Providing appropriate quantities of the right clinical equipment to the scene of a mass casualties big bang event is vital. But planning for such emergencies is challenging. This study has, for the first time, produced an expert consensus on the items and quantities of clinical equipment that are required to treat 100 people at the scene of a big bang mass casualties event. The results of this study clearly identify the equipment that is deemed of greatest importance. Items of clinical equipment that are highly likely to be required in the immediate care and treatment of trauma patients (e.g. morphine, oxygen, stretchers and carry sheets) were rated of greater importance than other items which may not be as essential (e.g. Hydrocortisone, Aspirin, Maternity pack, or Pillows).
The study used a specifically developed website for data collection. This enabled data to be presented to participants visually, overcoming some of the known limitations of using measures of central tendency [7, 18, 21, 22] when feeding back results to participants between rounds. The web-based platform also reduced the length of time required to conduct the study: analysis of each round’s findings occurred automatically at a time appointed by the researchers and participants were immediately able to commence a further round of data collection. As data was stored electronically, the likelihood of human inputting error was also low and preliminary data analysis occurred automatically. However use of a web-based platform did not increase rates of study participation. Given its advantages, the researchers are now exploring the potential use of the web-based platform in future Delphi studies.
Consensus is reached in a Delphi study when a pre-agreed percentage of participants have rated items similarly. Ironically, there is little agreement within the Delphi study literature as to what constitutes a ‘correct’ percentage level of consensus. Previous Delphi studies show marked variation (between 50% - 100%) in consensus levels [18, 23]. A consensus level of 80% was selected in this study as it marked a clear majority opinion and has been used in previous emergency medicine research using the Delphi method . However, in recognition that the 80% cut-off selected was a strict and somewhat arbitrary definition of consensus, the items that gained at least 70% were highlighted in the study findings as items that were nearing the pre-set cut off. Ultimately the decision as to what is a sufficient degree of consensus is taken by those who use the study findings within their specific context, and not by the research team themselves.
In mass casualty big bang scenarios emergency equipment is transported to the scene as quickly as possible. The logistics of achieving this are considerable and space and resources are limited. This study provides policy makers and planners with information to support them in making an informed decision about the type and quantity of equipment that is immediately required at the scene to supplement the equipment, consumables and drugs that will already be on the ambulance vehicles attending the incident. The study results will help to minimize wastage from unnecessary equipment ordering, and to increase efficiency in routine equipment audits. The study findings were presented, at a UK wide implementation meeting, to a stakeholder group that comprised senior ambulance service clincians and managers, consultant emergency physicians, representatives from NHS England, NHS Wales and NHS Scotland, as well as English and Scottish Government policymakers and advisers. Structured discussions took place regarding the study findings. Services within NHS Scotland have since agreed to reconfigure their service and adopt the items that reached consensus, and their quantities, as the basis for stockpiling clinical equipment for mass casualties big bang incidents. In England and Wales, the output from the study and the implementation meeting have been presented to the NHS England Clinical Reference Group for Emergency Preparedness and are being used to redesign the contents of the English mass casualty vehicles.
Delphi studies are onerous tasks, both in terms of activity required and the duration of involvement. Consequently participant attrition is a known limitation, but often hard to accurately calculate as the numbers of people invited to participate are rarely reported. In this study only 32 (23%) of people invited to participate agreed and completed data collection over the three rounds. The study required participants to complete two tasks, a rating of an item’s importance, and a quantification of how many of each item was required, instead of one. The former task is one for which Delphi method is well suited and traditionally employed; the latter arguably less so. This dual task and the sheer number of items to be rated (n = 248) may have contributed to participant attrition during the study. Whilst this is disappointing, the actual numbers of participants who completed all three rounds (n = 32) is similar to other Delphi pre-hospital emergency care research [14, 16]. Researchers undertaking future Delphi studies in pre-hospital emergency care should aim to keep the number of items as manageable as possible, and estimate that they will need to invite approximately five times more individuals than the number they wish to participate. Attaching two rating tasks per item, whilst feasible, is undesirable due to the potential negative effect this may have on participant retention.
Only 54% of items reached consensus by the end of the study. This figure may have improved had further rounds of data collection occurred, but as the consensus between rounds two and three was not statistically significantly different, it is not very likely. In any case, the potential benefits of further data collection had to be balanced against the potential of participant fatigue and the potential of decreasing response rates . Delphi studies accept participant responses at face value. As elsewhere  this study would benefit from further qualitative investigation to understand participants’ responses in greater depth.
Whilst the median quantities of items are a useful starting point for services planning their resources, these need to be considered together with the local context that the service is working in: urban settings may have a higher frequency of standard emergency ambulances with a base loading of equipment, whereas more rural environments will have different considerations and require to factor in the longer distances that patients will be travelling to hospital following treatment on scene. Variation in the number of items required may have resulted from some participants misunderstanding the instructions, and consequently listing how many of an item they would wish to bring to an incident, rather than the number of items that they felt would actually be required on-scene.