Recognition of deterioration: track-and-trigger and early-warning scores

The recognition of deterioration is the single most improvable step in the chain of survival for the acutely ill patient. In-hospital cardiac arrest is rarely a sudden, unpredictable event: it is usually preceded by measurable physiological deterioration over hours, much of it documented in the observation chart before the arrest. The premise of track-and-trigger systems is that this deterioration is detectable, that it follows recognisable patterns, and that a standardised, graded response to it prevents avoidable arrests and admissions to critical care. For the emergency physician the same principle governs the front door: the patient who arrives abnormal, or who trends abnormal while waiting, is the patient about to deteriorate, and a structured early-warning assessment converts that signal into an action. [1]

The antecedents of arrest

Studies of in-hospital cardiac arrest consistently show antecedent physiological abnormality in the hours before the event, and systematic review confirms that physiological track-and-trigger warning systems reliably identify at-risk patients on the ward.[1] The observations most commonly deranged before arrest are the respiratory rate, the heart rate, the blood pressure, the conscious level and the temperature — exactly the parameters aggregated in the modern early-warning scores. Critically, the abnormality is often present and recorded but not acted upon, which frames deterioration as a failure of the response system rather than of detection alone.

The intervention works. Introducing a Modified Early Warning Score into acute medical admissions was associated with a reduction in cardiopulmonary arrests and in intensive care utilisation, evidence that earlier, structured recognition changes outcome.[2] Aggregate scoring is also more consistent than unstructured judgement: track-and-trigger systems are reproducible across observers and shifts, which is what allows them to drive a standard, protocolised escalation.[3]

The physiology of deterioration and the order of the signs

Understanding which parameter deranges first, and why, makes the scores intelligible rather than a rote table. The body responds to a falling cardiac output or rising metabolic demand through sympathetic compensation: tachycardia and tachypnoea to preserve oxygen delivery, and vasoconstriction to preserve blood pressure. Compensation holds the measurable end-points — blood pressure and conscious level — until reserve is exhausted, after which decompensation is abrupt. [1]

The clinical consequence is the order in which the observations change. The respiratory rate is usually the first and most sensitive marker — it rises in metabolic acidosis (the compensatory hyperventilation of shock, sepsis and salicylate toxicity), in hypoxaemia, and in respiratory and neurological deterioration — and it is the most frequently omitted or estimated vital sign. The heart rate and the temperature follow; the blood pressure and the conscious level are late signs, preserved by compensation until the patient is on the cusp of collapse. A patient who looks well but has a rising respiratory rate and heart rate is often compensating, and the aggregate score exists to force an action at exactly that compensated stage, before decompensation. A normal blood pressure must never reassure the clinician into inaction when the trend and the compensatory signs are abnormal. [1]

Track-and-trigger systems: the three types

Track-and-trigger systems fall into three broad designs, distinguished in the systematic review literature.[1] A single-parameter system calls for review when any one of a set of observations breaches a threshold — simple, but blunt. A multiple-parameter system requires several derangements before triggering, trading sensitivity for specificity. An aggregate-weighted scoring system — the design of MEWS, NEWS and NEWS2 — assigns each parameter a weighted score and sums them into a single value that drives a graded response. The aggregate-weighted design is preferred because it captures severity across systems, it is reproducible, and it generates a continuous score rather than a binary alarm, allowing a proportionate escalation.

The aggregate early-warning score: from MEWS to NEWS2

The Modified Early Warning Score introduced the principle that a small number of routine observations, weighted and summed, predict deterioration.[2] Its successor, the Royal College of Physicians' National Early Warning Score, refined the weighting and standardised the parameters, and the current version, NEWS2, added a second oxygen-saturation scale for the patient at risk of hypercapnic respiratory failure and replaced the AVPU with the ACVPU conscious level (to capture new confusion as a marker of deterioration).[6] NEWS2 scores seven parameters — respiratory rate, oxygen saturation, supplemental oxygen, systolic blood pressure, pulse rate, conscious level and temperature — into an aggregate that drives a defined, graded clinical response.[1] Its performance has been validated across cohorts, including in patients with COVID-19, in whom it retained discriminative value for deterioration.[5]

NEWS2 in detail: parameters, bands and escalation

The NEWS2 parameters and their scoring bands are reproduced below, and a Fellowship candidate should be able to reproduce them. Each parameter is scored 0 to 3; the higher the deviation from the normal range, the higher the score. [1]

Respiratory rate (per minute): a rate of 12 to 20 scores 0; 9 to 11 scores 1; 21 to 24 scores 2; and a rate of 8 or fewer, or 25 or more, scores 3. Oxygen saturation (scale 1, the default): 96 per cent or above scores 0; 94 to 95 per cent scores 1; 92 to 93 per cent scores 2; and 91 per cent or below scores 3. For the patient with hypercapnic respiratory failure, scale 2 is used, which permits a lower target range (88 to 92 per cent) so that oxygen is not over-prescribed. Supplemental oxygen: air scores 0 and any supplemental oxygen scores 2, because the need for oxygen is itself a marker of severity. Systolic blood pressure (mmHg): 111 to 219 scores 0; 101 to 110 scores 1; 91 to 100 scores 2; and 90 or below, or 220 or above, scores 3. Pulse rate (per minute): 51 to 90 scores 0; 41 to 50 or 91 to 110 scores 1; 111 to 130 scores 2; and 40 or below, or 131 or above, scores 3. Consciousness (ACVPU): Alert scores 0; new Confusion, Voice, Pain or Unresponsive scores 3 — the addition of new confusion to the older AVPU captures an important, easily-missed sign. Temperature (degrees Celsius): 36.1 to 38.0 scores 0; 35.1 to 36.0 or 38.1 to 39.0 scores 1; 39.1 or above scores 2; and 35.0 or below scores 3.[6][1]

The aggregate score drives a graded escalation. A total of 0 is low-risk and suitable for routine monitoring at a minimum frequency of 12-hourly. A total of 1 to 4 is low-medium risk and prompts review by a registered nurse with an increased monitoring frequency. A single parameter scoring 3 is low-medium risk but mandates an urgent ward-based review by a competent clinician. A total of 5 or 6 is the key threshold — it prompts urgent review by a clinician with core competencies in assessing acute illness and a consideration of whether a critical-care referral is needed, and it is the score at which escalation is clearly indicated. A total of 7 or more is a clinical emergency demanding an immediate response, usually a rapid assessment by a critical-care team and consideration of transfer to a higher level of care.[6][1]

MEWS in detail: the original aggregate-weighted score

The Modified Early Warning Score, derived from the work of Morgan, Subbe and colleagues, demonstrated that a weighted sum of routine bedside observations predicts the deteriorating ward patient.[2] Its parameters and bands differ from NEWS2 in their detail, and a Fellowship candidate should know both the principle and the points at which MEWS and NEWS2 diverge. A MEWS of 4 or more is the classical trigger for urgent review; a MEWS of 5 or more carries a sharply higher risk of death, unanticipated intensive care admission and cardiac arrest.

NEWS2 at the bedside — the worked thresholds

The aggregate score drives a graded escalation, but the candidate must be able to apply it to a case, not merely recite the bands. The thresholds and their mandated responses are the single most testable element of this topic. [1]

Afferent and efferent limb failure

A track-and-trigger system has two limbs, and either can fail. The afferent limb is the detection: the observation is taken, scored and recorded. The efferent limb is the response: a clinician is summoned, attends, and acts. Afferent-limb failure is the missed or estimated vital sign — above all the respiratory rate — and the unrecorded observation. Efferent-limb failure is the abnormal observation that is charted but does not trigger an escalation, the review that does not happen, or the review that happens but does not act. [1]

The efferent limb is where most preventable arrests are lost. Retrospective study shows that a failure to respond to deranged physiology — not the absence of the abnormality — is the recurring system error, and that this failure is associated with modifiable factors including staffing.[4] National guidance therefore requires not only that acutely ill adults are monitored with a track-and-trigger system but that an agreed, graded physiological response is triggered automatically, independent of a subjective judgement that the patient looks well.[1] The score is a prompt to act, not a substitute for clinical assessment — but when it is high, the action must follow whether or not the clinician is reassured by the patient's appearance.

The Medical Emergency Team (MET) and the rapid response team

The detection of deterioration is worthless without a defined response capability that arrives at the bedside in minutes. This is the efferent limb made explicit: a team summoned by a physiological trigger that brings critical-care skills to the ward patient before the arrest. The terminology varies by region — the Medical Emergency Team (MET, the Australasian term), the Rapid Response Team (RRT, the North American term), the Critical Care Outreach Team (the UK term) — but the principle is identical: a single call activates a team of clinicians with the skills to resuscitate, reassess and relocate the deteriorating patient. [1]

The MET calling criteria combine single-parameter thresholds — any one of which triggers the call — with the explicit worried criterion that no score can encode. The genius of the worried criterion is that it delegates escalation authority to the bedside clinician: the person closest to the patient is empowered to call without fear of being wrong. The cultural change — the licence to call — is as important to the system as the numerical thresholds, and it is the step the MERIT trial identified as the rate-limiting one. [1]

Failure to rescue

Failure to rescue is the concept that frames why patients still die of deterioration despite a documented abnormality: it is not the absence of detection but the absence of an adequate response. The term, borrowed from the surgical-outcomes literature (where it denotes death after a treatable complication), has been adopted for the deteriorating ward patient to capture the recurring system failure — the abnormal observation that is charted but not escalated, the review that does not happen, the call that is not made. [1]

The landmark antecedent work established the scale of the problem. Schein and colleagues' seminal study of in-hospital cardiopulmonary arrest found that in the hours before the arrest the large majority of patients had a documented physiological deterioration, and that in the majority a clinician had documented the abnormality but had not acted on it adequately.[12] This single observation is the foundation of the entire track-and-trigger movement: the abnormality is seen; the response fails.

The evidence base and the comparison of scores

The aggregate early-warning scores have been validated repeatedly for the prediction of death, cardiac arrest, unanticipated intensive care admission, and the combined outcome. Their discriminative performance is moderate-to-good and is robust across populations and over time.[5] Head-to-head, an emergency-department prospective comparison of SIRS, MEWS, NEWS2 and qSOFA in predicting intensive care admission and mortality found NEWS2 to perform at least as well as the alternatives for general deterioration, while the quick Sequential Organ Failure Assessment, derived from the Sepsis-3 consensus as a bedside prompt for possible sepsis, is narrower in purpose.[7][8]

An important nuance, and a favourite examiner point, is what each score is for. qSOFA (respiratory rate of 22 or more, systolic blood pressure of 100 mmHg or less, altered mentation) was proposed in Sepsis-3 as a prompt to consider sepsis and to predict poor outcome in patients with suspected infection; it is not a general deterioration score and it is not a diagnostic criterion for sepsis. NEWS2 is a general acute-illness severity score that performs across all causes of deterioration. The two answer different questions, and conflating them is a common error. [1]

Limitations and controversies

Early-warning scores are decision aids, not diagnoses. Their sensitivity and specificity are imperfect: a low score does not exclude deterioration (the patient may be compensating or the deterioration may be in a parameter the score does not capture, such as pain, urine output or a subtle neurological change), and a high score is not always actionable on its own. Over-reliance on the number — "the score is only 2, so the patient is fine" — is a recognised cognitive trap; the score augments but does not replace the clinical assessment and the trend. The trend matters more than any single reading: a patient moving from a NEWS of 1 to a NEWS of 3 over four hours is deteriorating even though 3 is not a "trigger." Scores must be interpreted in context, with attention to the individual parameter scoring 3, which the aggregate can mask. The future lies in the electronic health record and in machine-learning models trained on continuous vital-sign data, which can outperform static scores — but these augment rather than abolish the need for a structured, acted-upon bedside assessment. [1]

Application in the emergency department

In the emergency department the same principles apply at the front door, integrated with triage. The Australasian Triage Scale allocates a triage category by the urgency of the presenting problem, but it is a triage tool rather than an illness-severity score; pairing it with an early-warning score captures the patient who is sicker than their triage category suggests. The deteriorating patient in the waiting room — abnormal observations on arrival, or a worsening trend while waiting — is exactly the patient a track-and-trigger system is designed to catch, and a protocolised re-triage and escalation prevents the avoidable collapse. Senior review and a low threshold to move the deteriorating patient to the resus bay should follow the score, not wait for the collapse. [1]

Management — the drug doses for the deteriorating patient

The deteriorating patient is managed with the targeted intervention: fluid resuscitation 500 mL bolus of balanced crystalloid for the hypovolaemic or septic patient; noradrenaline 0.05 to 0.5 mcg/kg/min for the vasodilatory shock; adrenaline 500 mcg IM for the anaphylaxis (or 1 mg IV in the arrest); ceftriaxone 2 g IV for the suspected sepsis; salbutamol 5 mg nebulised for the bronchospasm; naloxone 400 mcg IV or 800 mcg IM for the opioid toxicity; glucose 25 g IV (50 mL of 50 per cent) for the hypoglycaemia; and flumazenil 200 mcg IV for the benzodiazepine overdose. The key principle is the treat-the-cause identified by the track-and-trigger score. [1]

Differential diagnosis — the cause of the deterioration

• Sepsis — the raised temperature, the tachycardia, the raised lactate; treated with the antibiotic, the fluid, the vasopressor.
• Hypovolaemia — the haemorrhage, the dehydration, the burns; treated with the fluid, the blood.
• Cardiac — the arrhythmia, the MI, the heart failure; treated with the rate control, the reperfusion, the diuresis.
• Respiratory — the pneumonia, the pulmonary embolism, the pneumothorax; treated with the oxygen, the antibiotic, the anticoagulation.
• Neurological — the stroke, the seizure, the encephalopathy; treated with the thrombolysis, the anticonvulsant, the specific therapy.
• Metabolic — the hypoglycaemia, the hyperkalaemia, the acidosis; treated with the glucose, the calcium, the insulin-dextrose.

The escalation pathway and communication [1]

Detection is worthless without a response. The escalation pathway is a defined chain: from the bedside nurse who records the score, to the ward or emergency clinician who reviews the single parameter scoring 3 or the aggregate of 5 to 6, to the rapid-response or medical emergency team and critical care for the aggregate of 7 or more. National guidance embeds this in the requirement for an agreed physiological response to a defined trigger.[1] Communication at each handover should be structured — SBAR (Situation, Background, Assessment, Recommendation) or ISBAR — so the responder inherits the score, the trend, the examination and the working concern without ambiguity, and so the call for help is made early rather than after resuscitation has failed.

ISBAR — structured communication for the escalation

The escalation is only as good as the handover that requests it. An unstructured call — the patient in bed 6 looks bad, can you come? — loses the information the responder needs to triage and to prepare. The structured communication tool — SBAR (Situation, Background, Assessment, Recommendation) or ISBAR (with an added Identify step) — forces the caller to deliver the score, the trend, the examination, the working concern and the ask, in that order, in under a minute. The teaching of a structured tool measurably improves the clarity and the content of the interprofessional handover.[13]

Special populations

Paediatric deterioration is scored with the paediatric early-warning scores, which apply age-specific thresholds because a child's normal ranges and the meaning of abnormality change with age, and because children compensate superbly until they decompensate abruptly. The pregnant patient has altered normal physiology — a mild tachycardia, an increased minute ventilation, and a relative anaemia — that must not be mistaken for deterioration, yet pregnancy also reduces the reserve to decompensation; maternity early-warning scores adapt the parameters. The older patient frequently presents atypically — with confusion or a fall rather than classical symptoms — and may not mount the expected tachycardia (for example, the patient on a beta-blocker), so an unimpressive heart rate must not be reassuring. The patient with a baseline cognitive impairment or chronic physiological abnormality requires interpretation of the score against their individual baseline. [1]

Regional systems

While the principle is global, the implementation is regional. The United Kingdom standardised on NEWS2 through the Royal College of Physicians, and it is increasingly adopted internationally.[1] In Australia the standardised observation-and-response system in New South Wales, "Between the Flags", embeds colour-coded physiological zones (blue, yellow and red) on the standard observation chart with a defined Clinical Emergency Response, illustrating the same track-and-trigger principle in a different implementation; other jurisdictions use variants. The examiner expects the candidate to know the principle and the local implementation, and to recognise that the score is a means to a graded response, not an end in itself.

Between the Flags — the Australasian observation-and-response system

The New South Wales Between the Flags programme is the regional implementation the Australasian candidate is expected to know in detail. It embeds colour-coded physiological zones directly onto the standard observation chart, so that the deterioration is visually obvious on the chart itself — the track is built into the documentation, and the trigger is the crossing of a coloured line. The programme was deployed across the NSW public hospitals from 2010, and it is one of the most widely used standardised track-and-trigger implementations in the world. The design lesson is general: bring the trigger as close to the documentation as possible. [1]

The chart is divided into coloured bands. Green is the safe zone — the observation is within the normal range. Yellow is the clinical review zone — the observation is abnormal and prompts a review by the treating team. Red is the rapid-response zone — the observation is critically abnormal and triggers a Clinical Emergency Response (the call to the MET or the rapid-response team). A separate blue zone on the paediatric and maternity charts carries the age-specific normal ranges. The essential design feature is that the response is mandated by the colour the observation falls in, not by a clinician's subjective judgement that the patient looks well. [1]

Common pitfalls

The recurring failures are: omitting or estimating the respiratory rate, the most sensitive and most under-recorded sign; charting the abnormality without escalating it, the efferent-limb failure that precedes most avoidable arrests;[4] being reassured by a normal blood pressure or a well appearance in a patient whose trend and compensatory signs are abnormal; treating the score as a diagnosis rather than a prompt, or ignoring the trend in favour of a single number; masking a single parameter scoring 3 within an acceptable aggregate; and confusing qSOFA (a sepsis prompt) with NEWS2 (a general deterioration score).

The scores compared — a Fellowship reference table

The candidate should be able to place each scoring system by its purpose, its parameters and its trigger, because the examination rewards the candidate who knows what each score is FOR, not merely its arithmetic. The table below sets the major systems side by side. [1]

Red flags

The following features identify the patient at immediate risk of deterioration, in whom an urgent, structured response is required: [1]