone, it is critical that emergency dis-
 Other 24 (12–51) 52 patch systems monitor and evaluate all incidents involving
 Total call circumstance delays 38 (20–71) 252 a cardiac arrest. Ideally, this would entail the identification
Caller delays of all incidents that are treated as cardiac arrests by EMS
 Caller left phone 43 (20–110) 27 personnel followed by the evaluation of dispatch record-
 Hesitant to move the patient 19 (10–37) 109 ings and written documentation for all eligible events. This
 Caller provides unnecessary information or is 13 (7–22) 107 should be carried out in partnership with local EMS agen-
uncooperative cies, and key incident details should be noted for each case,
 Emotional state 13 (7–25) 100 as previously described.5,19
 Other 13 (8–20) 52 There are currently no national performance standards for
 Total caller delays 26 (14–55) 244 the recognition of cardiac arrest and delivery of CPR instruc-
Dispatcher delays tions by emergency dispatchers. It has been the experience
 Asks medical history questions 21 (11–36) 26
of the EMS system participating in this study that quality
improvement is best achieved through identification of key
 Asks unnecessary incident questions 17 (10–31) 117
performance standards, measurement, and ongoing training
 Asks unnecessary breathing or consciousness 14 (8–34) 112
aimed at improving performance. We propose that there are
questions
4 key performance standards that dispatch systems should
 Not engaging caller 13 (8–23) 22
strive to meet for EMS-treated cardiac arrest (see Figure 2).
 Asks patient age, sex, or chief complaint 10 (6–15) 126
We suggest that it is reasonable to expect dispatchers to iden-
 Other 19 (10–30) 109
tify 95% of EMS-treated cardiac arrests when given the oppor-
 Total dispatcher delays 28 (14–51) 261 tunity to assess for consciousness and breathing. Additionally,
Total delays 78 (43–138) 354 we suggest that DA-CPR chest compressions should be pro-
DA-CPR indicates dispatcher-assisted cardiopulmonary resuscitation. vided in 75% of cases in which bystander CPR is not already in
 1528  Circulation  October 1, 2013
Performance Standards
1) Recognion of cardiac arrest in 95% of cases in which the dispatcher has the opportunity to assess
consciousness and breathing
2) Recognion of cardiac arrest within one minute of the start of the call
3) Delivery of DA-CPR chest compressions in 75% of cases where the dispatcher has the opportunity to assess
consciousness and breathing and bystander CPR is not already in progress
4) Delivery of the first DA-CPR chest compression within two minutes of the start of the call.
Definions
Recognion of cardiac arrest is said to occur when:
1) The dispatcher gives CPR instrucons
2) The dispatcher indicates that an arrest is suspected (“We need to do CPR”)
3) Cardiac arrest dispatch protocols are implemented
Dispatcher-Assisted cardiopulmonary resuscitaon (DA-CPR) chest compressions are said to occur when:
1) The rescuer can be heard counng compressions
2) The caller verbally confirms that chest compression are being done
3) Wrien records clearly indicate that CPR due to dispatcher instrucons is provided (“DA-CPR in
progress”)
progress. In those EMS systems that fall below this benchmark,
we suggest a training program aimed at improving recognition
of agonal or abnormal breathing because this is a major barrier
to recognition of cardiac arrest.20,21 Training in the recognition
of agonal respirations has been demonstrated to improve the
detection of cardiac arrest.20,22–24
Not only is it imperative that dispatchers recognize car-
diac arrest and provide CPR instructions, but it is important
Downloaded from http://ahajournals.org by on August 22, 2022
that they do so in a timely manner. We found that dispatchers
were able to recognize cardiac arrest in a median time of 75
seconds. The median time to delivery of the first DA-CPR
chest compression was 2 minutes 56 seconds. There is wide
variation in the time to recognition of cardiac arrest by dis-
patchers reported in previous studies, from 50 seconds to 2
minutes 38 seconds.6–9 These studies took place when ven-
tilations were included in DA-CPR instructions. We found
only 1 study that reported the time to first chest compression
when ventilation instructions were not given (4 minutes in
a Medical Priority Dispatch system).10 We propose that the
time to recognition of the need for CPR and the time to first
DA-CPR chest compression are important benchmarks that
should be measured and minimized.
Although time to recognition of cardiac arrest was rea-
sonable in most cases, we found that delays in delivery of
DA-CPR chest compressions were common, with delays ≥5
seconds occurring in 92.9% of cases. Individual delays were
short in duration but cumulatively accounted for nearly half of
the time to delivery of the first chest compression. A propor-
tion of the delays could be attributed to the dispatcher ask-
ing superfluous incident and medical history questions after
it was established that the patient was unconscious and not
breathing. We believe that delays caused by unnecessary ques-
tioning can be eliminated through training and by making the
dispatcher aware of the impact of such deviations from proto-
col on the overall time to delivery of CPR.
Most of the delays documented during this study were
attributable to caller- and call circumstance–related factors
such as the emotional state of the caller and difficulty moving
Figure 2. Suggested performance standards for
emergency medical services (EMS)–confirmed
cardiac arrests. The 9-1-1 audio recordings and
written dispatch center documents should be
reviewed for all cases of EMS-treated cardiac
arrest. Whenever possible, review should be limited
to records from the dispatch agency. In cases
when a final determination cannot be made on
the basis of a review of dispatch records alone,
emergency medical technician and paramedic
records may be reviewed. CPR indicates
cardiopulmonary resuscitation; and DA-CPR,
dispatcher-assisted cardiopulmonary resuscitation.
the patient to a location where chest compressions could be
performed. Often there are factors beyond the control of dis-
patchers. Circumstances may dictate that deviation from pro-
tocol is inevitable, but dispatchers may be able to control the
extent and the duration of the deviation with vigilance and
assertive questioning. We are not aware of another study that
has documented to this level of detail the factors that lead to a
delay in the provision of chest compressions. We hypothesize
that these delays are not unique to our system and hope that
information on the frequency and duration of these delays will
help other programs design quality improvement protocols to
meet the performance standards we suggest.
This study is subject to several limitations. We were not able
to review 9-1-1 recordings for ≈20% of the cases that met the
inclusion criteria. However, we found that patient characteris-
tics were similar between cases for which the recording was
reviewed and those for which it was not reviewed (Table 1).
Additionally, we could not detect any systematic reasons for
missing recordings that may have biased the outcomes of this
study such as recording availability being based on the perfor-
mance of the dispatcher or the difficulty in identifying cardiac
arrest. Rather, recordings were missing as a result of technical
problems that occurred throughout the study period.
We excluded from our analysis patients <18 years of age;
consequently, our results apply only to the adult population. We
excluded younger patients for several reasons. First, the cause of
their cardiac arrest differs from that of the adult population.25,26
It follows that the reasons that dispatchers do not identify car-
diac arrest in young patients compared with adults may not be
the same. Second, the CPR instructions provided by dispatchers
specified that children should receive 2 ventilations before chest
compressions, whereas chest compressions were started imme-
diately for adults. As a result, the time to first chest compression
would be artificially inflated in the pediatric group.
We had to make assumptions about what the dispatcher was
thinking or what the rescuer was doing based on what we could
hear on the 9-1-1 recording. In many cases, the dispatcher
would state, “We need to do CPR” or the rescuer would count
 Lewis et al   Timeliness of Dispatcher-Assisted CPR   1529
out loud when initiating chest compressions, which allowed
us to determine when key events occurred. However, in cases
without direct audible evidence, we defined events that should
have closely followed the key event we were interested in. For
example, the first CPR instruction by the dispatch could be
used to estimate the time that the dispatcher determined that
CPR was necessary. When the rescuer did not count out loud
during chest compressions, we listened for other audible signs
that CPR was being done such as grunting or the rescuer say-
ing that he or she was performing chest compressions.
We did not assess the number of subjects who received
chest compressions when they were not in cardiac arrest and
whether these subjects suffered any injuries attributed to CPR.
This is a valid concern when emergency dispatchers diagnose
cardiac arrest over the telephone. However, a previous study
of DA-CPR patients not in cardiac arrest and involving the
same 2 communication centers that participated in our study
found that no deaths were attributed to bystander CPR and
that ≈2% of subjects had injuries attributed to bystander CPR
on review of hospital records.27 The authors concluded that an
aggressive approach to DA-CPR was warranted and that the
risk of injury was low.
As with all observational studies, we could not assess causa-
tion. However, our purpose was to describe factors that prevent
or delay the identification of cardiac arrest in this population,
to generate hypotheses that could guide future dispatcher train-
ing activities, and to suggest performance standards that may
be used to guide future quality improvement activities.
Downloaded from http://ahajournals.org by on August 22, 2022
Conclusions
This study demonstrates that dispatchers are able to accurately
diagnose cardiac arrest over the telephone but that recognition
is likely not possible in all circumstances. In some cases, rec-
ognition of cardiac arrest may be improved through ongoing
training in detection of agonal respirations. Similarly, delays
in delivery of DA-CPR chest compressions are common and
often attributable to factors beyond the control of the dis-
patcher. However, in many cases, unnecessary questioning by
the dispatcher creates delays. Dispatchers should be informed
of the impact that these seemingly minor delays have on the
overall time to delivery of chest compressions, and every
effort should be made to minimize deviations from protocol.
Performance standards for the successful and quick recogni-
tion of cardiac arrest and delivery of the first chest compres-
sion should be adopted as metrics against which EMS systems
can measure their performance.
Acknowledgments
We acknowledge the call receivers and dispatchers of King County,
Washington, whose professional and exemplary performance under
highly stressful conditions is truly remarkable. We also applaud their
willingness to continually learn and strive for perfection. We also
thank Megan Bloomingdale for her invaluable assistance in identify-
ing cases and obtaining dispatch recordings.
Disclosures
None.
References
1. Sasson C, Rogers MA, Dahl J, Kellermann AL. Predictors of survival from
out-of-hospital cardiac arrest: a systematic review and meta-analysis. Circ
Cardiovasc Qual Outcomes. 2010;3:63–81.
2. McNally B, Robb R, Mehta M, Vellano K, Valderrama AL, Yoon PW, Sasson
C, Crouch A, Perez AB, Merritt R, Kellermann A; Centers for Disease
Control and Prevention. Out-of-hospital cardiac arrest surveillance: Cardiac
Arrest Registry to Enhance Survival (CARES), United States, October 1,
2005–December 31, 2010. MMWR Surveill Summ. 2011;60:1–19.
3. Nichol G, Thomas E, Callaway CW, Hedges J, Powell JL, Aufderheide TP,
Rea T, Lowe R, Brown T, Dreyer J, Davis D, Idris A, Stiell I; Resuscitation
Outcomes Consortium Investigators. Regional variation in out-of-hospital
cardiac arrest incidence and outcome. JAMA. 2008;300:1423–1431.
4. Rea TD, Eisenberg MS, Culley LL, Becker L. Dispatcher-assisted car-
diopulmonary resuscitation and survival in cardiac arrest. Circulation.
2001;104:2513–2516.
5. Lerner EB, Rea TD, Bobrow BJ, Acker JE 3rd, Berg RA, Brooks SC, Cone
DC, Gay M, Gent LM, Mears G, Nadkarni VM, O’Connor RE, Potts J,
Sayre MR, Swor RA, Travers AH; American Heart Association Emergency
Cardiovascular Care Committee; Council on Cardiopulmonary, Critical
Care, Perioperative and Resuscitation. Emergency medical service dis-
patch cardiopulmonary resuscitation prearrival instructions to improve
survival from out-of-hospital cardiac arrest: a scientific statement from the
American Heart Association. Circulation. 2012;125:648–655.
6. Berdowski J, Beekhuis F, Zwinderman AH, Tijssen JG, Koster RW.
Importance of the first link: description and recognition of an out-of-
hospital cardiac arrest in an emergency call. Circulation. 2009;119:
2096–2102.
7. Culley LL, Clark JJ, Eisenberg MS, Larsen MP. Dispatcher-assisted tele-
phone CPR: common delays and time standards for delivery. Ann Emerg
Med. 1991;20:362–366.
8. Heward A, Donohoe RT, Whitbread M. Retrospective study into the deliv-
ery of telephone cardiopulmonary resuscitation to “999” callers. Emerg
Med J. 2004;21:233–234.
9. Vaillancourt C, Verma A, Trickett J, Crete D, Beaudoin T, Nesbitt L,
Wells GA, Stiell IG. Evaluating the effectiveness of dispatch-assisted
cardiopulmonary resuscitation instructions. Acad Emerg Med. 2007;14:
877–883.
10. Van Vleet LM, Hubble MW. Time to first compression using Medical
Priority Dispatch System compression-first dispatcher-assisted cardiopul-
monary resuscitation protocols. Prehosp Emerg Care. 2012;16:242–250.
11. Eisenberg MS, Hallstrom AP, Carter WB, Cummins RO, Bergner L,
Pierce J. Emergency CPR instruction via telephone. Am J Public Health.
1985;75:47–50.
12. Eisenberg MS. Resuscitate! How Your Community Can Improve Survival
From Sudden Cardiac Arrest. Seattle, WA: University of Washington
Press; 2009.
13. Larsen MP, Eisenberg MS, Cummins RO, Hallstrom AP. Predicting sur-
vival from out-of-hospital cardiac arrest: a graphic model. Ann Emerg
Med. 1993;22:1652–1658.
14. Bohm K, Rosenqvist M, Hollenberg J, Biber B, Engerström L, Svensson
L. Dispatcher-assisted telephone-guided cardiopulmonary resuscitation:
an underused lifesaving system. Eur J Emerg Med. 2007;14:256–259.
15. Dami F, Fuchs V, Praz L, Vader JP. Introducing systematic dispatcher-
assisted cardiopulmonary resuscitation (telephone-CPR) in a non-
Advanced Medical Priority Dispatch System (AMPDS): implementation
process and costs. Resuscitation. 2010;81:848–852.
16. Garza AG, Gratton MC, Chen JJ, Carlson B. The accuracy of predicting
cardiac arrest by emergency medical services dispatchers: the calling party
effect. Acad Emerg Med. 2003;10:955–960.
17. Ma MH, Lu TC, Ng JC, Lin CH, Chiang WC, Ko PC, Shih FY, Huang
CH, Hsiung KH, Chen SC, Chen WJ. Evaluation of emergency medi-
cal dispatch in out-of-hospital cardiac arrest in Taipei. Resuscitation.
2007;73:236–245.
18. Vaillancourt C, Charette ML, Bohm K, Dunford J, Castrén M. In out-
of-hospital cardiac arrest patients, does the description of any specific
symptoms to the emergency medical dispatcher improve the accuracy
of the diagnosis of cardiac arrest: a systematic review of the literature.
Resuscitation. 2011;82:1483–1489.
19. Jacobs I, Nadkarni V, Bahr J, Berg RA, Billi JE, Bossaert L, Cassan P,
Coovadia A, D’Este K, Finn J, Halperin H, Handley A, Herlitz J, Hickey
R, Idris A, Kloeck W, Larkin GL, Mancini ME, Mason P, Mears G,
Monsieurs K, Montgomery W, Morley P, Nichol G, Nolan J, Okada K,
Perlman J, Shuster M, Steen PA, Sterz F, Tibballs J, Timerman S, Truitt
T, Zideman D. Cardiac arrest and cardiopulmonary resuscitation outcome
 1530  Circulation  October 1, 2013

reports: update and simplification of the Utstein templates for resuscita-
tion registries: a statement for healthcare professionals from a task force
of the International Liaison Committee on Resuscitation (American Heart
Association, European Resuscitation Council, Australian Resuscitation
Council, New Zealand Resuscitation Council, Heart and Stroke
Foundation of Canada, Interamerican Heart Foundation, Resuscitation
Councils of Southern Africa). Circulation. 2004;110:3385–3397.
20. Bång A, Herlitz J, Martinell S. Interaction between emergency medical
dispatcher and caller in suspected out-of-hospital cardiac arrest calls with
focus on agonal breathing; a review of 100 tape recordings of true cardiac
arrest cases. Resuscitation. 2003;56:25–34.
21. Hauff SR, Rea TD, Culley LL, Kerry F, Becker L, Eisenberg MS. Factors
impeding dispatcher-assisted telephone cardiopulmonary resuscitation.
Ann Emerg Med. 2003;42:731–737.
22. Bohm K, Stålhandske B, Rosenqvist M, Ulfvarson J, Hollenberg
J, Svensson L. Tuition of emergency medical dispatchers in the
recognition of agonal respiration increases the use of telephone assisted
CPR. Resuscitation. 2009;80:1025–1028.
23. Perkins GD, Walker G, Christensen K, Hulme J, Monsieurs KG. Teaching
recognition of agonal breathing improves accuracy of diagnosing cardiac
arrest. Resuscitation. 2006;70:432–437.
24. Roppolo LP, Westfall A, Pepe PE, Nobel LL, Cowan J, Kay JJ, Idris AH.
Dispatcher assessments for agonal breathing improve detection of cardiac
arrest. Resuscitation. 2009;80:769–772.
25. Topjian AA, Berg RA. Pediatric out-of-hospital cardiac arrest. Circulation.
2012;125:2374–2378.
26. Young KD, Gausche-Hill M, McClung CD, Lewis RJ. A prospective, pop-
ulation-based study of the epidemiology and outcome of out-of-hospital
pediatric cardiopulmonary arrest. Pediatrics. 2004;114:157–164.
27. White L, Rogers J, Bloomingdale M, Fahrenbruch C, Culley L, Subido C,
Eisenberg M, Rea T. Dispatcher-assisted cardiopulmonary resuscitation:
risks for patients not in cardiac arrest. Circulation. 2010;121:91–97.

Clinical Perspective
For out-of-hospital cardiac arrest, the time interval from collapse to the start of chest compressions is a strong predictor of
survival. When cardiopulmonary resuscitation (CPR) is started by bystanders, the odds for survival increases 200% to 300%
compared with survival when CPR is started by emergency medical services personnel. Efforts to train the general public
in CPR go some way toward ensuring rapid delivery of CPR, but a more powerful strategy is for emergency dispatchers to
deliver instant chest compression instructions to callers reporting a cardiac arrest. This study identifies the barriers to rapid
identification of the need for CPR by emergency dispatchers and the reasons for delay in delivering chest compression
instructions. In the population studied, cardiac arrest was correctly identified by the dispatcher in 80% of cases in a median
time of 75 seconds. Chest compressions following dispatcher instructions occurred in 62% of cases in a median time of
176 seconds. We suggest performance standards for the recognition of CPR and time to deliver the first compression that
seek to improve on these findings. If these performance standards are adopted on a national level and measured by local
emergency medical services agencies, rates of bystander CPR might increase, potentially leading to greater survival from
out-of-hospital cardiac arrest.
Downloaded from http://ahajournals.org by on August 22, 2022

Go to http://cme.ahajournals.org to take the CME quiz for this article.