Page 43 - Journal of Special Operations Medicine - Spring 2014

complicating factor could be that if the resource limita- confusion on the field. Either immediate patients have
tions (S) are at a medium level, providers must switch priority over delayed patients all the time or delayed pa-
priorities between the immediate and the delayed at tients have priority over the immediate patients all the
some time during the response effort. It may be the case time. Second, the formulas that determine which one of
that a policy that maintains the same priority all the the two prioritization scheme to be used involves the
time would be easier to implement. To give providers total number of critical patients (I+D), not necessarily
this option, we propose Simple-ReSTART, which is de- the respective numbers in each class, making the method
termined from ReSTART by dividing region C, which more robust with respect to classification errors across
corresponds to medium level of resource scarcity (region the two classes, which can lead to incorrect estimates for
C in Figure 2) into two, designating one part as belong- I and D, but not their sum I + D.
ing to region A, where resources are severely limited,
and the other as belonging to region B, where resources Demonstration of Prioritization
are not restrictive. See Figure 4 for a visual depiction Policies via an Example
of the model behind Simple-ReSTART. Specifically, the We now use an example to compare three prioritization
steps for applying Simple-ReSTART are as follows: methods: START, ReSTART, and Simple-ReSTART. We
deliberately chose an example so that all three methods
Step 1: Classify patients according to the START criteria. will prescribe a different set of actions. In this hypothetical
example, police and emergency medical services pro-
Step 2: Determine the priorities among the immediate viders have responded to a mass shooting at a crowded
and delayed patients as follows:
theater where people are watching a play. The theater is
located 15 minutes from the only trauma center in the
(I) If I + D > 2VT/R, transport all the delayed pa- area. Ten ambulances are able to stage adjacent to the
tients first and then the immediate.
theater, but it takes ten minutes to secure the area and
(II) If I + D ≤ 2VT/R, transport all the immediate gain access to the theater. Two mutual aid ambulances
patients first and then the delayed. are also dispatched at the time of the incident, but it will
take about 30 minutes for them to arrive. About 200 ad-
Figure 4 A simpler model for taking resource limitations into ditional people are able to walk away and are directed
account when determining priorities. out of the theater. Among those who remain, EMS pro-
viders find 20 dead or expectant patients, 30 immediate,
30 delayed, and 20 with minor injuries.
Prioritize immediate.
many immediate Prioritize delayed. We will assume that it has been estimated that T, the time

at which the survival rates for delayed patients start declin-
A
ing faster than those for the immediate patients is 67 min-
few immediate patients B utes from the time of the shooting based on the survival
probabilities given in Figure 5. Note that these survival
probability functions given in the figure are for penetrat-
ing injuries, and were obtained by averaging the survival
probabilities corresponding to Sacco et al.’s RPM scores
9–12 for delayed and 4–8 for immediate. Based on the
6
few delayed patients many delayed patients other information given above, we have I = 30 immediate
patients, D = 30 delayed patients, V = 12 ambulances, and
R = 30 minutes. Then, the ReSTART parameter S equals
Like ReSTART, Simple-ReSTART also is an intuitively 29.5, indicating a medium level of resource scarcity and
reasonable policy: when there are few patients or more according to the ReSTART policy we should prioritize im-
ambulances, or when travel times are shorter, resources mediate patients only for the first 29.5 minutes after triage
are less likely to be restricted and the inequality in case starts. The Simple-ReSTART policy, on the other hand,
II is more likely to be true, in which case we use the recommends always prioritizing delayed patients. For this
standard practice of prioritizing the immediate patients. example, Table 1 shows how the value of S, the priority
Otherwise (in case I), prioritizing the immediate patients suggested by ReSTART, and the priority suggested by Sim-
would increase the response time for the delayed patients ple-ReSTART change based on different numbers of am-
so much that they could not benefit from their initially bulances (i.e., V). As shown in the table, as the number of
high chances of survival. In that case, priority is given to vehicles becomes more restricted (i.e., V becomes smaller),
the delayed patients. Simple-ReSTART is simpler than both ReSTART and Simple-ReSTART give higher priority
ReSTART in two respects. First, it always prescribes a to delayed patients, while START continues prioritizing
fixed prioritization scheme, eliminating any potential immediate patients.

ReSTART: Resource-Based Triage in Mass-Casualty Events 35