@article{ref1,
title="Does START triage work? An outcomes-level assessment of use at a mass casualty event [conference abstract]",
journal="Academic emergency medicine",
year="2007",
author="Kahn, Christopher and Schultz, Carl and Miller, Ken and Anderson, Craig Lewis",
volume="14",
number="Suppl 1",
pages="S12–S13-S12–S13",
abstract="2007 SAEM Annual Meeting AbstractsObjectives:The START system for mass casualty triagehas been widely used since 1983, but never field-testedfor accuracy. This study is the first outcomes-level assess-ment ever conducted following an actual mass casualtyevent to determine whether assigned triage levels corre-late with patient outcomes. The authors hypothesizedthat START achieved 90% sensitivity and 90% specificityfor each triage level.<br><br>METHODS:All triaged victims of an April 2002 train crashwere studied retrospectively. Patient records were re-viewed by the same two independent reviewers at eachhospital using a standardized tool and compared witha priori measures (modified from the Baxt criteria) defin-ing the appropriateness of each assigned triage levelusing an intention-to-treat analysis.<br><br>RESULTS:A total of 132 patients were seen at 9 hospitals,comprising 15 red, 60 yellow, and 57 green patients.Sixty-four (64) patients were triaged correctly as definedby outcomes criteria. Sixty (60) patients were overtriagedby 1 level, and 5 patients were overtriaged by 2 levels.Three (3) patients were undertriaged by 1 level. Some ofthe assigned triage levels differed from those whichshould have been assigned via strict application of STARTcriteria, based on review of out-of-hospital records. Sen-sitivity, specificity, and positive and negative predictivevalues (and 95% CIs) for each triage level were... Conclusions:This analysis demonstrates variable agree-ment between triage levels assigned by START and a priori outcomes criteria for each level. The ''walking fil-ter'' defining the green triage level functions well in rulingout serious injury. The red criteria approach hypothe-sized accuracy. Yellow criteria functioned poorly. A limi-tation is that the study could not differentiate betweenfailures of the methodology versus its inaccurate imple-mentation at the scene.<p /><p>Language: en</p>",
language="en",
issn="1069-6563",
doi="10.1197/j.aem.2007.03.704",
url="http://dx.doi.org/10.1197/j.aem.2007.03.704"
}