Page 38 - Journal of Special Operations Medicine - Spring 2014

ReSTART:
A Novel Framework for Resource-Based Triage

in Mass-Casualty Events

Alex F. Mills, PhD; Nilay T. Argon, PhD; Serhan Ziya, PhD;
Brian Hiestand, MD, MPH; James Winslow, MD

ABSTRACT
Objective: Current guidelines for mass-casualty triage Further validation is required before field implemen-
do not explicitly use information about resource avail- tation; however, the framework proposed in here can
ability. Even though this limitation has been widely serve as the foundation for future work in this area.
recognized, how it should be addressed remains largely
unexplored. The authors present a novel framework de- Keywords: triage, mass-casualty event, prioritization
veloped using operations research methods to account
for resource limitations when determining priorities for
transportation of critically injured patients. To illustrate
how this framework can be used, they also develop two Introduction
specific example methods, named ReSTART and Sim- There is very little existing research to validate the reli-
ple-ReSTART, both of which extend the widely adopted ability and effectiveness of triage systems and algorithms
triage protocol Simple Triage and Rapid Treatment currently in use. In a mass-casualty situation, triage is
1
(START) by using a simple calculation to determine pri- required to maximize the delivery of limited resources so
orities based on the relative scarcity of transportation as to benefit injured patients to the greatest extent pos-
resources. Methods: The framework is supported by sible. For example, within the START triage guidelines,
three techniques from operations research: mathemati- patients are classified as expectant (not expected to sur-
cal analysis, optimization, and discrete-event simulation. vive injuries given severity or care available), immediate
The authors’ algorithms were developed using mathe- (survivable injuries but requires definitive medical treat-
matical analysis and optimization and then extensively ment within 1 hour to survive), delayed (potentially seri-
tested using 9,000 discrete-event simulations on three ous but not expected to deteriorate over the next several
distributions of patient severity (representing low, ran- hours), and minor. In the START framework, all im-
2
dom, and high acuity). For each incident, the expected mediate patients in a mass-casualty incident should be
number of survivors was calculated under START, transported before delayed patients, regardless of the
ReSTART, and Simple-ReSTART. A web-based decision resource availability. Previous authors have suggested,
support tool was constructed to help providers make pri- however, that there is potential to improve outcomes
oritization decisions in the aftermath of mass- casualty if prioritization decisions explicitly account for the re-
incidents based on ReSTART. Results: In simulations, source limitations that may arise in various stages of the
ReSTART resulted in significantly lower mortality than response effort. 1-5 Research to date has not addressed
START regardless of which severity distribution was whether there is evidence to support this claim or how
used (paired t test, p < .01). Mean decrease in critical existing triage protocols such as START could be ex-
mortality, the percentage of immediate and delayed pa- tended in a way that takes into account resource limi-
tients who die, was 8.5% for low-acuity distribution tations. Despite the fact that one of the core concepts
(range –2.2% to 21.1%), 9.3% for random distribution of mass-casualty triage is that triage systems should be
(range –0.2% to 21.2%), and 9.1% for high-acuity dis- resource dependent, existing triage guidelines fail to de-
3
tribution (range –0.7% to 21.1%). Although the critical scribe how “the decision of whom to treat and/or trans-
mortality improvement due to ReSTART was different port first and how best to use the resources on hand,”
for each of the three severity distributions, the varia- should be made. Currently, there is no guideline or
2
tion was less than 1 percentage point, indicating that the rule-of-thumb to help emergency responders make such
ReSTART policy is relatively robust to different severity decisions based on resource limitations. To fill this gap,
distributions. Conclusions: Taking resource limitations we report the findings of a research project that uses op-
into account in mass-casualty situations, triage has the erations research methodologies including mathematical
potential to increase the expected number of survivors. analysis, optimization, and discrete-event simulation.

30