Over a relatively short period in the evolution of hospital practice, some 45 years, cardiopulmonary resuscitation (CPR) has moved from its historical position as a new experimental technique in adults¹ to its current, clinically pervasive status where we expect all hospital based paediatricians to be proficient and competent in life saving procedures.^2,3 In this issue, Tibballs et al from the Royal Children’s Hospital (RCH), Melbourne, present a new development in this story.⁴ That is, the introduction of a specialist paediatric emergency team charged with improving in-patient safety and providing urgent assistance whenever it is requested—not just for cardiac arrest. On initial reading you may wonder, “isn’t that what we did when we were residents on-call?”. In the following commentary we will explore the arguments and observations that indicate the importance of this Australian development.

Unexpected cardiac arrest is a rare event in paediatric in-patients. Tibballs et al observed 20 instances in almost 105 000 admissions over a period of 41 months—that is, ∼1 per 5000 admissions. If we translate this figure to a typical UK district general hospital paediatric unit (and set aside issues such as potential differences in patient acuity and case mix) then we should have no more than one or two unexpected in-patient cardiac arrests a year. This estimate does raise an important question that will warrant local audit and investigation: if we see much more than 1 cardiac arrest per 5000 admissions, then we should be asking ourselves why.⁵ The answer could lie in a change in pattern of disease or practice, unappreciated severity of illness, or unusual circumstances—sometimes even criminal. In regard to the issue of whether we can do better with CPR there is a more pertinent question though: is this activity frequent enough to maintain skills in all attendant clinical staff working in a particular unit? One response to this question may be to think that since there are educational systems in place to ensure all clinicians have annual training in resuscitation skills (for example, Basic Life Support, Advanced Paediatric Life Support, or Pediatric Advanced Life Support (PALS) courses), then it doesn’t matter that CPR is needed rarely. Paediatricians are well prepared on surrogates of this activity with life-like clinical scenarios and manikins. There are two indications that this argument may not be correct. First, there is the influence of “self-efficacy” on performing CPR.⁶ This is not the same as self-confidence which is a relatively stable personality trait that may or may not be founded in reality. Self-efficacy varies; it may mean that despite appropriate knowledge and training, in an emergency, resuscitation techniques and practical procedures may fail to be skilfully applied unless the operator also has an adequately strong belief in their capability. For example, Simon and Sullivan⁷ examined “confidence in performance” of paediatric emergency procedures in 117 suitably qualified physicians. Over one quarter of them were uncomfortable with performing certain essential life-saving procedures (for example, tracheotomy tube change) that the American Academy of Pediatrics recommended competence in for paediatric emergency specialists.

The second indication that there could be a problem with our current system of maintaining “expertise” was identified earlier this year by Abella and colleagues.⁸ These researchers observed performance of resuscitation teams doing CPR in 67 adult, in-hospital, cardiac arrests by using a special defibrillator with monitoring features. A number of parameters of real-time CPR quality (for example, chest compression rate and depth, ventilation rate, and the fraction of arrest time without chest compressions) were followed; the authors came to the conclusion that CPR delivered at the bedside “was inconsistent and did not meet published guideline recommendations, even when performed by well-trained hospital staff”. This is not an unexpected conclusion, particularly when we consider the complexity of CPR.⁹ There is also no reason to suppose that a study of paediatric CPR teams would be any different. For example, in a resuscitation scenario requiring emergency defibrillation, White et al found that 90% of PALS trained, paediatric residents in a university based training programme discharged the defibrillator only after a dangerously long pause—on average after 149 seconds.¹⁰ Therefore if education, practice, and clinical opportunity cannot make us perform adequate CPR in the real world, then is there a better strategy?

Yes there is—we may be able to prevent cardiac arrests from occurring. The international ACADEMIA (a comparison of Antecedents to Cardiac Arrests, Deaths and EMergency Intensive care Admissions) study of 638 in-hospital emergency events, in 627 adults, found that in 60% of these instances there were significant preceding derangements in vital signs.¹¹ These so-called “antecedents” were categorised as follows: threatened airway obstruction, abnormality in breathing rate and pulse rate (too fast or too slow), low blood pressure, and altered mental state (fall in Glasgow Coma Scale (GCS) score by two or more points). Almost 500 antecedents were present in the period from 15 minutes to 24 hours before the emergency event—the most common being hypotension and fall in GCS. This finding should not surprise us—after all there is a reason why we monitor vital signs in hospital practice.¹² Over a century ago Carl Wunderlich (1815–77) and Etienne-Jules Marey (1830–1904) each developed sequential, clinical recording of patient temperature, pulse rate, and respiration. Marey even considered that the best expression of change in medical state was to have “…all the phenomena of life—movements which are so light and fleeting, changes of form so slow or rapid, that they escape the senses—an objective form must be given to them…” The work from the RCH follows in this tradition, and extends the findings of the ACADEMIA study. The assumption being: since in-patient cardiac arrests are preceded by deterioration in vital signs, then certain criteria should be used for gaining “urgent assistance”. The criteria used by Tibballs et al included paediatric thresholds for antecedents identified in the ACADEMIA study—threatened airway obstruction, tachypnoea, tachycardia or bradycardia, hypotension, and altered mental state—as well as low pulse oximetry values and whether the attendant was “worried” about the child’s state. Before using this system of emergency in-patient care Tibballs et al found that there were 20 cardiac arrests in almost 105 000 admissions, and 13 of these 20 children died. In the year after introduction of the “urgent assistance” system, there were four cardiac arrests (with two deaths) in almost 36 000 admissions (that is, ∼0.6 per 5000 admissions), and 181 calls requesting the emergency team (that is, 25 calls per 5000 admissions). The difference in rate of cardiac arrest was not statistically significant in this before-and-after study, but the direction of trend in a large number of patients is encouraging. Early identification of at-risk patients may therefore be the answer to improved in-patient safety and outcome. That is, doing better by trying to prevent in-patient cardiac arrest from occurring.

The implementation of the new system of “urgent assistance” in the RCH required what the authors called an “organisational change”. The last question we will consider in this commentary then is whether the organisational model Tibballs et al describe is applicable and practicable in places other than a children’s hospital with ∼36 000 in-patients a year; and if not, are there improvements that we can make in our own centres? Tibballs et al give an account of two changes they brought about. First, they changed the prevailing medical and nursing culture in their centre and empowered all staff to call for urgent assistance when required. Over a period of three months they instituted a hospital-wide educational campaign—Sick Child Workshops—on the clinical features of serious illness. We can all learn from this institutional commitment to patient safety and education. Our own centres and personnel would clearly benefit from such investment. Second, they built a multidisciplinary, in-patient “urgent assistance” team—comprising paediatric intensive care unit (PICU) consultant or registrar, PICU nurse, emergency department (ED) senior doctor, and medical registrar—to support bedside nurses and doctors. Few centres can provide an emergency team of similar composition to the RCH team. However, there may be an alternative solution. In their first year, the RCH team dealt with ∼1 cardiac arrest and ∼50 emergency calls per 10 000 admissions. In a UK paediatric unit, these figures could be translated into in-patient cardiac arrests being a rare event at the cost of, on average, one “urgent assistance” call per week. But there does need to be senior expertise, technical ability, and support readily available to provide the level of assistance required: a leader and team that is in some way equivalent to the RCH combined PICU, ED, and medical “urgent assistance” team, and not solely reliant on paediatric residents or specialist registrars (see above). A possible solution in the UK, identified by the Royal College of Paediatrics and Child Health, is to have acute consultants being resident on-call.¹³ Robert Scott-Jupp recently described his experience of sleeping in and what it meant for him.¹⁴ The bottom-line is that “sicker children get a speedy, mature assessment and a decision on transfer for intensive care where appropriate”. As far as he is concerned, his patients are safer.

Providing better paediatric in-patient resuscitation is an issue filled with complexity. At one level we can be better educated and practiced. However, in the real world of emergency care, Abella et al have shown that we need to improve our delivery of CPR.⁸ We could try and improve our technique and team performance, but if we really want to improve the effectiveness of in-patient CPR, perhaps it is time to re-explore mechanical aids that can undertake adequate and consistent repetitive, resuscitative actions.^1,9 At another level—preventing the need for CPR—the preliminary report from the RCH describes an important development in hospital paediatrics. Tibballs et al suggest that our system for urgent assistance of in-patients also needs to change. The exact solution will depend on our practice and place of work, and our human resources. Whichever healthcare model we choose it will, by necessity, require more involvement of senior clinicians—perhaps a new breed or generation of them.^13–,¹⁵