We thank Kelly et al. for their supportive comments and interest in the National Tracheostomy Safety Project (NTSP) multidisciplinary guidelines paper [1]. As she states, a huge amount of time was spent reviewing existing guidelines, algorithms (including the ‘Bath Algorithm’ supplied with their letter), published critical incident reports, and coroner enquiries, together with input from the many specialities that are involved in tracheostomy care. As stated in our paper, whilst it is clear that many different tracheostomy or laryngectomy critical incident scenarios are possible, there are clear themes to these airway incidents. These led us to develop a single, generic algorithm for tracheostomy-related emergencies, that could use the principles of airway management and apply them to any situation. This approach was supported and endorsed by the stakeholder organisations, including representation from anaesthesia and critical care.

Multiple algorithms for multiple situations, locations or institutions can cause confusion with emergency management strategies – a feature noted in National Patient Safety Agency critical incident reviews [2, 3]. This is especially true for new or rotating medical, nursing or allied health staff. The NTSP working party agreed that a universal algorithm was more likely to direct responders to manage an emergency in a uniform manner, whilst allowing experts to use their experience where appropriate. This could be considered analogous to an expert managing a difficult airway or performing advanced cardiac life support with respect to published national difficult airway or resuscitation algorithms. Kelly et al.'s proposed alternative algorithm states it is to be used in the intensive care unit (ICU) where tracheostomy tube displacement is suspected, in patients without an ICU airway alert. What about a blocked tracheostomy, one that bleeds, cuff failure, pneumothorax, the known difficult native airway, or the patient who was in ICU yesterday but is on the medical ward today [4]? Should we choose yet another algorithm? When the patient turns blue, there is no time to decide which piece of paper the responders should follow, especially for ventilator and oxygen dependant patients. For these reasons, the consensus view of the multidisciplinary representation of the NTSP working party did not agree with Kelly et al. that separate algorithms were required for ICU.

We recognised that special circumstances exist in some situations and that the features of a recently performed percutaneous tracheostomy, more likely to be displaced, would be more common in critical care areas than perhaps elsewhere. However, the same management principles can be applied as other situations: recognising and dealing with the commonest reversible problems and maintaining oxygenation by any means possible. Kelly et al. highlight the established role of capnography in monitoring airway device patency and assisting in the management of airway manipulations. The NTSP algorithms use capnography prominently and early in the decision tree (in the first box) with its use emphasised throughout the algorithms. Clearly, just having capnography available is not going to improve airway management. The NTSP have developed extensive resources to support teaching of the algorithms that detail the importance of understanding and interpreting capnography waveforms (described in the paper, with further resources available at and in the NTSP iPhone App, available at

Whether you are oxygen and/or ventilator dependant, managed on a critical care unit or a ward, we believe that the NTSP algorithms are universally applicable and we have devised, refined and tested them in multiple clinical and simulated scenarios, locations and countries to demonstrate this. We welcome further feedback such as that provided by Kelly et al. so that any updates and revisions to the NTSP algorithms are informed by clinical experience of their use in as many situations and scenarios as possible. Our feedback to date has involved the use of the NTSP algorithms in training and clinical situations by multidisciplinary professionals in theatres, wards, emergency departments, community settings and critical care units in the UK and beyond.

All of the features detailed in Kelly et al.'s supplied ‘Bath Algorithm’ are present in those published by the NTSP. We would suggest avoiding a ‘trial of ventilation’ with a potentially displaced tracheostomy tube, as discussed in our paper. We believe that standardising the response to tracheostomy and laryngectomy emergencies is beneficial for this vulnerable cohort of patients wherever they are managed and rather than developing a confusing variety of bespoke algorithms, we hope that clinicians will focus on adopting agreed, endorsed resources.


  1. Top of page
  2. References