Simple Triage And Rapid Treatment — sorting mass-casualty patients into Red / Yellow / Green / Black in under 60 seconds each
The Simple Triage And Rapid Treatment (START) algorithm was developed in 1983 at Hoag Hospital and the Newport Beach Fire Department in Newport Beach, California, to answer one brutal problem: when the number of casualties overwhelms the number of responders, how do you decide — in seconds, not minutes — who is treated first? START gives every first responder the same decision tree so triage stays consistent across dozens of people working a scene at once.
In everyday emergency medicine, the sickest patient gets the most attention. Mass casualty incidents (MCIs) invert that logic. When casualties outnumber responders, spending 20 minutes on one unsalvageable patient can cost several salvageable lives elsewhere. Triage reframes the ethical question from "how do I save this patient" to "how do I save the most patients possible with the resources on scene right now."
START was designed to be taught in a single in-service training session and executed reliably by firefighters, EMTs, police officers, and even trained bystanders — not just physicians. That simplicity is the entire point: a mnemonic-driven decision tree that produces the same tag regardless of who is holding the clipboard.
The very first action at a START scene is a loud, scene-wide announcement: "If you can hear my voice and you are able to walk, please stand up and walk to [landmark] now." Patients who get up and walk are tagged GREEN (Minor) on the spot — no further exam needed at this stage — and are escorted to a collection point away from hazards.
This single instruction typically removes half or more of all casualties from the "unknown" pool in under a minute, without a single responder having to touch a patient. It also indirectly screens for airway, breathing, circulation, and mental status all at once: a person who can stand, balance, and walk purposefully across a chaotic scene has, by definition, adequate perfusion and mental status. Everyone who does NOT get up becomes the priority pool for the RPM assessment carried out in the next three stages.
Children are not small adults: normal pediatric respiratory rates are faster, and children more often suffer respiratory rather than primary cardiac arrest. In 1995, pediatrician Lou Romig, MD developed JumpSTART, a modified algorithm for patients who appear to be under about 8 years old. It adjusts the respiratory rate thresholds to 15–45 breaths/min, adds a step where an apneic child with a palpable pulse receives five rescue breaths before being declared BLACK, and swaps the "follow commands" mental-status step for the AVPU scale (Alert / Voice / Pain / Unresponsive), since young children often cannot follow verbal commands even when neurologically intact.
Everyone who did not get up and walk now enters the RPM assessment: Respiration, Perfusion, Mental status — three checks, each with a single binary branch point, executed in that fixed order. Respiration comes first because an airway problem kills fastest and is often reversible with a single maneuver that takes only seconds.
For each non-ambulatory patient: is the patient breathing? If NO — open the airway once with a head-tilt/chin-lift (or jaw-thrust if spinal injury is suspected) and reassess immediately. If breathing resumes, tag RED (an airway that had to be manually opened is inherently unstable and time-critical). If the patient is still apneic after one reposition attempt, tag BLACK — expectant/deceased — and move on.
If the patient IS already breathing on first check, count the respiratory rate for roughly 15 seconds. A rate greater than 30 breaths per minute (tachypnea) indicates the body is compensating for hypoxia, shock, or chest injury — tag RED immediately without proceeding to perfusion or mental status. A rate of 30 or below allows the patient to proceed to the perfusion check in Stage 3.
This is the most ethically difficult and most frequently misunderstood step in START. In routine EMS, an apneic pulseless patient triggers CPR. In an MCI with a true resource shortfall, START explicitly withholds CPR at the initial triage pass: a single rescuer spending 8–10+ minutes on a resuscitation with a low probability of success can mean two, three, or more salvageable RED patients elsewhere deteriorate and die without ever being reached.
This is a utilitarian allocation rule, not a clinical judgment that the patient cannot be saved under any circumstance — it is a judgment that, given the responder-to-casualty ratio in front of them right now, resuscitation attempts do not maximize total lives saved. BLACK patients are reassessed as more responders and resources arrive and the RED/YELLOW backlog clears.
A blocked airway from an unconscious patient's relaxed tongue, blood, or debris is one of the most common and most reversible causes of preventable death at a trauma scene. Studies of battlefield and civilian trauma mortality consistently rank airway obstruction among the leading causes of death that could have been prevented with a maneuver taking under 10 seconds. That is precisely why respiration — not perfusion, not mental status — is checked first in RPM: it is the fastest-killing, fastest-fixing problem on the list.
START allows exactly one airway reposition attempt per patient at the initial triage pass. This single rule — check breathing, reposition once, reassess — routinely converts a fraction of apparently BLACK patients into RED (salvageable) without costing the responder more than a few extra seconds per patient.
Patients who are breathing at 30/min or slower move on to the second branch of RPM: perfusion. The question is simple — is enough blood reaching the periphery to sustain life in the next few minutes? The original 1983 protocol used capillary refill; most modern EMS systems have shifted to the radial pulse check because it is far more reliable outside a warm, well-lit hospital room.
Capillary refill (pressing a nail bed and timing the color return) was the original 1983 START criterion: refill of 2 seconds or longer indicated inadequate perfusion. In practice, cold ambient temperature, dim or flashing emergency lighting, dark nail polish, and skin pigmentation all degrade the reliability of a visual color-return test performed by a stressed responder in the first minutes of a chaotic scene.
Because of this, most current START training instead — or additionally — checks for a palpable radial pulse at the wrist. A palpable radial pulse roughly correlates with a systolic blood pressure of at least 80–90 mmHg, a reasonable proxy for perfusion adequate to reach the brain and other vital organs in the short term. No palpable radial pulse (or capillary refill ≥2 seconds where that criterion is used) means the patient is tagged RED immediately.
A patient tagged RED for absent perfusion is not simply labeled and left — START directs the responder to apply immediate hemorrhage control before continuing to the next patient: direct pressure, a hemostatic dressing, or a tourniquet on a extremity wound with life-threatening bleeding. This is the one "treatment" folded into the otherwise rapid-sort algorithm, because uncontrolled hemorrhage is both immediately reversible and rapidly fatal — exactly the profile of intervention worth the few seconds it costs during initial triage.
Newer national frameworks such as SALT triage push this idea further, moving lifesaving interventions (hemorrhage control, airway opening, chest decompression, antidote auto-injectors) to occur during the very first pass across all patients, before categorization is even finalized.
Hemorrhagic shock is commonly staged in four classes. Class I (up to ~15% blood volume lost) produces few outward signs. Class II (15–30%) begins to narrow pulse pressure. Class III (30–40%) is where systolic pressure starts to fall and the radial pulse becomes difficult or impossible to palpate — the threshold START is built to catch. Class IV (>40%) is immediately life-threatening. Because compensatory mechanisms (tachycardia, vasoconstriction) can mask early shock, a responder cannot rely on "the patient looks okay" — the radial pulse check is a fast, hands-on physiological cutoff that catches decompensation before it becomes obvious.
The final branch of RPM applies only to patients who are breathing at 30/min or less AND have a palpable pulse. START asks one more question: can this patient follow a simple command? The whole three-step decision tree is often taught with a single mnemonic — "30-2-Can Do" — referencing the respiratory rate cutoff, the capillary refill cutoff in seconds, and this final compliance check.
A patient can be breathing calmly at 18/min with a strong radial pulse and still be in serious danger if they cannot follow a simple command such as "squeeze my two fingers" or "open and close your eyes." Altered mental status in this setting usually reflects inadequate cerebral perfusion, significant head injury, or another life-threatening process that has not yet manifested in the respiratory or pulse exam. Because the brain is exquisitely sensitive to falling oxygen or blood pressure, an inability to follow commands is treated as an immediate red flag regardless of an otherwise reassuring RPM result up to that point — the patient is tagged RED.
Patients who ARE able to follow the command are tagged YELLOW (Delayed): seriously injured and requiring treatment, but not showing signs of imminent airway, breathing, circulation, or neurological collapse. YELLOW patients are re-triaged periodically, since a delayed patient can deteriorate into a RED one while waiting for transport.
By the end of this stage, every patient on scene has one of four tags. GREEN patients self-identified in Stage 1 by walking. BLACK and a subset of RED were identified by the respiration check in Stage 2. Another subset of RED came from the perfusion check in Stage 3. The remainder split into RED and YELLOW here, based purely on the ability to follow a command. This is the moment triage command gets a full, if provisional, picture of the incident's severity mix — the number that determines how many ambulances, helicopters, and receiving hospital beds need to be mobilized in Stage 5.
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| START | |||
| JumpSTART | |||
| SALT |
Triage tags are only useful if they drive action. The final stage of an MCI response is the transport officer's job: sequence patients onto ambulances and helicopters strictly by priority — RED first, then YELLOW, then GREEN — while distributing the load across multiple hospitals so no single emergency department is overwhelmed. BLACK patients are moved to a holding area, not to a hospital.
As patients reach the treatment area or ambulance loading zone, a more detailed secondary triage (vital signs, brief exam) often refines the initial START tag before transport. A transport officer tracks which hospitals have accepted how many patients and of what severity, deliberately spreading RED patients across several trauma centers rather than sending them all to the closest one — a single nearby hospital overwhelmed with every critical patient can begin failing the very people triage was meant to save.
Every patient leaves the scene with a triage tag (a color-coded card such as a METTAG) documenting the assigned category, time of assessment, and any interventions performed — this travels with the patient to the receiving hospital and becomes part of the incident's medical and legal record.
Two error types matter in triage quality review. Under-triage means labeling a critically injured patient as less severe than they are — this is the dangerous error, since it can delay life-saving care to someone who needed it immediately; the American College of Surgeons Committee on Trauma (ACS-COT) benchmark targets an under-triage rate below 5%. Over-triage means labeling a stable patient as more severe than they are — costly and resource-straining, but rarely directly fatal. Research on START in real incidents and simulations has repeatedly found it skews toward substantial over-triage, commonly in the 30–50% range, reflecting the algorithm's deliberate conservative bias: when in doubt, tag it RED.
START has been criticized on several fronts. Inter-rater reliability studies show meaningful disagreement between responders triaging the same patient, particularly at the RED/YELLOW boundary. It was designed primarily around blunt trauma physiology and adapts less cleanly to blast injuries, burns, or chemical/radiological exposure, where injury patterns do not map neatly onto the RPM tree. It also does not explicitly address ongoing hemorrhage control as thoroughly as newer frameworks — a gap SALT triage was partly designed to close. And START gives no formal weight to a patient's trajectory over time; a single point-in-time tag can go stale in a long-duration incident, which is why continuous re-triage of YELLOW and GREEN patients is considered essential rather than optional.
Large-scale incidents such as the 2013 Boston Marathon bombing (over 260 injured) and the 2017 Las Vegas mass shooting (hundreds of gunshot and stampede injuries in minutes) are frequently studied as case examples of triage and transport systems under extreme, sudden load. After-action reviews of well-run responses to such events consistently credit rapid, disciplined field triage — getting the right patient to the right hospital in the right order — as a major factor in unusually low rates of preventable death among patients who reached definitive care alive.
The entire purpose of a four-color, single-pass algorithm like START is speed under overwhelming demand: a rate of roughly 60 seconds per patient is what allows a handful of first-arriving responders to sort dozens of casualties before the first ambulance even reaches the scene.