Is the Modified Early Warning Score (MEWS) superior to clinician judgement in detecting critical illness in the pre-hospital environment?

Abstract Aim Physiological track and trigger scores have an established role in enhancing the detection of critical illness in hospitalized patients. Their potential to identify individuals at risk of clinical deterioration in the pre-hospital environment is unknown. This study compared the predictive accuracy of the Modified Early Warning Score (MEWS) with current clinical practice. Methods A retrospective observational cohort study of consecutive adult (≥16 yrs) emergency department attendances to a single centre over a two-month period. The outcome of interest was the occurrence or not of an adverse event within 24 h of admission. Hospital pre-alerting was used as a measure of current critical illness detection and its accuracy compared with MEWS scores calculated from pre-hospital observations. Results 3504 patients were included in the study. 76 (2.5%) suffered an adverse event within 24 h of admission. Paramedics pre-alerted the hospital in 224 cases (7.3%). Clinical judgement demonstrated a sensitivity of 61.8% (95% CI 51.0–72.8%) with a specificity of 94.1% (95% CI 93.2–94.9%). MEWS was a good predictor of adverse outcomes and hence critical illness detection (AUC 0.799, 95% CI 0.738–0.856). Combination systems of MEWS and clinical judgement may be effective MEWS ≥4 + clinical judgement: sensitivity 72.4% (95% CI 62.5–82.7%), specificity 84.8% (95% CI 83.52–86.1%). Conclusions Clinical judgement alone has a low sensitivity for critical illness in the pre-hospital environment. The addition of MEWS improves detection at the expense of reduced specificity. The optimal scoring system to be employed in this setting is yet to be elucidated.