Pediatric mass-casualty triage adaptation — how the JumpSTART algorithm corrects adult START for children's physiology
Adult mass-casualty triage systems like START (Simple Triage And Rapid Treatment) begin with a single question: "Can you walk to me?" For children, this instruction is unreliable in ways that directly cost lives — infants cannot walk under any circumstances, toddlers may be physically able but too frightened or developmentally unable to follow a verbal command from a stranger, and school-age children may simply be too scared to move even when uninjured. JumpSTART, developed specifically to correct these gaps, changes both the question and what is done with the answer.
START triage, developed in the 1980s for adult industrial and transportation disasters, uses ambulatory status as its very first sorting gate: anyone who can walk to a designated area is tagged GREEN (minor) and no further immediate assessment is performed. This works reasonably well in adults because walking under acute stress correlates fairly well with hemodynamic and neurologic stability.
In children the correlation breaks down for reasons that have nothing to do with injury severity:
• Infants (under ~1 year) cannot walk at baseline — ambulatory status is developmentally meaningless for them • Toddlers (roughly 1–3 years) may be physically capable but will not follow a verbal instruction from an unfamiliar adult in a chaotic, loud environment — stranger anxiety and situational terror override compliance • School-age children who are physically uninjured may freeze, cling to a caregiver's body, or hide rather than walk toward a stranger shouting instructions • Conversely, some injured children (e.g., with early compensated shock) can still walk briefly on adrenaline, masking a RED-level injury if ambulatory status alone is trusted
Lou Romig, a pediatric emergency physician, recognized in the mid-1990s that a straight adaptation of adult START would systematically miscount pediatric casualties. She created JumpSTART (JumpSTART Pediatric MCI Triage Tool) explicitly to preserve rapid triage speed while correcting for developmental and physiologic differences.
JumpSTART was developed by Lou Romig, MD, in 1995 following her observations of pediatric mis-triage in disaster drills, and formally published with validation criteria in 2002. It remains the most widely adopted pediatric-specific MCI triage algorithm in North America.
JumpSTART keeps the ambulatory screen for children developmentally capable of walking — school-age children who can and do walk to the collection point are still tagged GREEN immediately, exactly as in adult START, preserving triage speed for the majority of survivors in most incidents.
But for infants and children who are developmentally non-ambulatory, JumpSTART removes "can they walk" from the decision entirely. These children are physically carried to the secondary triage/treatment area by a rescuer or an accompanying caregiver, and are then run through the full physiologic exam (breathing, respiratory rate, pulse, mental status) exactly like any other non-ambulatory casualty — carrying is a matter of logistics, not a triage verdict.
This distinction matters enormously in practice: a mass-casualty scene with many young children (a school, daycare, or refugee camp) may have a large fraction of the pediatric cohort who are physically uninjured but developmentally unable to self-triage. Treating "cannot walk" as equivalent to "critically injured" would flood the RED category with false positives and slow care for children who are truly critical.
Beyond the ambulatory question, children differ from adults in ways that make every subsequent triage decision point age-sensitive:
• Higher metabolic rate: children consume oxygen roughly 2× faster per kilogram than adults, so hypoxia and physiologic decompensation occur faster and with less warning • Higher body-surface-area-to-mass ratio: a young child has roughly 2–3× the relative skin surface of an adult, causing faster hypothermia in cold/wet disaster environments and faster absorption of topical toxins, chemical agents, or radiologic contamination • Smaller circulating blood volume: an infant carries only ~80 mL/kg of blood (roughly 800 mL total for an 10 kg toddler) — a blood loss that would be trivial in an adult (200–300 mL) can represent 25%+ of total volume in a small child, triggering shock far earlier and with more compensation (normal-looking vitals until sudden collapse) • Compensated shock masking: children maintain blood pressure via tachycardia and vasoconstriction far longer than adults before decompensating — a "stable-looking" injured child can crash abruptly and with less warning than an adult
Every pediatric-specific decision point in JumpSTART — the 5 rescue breaths, the 15–45 RR window, peripheral pulse over capillary refill, and posturing on AVPU — exists specifically to compensate for one or more of these physiologic realities.
Adult START handles an apneic casualty in one move: reposition the airway once; if breathing does not resume, tag BLACK (deceased/expectant) and move on. JumpSTART inserts a deliberate pause here — 5 rescue breaths — before making that same irreversible call, because pediatric cardiac arrest is fundamentally different in origin from adult arrest and far more likely to respond to airway intervention alone.
Adult cardiac arrest is overwhelmingly primary-cardiac in origin — a sudden dysrhythmia (ventricular fibrillation, pulseless VT) in a heart with underlying coronary disease. Once that occurs, brief airway repositioning without breaths or compressions is unlikely to restart circulation, which is why adult START does not spend time on rescue breathing at the scene-triage stage.
Pediatric arrest is different in a way that matters enormously for triage: it is overwhelmingly a respiratory or hypoxic cascade. A child who stops breathing — from airway obstruction (blood, vomitus, a swollen or malpositioned tongue in an unconscious child), smoke inhalation, blast lung, or simple positional airway occlusion — will often have a heart that is still perfusing marginally. Restoring oxygenation with a few rescue breaths can be enough to restart spontaneous breathing and avert what would otherwise become a fatal downward spiral.
This is the single most consequential difference between the two algorithms: applying the adult "one reposition, then BLACK" rule to a child can and does condemn some savable children who would have breathed again after 5 rescue breaths.
JumpSTART's rationale is blunt in its clinical literature: because pediatric cardiopulmonary arrest is usually the end stage of respiratory failure rather than a primary cardiac event, a brief trial of rescue breathing before declaring a child expectant (BLACK) can convert an otherwise-fatal miscall into a RED (immediate, salvageable) tag.
For any non-ambulatory child reaching the breathing check:
1. Check for spontaneous breathing. If present, proceed directly to the respiratory rate step (no rescue breaths needed). 2. If absent, open the airway with a head-tilt/chin-lift (or jaw thrust if trauma is suspected) — this alone resolves many simple obstructions in an unconscious child. 3. Recheck breathing. If breathing resumes after airway positioning alone, tag RED and move on — no rescue breaths required in this branch. 4. If still apneic after repositioning, deliver 5 rescue breaths. 5. Recheck once more. If breathing resumes, tag RED (this child is salvageable and needs urgent ongoing airway management). 6. If still apneic after the 5 breaths, tag BLACK — expectant, given the scene-triage context where resources must go to salvageable casualties first.
This entire sequence adds only seconds per child but changes outcomes meaningfully at the population level in incidents with many pediatric casualties, such as school events or building collapses affecting daycare-age children.
The 5-breath step is not free — in a scene with dozens of casualties, spending 15–30 extra seconds per apneic child is a real allocation of scarce responder time during the highest-acuity phase of the response. JumpSTART's designers accepted this trade-off deliberately: the added time cost is justified by the substantially higher reversibility of pediatric respiratory arrest compared to adult cardiac arrest, meaning the expected number of lives saved per minute spent on this step is higher for children than the equivalent minute would be for a similarly apneic adult.
Training programs for pediatric disaster response emphasize that responders should not skip this step even under extreme time pressure — anecdotal and simulation data consistently show that responders trained only on adult START, when faced with an apneic child, default to the adult one-step rule and under-triage children who would otherwise have been reversible.
Adult START uses one number: if respiratory rate exceeds 30 breaths per minute, tag RED; below that, move to the next check. Applying a single adult threshold to children is a category error — normal pediatric respiratory rates vary enormously by age, and both too-slow and too-fast breathing carry independent danger signals that a single high-side cutoff cannot capture.
True pediatric normal respiratory rates shift continuously with age — a healthy newborn breathes 30–60 times per minute, a toddler 20–40, and an older child or adolescent approaches adult norms of 12–20. Building a scene-triage tool that requires responders to recall age-specific normal ranges for every casualty in a chaotic environment is impractical; JumpSTART's designers instead chose a single pragmatic band — 15 to 45 breaths per minute — that reasonably spans normal-to-mildly-abnormal breathing across the pediatric age range without requiring responders to estimate age precisely or consult a reference chart mid-triage.
The key structural difference from adult START is not just where the numbers sit, but that JumpSTART flags both directions of abnormality. Adult START asks only "is it too fast?" (>30). JumpSTART asks "is it outside the whole normal band?" — capturing dangerously slow breathing (bradypnea, often a pre-terminal sign of exhaustion, rising intracranial pressure, or opioid/toxic exposure) that a one-sided adult threshold would silently pass through as "fine."
Too fast (tachypnea, RR>45): commonly compensatory — the child is trying to maintain oxygenation or blow off excess CO2 in response to hypoxia, chest trauma (pneumothorax, flail segment, pulmonary contusion), fever, pain, or early shock (respiratory rate rises before blood pressure falls in pediatric compensated shock, making it an earlier warning sign than in adults).
Too slow (bradypnea, RR<15): typically a much more ominous sign in a child than in an adult, since children rarely bradypnea from simple fatigue the way adults might — pediatric bradypnea usually reflects exhaustion after a prolonged high-effort period (about to arrest), CNS depression (traumatic brain injury, toxic/opioid exposure), or a pre-arrest state. A slow-breathing child who "looks calm" is frequently closer to respiratory failure than a fast-breathing one.
Because both directions predict deterioration, JumpSTART tags RED for any child whose measured rate falls outside the 15–45 band, sending them straight to the immediate/RED treatment priority rather than continuing down the algorithm.
Respiratory rate is deceptively hard to count accurately in a screaming, crying, or combative child on a chaotic scene — crying itself dramatically elevates apparent rate and can trigger false RED tags if counted during an active cry rather than a calmer breathing cycle. JumpSTART training materials recommend counting for a full 30–60 seconds when possible (rather than extrapolating from a 6- or 10-second count, which pediatric variability makes especially error-prone) and, where feasible, observing chest rise during a brief pause in crying.
Field responders are also trained to distinguish effortful/labored breathing (retractions, nasal flaring, grunting — visible signs of respiratory distress) from the raw numeric rate; a child breathing within the 15–45 window but showing significant visible distress should still be flagged for expedited secondary assessment even though the strict JumpSTART algorithm would pass them to the next check.
Adult START commonly uses capillary refill time (press the nailbed, count seconds to color return) as a fast proxy for perfusion, with >2 seconds indicating inadequate circulation. JumpSTART deliberately swaps this for a palpable peripheral pulse check, because capillary refill is a notoriously unreliable measurement in exactly the conditions a mass-casualty scene guarantees: cold ambient temperature and a frightened, crying child.
Capillary refill time was adopted into triage protocols because it is fast and requires no equipment — press a fingertip or toe, release, count seconds until color returns. In a controlled clinical setting at room temperature, prolonged refill correlates reasonably with hypoperfusion.
A mass-casualty scene reproduces almost none of those controlled conditions. Ambient temperature is frequently cold (outdoor disaster sites, night-time incidents, winter events), and peripheral vasoconstriction from cold alone slows capillary refill in a perfectly well-perfused child, generating false-positive RED tags that would waste scarce resources on children who are not actually in shock. Fear and crying add further peripheral vasoconstriction via catecholamine release, compounding the problem specifically in children, who mount a more pronounced catecholamine-driven vasoconstrictive response to fear than adults do. Ambient lighting (fires, smoke, dusk, artificial lighting) further degrades a responder's ability to accurately judge subtle skin color change.
The net effect: capillary refill in exactly the population and environment where JumpSTART is used tends to systematically over-triage cold, frightened, but adequately perfused children as RED — degrading the entire triage system's ability to correctly prioritize truly critical casualties.
JumpSTART instead asks a binary, less environmentally confounded question: is a peripheral pulse (radial at the wrist, or pedal at the top of the foot for very young children where radial pulses are harder to locate) palpable at all? A palpable peripheral pulse implies a systolic blood pressure sufficient to perfuse distal extremities — roughly correlating with a systolic pressure above ~80–90 mmHg in a child, a threshold well above outright decompensated shock.
The absence of a palpable peripheral pulse is a much less ambiguous finding than a few extra seconds of capillary refill, and does not degrade as easily under cold, fear, or poor lighting — a responder either feels a pulse or does not. Children with no palpable peripheral pulse are tagged RED and moved to the front of the treatment queue, since this reliably reflects significant hypoperfusion or shock.
This swap illustrates a broader JumpSTART design principle: every adapted decision point favors the sign that remains reliable under field disaster conditions specific to children, even when the adult version of that sign is faster or requires less physical contact with the casualty.
Because children have proportionally smaller circulating blood volumes (~80 mL/kg), a hemorrhage that would be minor in an adult can represent a life-threatening fraction of total blood volume in a small child — making an accurate, field-reliable perfusion check disproportionately important in pediatric triage compared to adult triage.
Even a palpable pulse does not guarantee a child is out of danger. Children compensate for blood loss and hypoperfusion far more effectively than adults in the short term — via profound tachycardia and peripheral vasoconstriction — meaning blood pressure and peripheral pulses can remain deceptively normal until a sudden, late collapse once compensatory mechanisms are exhausted. This is why JumpSTART treats the pulse check as one gate among several (breathing, RR, pulse, AVPU) rather than a single definitive verdict, and why rapid reassessment and short reassessment intervals for YELLOW-tagged children are emphasized in JumpSTART training — a child who passes every gate at the moment of triage can still deteriorate within minutes.
The final JumpSTART gate assesses mental status using AVPU (Alert, responds to Voice, responds to Pain, Unresponsive) — but with a critical refinement adult protocols lack: a child who "responds" to a painful stimulus only with abnormal posturing (decorticate or decerebrate flexion/extension) is tagged RED, not treated as a reassuring response. This single refinement, combined with the cumulative effect of every prior pediatric-specific step, is what produces the largest gap between JumpSTART and a naive adult-START application to children.
A naive reading of "responds to pain" sounds reassuring — the casualty is not fully unresponsive. But JumpSTART distinguishes purposeful responses (localizing to pain, withdrawing a limb, crying/verbalizing appropriately) from abnormal posturing: decorticate posturing (arms flexed inward toward the core, legs extended) or decerebrate posturing (arms and legs both rigidly extended), both of which reflect significant brainstem or diffuse cortical injury, not intact neurologic function.
A responder without pediatric-specific training who sees a child's arms move in response to a pinch may reasonably (but incorrectly) read that as "responds to pain, therefore YELLOW" under a generic AVPU application. JumpSTART explicitly overrides that instinct: any child demonstrating inappropriate/abnormal posturing to a painful stimulus is tagged RED, on par with a fully unresponsive child, because posturing indicates a comparable severity of neurologic injury despite superficially "more responsive" appearance.
Beyond posturing, several other adult-START assumptions actively fail when applied to children, compounding across the full algorithm:
• "Obeys commands" is developmentally meaningless below roughly 2–3 years of age — an infant physically cannot obey a verbal command regardless of neurologic status, so an adult rater applying "obeys commands = not RED" to a preverbal child either has to guess or defaults to over-triaging every young child as RED, congesting the highest-priority category with children who are, in fact, neurologically intact • Skipping the 5 rescue breaths converts some savable apneic children (who would have resumed breathing) directly into BLACK/expectant under adult rules — a fatal, non-recoverable mis-tag rather than a resource-allocation inefficiency • The single 30/min RR cutoff misses dangerously slow breathing children entirely (RR<15, e.g. RR of 12) since adult START only checks the high side • Capillary refill in cold/frightened children generates false-positive RED tags, diluting the priority given to truly critical casualties and slowing overall scene throughput
Each of these produces measurable "mismatches" — cases where the two algorithms disagree — and several of them are not just inefficiencies but outright fatal errors (savable child tagged BLACK) or dangerous omissions (deteriorating child tagged GREEN/YELLOW).
Children are disproportionately represented in several categories of real mass-casualty events, making a validated pediatric-specific protocol a matter of population-level consequence, not a theoretical refinement:
• School-site incidents (shootings, building collapses, fires) concentrate large numbers of same-age-range children in one location, with responders often trained primarily on adult protocols • The 2004 Beslan school siege (Russia) remains one of the starkest examples: of roughly 334 people killed, 186 were children — a hostage-taking and rescue operation that overwhelmed responders with an overwhelmingly pediatric casualty population in a single building • Natural disasters and armed conflict disproportionately affect displaced and refugee populations, where UNHCR estimates children make up roughly 40% of the forcibly displaced population worldwide — meaning disaster medicine in refugee camp and post-disaster settings routinely triages majority-pediatric casualty groups • Daycare centers, pediatric hospitals, and school buses represent scene types with zero adult casualties, where a responder without pediatric triage training has no adult patients to "recalibrate" against and must rely entirely on correctly applying pediatric-specific criteria
JumpSTART, along with related tools (SALT, Pediatric Triage Tape), is now incorporated into disaster medicine curricula precisely because these scenarios are not rare edge cases — they are a predictable and recurring category of mass-casualty incident.
In the 2004 Beslan school siege, 186 of the roughly 334 people killed were children — a single incident illustrating why disaster medicine systems cannot treat pediatric-specific triage protocols as optional or secondary to adult-oriented mass-casualty planning.
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| Ambulatory screen | "Can you walk to me?" — sole gate to GREEN | Same for developmentally ambulatory children; non-ambulatory (infants/toddlers) are carried and screened by physiology instead of being defaulted to a tag | Prevents infants/toddlers from being mis-classified purely for being unable to walk |
| Apnea response | One airway reposition; no breath → BLACK | Reposition airway, then 5 rescue breaths before declaring BLACK — accounts for respiratory-origin pediatric arrest | Rescues children whose arrest is reversible with brief ventilation support |
| Respiratory rate | Single cutoff: RR > 30/min → RED | Normal band 15–45/min for all pediatric ages; RED if outside band in either direction | Catches dangerous bradypnea (RR<15) that a one-sided adult cutoff would miss entirely |
| Perfusion check | Capillary refill time, >2 sec → RED | Palpable peripheral (radial/pedal) pulse; absent → RED | Avoids false positives from cold- and fear-induced vasoconstriction common in children |
| Mental status | "Obeys commands" → not RED | AVPU with explicit posturing check; abnormal posturing (P) or unresponsive (U) → RED even if "responsive" | Correctly flags significant neurologic injury masked by a superficial pain response; valid for preverbal children |