Major Incident Triage — Sieve-to-Sort Sequencing, P1 Sensitivity Fences and Tactical Surgical Doctrine

The major incident is the moment casualties outnumber care. Triage is the process of prioritising patients on the basis of their clinical acuity, and the UK doctrine states its purpose plainly: a system of triage for casualties to prioritise treatment and ensure "the greatest good for the greatest number".^[3][7] The surgical version is stricter — a logical classification of victims according to severity of injury, the necessity of treatment and the degree of urgency, done early, done dynamically, and re-done as the scene evolves.^[11]

The numbers behind that doctrine are humbling. Across 195,709 UK trauma-registry patients only 11.3% met intervention-based P1 criteria — yet they died at 12.8% against 5.0%, needed intensive care at 52.4% against 5.0%, and carried a median ISS of 21 against 9.^[1] The best primary tool finds 70.4% of them while over-triaging 70.9%; the secondary tool the UK still teaches finds 15.7%.^[1][3] Three traps decide the viva: the sieve that is fast but insensitive, the Sort label that reassures falsely, and the index operation that tries to do too much while the corridor fills behind it.

Thirty casualties from a rail bombing, four resuscitation bays, two theatres, one surgeon on the floor. Walking wounded stream past, a silent teenager sits grey among the shouting, and the first ambulance carries a man with a belly full of blood. The examiner wants the primary sieve with its numbers, what the secondary Sort can and cannot do, which child gets which tool, and what the index operation is allowed to contain. This page answers each with the number from the paper beside it.^[6][3][12]

Overview & Definition — declare early, sort by severity-necessity-urgency

A major incident is declared when casualties threaten to overwhelm the care available — and triage should be early, dynamic, and lead to a categorization that optimally utilizes resources while ensuring efficient management.^[11] Medical and surgical triage allows a logical classification of victims according to severity of injury, the necessity of treatment and the degree of urgency — severity tells you who is dying, necessity tells you who benefits, urgency tells you who goes first.^[11]

Two category ladders run in parallel and the candidate must speak both. The UK registry ladder assigns P1, P2, P3, Expectant or Dead based on pre-defined, intervention-based criteria — P1 is the patient who needs the life-saving intervention, proven by 19.5% of a 127,233-patient cohort receiving at least one LSI as priority 1.^[1][3] The US SALT ladder sorts to T1/red/immediate, T2/yellow/delayed, T3/green/minimal, T4/black/dead — designed so that 16 of 52 drill victims were T1, 12 T2, 14 T3 and 10 T4.^[21] The registry-outcome ladder behind the American head-to-heads reads minor/green, delayed/yellow, immediate/red, dead/black — 60/5/29/6 across 100 charts.^[24]

Epidemiology & Risk — bombings, registries, and the 11% who matter

The modern civilian series is Mumbai: the attack lasted for 60 h with bullet and blast injuries; 271 casualties were encountered, 108 dead at admission, 163 triaged with 23 managed as outpatients; 140 admitted, 194 operations in 127 patients, six postoperative deaths.^[14] London's scale marker is the 7th July 2005 transport bombings — the largest number of casualties on mainland UK soil since World War 2 — with 203 Royal London records, outcome data for 166, and 8 critically injured.^[7] England's steady state is five terrorism-related major incidents in 2017 alone, approximately 40 fatalities and 400 injured.^[5]

The registry reframes the risk: 195,709 patients, 7.0% mortality, median ISS 9, 97.1% blunt — and the 11.3% who met P1 criteria carried nearly all the death, intensive care and injury burden.^[1] The second registry is older and sicker: 218,985 adults, 19.5% priority one, 70% male, median age 51, ISS 16.^[4] Children concentrate differently again: 31,292 paediatric cases, 3.3% died, 21.9% major trauma, nearly half aged 8 or under.^[19]

Pathophysiology — two prices: the missed P1 and the flooded bay

Insufficient triage kills two ways: the under-triaged P1 waits while the over-triaged P3 consumes the bay — and insufficient triage may lead to increased morbidity and mortality due to delayed evacuation and treatment of the most critically injured patients.^[15] The under-triage price is measured in deaths: the best-understood under-triaged population died at 5.7%, with thorax and head injuries the ones the sieve leaves behind.^[4] The over-triage price is measured in access: the best primary tool over-triages 70.9%, the NARU sieve 56.4%, and the most sensitive paediatric tool 75.0% — every false P1 is a true P1 delayed.^[1][18]

Berlin priced the hospital end: the primary triage category was allocated correctly to 61% of simulated severity; 24% overtriage and 16% undertriage were observed; 18% of life-threatening injuries were undertriaged — and of the 62% with secondary right allocation, re-triage was only used in 4%.^[26] The train-crash audit priced the field end: 148 records, field labels 22 red / 68 yellow / 58 green against outcome-based 2 / 26 / 120 — 79 overtriaged, 3 undertriaged, 66 matched.^[25] Overall the in-hospital verdict stands: the accuracy of in-hospital triage is low at 61%.^[26]

Presentation & the Primary-Secondary Sort — sieve at the scene, sort at the door

Primary triage happens at the scene in seconds. The UK Triage Sieve advocated by the MIMMS course categorises casualties by ability to walk, respiratory rate and heart rate or capillary refill time; the military version adds assessment of consciousness.^[6] In 1,657 military cases the military sieve was significantly more sensitive than the civilian sieve (59% versus 53%) with similar specificity (89 versus 88) — yet optimising every physiological cut-off lifted sensitivity only to 71%, so a purely physiological sieve may never reach an acceptable level.^[6] The MPTT-24 is likely the optimum physiological method for primary triage — but it needs its secondary process beside it, because no sieve finishes the job.^[5]

The American primary is SALT: sort, assess, life-saving interventions, treatment and/or transport — developed from the best available science and consensus opinion as a single overarching guide for unifying mass-casualty triage across the United States.^[20] It works after brief teaching: immediately following a 30-minute didactic session, trainees triaged an eight-victim mock incident — and accuracy, error patterns and triage time were similar between paramedic and fire students using SALT.^[29] The pilot numbers set the baseline: two triage officers applied SALT correctly to 41 of 52 patients (78.8%), with 13.5% overtriage and 3.8% undertriage.^[21]

Secondary triage allows a more detailed assessment of the patient — in the UK the Triage Sort is the preferred method, combining GCS, systolic BP and RR to categorise Priority 1 casualties — and the shock index (HR divided by SBP) is its challenger.^[16] Across 345 Camp Bastion traumas, 58.8% gold-standard P1, the Sort predicted LSI need at 58.6% sensitivity and 88.7% specificity — while a shock index above 0.75 reached 70.0% sensitivity at 74.7% specificity.^[16] That comparison carries its boundary condition in the same paper: whether it translates to civilian practice needs its own examination.^[16]

Scoring & Classification — quote sensitivity with its over-triage shadow

Classify by measured performance, never by habit. In 16-to-64-year-olds the BCD Sieve predicted P1 at 70.4% sensitivity with 70.9% over-triage; the NARU Sieve managed 44.9% sensitivity with 56.4% over-triage — and all tools performed poorly past 65, which is why the BCD Sieve should supersede the NARU Sieve as the UK primary tool.^[1] The machine-learning replication agrees on the order: BCD was the best existing tool at 68.2% sensitivity (AUC 0.688); inability to breathe spontaneously, chest injury and mental status predicted P1 best; the three-variable decision-tree primary reached 73.0% (AUC 0.782); the four-variable portable secondary reached 77.9% (AUC 0.817) and 97.6% (AUC 0.778) in 5,956 military validations.^[2]

Quote the secondary failure exactly. The Triage Sort had the lowest accuracy of all tools at 15.7% sensitivity with 84.3% undertriage — but the greatest specificity at 98.7% — so within a civilian registry its use as a secondary tool should be reviewed and the optimum secondary method remains an open research question.^[3] The MPTT is the counterweight: the lowest under-triage rate at 42.4%, with its under-triaged population dying at 5.7% and fewer serious thorax and head injuries missed than both the NARU and MIMMS sieves.^[4] Alberta confirms the ranking from the other side of the Atlantic: across 8,652 records MPTT was most sensitive at 0.76 while four of seven tools sat below 0.45 — and every tool should be employed with caution because all miss a large proportion of patients needing urgent life-saving interventions.^[17]

Quote START honestly on both sides of the ocean. The 32-study meta-analysis gives START 0.73 accuracy with 0.14 over-triage and 0.10 under-triage — and concludes START is not accurate enough to serve as a reliable disaster triage tool.^[9] The 13-study review agrees: every system sat below 90% on accuracy, sensitivity and specificity except Smart above 90%.^[28] The register meta-analysis explains why: the Triage Sieve significantly underperformed START and CareFlight on diagnostic odds (19.85 versus 13.23; 23.72 versus 12.83), with higher specificity but lower sensitivity — Sieve 96.1% specific against START 93.6%, but Sieve 34.8% sensitive against START 57.8%.^[8]

Quote SALT against START with all four head-to-heads. The ED study of 125 adults gives SALT 52% accuracy against 36–37% for START, CareFlight and Sieve, with SALT under-triaging 26% against 57–58% — at the price of the highest over-triage, 22% against 7%.^[22] The simulation study gives SALT r 0.860 and κ 0.632 with 100% agreement on black and green, undertriage 9% against START's 20% and field triage's 37% — SALT overall more accurate, specifically in delayed and immediate categories.^[23] The 100-chart study dissents and must be quoted: SALT 5 over / 30 under / 65 correct against START 12 / 33 / 55 — neither sensitive nor specific for predicting clinical outcome.^[24] The only real-disaster outcomes study dissents harder: no level met the 90%-and-90% bar, though red was 100% sensitive and green 89.3% specific.^[25] The 7th July validation reconciles them at small numbers: all three systems found the same three patients, 50% sensitive and 100% specific at the top priority, 75% and 99% across the top two.^[7]

Secondary Triage & Hospital Reception — the door is a second sieve, not a destination

Treat the hospital door as a second triage event, because the field label degrades in transit: primary category was correct in only 61% of Berlin's 601 training casualties, and re-triage was used in only 4% of those re-allocated — re-triage every arrival.^[26] The receiving hospital must physically become a triage machine: Mumbai converted the casualty receiving area into a triage zone with patients streamed to stations, and rotated personnel every 8 hours.^[14] The evidence behind hospital triage is thin — six MEDLINE studies across income settings, real-world application insufficiently studied — so run the local plan as the intervention and exercise it.^[27][10]

Surgical Sequencing — TASC serves the situation, damage control serves the patient

Expect the first arrivals to be the worst: the first patients taken to medical facilities present with uncontrollable bleeding from trunk and body-cavity injuries.^[12] For patients in extremis the index operation has exactly two aims — stop the bleeding and control the contamination — because unlike damage control surgery, which is tailored to the patient's condition, tactical abbreviated surgical care is first and foremost adapted to the overall situation.^[12] The three lawful options after that are early total care, damage control surgery, or tactical abbreviated surgical care — chosen against the number and distribution of arrivals, injury patterns, infrastructure and personnel, which is why the TDSC course trains clinical decision-makers rather than technicians.^[13]

Mumbai shows the throughput this doctrine buys: 140 admissions, 194 operations in 127 patients, six postoperative deaths — with personnel rotated every 8 hours to hold the pace.^[14] Resuscitation ratios, transfusion clocks and damage-control physiology are cited from the damage-control resuscitation and massive-transfusion topics beside this one — here only the triage priority and the abbreviated-operation boundary.^[12][13]

Complications, Thresholds & Special Groups — children need their own tool, the elderly defeat all of them

Children are 17.6% LSI among 15,133 UK paediatric traumas — and the adult tools misread them: the Sheffield paediatric tool reached 92.2% sensitivity at 75.0% over-triage, while the Paediatric Triage Tape managed 34.1% and JumpSTART 45.0%.^[18] The five-tool study agrees with finer grain: CareFlight predicted death best at 95.3% sensitivity and 80.4% specificity, while JumpSTART was superior under age 8 at 86.3% and 84.8%.^[19] Alberta adds the warning label: JumpSTART had the lowest sensitivity and the highest under-triage rate for paediatric patients.^[17] Counsel it straight: sieve the child with the paediatric tool, accept the over-triage as the price of the 92.2%, and never defend a JumpSTART negative in an under-8 without re-triage.^[18][19]

The elderly invert the bargain: all tools performed poorly amongst patients past 65 — so undertriage in the older casualty is assumed, never argued away.^[1] The MPTT analysis shows where the cost concentrates: serious thorax and head injuries under-triaged — fewer missed than with either sieve, but missed still at 42.4% under-triage overall.^[4]

Evidence, Guidelines & Regional Differences — who proved what

No single body owns this topic — no triage algorithm can be scientifically proven superior in all aspects, across 25 studies from 2009 to 2021 — which is why the German guideline's strong-consensus recommendation is procedural, not algorithmic: in-house exercises or virtual training with verified systems, and regular hospital-plan exercises.^[10] SALT owns the American consensus (best science plus opinion into one unifying national guide) with a pilot at 78.8% correct, 13.5% over and 3.8% under — teachable in 30 minutes to paramedics and firefighters alike.^[20][21][29] The BCD Sieve owns the UK registry argument (195,709 patients; best in class; supersede NARU) with MPTT as the validated alternative (lowest undertriage, lowest undertriaged mortality) and MPTT-24 as the likely optimum primary awaiting its secondary.^[1][4][5] START owns the global installed base and the hardest audit: 0.73 accuracy in meta-analysis, 79-over/3-under in its only real-disaster outcomes study, identical-to-rivals at 7th July small numbers.^[9][25][7] Live-simulation evidence owns the humility footnote: 15 studies testing 6 of 41 methods, author-defined standards, unclear-to-high bias throughout.^[15]

Exam Pearls — the one-liners that score

• Triage early and dynamic by severity, necessity and urgency — greatest good for the greatest number.^[11][7]
• P1 is 11.3% of the registry but carries 12.8% mortality and 52.4% ICU need against 5.0% and 5.0%.^[1]
• BCD Sieve 70.4% sensitive supersedes NARU 44.9% as the UK primary tool.^[1]
• MPTT under-triages least at 42.4% and its missed die at 5.7% — the validated alternative.^[4]
• Triage Sort finds 15.7% of LSI need with 84.3% undertriage — review its secondary use.^[3]
• Shock index above 0.75 beats the Sort at 70.0% against 58.6% — military proof, civilian translation pending.^[16]
• START runs 0.73 accurate with 0.14 over and 0.10 under — not reliable enough alone.^[9]
• SALT beats START on accuracy 52% against 36% and undertriage 26% against 57% — priced in 22% over-triage.^[22]
• SALT correlates r 0.860, κ 0.632, with 100% black-green agreement — undertriage 9% against 20%.^[23]
• The 100-chart dissent: neither SALT nor START predicts outcome sensitively — quote it before the examiner does.^[24]
• Mumbai: 271 encountered, 108 dead at admission, 163 triaged, 194 operations in 127, six deaths.^[14]
• Index surgery stops bleeding and controls contamination — TASC serves the situation, damage control the patient.^[12]
• Children: SPTT 92.2% sensitive at 75% over; tape 34.1%, JumpSTART 45.0%; under-8 JumpSTART 86.3/84.8.^[18][19]
• Elderly defeat every tool past 65 — assume undertriage, never argue it away.^[1]
• Berlin door: 61% correct, 18% of life-threats missed, re-triage in 4% — re-triage every arrival.^[26]

31,292 cases, 3.3% died, 21.9% major, 47% aged 8 or under?.^[19]