Previous work in animals has clearly demonstrated the adverse effects of short interruptions in cardiac massage [1-3]. Therefore, cardiopulmonary resuscitation should not be interrupted until the return of spontaneous circulation or the decision to terminate the resuscitation attempt. Analysing recordings of automated external defibrillators in out-of-hospital arrests, van Alem and colleagues observed that trained police and fire brigades acting as first responders failed to execute cardiac massage during 23% of the possible time . There are presently no data on how consistent cardiopulmonary resuscitation is performed in the in-hospital settings. Accordingly, the aim of the present study was to identify length and type of unnecessary interruptions, if any, in simulated cardiac arrests.
The study was approved by the local Ethics Committee. A commercially available patient simulator with the possibility of remote control of vital parameters was used (Human Patient Simulator, METI, USA). The simulator was placed in a room specially designed to resemble as closely as possible the working conditions of the real intensive care unit (ICU).
Twelve teams participated. Each team consisted of three registered nurses, one resident and one staff physician. All nurses had a diploma in intensive care. Residents had experienced at least 2 yr of training in internal medicine while staff physicians were board certified in internal medicine and were employed on a consultant level. At the time of the study, all participants were either working in the ICU (nurses, residents) or covering for the ICU during weekends or night shifts (staff physicians). Using a checklist, all participants were extensively instructed on technicalities of the patient simulator and the equipment available.
The scenario took place in the settings of an ICU. All participants were informed that the scenario assumes that they were, during their present shift, responsible for the care of a simulated patient. The ‘patient’ was a 52-year-old male with an acute myocardial infarction admitted to intensive care. The ‘patient’ was placed in a bed and connected to a monitor that displayed a continuous electrocardiogram (ECG), continuous pulse-oximetry and the non-invasive blood pressure (BP) at selected time intervals. A cannula was placed in a peripheral vein to allow for intravenous (i.v.) administration of drugs.
The scenario started with a witnessed cardiac arrest due to a pulseless ventricular tachycardia. After the third defibrillation ventricular fibrillation occurred followed by asystole. One minute after the injection of epinephrine asystole converted into ventricular tachycardia that in turn could be converted into sinus rhythm by defibrillation.
All data analysis was made post hoc using videotapes recorded during the simulation. For each 5 s of the videotaped events it was coded whether or not the teams had performed cardiac massage, mask ventilation, or defibrillation. The focus of the study was on the later phase of cardiopulmonary resuscitation, that is, after basic life support is fully established. Accordingly, the total time of possible cardiac massage was defined as period starting with the beginning of cardiac massage and ending with the return of spontaneous circulation.
Legitimate interruptions of cardiac massage were defined as a period of 10 s following defibrillation, and the actual time required for endotracheal intubation. Unnecessary interruptions of cardiac massage were defined as any interruption of cardiac massage of 10 s or more that did not fulfil the criteria of a legitimate interruption. The actually observed time of cardiac massage was expressed in percentage of the total time possible. The compression rate was obtained by counting the actual numbers of cardiac compressions delivered in 1 min. Data are means ± SD unless stated otherwise.
Prior to watching the videotapes, the participants were asked to recall the scenario and indicate what they as a team did well and what could have been done better.
The total possible time of cardiac massage of the 12 teams was 414 ± 125 s. The mean compression rate was 83 ± 12 compressions/minute. At least one unnecessary interruption of cardiac massage was observed in each team (median 3, range 1-5). The interruptions mounted up to 65 ± 40 s (range 20-155 s) or 16 ± 10% (range 5-41%) of the total time of possible cardiac massage. Failure to swiftly resume cardiac massage after an unsuccessful defibrillation accounted for 14 of 39 episodes and for 44 ± 40% of the time of unnecessary interruptions. Ten of 39 episodes of unnecessary interruptions of cardiac massage occurred during the first quarter of the time course of cardiopulmonary resuscitation, 8 during the second quarter, 10 during the third quarter and 11 during the fourth quarter, respectively (P = n.s. χ2-test). None of the participants had realized and/or recalled the unnecessary interruptions in cardiac massage. Accordingly, when watching the videotapes most participants were surprised about the number and length of the interruptions revealed.
The present study identified a significant amount of unnecessary interruptions of cardiac massage in simulated cardiac arrests. These interruptions were neither noticed nor recalled by the health-care workers involved. There was a large between-team variability with accumulated time of interruptions ranging from 5 to 41%.
It is well established that in cardiac arrests the length of the initial delay in starting basic life support is inversely correlated to outcome [1-3]. There is no reason to assume that interruptions in basic life support of any given period are less harmful than an initial delay of that very period. Thus, unnecessary interruption of cardiac massage should be regarded as clinically relevant and might contribute to the poor outcome of in-hospital resuscitation .
There are presently no data on unnecessary interruption in basic life support in the in-hospital settings. None of the participants of the present study did notice and/or recall their team's unnecessary interruptions. Thus self-reporting appears to be unsuitable to assess this topic. Some authors used video recording to evaluate the performance during cardiopulmonary resuscitations [6,7]. However, none of these studies focussed on interruptions of basic life support. Analysing recordings of automated external defibrillators in out-of-hospital arrests with trained police and fire brigades acting as first responders, van Alem and colleagues observed that cardiac massage was not executed because of performance of the first responders during 23% of the possible time . As in the present study, these unnecessary interruptions were mainly caused by a delay in resuming cardiac massage after defibrillation. However, with a median period of 7 s (25th-75th percentile 2-15 s) the interruptions in the van Alem study were significantly shorter as in the present study (median 20 s; range 11-77 s). This difference is most likely explained by the voice prompts of the automatic defibrillator advising to perform basic life support after an unsuccessful defibrillation. It is intriguing to note that the process of self-monitoring of one's own team activity within teams of experienced health-care workers is far less effective than automatic voice prompts of an automatic defibrillator.
As all participants were, at least partly, working in the ICU, every team studied might have faced a real resuscitation in a real patient at any time during the study period. Our results may therefore be regarded as representative for the quality of cardiopulmonary resuscitation in intensive care.
The two main strengths of simulator-based studies are that (1) identical conditions for all teams can be achieved and (2) objective data from both ‘patient’ and participants can be recorded simultaneously. Both conditions are virtually impossible to achieve in real patients. A potential limitation of simulator-based studies is that participants might regard an emergency situation in a manikin less seriously than in a real patient. However, the behaviour during the simulation and while watching video recordings did indicate that they were trying to do their best to save their ‘patient’.
In conclusion, in a scenario of simulated witnessed cardiac arrest a significant amount of unnecessary interruptions of cardiac massage occurred that were not noticed by the health-care workers involved. Given the potentially serious consequences of these interruptions, further research is necessary to determine on how to improve team performance during cardiopulmonary resuscitation.
The study was supported by a grant of the Swiss National Foundation for Medical Sciences.
1. Kern KB, Hilwig RW, Berg RA, Sanders AB, Ewy GA. Importance of continuous chest compressions during cardiopulmonary resuscitation
: improved outcome during a simulated single lay-rescuer scenario. Circulation
2. Sato Y, Weil MH, Sun S et al
. Adverse effects of interrupting precordial compression during cardiopulmonary resuscitation
. Crit Care Med
3. Yu T, Weil MH, Tang W et al
. Adverse outcomes of interrupted precordial compression during automated defibrillation. Circulation
4. van Alem AP, Sanou BT, Koster RW. Interruption of cardiopulmonary resuscitation
with the use of the automated external defibrillator in out-of-hospital cardiac arrest. Ann Emerg Med
5. Peberdy MA, Kaye W, Ornato JP et al
. Cardiopulmonary resuscitation
of adults in the hospital: a report of 14720 cardiac arrests from the National Registry of Cardiopulmonary Resuscitation
6. Cooper S, Wakelam A. Leadership of resuscitation
teams: ‘Lighthouse Leadership’. Resuscitation
7. Mann CJ, Heyworth J. Comparison of cardiopulmonary resuscitation
techniques using video camera recordings. Emerg Med J