What Is the MEWS Score and How Is It Calculated?

The Modified Early Warning Score, or MEWS, is a bedside tool that converts a handful of routine vital signs into a single number reflecting how sick a hospital patient is right now. Nurses or other clinicians measure things like heart rate, blood pressure, breathing rate, temperature, and level of consciousness, assign each a small numerical score based on how far it strays from normal, and add them up. The total flags patients whose condition is worsening so that help can arrive before a crisis. The concept is deliberately simple, designed so that any ward nurse can calculate it quickly without specialized equipment, yet the evidence behind it shows genuine predictive power for outcomes like intensive-care admission and death.

What MEWS Measures

The original MEWS uses five core physiological parameters: systolic blood pressure, heart rate, respiratory rate, temperature, and level of consciousness. Many hospital versions add oxygen saturation and urine output, bringing the total to seven parameters. A development study in South Africa, for example, built its MEWS chart around all seven of these variables and found the expanded set useful for general wards in resource-limited settings.¹ Urine output, though clinically valuable as an early sign of circulatory compromise, is frequently left unrecorded. In one study, it was missing in over 97% of observation sets, which is why some institutions drop it from their version of the chart.¹

Respiratory rate deserves special attention because it is often the first vital sign to change when a patient starts to deteriorate, sometimes hours before other numbers shift.² Despite this, respiratory rate is the vital sign most frequently skipped or estimated rather than counted, which can undermine the entire score. Level of consciousness is captured using the AVPU scale: Alert, responding to Voice, responding to Pain, or Unresponsive. Some institutions derive this from the more detailed Glasgow Coma Scale, mapping the numerical GCS ranges onto the four AVPU categories.³

How the Score Is Calculated

Each parameter is assigned a value of 0, 1, 2, or 3 based on how far the reading deviates from a normal range. A score of 0 means the value sits comfortably within what is expected for a stable patient. As the reading drifts further away from normal in either direction, the score climbs. A very high or very low heart rate, for instance, earns a higher component score than one just slightly outside the normal band. The individual scores are then added together to produce a total MEWS.⁴

A typical five-parameter MEWS can range from 0 to 15, while a seven-parameter version extends the possible maximum. The scoring thresholds vary slightly from hospital to hospital, but a representative example looks like this:

Heart rate: 51–100 beats per minute scores 0; 101–110 or 41–50 scores 1; above 130 or below 40 scores 3.
Systolic blood pressure: 101–199 mmHg scores 0; 80–100 scores 2; below 70 or above 200 scores 3.
Respiratory rate: 9–14 breaths per minute scores 0; 15–20 scores 1; 21–29 scores 2; above 30 or below 9 scores 3.
Temperature: 35–38.4°C scores 0; values above or below that range score progressively higher.
Consciousness: Alert scores 0; responds to voice scores 1; responds to pain scores 2; unresponsive scores 3.

Most MEWS systems also incorporate a “single-parameter trigger.” Even if the total score remains low, any one parameter hitting a 3 can automatically trigger an escalation response. This combination approach catches the patient whose blood pressure is crashing but whose other numbers happen to still look reasonable.⁴

What the Numbers Mean at the Bedside

The clinical action triggered by a MEWS depends on the total and on local hospital policy, but a commonly used threshold is a score of 4 or more. In a prospective study of surgical ward patients, a MEWS of 4 or above was 75% sensitive and 83% specific for identifying patients who needed transfer to intensive care or a high-dependency unit.⁵ In that study, about 17% of ward patients triggered the call-out algorithm.

Other institutions set the threshold at 3 or at 5, depending on their patient population and staffing. An emergency department study found that a MEWS cutoff of 3 at admission had a sensitivity of about 64% and specificity of about 70% for predicting short-term critical illness.⁶ That same study showed that a rising MEWS over time was itself a danger sign: patients whose score increased between the pre-hospital reading and one hour after emergency department arrival had roughly three times the incidence of critical illness compared to those whose score stayed flat or fell.

Typical escalation protocols look something like this: a score of 0–1 means continue routine monitoring, a score of 2–3 means increase the frequency of observations and inform the charge nurse, a score of 4 or more means contact the physician or activate a rapid response team, and a very high score (7 or above) typically calls for immediate senior review or intensive care consultation. The exact cutoffs are locally defined, so the numbers you see on the wall chart at one hospital may not match another.

How Well MEWS Predicts Bad Outcomes

The evidence consistently shows that higher MEWS values correlate with worse outcomes, though the precision of the tool varies across populations and settings. A prospective study tracking patients after intensive-care discharge found that the MEWS had strong discriminatory power for predicting mortality, with an area under the curve of 0.88.⁷ An Indian study found that a MEWS of 5 or higher at 24 hours after admission predicted in-hospital mortality with 78% sensitivity and 94% specificity.⁸ In a Ghanaian hospital, patients with elevated MEWS values had roughly six times the adjusted odds of dying compared to those with low scores.⁹

One pattern worth noting is that MEWS tends to be more useful as a serial measurement than as a one-time snapshot. A single reading can miss a patient whose vital signs are just starting to drift. Repeated measurements that show a climbing score carry more predictive weight than any single total.

How MEWS Compares to Other Scoring Systems

MEWS is not the only early warning score in clinical use. The National Early Warning Score (NEWS), its update NEWS2, and the quick Sequential Organ Failure Assessment (qSOFA) are common alternatives, and researchers regularly pit them against each other. In the study of post-ICU patients mentioned above, the NEWS had a slightly higher area under the curve (0.91 versus 0.88) for predicting mortality, but the difference was not statistically significant.⁷

A real-time emergency department evaluation comparing MEWS, NEWS2, qSOFA, and systemic inflammatory response syndrome (SIRS) criteria found more nuanced results. For predicting 28-day in-hospital mortality, qSOFA and NEWS2 significantly outperformed MEWS. NEWS2 also maintained the highest sensitivity and negative predictive value across measures.¹⁰ This finding has driven many healthcare systems, particularly in the UK and parts of Europe, to standardize on NEWS2 rather than MEWS for inpatient monitoring.

Still, MEWS retains real-world advantages. It requires fewer parameters than NEWS (which adds oxygen supplementation status and a more granular consciousness scale), making it faster to calculate without electronic tools. In settings where staffing is thin or electronic health records are unavailable, MEWS remains a practical frontline option.

Using MEWS Outside the Hospital Ward

Paramedics and emergency triage nurses have adopted MEWS for pre-hospital and emergency department settings. A study of pre-hospital patients found that MEWS was a good predictor of adverse outcomes, with an area under the curve of about 0.80. Combining MEWS with the paramedic’s own clinical judgment performed even better, reaching 72% sensitivity and 85% specificity.¹¹ The takeaway there is that the score works best as a supplement to experienced assessment, not a replacement for it.

In the context of sepsis, pre-hospital MEWS values have also shown predictive value. Patients arriving with high MEWS scores had about 1.24 times the mortality rate and about 1.34 times the likelihood of requiring ICU admission compared to those with lower scores.¹² Those effect sizes are modest on a per-patient basis, but when applied across thousands of emergency department arrivals, they help triage teams allocate attention more efficiently.

Specialized Versions for Specific Populations

The standard adult MEWS does not translate well to every patient group, so several specialized variants exist. The Pediatric Early Warning Score (PEWS) adapts the concept for children, whose normal vital sign ranges differ substantially from adults. In a study of pediatric emergency department patients, PEWS scores were significantly higher in children who eventually needed ICU admission. For every one-point increase in the PEWS score, the odds of ICU admission roughly doubled.¹³

Pregnant and postpartum patients present another challenge, because normal pregnancy shifts vital sign baselines. The Modified Early Obstetric Warning System (MEOWS) uses a color-coded chart with different trigger thresholds from the adult MEWS, and includes additional parameters such as proteinuria, color of amniotic fluid, and lochia appearance. The “white zone” represents the normal range, “yellow” is a single-trigger alert, and “red” indicates a critical deviation.¹⁴ A pregnant woman’s resting heart rate, for instance, sits higher than a non-pregnant adult’s, so the standard MEWS thresholds would generate a flood of false alarms on a labor ward.

The Problem of Incomplete and Inaccurate Scoring

A tool is only as good as the data feeding it, and this is where MEWS runs into serious practical trouble. A large retrospective study of over 1.6 million vital sign measurements found that all the items needed to calculate a complete MEWS were present in only 76% of cases. One in four assessments was incomplete.¹⁵ Worse, the problem is not random: incomplete observation sets were more likely to contain readings that should have triggered an alert than complete ones. In other words, the sickest patients were the most likely to have gaps in their records.¹⁶

Paper-based scoring introduces another layer of error. When observations that should have triggered an alert were present, the alerting score was correctly calculated only about 39% of the time in incomplete sets and 30% in complete ones.¹⁶ That is a startling gap between theoretical performance and real-world reliability. It means a significant fraction of patients who should have had rapid response teams called never actually had the alarm raised, not because the scoring system failed conceptually but because the arithmetic was done wrong or the chart was filled in carelessly.

Automating the Score

Electronic health record systems and bedside monitoring devices increasingly calculate MEWS automatically, eliminating arithmetic errors and ensuring that the score updates every time a new vital sign is entered. One hospital that implemented automated MEWS-based alerts found that the median time from clinical deterioration to activation of the medical emergency team dropped from 60 minutes to 34 minutes.¹⁷ That 26-minute improvement matters in acute care, where delays in escalation are linked to worse outcomes.

Automation brings its own headache, though: false alarms. When monitoring devices continuously feed data into an automated MEWS calculator, artifact from motion, cuff displacement, or a poorly placed oxygen sensor can generate scores that do not reflect the patient’s actual physiology. Alarm fatigue, where staff begin to ignore or silence frequent false alerts, is a recognized patient-safety threat.¹⁸ Some systems have attempted to address this by incorporating additional parameters like end-tidal carbon dioxide into the automated score, hoping that more data points reduce the false-positive rate.¹⁹

When Nurse Intuition Fills the Gaps

Vital signs capture only part of the clinical picture. Experienced nurses frequently sense that a patient “doesn’t look right” before the numbers confirm it, and researchers have tried to formalize that instinct. The Nurse Intuition Patient Deterioration Scale (NIPDS) is one such instrument. A prospective study found that combining a structured nurse-intuition score with a standard early warning score provided more net benefit in predicting serious adverse events than either tool alone.²⁰

A similar approach in an emergency observation unit combined MEWS with the NIPDS and found that the combined model significantly outperformed MEWS alone in predicting clinical deterioration within 24 hours.²¹ This underscores the idea that MEWS works best not as a standalone oracle but as one input in a broader clinical picture. The score catches the physiological drift; the nurse catches the pale, anxious look on the patient’s face that no number describes.

Cost and Resource Effects

Early detection of deterioration should, in theory, save money by preventing expensive ICU stays and reducing hospital length of stay. The evidence supports this, at least when the system is implemented well. One US hospital study found that after introducing an early deterioration detection system for general-care patients, the average cost per discharge fell by about 18% and the average length of stay dropped by about a day.²²

A separate study focused on sepsis patients found that a MEWS-based clinical decision support tool was associated with a roughly 23% decrease in median total direct costs and a trend toward shorter hospital stays.²³ That study did not find a significant change in mortality, suggesting the main benefit was getting patients the right care faster rather than pulling them back from the brink. From a hospital operations standpoint, even modest reductions in length of stay free up bed capacity and smooth patient flow, effects that compound across thousands of admissions per year.

What MEWS Cannot Do

No early warning score can diagnose a disease. MEWS tells you that something is going wrong physiologically; it does not tell you what. A rising score triggers the response team, but the team still needs to figure out whether the patient is bleeding, going septic, having a cardiac event, or experiencing a drug reaction. The score is a fire alarm, not a fire investigator.

MEWS also has a ceiling of complexity. Patients with chronic conditions may have baseline vital signs that are permanently abnormal, such as someone with chronic obstructive lung disease whose resting respiratory rate is always slightly elevated. These patients can score higher at baseline, creating a risk of “alert fatigue” for the team, or they can be habituated to a score that masks true deterioration because it does not look dramatically different from their usual numbers. Adjusting MEWS thresholds for individual patients is sometimes done informally but is not built into the standard scoring framework.

Post-operative patients present another edge case. A study of surgical patients in Cape Town found that while a fast heart rate and low blood pressure during the first eight post-operative hours were associated with death, most other vital sign parameters recorded on admission were not independently predictive.²⁴ This highlights that MEWS works differently across clinical contexts. Its power depends on which parameters are doing the heavy lifting, and that shifts depending on the patient’s situation.