How Robust Are Early Warning Scores for Predicting Patient Deterioration?

A systematic review of early warning scores showed widespread methodological weaknesses.

Early warning scores are used widely as clinical prediction models; they use vital signs and other historical data integrated into the electronic medical record to identify clinical deterioration in hospitalized patients; examples include the modified early warning score (MEWS) and the national early warning score (NEWS). In a systematic review of 95 articles (mostly retrospective cohort studies), published between 2001 and 2019, researchers sought to determine how these data are analyzed and which outcomes are used by hospitals worldwide.

Death was the most commonly predicted outcome, followed by cardiac arrest and intensive care unit admission with various time horizons (the most common was 24 hours). The most frequently used predictors we…

Reviewing Author

Andrew S. Parsons, MD, MPH

Disclosures

Consultant/Advisory BoardNEJM Healer Advisory Group; Aquifer Clinical Excellence; NBME Clinical Reasoning

Grant/Research SupportSouthern Group on Educational Affairs (SGEA)

Editorial BoardsDiagnosis

Leadership Positions in Professional SocietiesUndergraduate Medical Education (UME) Section Chair, Southern Group on Educational Affairs (SGEA); Chair of Early Career Physicians, American College of Physicians (ACP), Virginia Chapter

Andrew S. Parsons, MD, MPH

Disclosures

Consultant/Advisory BoardNEJM Healer Advisory Group; Aquifer Clinical Excellence; NBME Clinical Reasoning

Grant/Research SupportSouthern Group on Educational Affairs (SGEA)

Editorial BoardsDiagnosis

How Robust Are Early Warning Scores for Predicting Patient Deterioration?

Topics

Reviewing Author