Article Text

Download PDFPDF
Comparative quality analysis of hands-off time in simulated basic and advanced life support following European Resuscitation Council 2000 and 2005 guidelines
  1. Hendrik Ilper1,
  2. Tina Kunz1,
  3. Holger Pfleger2,
  4. Richard Schalk1,
  5. Christian Byhahn1,
  6. Hanns Ackermann3,
  7. Raoul Breitkreutz1,4
  1. 1Clinics of Anaesthesiology, Intensive Care Medicine and Pain Therapy, Hospital of the Johann Wolfgang Goethe-University, Frankfurt am Main, Germany
  2. 2Frankfurter Institut für Rettungsmedizin und Notfallversorgung (FIRN), Professional Fire Department, Frankfurt am Main, Germany
  3. 3Institute for Biostatistics and mathematical modelling of the Johann Wolfgang Goethe-University, Frankfurt am Main, Germany
  4. 4Department of Anaesthesiology, Intensive Care Medicine and Pain Therapy, University Hospital and Medical Faculty of the Saarland, Homburg (Saar), Germany
  1. Correspondence to Dr Raoul Breitkreutz, Department of Anaesthesiology, Intensive Care Medicine and Pain Therapy, Kirrbergerstr, D-66421 Homburg (Saar), Germany; raoul.breitkreutz{at}


Aim To compare hands-off time (HOT) in simulated advanced life support (ALS) following European Resuscitation Council (ERC) 2005 guidelines and ERC 2000 and to provide quantitative data on workflow.

Subjects and Methods Observations with 18 professional paramedics, performing 39 megacodes (mega-code training; MCT) were videotaped during ALS re-certification. Teams were randomly assigned to train according to ERC 2000 or ERC 2005. HOT, hands-off intervals (HOI) and other variables describing interventions and workflow were analysed.

Results In group ERC 2000 17±3 HOI appeared with a mean duration of 17.5±10.8 s (mean±SD). Overall HOT was 382±47 s, equivalent to a mean hands-off fraction (HOF) of 0.45±0.05. 15±5 ventilation-free intervals (VFI) were observed, with a mean duration of 21±10 s. In contrast after ERC 2005 variables resulted in 18±3 HOI with a mean duration of 10.0±4.0 s (p<0.001 vs ERC 2000), overall HOT 196±33 s (HOF 0.23±0.04; p<0.001), 24±12 VFI with a duration of 24±7 s (p<0.05). The first HOI lasted for 60.4±33.1 s in ERC 2000 and 17.6±4.3 s in ERC 2005 (p<0.001). In ERC 2000 6.1±2.6 interruptions for two bag/mask ventilations (BMV) lasted for 5.4±0.8 s, whereas in ERC 2005 9.6±3.1 interruptions for two BMV took 6.5±2.2 s (p<0.001). In both groups HOI were used thoroughly for basic life support/ALS-based interventions.

Conclusion The application of ERC guidelines of 2005 markedly reduced the first HOI and mean duration of HOI at the cost of delayed secure airway management and ECG analysis in this MCT model.

  • ALS
  • BLS
  • clinical assessment
  • CPR
  • ECG
  • education
  • emergency department management
  • hands-off interval
  • paramedic
  • prehospital care
  • resuscitation
  • time
  • training
  • work-flow

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

The new European Resuscitation Council (ERC)1 2 guidelines of 2005 recommended a change in paradigm of resuscitation strategy. The chain of survival set the emphasis on early recognition, activation and cardiopulmonary resuscitation (CPR). The former intention of early defibrillation3 was changed to concentrating on basic resuscitation efforts such as thorax compressions and ventilations. The advanced life support (ALS) universal algorithm of 2005 was designed to reduce the duration of interruptions, to increase the number of thorax compressions and to simplify teaching.1 Basic resuscitation should be started as early as possible after initial evaluation. Any interruption of chest compressions must now be followed by 2 min of uninterrupted CPR and non-basic life support (BLS)/ALS-based interruptions should be avoided.1 The number of defibrillations was reduced to one single shock, and CPR should be continued immediately after shock delivery. Hands-off time (HOT) of 17–40 s was related to series of defibrillations and the use of automated external defibrillation (AED) devices.4–6 Stopping chest compressions for even a short period was shown to require further compressions to re-establish brain and coronary perfusion.2 7 Furthermore, it has been shown that an AED-based protocol using a single shock without rhythm re-analysis, stacked shocks or post-defibrillation pulse checks and an extension of the period of CPR between rhythm analyses from 1 to 2 min can increase survival rates.8 One major aim of this new ERC 2005 strategy was therefore to reduce no-flow intervals (NFI) in BLS/ALS; however, there are still intervals, for example initial BLS evaluation, initial bag/mask ventilations (BMV), analysis of an AED and defibrillation, orotracheal intubation/airway management, pulse checks and non-BLS/ALS-related interventions in which NFI can occur.9–13

To characterise the influence of ERC 2005 on HOT, we studied in a standardised ALS mega-code training (MCT) model13 with a manikin as the gold standard of CPR simulation the training of professional paramedics to analyse workflow and HOT. The advantage of MCT with paramedics was the possibility of studying timing and performance-related interruptions in more precise detail related to the training of paramedics and German prehospital work than would be possible in real scenarios such as an emergency room setting. Furthermore, target ECG rhythms could be implemented to facilitate comparison between groups. Because survival data are unavailable in an MCT, we measured the number and duration of hands-off intervals (HOI) to describe differences. HOI/HOT in an MCT can resemble a surrogate parameter to NFI/no-flow time of real scenarios.13

Subjects, materials and methods

Subjects and inclusion criteria

Professional paramedics (n=18) were included in the study protocol. The mean age of paramedics was 30.2±5.1 years (14 men, four women). They were experienced in BLS/ALS and familiar with MCT scenarios at regular emergency medical system (EMS) work. Obtaining annual recertification is mandatory to work in EMS of the city of Frankfurt am Main, Germany. Professional experience was 10.3±4.5 years. Participation within the study was voluntary, no personal advantages or disadvantages followed and the results were kept confidential. Informed consent was obtained from every participant including videotaping. Paramedics were informed of being filmed and of being part of a study. The camera was visible at all times. To ensure quality of data acquisition, only those MCT trials were included if no technical problems occurred, if it completely conformed to ERC 2000 or 2005 and if no other events occurrred that could influence HOI. As a result, 39/39 (100%) MCT trials could be included in the analysis.

Study design

The study was carried out after implementation of ERC 2005 in Frankfurt EMS in March 2006 to September 2006. It was a randomised trial of a two-rescuer scenario. Institutional review board approval of the University of Frankfurt am Main, Germany, was obtained for studies related to ALS training and emergency ultrasound. We aimed to study the influence of paramedic's behaviour when inserting focused ultrasound procedures into sensitive processes such as ALS. As no data on times for selected ALS procedures and real-time workflow were available from scientific sources, we studied this basic information first. The study was later extended to observe the influence of focused ultrasound. The present study was thus solely performed to establish reference data on HOT/HOI and to identify natural interruptions of chest compressions.

The fire department demands MCT annually from every paramedic for educational purposes. Paramedics were blinded for aims and outcome variables of the study. Immediately before an MCT trial, a team of two paramedics was randomly assigned into two groups. We chose block randomisation with two permutated blocks of identical length each. This was controlled by our statistics department.

Group ‘ERC 2000’ was scheduled to train according to ERC 2000 guidelines and group ‘ERC 2005’ applied the new ERC 2005 guidelines. The theory of both types of guidelines was repeated as a brief lecture 1 h before each trial and every paramedic was equipped with a paper copy of the BLS/ALS algorithm. According to Frankfurt EMS procedures, a keyword for the assignment of a paramedic unit was set: ‘Unconscious person. No further information or persons on scene’.

Paramedics were asked to work as they would in a real scenario. Briefing included uninterrupted chest compressions once the airway was secured by orotracheal intubation. Intubation was required because at this time laryngeal tubes or laryngeal mask airways were not yet standard equipment in every paramedic unit.

A predefined sequence of events was set and ran automatically (Ambu MegaCode Simulator Software 3.0) in every MCT: seconds 0–180; ventricular fibrillation, 180–360; asystole, 360–540; ventricular fibrillation, 540–720; asystole, 720–840; pulseless electrical activity, greater than 840; sinus rhythm and carotid pulse palpable after 5 s and controlled by an independent staff member. End of MCT was scheduled at 845 s as sinus rhythm with a palpable pulse and was established by the simulator software. Every paramedic team had to perform a complete BLS and ALS, including predefined actions of defibrillation, orotracheal intubation, intravenous access and administration of drugs such as atropine and epinephrine. After enrolment, every team completed the MCT once and then changed team leader for an additional crossover training of the same scheme to reduce bias. Six paramedics repeated three megacodes with a different colleague by building up a new team. MCT were video-recorded, stored on DVD and later analysed anonymously twice until 900 s by a single observer to reduce the intra-observer bias. Outcome variables were defined prospectively. In addition, 10 randomly selected megacodes were analysed for five arbitrarily selected HOI per video by a second observer to crosscheck results tentatively and resulted in good agreement. However, we did not take further quantitative results from additional observers and therefore could not prevent inter-observer bias.


A CPR training manikin with full-scale simulation (Ambu MegaCode Simulation System, Friedberg, Germany) and an ECG/defibrillator (M-Series; Zoll Medical, Cologne, Germany) were used. Equipment contained standard materials of Frankfurt EMS, that is emergency backpack, intravenous access and infusions (B. Braun, Melsungen, Germany), laryngoscope (KaWe, Asperg, Germany), orotracheal tubes, (Mallinckrodt, Ireland) and emergency respirator Oxylog 1000 (Draeger, Hamburg, Germany). Paramedics were familiar with this equipment and were trained to use them as suggested by the EMS director of Frankfurt.

Outcome variables

Primary outcome variables were the number and duration of HOI and HOT. Secondary outcome variables included ventilation-free intervals (VFI), any other interruptions or event-related variables (table 1). Corresponding with former publications,8 14 HOI were defined as the time between the end of chest compressions (hands-off) and the beginning of the next series of chest compressions (hands-on) of the next CPR cycle. HOI were rounded up to full seconds. No HOI was below 1 s as reviewed by the observer. Hands-off fraction (HOF) was defined as the ratio of mean HOT to the end of total experimental observation time until the pulse was arbitrarily set to be palpable at 845 s. VFI were defined as the end of two BMV until the beginning of the next BMV of the next CPR cycle. Beginning of BMV was defined as reaching out for bag and ending with lifting the mask from the manikin's face.

Table 1

Detailed analysis of time-related variables and events in BLS/ALS mega code training of experienced paramedics training in either ERC 2000 or ERC 2005

Statistical methods

For sample size planning, error of the first kind was set to α=0.05, power 0.90, estimation of a relevant difference of δ=3 s and, resulting from a hypothesised range of 10 s, SD=2.5. This resulted in a sample size of 16 in both groups. Data analysis and sample size planning was done using BIAS 9.02 (BIAS; epsilon Verlag, Frankfurt, Germany) and used for descriptive statistics. The Mann–Whitney U test was used for examination of group differences. A p value of less than 0.05 was considered, indicating significant differences between groups. Distributions of variables are given as mean±SD with CI and/or median and 25th/75th percentile when assuming non-normal distribution. If not indicated otherwise, all timings were stated in seconds.


Time until chest compressions were initiated, ie, the first HOI, was markedly reduced (table 1 figure 1) when applying BLS/ALS according to ERC 2005 guidelines. In ERC 2005, only two HOI were longer than 10 s, whereas in ERC 2000 all HOI except two exceeded this value. As intended in the ERC 2005 guidelines, the mean duration of HOI, overall HOT and HOF decreased in group ERC 2005, whereas instances of HOI were comparable in both groups (table 1 figure 2). Results of group ERC 2005 were closer to calculate ‘ideal’ HOT and HOF (table 1). Consequently, the mean duration of VFI increased in group ERC 2005.

Figure 1

Reduction of the first hands-off intervals (HOI) following basic life support/advanced life support training of experienced paramedics according to European Resuscitation Council (ERC) 2005 in comparison with ERC 2000. The first HOI was defined as the time between the beginning of the scenario until the first application of chest compression. Box plots indicate 25th and 75th percentile with median and outliers. Mean is indicated by dashed line. p Value indicates statistical differences between the groups.

Figure 2

Quantity and pattern of occurrence of hands-off-intervals (HOI) in simulated cardiopulmonary resuscitation. Upper line marks the sequence of ECG events during mega code. SR, sinus rhythm. The mean duration of HOI is indicated as black vertical bars. Periods including chest compressions are shown as intermediated white vertical bars. Grey zones indicate phases when both HOI and chest compressions occurred between training groups. The upper part depicts HOI of basic life support (BLS)/advanced life support (ALS) training according to European Resuscitation Council (ERC) 2000; the lower part represents HOI according to ERC 2005 guidelines. Counts of HOI did not differ, but cumulative duration of HOI was markedly reduced (p<0.0001) following ERC 2005. The dashed line indicates the time point of the end of each trial at 845 s, when a pulse was set to be palpable.

In contrast, the count and time for two BMV increased in ERC 2005 (table 1) and time until successfully secured airway was markedly extended (table 1) in comparison with ERC 2000. However, HOI caused by intubation manoeuvres were reduced by more than 50% of HOT in group ERC 2005 (table 1). As a consequence of earlier chest compressions, completion of first ECG analysis was also delayed and HOT for defibrillation was markedly reduced in group ERC 2005 (table 1). There were no interruptions without relation to BLS or ALS procedures.


The most prominent finding of our MCT simulation study was the first HOI to be markedly reduced when training with respect to the ERC 2005 guidelines in comparison with the old 2000 guidelines. Our model data further indicate that ERC 2005, as expected from doubling the number of chest compressions, supports earlier coronary and cerebral perfusion and reduces the incidence of defibrillation.8 14 From a retrospective analysis, it suggests that high-quality CPR was achievable in out-of-hospital cardiac arrest situations and that HOT has been partly reduced.14 However, one should keep in mind that training after ERC 2005 also resulted in a delay of initial ECG analysis and secured airway, as indicated by our model data.

As a consequence, paramedics in our MCT study caused only 23% of interruptions of continuous chest compressions due to better parallel insertion of airway and workflow. While no non-BLS/ALS-based interruptions occurred when training after ERC 2005, those 23% of HOI seem to be unpreventable when applying ERC 2005 guidelines. However, paramedic teams did not reach ideal HOT/HOI (table 1). One possible explanation for this may be that this study was carried out just after the ERC 2005 guidelines were implemented in the Frankfurt EMS.

Following former 2000 guidelines, Wik et al10 and Valenzuela et al11 observed that more than 51% of resuscitation time was not used for chest compressions. Regarding in-hospital arrests, however, a HOF of 0.24 can be observed in certified CPR performers training after the 2000 guidelines,9 although this naturally encompasses HOT during AED use and ECG analysis.15

ERC and American Heart Association 2005 guidelines recommend two BMV of 1 s each. In our MCT model, this approach resulted in mean interruptions of 6 s. BMV according to ERC 2000 guidelines were performed even slower in our MCT model. We interpret those findings as being caused by fatigue due to the increased number of chest compressions. In contrast, our findings on the reduction of HOI for intubation when applying ERC 2005 can be viewed as an improvement in workflow overlapping, resulting in better teamwork of paramedics. Our data are in line with Assar et al,16 who found in trained lay rescuers pauses of chest compressions for 16 s during conventional CPR ventilation. Ødegaard et al17 described a mean of 5±1 s for two BMV in a manikin study and a median of 5.5 s in out-of-hospital cardiac arrest.18 According to ERC 2005, 1 s is required to interrupt chest compressions, 2 s to perform two BMV, and 1 s to continue chest compressions, resulting in a total time of 4 s necessary for each interruption of BMV. American Heart Association 2005 indicates interruption time for ventilation can require 3–4 s.19

In our two professional rescuer MCT models, we were able to identify remaining reasons for CPR interruptions according to ERC 2005. These were: BLS, BMV until secure airway, defibrillator charging and defibrillation and establishing secure airway. We also found BMV to be prolonged and airway management to be markedly delayed. One may thus consider the use of alternative techniques, such as the laryngeal mask airway or laryngeal tube, which can be used to secure the airway faster and potentially earlier in ALS than orotracheal intubation20 21 to reduce HOT. Our data may serve for quality assessment and comparative analysis in resuscitation training. Although this result cannot be shown in objective data, workflow during ERC 2005 occurred in a more structured and organised manner than during ERC 2000. Because of the increased number of chest compressions, one team member is required to perform those compressions most of the CPR time; meanwhile, the other has to prepare and carry out the ALS almost alone. This makes the ERC 2005 look more structured and less chaotic.


As a limitation of our study we must state that our results are mostly based on a single viewer's findings and only partly cross-checked by another observer as our budget did not allow another overall analysis by a second observer.

Furthermore, there may have been an influence of the observation itself as known from the Hawthorne experiments22 and this can be related to prehospital research.23 One of the Hawthorne effects (experiment 1; improving illumination at work in one group and observation of both control or study group resulting in higher productivity of both groups) is likely to have influenced the results. However, as both of our study groups (ERC 2000 and 2005) were observed in a similar way, both groups may have performed optimally (and may be much better than in real life). In contrast, experiment 2 of the Hawthorne experiments (more money for one group, change of the behaviour of team leaders resulting in higher productivity) does not apply to our results for there was no money or other compensation given to the selected study participants.

This was not a multicentre study as all participating paramedics were trained in Frankfurt. The results; therefore, although based on the universal ERC 2005 may vary in other EMS areas in Europe. The results only represent the situation mentioned in the methods section; if the ECG findings are different, times and HOI may vary significantly, especially in the case of asystole, when no defibrillation is indicated. Last but not least, this is a simulator-based evaluation that may differ from real CPR although the setting may be the same.


The application of the ERC guidelines of 2005 in a MCT model markedly reduced the first HOI and the mean duration of HOI of a professional two-rescuer scenario. One of the key intentions of the ERC 2005 guidelines was thus formally confirmed in this study. However, this was at the cost of delayed secure airway management and crucial ECG analysis or shock delivery in this MCT model.


The authors would like to thank the participating paramedics for their generous work, Professor Dr Leo Latasch, Stadtgesundheitsamt, Frankfurt am Main, Germany, for continuous support as well as Volker Wilken, Dieter Oberndörfer, Bernd Jantke and Oliver Haller from the Professional Fire Department, Frankfurt am Main, Germany and Colleen Cuca for editing.



  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the ethics commitee of the Goethe-University of Frankfurt am Main in the application context of prehospital ultrasound exam studies.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles

  • Primary survey
    Darren Walter