The aim of our prospective study was to assess the structural and proc
edural quality of an urban emergency medical services (EMS) system pro
viding prehospital basic and advanced cardiac life support (BLS/ACLS),
to compare the on-site performance of physicians and non-physicians i
n ECG diagnosis and defibrillation, and to identify incidence and caus
es of avoidable delays in the initial treatment sequences. Methods: Be
tween 1 February 1991 and 1 July 1992, 162 on-line tape recordings of
prehospital cardiopulmonary resuscitation (CPR) efforts performed by t
he staff of the EMS system of the city of Mainz were evaluated. After
arrival at the patient's side, time intervals to initial ACLS steps (f
irst ECG-diagnosis, first defibrillation, endotracheal intubation, fir
st epinephrine administration) were measured. Times to rhythm identifi
cation and countershock by EMT-Ds vs. physicians were compared (Mann-W
hitney U-test). Time intervals are presented as median values. One-hun
dred sixty-two adult patients with out-of-hospital cardiac arrests (ve
ntricular fibrillation [VF] or ventricular tachycardia [VT], 72; asyst
ole or electromechanical dissociation [EMD], 90) receiving CPR by EMTs
, EMT-Ds, and physicians of the Mainz EMS were included. Patients with
arrests due to non-cardiac aetiologies were excluded. Results: After
arrival at the patient's side, for patients with VF/VT, the EMT-Ds too
k 1:36 min and the physicians took 1:00 min to obtain the first ECG di
agnosis (P = 0.004). The first countershock was delivered within 1:42
min by both EMT-Ds and physicians of the mobile intensive care unit (M
ICU). After diagnosis was established, the EMT-Ds took 0:08 min to def
ibrillate, whereas the physicians took 0:36 min (P = 0.0001). Endotrac
heal intubation was performed within 3:30 min, and epinephrine was adm
inistered within 3:56 min. In patients with asystole or EMD, the first
ECG-diagnosis was obtained within 0:48 min by both EMT-Ds and physici
ans. Patients were intubated within 2:44 min, and received epinephrine
within 3:24 min. Conclusions: Since there is no consensus yet about t
he methods of EMS evaluation and comparison of ACLS time intervals, re
liable reference data are lacking. Yet we found that in cases of VF/NT
, physicians equipped with manual defibrillators obtain the diagnosis
earlier than EMT-Ds equipped with semi-automatic devices, but deliver
the first shock later. Another delay manifested itself in cases of asy
stole or EMD, between ECG-diagnosis and intubation or epinephrine admi
nistration. Possible causes for such delays in ACLS sequences are prob
lems in communication between physicians and EMTs. As a result of this
study, our EMS system has implemented rehearsals of taped real codes
in our megacode training.