Every year, approximately 30 000 people in the UK have an unexpected cardiac arrest in hospital. Despite significant advances in resuscitation research, survival to hospital discharge following cardiac arrest in adults remains poor . The survival benefit of well-performed cardiopulmonary resuscitation (CPR) is well documented. Recent evidence from both resuscitation training and in-hospital cardiac arrest suggests that CPR quality is suboptimal [2, 3].
Human factors affect the quality of CPR and disparity exists between resuscitation theory and its practical application – even experienced teams often perform sub-optimally in simulated resuscitation scenarios . Possible explanations for this include the high-stress environment resulting in poor leadership behaviour, failure to delegate tasks explicitly, poor recall of knowledge and inevitable skill decay [5–7].
Resuscitation feedback devices and cognitive aids (visual or auditory prompts to aid recall) can improve the quality of CPR during training and simulated cardiac arrests . Smart phones downloaded with resuscitation algorithms have the potential to improve performance in emergency scenarios. To date, randomised clinical trials have shown them to be of significant benefit only in the training of non-medical ‘bystanders’ in basic life support [8, 9].
The aim of this study was to ascertain whether providing appropriate prompts in a portable, user-friendly format produces better results in a simulated medical emergency than relying upon memory alone.
This study was assessed by the chair of the Royal United Hospital Bath Research and Ethics Committee, who considered full ethical review to be unnecessary. This was an open label, randomised controlled trial using junior doctor volunteers, conducted in the Education Centre at the Royal United Hospital. The study took place over three evening sessions, from February to March 2010, to enable as many junior doctors as possible to attend. While the study groups were waiting for their advanced life support (ALS) scenario, they participated in clinical skills workshops. Participants were randomly assigned to either the intervention arm or the control arm in a 1:1 ratio by receiving shuffled opaque-sealed envelopes.
Eligible participants were Resuscitation Council (UK) ALS-trained doctors (within 5 years of qualification) working in either the Royal United Hospital, Bath or Southmead Hospital, Bristol, at the time the study took place. They were recruited via poster and email. Written, informed consent was obtained from all the participants. They were not blinded as to the purpose of the study. Those who had not completed a Resuscitation Council (UK) ALS course within the last 4 years were excluded.
Participants were invited to join a resuscitation training session after normal working hours. All participants were briefed in pairs for 10 min, and shown how to use the iResus ‘app’ (Version 1.0) preloaded on an Apple iPhone. The participants were taught how to operate the smart phone and navigate the ‘app’, but were not exposed to the bradycardia algorithm that would form the basis of their assessment (they were taught how to switch to the advanced life support algorithm as part of the training). After randomisation, participants in the intervention arm were given an iPhone and encouraged to use the iResus ‘app’ during the scenario; the participants randomly assigned to the control arm did not have access to any cognitive aids.
Participants were assessed using one of the Resuscitation Council (UK) cardiac arrest simulation tests (CASTest). The CASTest tests the application of resuscitation knowledge and skills during a simulated cardiac emergency. During the assessment, the participant had initially a single nurse helper who would locate equipment and follow instructions; as the scenario progressed, an additional helper became available who could perform CPR, give drugs and defibrillate. Resuscitation equipment and drugs were set out in a standardised ALS scenario assessment format. All candidates had the same scenario (a patient with a recent inferior myocardial infarction, complicated by compromising complete heart block at a rate of 40–50 beats.min−1 who deteriorated to cardiac arrest – pulseless electrical activity and then ventricular fibrillation, which, if treated successfully, would revert to a perfusing sinus rhythm). A SimMan™ simulator (Laerdal, Stavanger, Norway) was used with full defibrillation capability.
Performance during CASTest was measured using the validated CASTest scoring system . The score sheet contains four domains, each with performance criteria within them to characterise the quality of participant performance in further detail (Table 1 and Appendix).
Table 1. Advanced life support scenario evaluation criteria scoring system.
The 24 performance criteria are individually scored out of a maximum of four (1 = unacceptable; 2 = borderline; 3 = acceptable; 4 = excellent). The maximum score was 96. There were at least two assessors for each scenario and scoring was done by consensus between assessors.
Following their resuscitation assessment, the participants were asked to complete a questionnaire. Questions included whether the participants owned an iPhone, (and if they had already downloaded iResus). In addition, regarding the use of the iResus ‘app’ and their attitude and perceived attitudes of using iResus in clinical situations, the participants were asked to score the statements from 1 (strongly disagree) to 10 (strongly agree).
The primary outcome measure was the score on the CASTest in the two groups. Secondary outcome measures were the participants’ iPhone ownership and attitudes towards using the ‘app’ as assessed by a questionnaire.
A power calculation suggested that 15 candidates in each group would provide 80% power for demonstrating a 20% improvement in CASTest performance. The likely performance of the control group (mean score 57.1, SD 10.5) was estimated using data from a cardiac arrest simulation test (CASTest) scoring study using data from candidates who were retested 12 months after their ALS certification (data provided by Resuscitation Council (UK)).
Subjects were allocated to their group using opaque-sealed envelopes, containing a folded strip of paper with ‘iPhone’ or ‘No iPhone’ written on it. There were equal numbers of envelopes for the two groups. The envelopes were subjected to simple randomisation. Researchers who were not involved in the scenario assessment process performed sealed envelope preparation and allocation.
It was not possible to blind assessors to allocation groups.
Results were analysed using STATA 10.1 for Macintosh (STATA Corp., College Station, TX, USA). The CASTest scores had a non-normal distribution so the two groups were compared using the Wilcoxon rank sum test (Mann–Whitney).
Out of the 47 potentially eligible junior doctors from all specialties who responded to email invitation from January to March 2010, seven were excluded as they had not passed a Resuscitation Council (UK) ALS course within the last 4 years, and an additional nine, who initially expressed interest, failed to attend any of the three organised sessions. Sixteen out of the 31 participants were randomly assigned to perform their ALS scenario with a smart phone (with the iResus ‘app’ downloaded), and 15 without (see Fig. 1). All the participants completed a feedback questionnaire. There were no losses to follow-up.
The two groups had similar baseline characteristics (Table 2).The CASTest score for the participants in both groups was not normally distributed (Fig. 2), but similar in spread. There was a significant difference in the CASTest scores between the two groups. The median (IQR [range]) CASTest score in the smart phone (iResus) group was 84.5 (75.5–92.5 [64–96]) compared with a score of 72 (62–87 [52–95]) for the control group (p = 0.02).
Table 2. Baseline characteristics for the study participants. Data are mean (SD) or median (IQR [range]).
Smart phone (iResus)
No smart phone
ALS, advanced life support.
Months post medical degree qualification
Months post ALS certification
10 (6–23 [1–40])
11 (5–32 [3–44])
Already own an iPhone
There were 11 (35%) existing iPhone users in our study population; these were equally distributed (six were randomly assigned to the iResus group and five to the no iResus group). Out of those who already owned an iPhone, only one (9%) already had iResus downloaded on their phone. This participant had been randomly assigned to the smart phone group.
The participant scores are summarised in Tables 3 and 4. Participants stated that the iResus ‘app’ was easy to use, increased their confidence in making decisions, and that they would be prepared to use it in real clinical emergencies. From their own perspective, they did not think using such an ‘app’ would be unprofessional or indicate poor training. They expressed a neutral response when asked if the public or other healthcare professionals would view usage in these negative terms.
Table 3. Participants' views on usage of the iResus ‘app’ based on scores of 1 (strongly disagree) to 10 (strongly agree). Values are median (IQR [range]).
No smart phone (n = 15)
Smart phone (iResus) (n = 16)
I found the ‘app’ easy to use
8 (7.5–8.5 [7–10])
8 (7–8.25 [3–10])
In a clinical emergency would increase confidence in decision-making
8 (7–8.5 [4–10])
7.5 (7–9 [3–10])
Happy to use in a real clinical emergency
8 (6.5–9 [4–10])
7.5 (5–9 [4–10])
Table 4. Participants’ attitudes and perceived attitudes towards the iResus applications based on score of 1 (strongly disagree) to 10 (strongly agree). Values are median (IQR [range]).
All participants (n = 31)
‘Doctors using iResus in a real clinical situation are unprofessional and poorly trained’
3 (2–4 [1–8])
‘Public will perceive doctors who use iResus in a real clinical situation to be unprofessional and poorly trained’
5 (4–7 [1–10])
‘Other healthcare professionals will perceive doctors who use iResus in a real clinical situation to be unprofessional and poorly trained’
5 (4–7 [1–8])
iResus improved junior doctors’ CASTest scores during a standardised simulated cardiac arrest scenario when compared with those applying purely their own knowledge and experience.
The junior doctors in our study found iResus easy to use and felt that it would provide them with an increased level of confidence in a stressful emergency scenario. They did not consider the use of iResus to be unprofessional or reflect a poor level of training in a real clinical situation. This contrasts findings from a previous study where junior doctors felt that using a cognitive aid would show a lack of knowledge and professionalism .
iResus provides almost instant access to the appropriate algorithms and drug doses for resuscitation situations that may be managed incorrectly if memory alone is relied upon; it has been shown that stressful situations make errors more likely . There is growing literature to support the use of cognitive aids in resuscitation. Two randomised controlled studies involving simulated basic life support (BLS) by laypeople without prior CPR training showed improved performance with the use of mobile phone CPR programmes [8, 9]. A recent review of CPR feedback/prompt devices (including cognitive aids) used during training and CPR performance found no randomised controlled trials of their use during actual cardiac arrests .
Our findings are similar to a previous study of animation assisted CPR among laypeople 6 months after CPR training: those in the animation assisted group had higher checklist scores, demonstrating that those trained with cognitive aids perform better . In another study, medical students who had received CPR training 2 months before were divided into three groups: a control group; a short CPR checklist; and a longer, more detailed version – those in the longer checklist arm performed best . Other groups have investigated the use of cognitive aids in life-threatening ‘peri-arrest’ scenarios (paediatric anaphylaxis and malignant hyperpyrexia), also with supporting outcomes [11, 14, 15].
Improved recall of factual information is also important for effective ALS; studies undertaken before the development of personal digital assistants (PDAs) PDAs and the iPhone have demonstrated that more simplistic aide memoires are also effective [16, 17].
Contrasting evidence also exists. One study of neonatal resuscitation used a poster with resuscitation guidelines as the intervention (versus control – no poster); none of the study participants in either group was able to perform adequately to pass a neonatal resuscitation programme test (based on completion of five key steps displayed on the poster). It was postulated that infrequent use of cognitive aids had contributed to their results . Potential harm was reported in two further studies, including delay in initiation of CPR and the use of incorrect algorithms; thus the outcome of using a cognitive aid such as a checklist may be specific to the aid or the situation [19, 20]. A recent randomised controlled study using a mobile phone with audio CPR prompts did improve CPR quality (better hand position, compression rate and depth and fewer pauses) in lay rescuers; however, there was a 30-s delay to initiation of CPR .
In other high-risk industries (such as aviation or nuclear power), the use of cognitive aids is integral to their standard operating procedures. The safety culture in medicine has changed with respect to checklists, for example the recent wide implementation of the World Health Organization (WHO) surgical checklist within operating theatres. However, a culture still remains in which doctors may be reluctant to use cognitive aids for fear of appearing incompetent. Improved team performance in a simulated anaesthetic emergency relating to the use of cognitive aids has been demonstrated, and the investigators commented on the need to confront negative attitudes within healthcare towards the use of such aids .
Our study provides further additional support to the current evidence, suggesting that CPR prompt devices improve skills and therefore potentially, patient outcome.
It is possible that the improved performance was because of the rapid availability of the ALS algorithms and drug doses, rather that the medium with which it was presented (smart phone and iResus), although this is difficult to prove unequivocally. A similar improvement may have been seen with the use of a wall poster. However, as iResus is supported by the Resuscitation Council (UK), the iResus ‘app’ will always provide the most up-to-date UK guidelines. We chose not to compare iResus with another type of cognitive aid because most cardiac arrests do not occur near a wall poster, and most healthcare professionals do not carry card-based cognitive aids. Thus, we considered a study group using iResus compared with a control group with no cognitive aids to be most representative of the real world.
There was no pre-testing of candidates and therefore we cannot be absolutely certain that the two groups were of equal ability; however, the baseline data suggest that they are similar in terms of clinical experience and time since ALS certification. A potential limitation is that both groups received training on the ‘app’. This could serve as revision of the guidelines though neither group trained using the bradycardia algorithm. If training had not occurred in the control group, we would not be able to discern the effect of the ‘app’ as opposed to teaching/revision of algorithms during the training. The subjects were informed that even though they received training, they might not have access to the ‘app’ during their assessment scenario.
A further limitation of the study is that the assessors were not blinded to the study group during scenario performance. Blinding would be very difficult to achieve in these circumstances because it was immediately apparent to the assessor when the participant was referring to iResus. An option would have been to make the assessors unaware of the purpose of the study, e.g. they might have been told ‘candidates are free to use an ‘app’ or not, as they wish’. However, the authors were keen to investigate how even a very limited training period on this ‘app’ (none of the doctors had been exposed to the bradycardia algorithm on the ‘app’) could affect the performance of the doctors. Hence, those randomly assigned to the iResus group were strongly encouraged to use the ‘app’; although it was not mandatory, they uniformly referred to the ‘app’ and were able to navigate an algorithm that was unfamiliar to them.
Whilst simulation may be a useful training tool, the extent to which results from simulation studies can be extrapolated into clinical practice remains largely unknown. However, one study has shown that a simulation-based educational programme significantly improved performance during cardiac arrest .
New technology has changed the way in which we access information. Further studies are required to investigate the effect of cognitive aids on CPR quality (interruption to chest compressions). Further evaluation of the iResus ‘app’ in real-life emergency scenarios is required.
We are very grateful to Iain Smith of the simulation laboratory at the Royal United Hospital, Bath, for the loan of simulators, resuscitation training equipment, and for his help with the set up for the assessments.
Daniel Low, Medical Director of iMobileMedic.com, was commissioned by Resuscitation Council (UK) to develop iResus. During the study, he was not involved with assessment, data collection or analysis, nor did or does he receive any financial gain from the ‘app’. iMobileMedic.com received a licence fee from the Resuscitation Council (UK) for the first 50 000 downloads. Jasmeet Soar is Chair, Resuscitation Council (UK), Editor of Resuscitation and co-chair of the Education, Implementation and Teams Task Force of the International Liaison Committee on Resuscitation. Jerry Nolan is Editor-in-Chief of Resuscitation, and Co-chair of the International Liaison Committee on Resuscitation. Natasha Clark, Adam Stoneham, Andrew Padkin and Gavin Perkins declare no competing interest.
Appendix Cardiac arrest scenario.
Candidate is given the following briefing
You are called to see a 60-year-old patient who has developed complete heart block (CHB) after an acute inferior myocardial infarction.
Scenario develops as follows
Initially, reduced conscious level, bradycardic, hypotensive, unresponsive to treatment with atropine. Patient deteriorates and collapses. Pulseless electrical activity (PEA) arrest complete heart block (CHB) rate 40–50 min−1) that changes into ventricular fibrillation (VF) arrest. Patient reverts to spontaneous circulation after 2nd shock, begins to breathe, but remains hypotensive.
CAS test scenario – marking criteria
NSR, Normal sinus rhythm. Each criteria is graded 1–4 (1 = unacceptable, 2 = borderline, 3 = acceptable, 4 = excellent).
(CHB) ABCDE approach
Oxygen, IV access
Recognise compromised bradycardia
Atropine 0.5 mg IV
Consider up to 3 mg atropine IV
Request transcutaneous pacing
Cardiac arrest management
(PEA) Check patient (breathing/circulation)
Call resuscitation team/help
2 min CPR (30:2)
Attach ECG monitoring (if not already)
Give adrenaline 1 mg IV
Recognise and treat relevant reversible causes (drugs/electrolyte disturbances)
(VF) Check monitor/confirm rhythm
First shock (150–200 J biphasic or 360 J monophasic)
2 min CPR (continuous chest compression/ventilation)
(VF) Check monitor/confirm rhythm
Give further adrenaline after 3–5 min
Minimise interruptions in CPR
Second shock (150–200 J biphasic or 360 J monophasic)
2 min CPR (continuous chest compression/ventilation)