Written and maintained by CASRAI Editorial Board
Last updated
The Modified Early Warning Score (MEWS) is a bedside aggregate scoring tool that turns five routinely recorded vital signs — systolic blood pressure, heart rate, respiratory rate, temperature, and level of consciousness — into a single number flagging a hospitalized patient at risk of clinical deterioration. It was developed and validated by Subbe CP, Kruger M, Rutherford P, and Gemmel L (“Validation of a Modified Early Warning Score in medical admissions,” QJM, 2001;94(10):521–526), whose original 709-patient validation cohort found a MEWS of 5 or higher associated with mortality and ICU admission. MEWS predates, and is conceptually the direct ancestor of, the UK’s Royal College of Physicians NEWS2 — see below for exactly how the two differ, because the difference matters for anyone comparing published MEWS numbers across hospitals.
The five parameters and how the score is built
Each of the five vital-sign parameters is assigned a point value, typically on a 0–3 scale, based on how far it strays from a normal range in either direction; the five parameter scores are summed into a total, with published implementations generally allowing a range up to 14. The core structure — independently reproduced across several sources describing MEWS — is:
| Parameter | What earns points |
|---|---|
| Systolic blood pressure | Points increase as SBP falls below roughly 100–80 mmHg, or rises markedly above the normal range |
| Heart rate | Points increase for rates below roughly 50 bpm or above 100–130 bpm, with the highest points at the extremes |
| Respiratory rate | Points increase for rates below roughly 9 breaths/min or above 20–30 breaths/min |
| Temperature | Points added for values below roughly 35–35.5°C or above 38–38.5°C |
| Level of consciousness (AVPU) | Alert scores 0; responsive to Voice, then Pain, then Unresponsive each add successively more points |
The exact numeric cutoffs in that table are not standardized across sources, and that is a genuine, documented feature of MEWS rather than an error in any one of them. Unlike NEWS2 — a single instrument published and version-controlled by one body, the Royal College of Physicians, with a single official scoring chart — MEWS was never centrally standardized after Subbe’s original 2001 publication. Hospitals, EHR vendors, and published calculators have each implemented their own version of the same five-parameter concept, and the boundary values differ from one implementation to the next: some add urine output or oxygen saturation as a sixth parameter, some use different bpm/mmHg breakpoints, and some cap individual-parameter points at 2 rather than 3. If you are comparing a MEWS value reported by one unit or hospital against another, confirm both are using the same locally-adopted scoring table before treating the numbers as equivalent — the name “MEWS” alone does not guarantee two hospitals are scoring the same way.
MEWS vs. NEWS2: why CASRAI covers them as related but distinct scores
MEWS and NEWS2 solve the same clinical problem — aggregate multiple vital signs into one escalation-triggering number — and NEWS2 is best understood as the standardized, single-governing-body successor to the same idea MEWS pioneered. The practical differences that matter for a hospital choosing or auditing either instrument:
- Governance. NEWS2 has one authoritative published chart (RCP, 2017) and NHS England endorsement; MEWS has no equivalent single governing publication after Subbe 2001, so “MEWS” in practice means whichever locally-adapted version a given hospital or vendor built.
- Parameters. NEWS2 uses six parameters, explicitly including oxygen saturation with a separate scoring scale for patients on supplemental oxygen versus room air. The MEWS parameter sets in wide circulation typically use five (occasionally adding urine output instead of, or alongside, oxygen saturation), without NEWS2’s supplemental-oxygen distinction.
- Escalation thresholds. NEWS2’s RCP chart defines explicit, graded response bands (routine monitoring at 0, ward-nurse review at 1–4, urgent review at 5 or more in aggregate or 3 in a single parameter, emergency/critical-care-team review at 7 or more). MEWS escalation practice varies by implementation; Subbe’s original validation flagged ≥5 as associated with mortality and ICU admission, but individual hospitals have adopted lower or higher local trigger thresholds depending on their patient population and the alarm burden a given threshold produces.
Neither point makes one score simply “better” — a hospital already running a mature, well-calibrated MEWS implementation does not automatically improve outcomes by switching instruments, and our Early Warning Score Implementation guide covers the actual decision (which score to run, how to calibrate the trigger threshold, and how to manage the resulting alarm burden) in more depth than a single-score page like this one can. What matters here is not conflating the two when reading published literature or comparing hospitals: a “MEWS of 4” and a “NEWS2 of 4” are not measuring deterioration risk on the same scale.
Where MEWS sits in escalation and rapid-response practice
MEWS is designed to be recorded at routine vital-sign checks and trended over time, not just read as a single snapshot — a rising trend can matter as much as the absolute value. A high or rapidly rising score is the trigger most hospital rapid response team activation criteria policies point to as one qualifying pathway alongside direct clinical concern (a bedside nurse or family member’s judgment that something is wrong, independent of any number). Because MEWS shares its consciousness-level component conceptually with fuller neurological assessment, some institutions cross-reference a deteriorating Glasgow Coma Scale trend alongside a rising MEWS rather than relying on AVPU alone for patients with a primary neurological concern. And because a MEWS-driven alert is only useful if the alarm actually reaches the right person at a sustainable rate, calibrating the trigger threshold is inseparable from the broader work covered in our Clinical Alarm Management Program guide — a threshold set too low against a given unit’s baseline produces alarm fatigue that defeats the score’s purpose.
Frequently asked questions
What is a normal MEWS score?
Most implementations treat a total score of 0–1 as low risk requiring only routine monitoring, though the exact “normal” cutoff depends on which locally-adopted version of MEWS a given hospital runs — there is no single centrally published definition of “normal” the way there is for NEWS2.
What does a MEWS score of 5 mean?
In Subbe et al.’s original 2001 validation cohort, a MEWS of 5 or higher was associated with mortality and ICU admission and is the threshold most commonly cited as clinically significant in the literature that follows that publication. Some hospitals apply a lower local trigger (the score’s own developers found a lower cutoff, around 2, better suited to a lower-risk ward-transfer population in a later study) — check your facility’s adopted policy rather than assuming the original 2001 threshold is what your institution actually escalates on.
Is MEWS the same as NEWS2?
No, though they measure the same underlying concept. NEWS2 is a single, centrally published, RCP-governed six-parameter instrument with an official UK national chart. MEWS is an earlier, five-parameter concept from the same research tradition that was never centrally standardized after its original 2001 publication, so “MEWS” in practice refers to whichever locally-adapted version a given hospital or EHR vendor implemented — the two produce scores on different scales and should not be treated as interchangeable numbers.
Why do different sources show different MEWS scoring tables?
Because MEWS was never centrally standardized the way NEWS2 was. Every published calculator, EHR vendor implementation, and hospital policy has built its own version of Subbe’s original five-parameter concept, with different numeric breakpoints, and some add a sixth parameter (commonly urine output). Confirm which specific version a source or your own institution uses before comparing MEWS values across settings.
