Major Incidents, Mass Casualty Incidents, Patient Care

A Scientific Look at START and Our Ability to Do It

Issue 5 and Volume 40.

Tragic scenes of mass casualty incidents (MCIs) and patient suffering currently saturate the media and America’s collective conscience. Regular images of MCIs, such as the Virginia Tech and Sandy Hook Elementary school shootings, the Boston Marathon bombings, earthquakes, fires, floods and tornados, parade through our daily lives.

These events take many forms and are derived from a plethora of catalysts. Regardless of the character of the incident, anywhere bullets or debris fly, citizens will bear the damaging effects of these forces. A common thread linking any event where multitudes of people lay injured or dying is EMTs and paramedics performing heroic acts to save lives and limit suffering.

EMS personnel on scene at any MCI dutifully begin the harrowing process of sorting through the injured and determining which patients they can save, which patients they must pass over and which patients should leave the scene rapidly on a high-priority basis.

This process collectively is known as triage. This article will dissect the most common system of triage, known as the START triage method. Incidents of START triage application in recent MCIs will be analyzed to determine how efficient, effective and consistently it’s performed and discuss times when personnel must use other parameters and considerations (such as resources available and hospital capabilities) to perform triage and patient distribution in an optimal manner.

Every patient at an MCI needs to be tagged or marked in a prominent way that ensures they’re treated, transported and properly logged and accounted for. Photo courtesy Laura Bunch/Cabarrus County EMS


MCI Defined

An MCI is defined by its impact on emergency resources. It might be a three-car accident with four patients in a rural area with limited resources, or a commuter train crash in an urban area with significant emergency resources. The gross number of patients alone doesn’t define an MCI–the degree to which the responding emergency services are overextended and have to absorb the patient influx does.

MCIs are further categorized depending on the ramifications they have on available resources. Low-impact incidents are those that can be handled with available local resources without need for outside assistance. High- impact incidents overwhelm local resources and require assistance from outside the immediate jurisdiction or geographical area.

The final category refers to an MCI that’s so large as to necessitate regional or federal resources, multiple layers of command, an emergency operations center and is often a protracted event. This is known as a disaster/catastrophic incident.1

Consideration from a global perspective is daunting. It’s estimated that 225 million people are affected by disasters annually.2 At any MCI, regardless of the degree of impact, numerous patients will require immediate triage and treatment.

Serious injuries that can be aggressively managed on scene can be triaged to the Priority 2 level so that Priority 1 patients can leave the scene before them. Photo courtesy Ron Tencati/Cabarrus County EMS


START Triage

Triage is derived from the French word trier, which means “to sort.”3 Resulting from military situations in which hundreds or thousands of soldiers laid injured on the battlefield, this process involves responders systematically assessing and determining which patients are most injured and which patients are least injured.

In order to provide the most help to the most soldiers, military doctors began ascertaining methods to prioritize patients, hoping to ensure the most severely injured are treated first and that those beyond hope of salvation don’t absorb precious resources in a futile attempt.

The first documented use of organized triage was during the Napoleonic Wars.3 This process has evolved over centuries, yet remains a difficult course of action.

In order to simplify this brutally visceral process, multiple systems for performing triage have been devised. In the United States, the system known as “Simple Triage and Rapid Treatment,” or START, is by far the most commonly used system in EMS.1

The START system is designed to direct responders in the rapid assessment of victims’ abilities to walk, their mental status, hemodynamic status and respiratory drive within 60 seconds of contact.4

Based on the assessment performed during triage, patients are then categorized into one of four categories. (See Figure 1, below.) The most severely injured patients who will receive the most care and fastest transport from the scene are classified “Immediate” and color-coded red.

The second category applies to those patients who are injured, yet whose condition will remain stable without being treated or transported until all more serious patients are addressed. These patients are assigned the color yellow and titled “Delayed.”

The third category addresses the patients who have minor injuries. These patients often represent the “walking wounded” and are classified as “Minor” and assigned the color green. Finally, victims killed during the incident, or who are so severely injured they can’t possibly be resuscitated with the resources available, are assigned the color black and are termed “Deceased or Expectant.”

To assist with patient tracking and reassessment, each casualty triaged is assigned a color-coded tag or other standardized, identifiable triage device that’s physically placed on their person. The triage tag or device color corresponds to their triage category and will vary in verbiage based on the manufacturer (e.g., “Immediate” or “Priority 1”).

Scientific Limitations

The START method is popular due to a variety of factors. It’s been in place across the U.S. since the early 1980s and enjoys almost universal usage among agencies. But, surprisingly, its use is based on scant scientific evidence or academic evaluation.2,5

The lack of conclusive scientific study into a critical system used in high-stress situations represents a significant deficiency in emergency medicine. In one recent study, the authors concluded, “There is no existing measure in which to judge the accuracy or appropriateness of mass casualty triage decisions in either trauma or disaster literature.”6 Obvious difficulties exist in testing any triage method during an actual MCI.

Published literature is by no means the only method to evaluate the efficacy of the system. It is, of course, a system that’s currently in place and is used each day somewhere in the U.S.

One of the most controllable methods to systematically test the system is by staging disaster drills. The New York City Fire Department (FDNY) participated in such a drill in 2006. The established scenario for the drill involved a train collision with resulting chemical release and fire. FDNY EMS personnel then triaged 130 patient-actors achieving an overall triage accuracy of 78%.7 Successful evidence of START triage in practice in an observed setting certainly seems to indicate that the system is a valid process.

Additionally, one must take into consideration that any triage system is only as good as the human wielding its guidelines. Stress, panic, poor visibility, confusion and inexperience are all potential distractions to practitioners performing triage.

MCIs are often once-in-a-career events and EMTs, paramedics or firefighters may never have sufficient training or experience mitigating to be fully prepared. In 2001, the Annals of Emergency Medicine published a paper studying medics’ ability to triage non-disaster patients into one of four categories in order to study their clinical judgment skills. The study indicated that paramedics’ triage skills fell short of accepted standards and proposed that clear triage guidelines and criteria were critical to a successful triage system.8

A tourniquet can quickly help you triage a patient as Priority 2. Photo courtesy Ron Tencati/Cabarrus County EMS


Patient Outcome & Transportation

Placement of triage tags on patients doesn’t end the triage process. Once triaged, emergency treatment and transportation to the hospital follows in accordance to the assigned priority. However, an evaluation of available studies indicates, “…there are no data to show whether correctly sorting patients into categories set forth by any particular triage system results in improved outcomes, either for an individual patient or for a group of patients as a whole.”5

It seems difficult to accept the statement that triage doesn’t appear to improve patient survivability because there has to be some manner for EMS to address multiple patients.

So the question is, can START be numerically linked to the improvement in patient outcomes? The answer is “no,” not without proper communication with the receiving facility and optimization of the continuum of care. EMS field triage must be an ongoing process that seamlessly integrates into the larger healthcare arena.

An example of this statement in action is evidenced by the 2005 train disaster in Los Angeles. During the incident, 241 passengers were triaged. Of these, 11 were fatally injured, 129 patients were transported, and the rest didn’t require attention.9

EMS units transported the majority of critical patients to community hospitals without fully utilizing nearby trauma centers. In one study, confusion over START triage nomenclature was blamed for the misdistribution of patients. The authors concluded that the categorization of patients as “delayed,” “immediate,” etc. led to confusion that ultimately placed doubt in the triage system’s “utility and accuracy.”9

This case could be more descriptive of an EMS system/hospital communication failure rather than implying an inherent weakness of the START algorithm. Contradictory research conducted involving disaster medical assistance teams (DMAT) during four separate deployments found that, “… a higher-acuity START category was associated with an increased likelihood of transfer to a hospital and may continue to be a useful tool in this respect for future DMAT deployments.”10 Further research into the EMS/hospital interface specific to triage operations is warranted.

Triage is a fast, challenging and unforgiving dance with life and death. Photo courtesy Ron Tencati/Cabarrus County EMS



In order to gauge whether START is effective and consistent, we must look at academic and empirical data in conjunction with personal accounts.

In one study, researchers analyzed data from a train crash involving 163 triaged patients. Interestingly, the jurisdiction where the incident occurred was in the process of staging for an MCI drill when the actual emergency occurred.2

The analysis found that “START ensured acceptable levels of under triage (100% red sensitivity and 89% green specificity) but incorporated a substantial amount of over triage.”2

Essentially, START correctly identified all critical patients as at least “Immediate”; however, some of those identified were actually not that severe. This is a problem since resources will be stretched thin and over-triage, while often encouraged in some EMS systems, could lead to a shortage of resources directed to critical patients.

Over-triage is preferable as compared to under-triage, which could ultimately lead to patient death due to a lack of attention, care and transportation. However, the report demonstrated an acceptable level of under-triage while noting a frequency of over-triage in 79 out of 148 patients (53%).2

The Virginia Tech shootings on April 16, 2007, where a student shot 49 people and killed 32, is an example of appropriately utilizing START at an active shooter/MCI incident.11 A panel report detailing the event found that using START, the initial tactical medics were able to correctly identify expectant patients and overall triage was appropriate.12

This is an example of how, with proper training and implementation, the triage system provides a foundation on which emergency responders are able to accurately sort and treat victims, providing timely care for patients who are salvageable.

However, one problem with implementation at the incident was demonstrated. Interviews with responders indicate that, while some patients were assigned a triage tag, others weren’t, which led to confusion about transported patients.11

The use of triage tags is vital to the successful execution of the system and is dependent on first responders to ensure their placement in order to optimize the system. Mirroring the previously mentioned train incident, “The under-triage rate at Virginia Tech was 10%, with a correspondingly over-triage rate of 69%.”13 Again, the disparity among the triage rates could easily equate to an inequality among resource allocation.

Triage is an unimaginable job, in a chaotic situation, yet someone has to perform the methodical triage process. A personal perspective detailing the inherent difficulty with EMS field triage can be found in the shooting of Rep. Gabrielle Giffords on Jan. 8, 2011, in Tucson, Ariz. In this case, a lone gunman opened fire into a crowd gathered to meet with Giffords.

The ensuing carnage killed six (including a young child) and wounded 13 others. In an interview, paramedic Tony Compagno, one of the first-arriving paramedics on the scene, stated, “Lots of people were laying on the ground … I could tell the [congresswoman] was still alive. People were giving a little girl CPR. My mind went away. I started counting and then I thought, what am I counting, injured or dead?”14

At this incident, the lack of triage tag use and application increased confusion on scene and at receiving facilities. Additionally, patients who were clearly dead and unsalvageable at this fast-moving MCI were triaged as “Red” or “Immediate” due to human emotional attachments. The 9-year-old girl who was receiving CPR on scene should have been classified as “Deceased.” However, Compagno stated, “The little girl, I counted her as ‘Immediate.'”14

Presenting these deviations from START guidelines aren’t meant to criticize or judge the responders, but rather to illustrate how difficult and dynamic the triage process can be, and how the inherent human elements plays a significant role that can’t be quantified with research.

START Deviations

It must be emphasized that the START triage method is a tool meant to provide responders with a framework to begin a difficult process. In addition, adult algorithms are different from pediatrics, particularly as they relate to respiratory rate tolerances and initial resuscitation attempts. Therefore, triage must be looked at as a dynamic process in which a patient’s category may be upgraded or downgraded as their presentation and the scene develops.

Certain pathophysiologies also alter the process. Burns, for example, don’t fit neatly into START categories. A critical burn may initially be classified as “Delayed” if there’s no airway involvement or if the pulse and respirations aren’t out of appropriate ranges.

Mass violence/active shooter instances may also cause need for deviation from standard triage processes as limited patient egress and resources may necessitate early evacuation based on ability to move patients from a location rather than by their triage category.

You must also consider patients with special needs or chronic medical conditions that may alter the presentation of a patient baseline, for example dementia or non-ambulatory status.

Best practice analysis indicates MCI scenes may be initially managed with standard START application, with secondary triage and provider discretion guiding ongoing triage and treatment.


Triage is by nature a dynamic and ambiguous process. MCIs and chaotic rescue scenes don’t lend themselves to the studious collection of data. Scientific evidence is lacking on triage systems as well as alternative strategies. Further study into triage systems would enrich the practice of emergency and disaster response.

Common characteristics at MCIs were found to be pertinent. In every event researched, the tagging of patients who were “walking wounded” was judged to be highly accurate and effective at removing these victims from the triage continuum. EMS personnel also consistently properly identified critical patients with an acceptable level of over-triage.

The difficulty seen was differentiating “Delayed” patients from “Immediate” patients with consistency.2 Failure to physically apply the triage tags was also a recurrent complicating factor.

The critical nature of proper field triage is perhaps best exemplified by Nicholas Senn, MD, who stated, “The fate of the wounded lies in the hands of the ones who apply the first dressing.”15 Consideration must be given to the fact that the human element inherent in this system is ultimately fallible.

Triage is a fast, challenging and unforgiving dance with life and death. Those who have the task will carry the memory of the decisions they were forced to make forever. START is, at best, an imperfect solution to an almost impossible problem.



1. Bledsoe BE, Porter RS, Shade BR. Paramedic emergency care, 3rd edition. Prentice Hall: Upper Saddle River, N.J., 1997.

2. Kahn CA, Schultz CH, Miller KT, et al. Does START triage work? An outcomes assessment after a disaster. Ann Emerg Med. 2009;54(3):424–430.

3. Owens K. EMS triage: Sorting through the maze. Fire Engineering. 2008;161(3):155–162.

4. Cook L. The World Trade Center attack. The paramedic response: An insider’s view. Crit Care. 2001 Dec;5(6):301–303.

5. Cone DC, MacMillan DS. Mass casualty triage systems: A hint of science. Acad Emerg Med. 2005;12(8):739–741.

6. Lerner EB, Schwartz RB, Coule PL, et al. Mass casualty triage: An evaluation of the data and development of a proposed national guideline. Disaster Med Public Health Prep. 2008;2(Suppl 1):S25–S34.

7. Schenker JD, Goldstein S, Braun J, et al. Triage accuracy at a multiple casualty incident disaster drill: The emergency medical service, Fire Department of New York City experience. J Burn Care Res. 2006;27(5):570–575.

8. Pointer JE, Levitt MA, Young JC, et al. Can paramedics using guidelines accurately triage patients? Ann Emerg Med. 2001;38(3):268–277.

9. Zoraster RM, Chidester C, Koenig W. Field triage and patient misdistribution in a mass casualty incident. Prehosp Disaster Med. 2007;22(3):224–229.

10. Richards M, Nufer K. Simple triage and rapid transport: Does it predict transportation and referral needs in patients evaluated by Disaster Medical Assistance Teams? Ann Emerg Med. 2004;44(4 Suppl 1): S33–S34.

11. Perkins T. Virginia Tech mass shooting review panel report. EMS Magazine. 2007;36(10):34–36.

12. TriData Division. Mass shootings at Virginia Tech: Addendum to the report of the review panel. System Planning Corporation: Arlington, Va., pp. 89–105, 2009.

13. Armstrong JH, Frykberg ER. Lessons learned from the response to the Virginia Tech shootings. Disaster Med Public Health Prep. 2007;1(1 Suppl):S7–S8.

14. Burkhart F. (Jan. 15, 2011.) From bloody scene to E.R., Lifesaving choices in Tucson. The New York Times. Retrieved Jan. 27, 2015, from

15. Counter Narcotics and Terrorism Operational Medical Support. (n.d.) Tactical first responder (TFR) course. Retrieved Nov. 9, 2014, from