7 Mass casualty incidents: a review of triage severity planning assumptions

Background Recent events involving a significant number of casualties have emphasised the importance of appropriate preparation for receiving hospitals, especially Emergency Departments, during the initial response phase of a major incident. Development of a mass casualty resilience and response framework in the Northern Trauma Network included a review of existing planning assumptions in order to ensure effective resource allocation, both in local receiving hospitals and system-wide. Existing planning assumptions regarding categorisation by triage level are generally stated as a ratio for P1:P2:P3 of 25%:25%:50% of the total number of injured survivors. This may significantly over-, or underestimate, the number in each level of severity in the case of a large-scale incident. Methods A pilot literature review was conducted of the available evidence from historical incidents in order to gather data regarding the confirmed number of overall casualties, ‘critical’ cases, admitted cases, and non-urgent or discharged cases. This data was collated and grouped by mechanism in order to calculate an appropriate severity ratio for each incident type. Results 12 articles regarding mass casualty incidents from the last two decades were identified covering three main incident types: (1) Mass transportation crash, (2) Building fire, and (3) Bomb and related terrorist attacks and involving a total of 3615 injured casualties. The overall mortality rate was calculated as 12.3%. Table 1 summarises the available patient casualty data from each of the specific incidents reported and calculated proportions of critical (‘P1’), admitted (‘P2’), and non-urgent or ambulatory cases (‘P3’). Conclusions Despite the heterogeneity of data and range of incident type there is sufficient evidence to suggest that current planning assumptions are incorrect and a more refined model is required. An important finding is the variation in proportion of critical cases depending upon the mechanism. For example, a greater than expected proportion results from incidents involving a building fire whereas the existing model may over-estimate critical caseload in more ‘conventional’ incidents such as a transportation accident or even in terrorism-related incidents. A new model suggesting the proportions of casualties expected by severity categorisation and incident type is shown in table 2. A more detailed investigation is planned to further refine and develop this model. Table 1 Table 2