Reliability Testing of a Modified Early Warning Score (MEWS) System

Purpose: Internationally there is a lack of consensus on what should be included in modified early warning scoring (MEWS) tools. Furthermore, few studies have reported reliability and validity testing of MEWS tools. Further, few studies have reported reliability and validity testing of MEWS tools.  The purpose of this study researchers sought to report reliability testing of the Covenant Health MEWS tool for use in an adult medical-surgical population.

Background: Vital sign and assessment findings recording by nurses do not necessarily translate into rapid recognition of a patient’s deteriorating condition. With signs of health condition declining evident hours before an adverse event occurrence, clinical MEWS tools quantify and rank physiological assessment findings triggering life-saving care. MEWS assessment tools can assist healthcare providers in the rapid detection and recognition of subtle changes in patient condition signaling clinical deterioration. Comprehensive literature review findings indicated the need for more rigorous testing and reporting of the reliability of MEWS tools in relationship to patient outcomes.  The development of the MEWS instrument was initiated from shared governance committee responsible for patient safety. The MEWS tool was developed following a comprehensive literature review and a plan-do-check-act (PDCA) cycle process for improvement and preliminary testing of the tool.  The purpose of this study is to test reproducibility of the Covenant Health MEWS tool using a test-retest design.

Methods: A convenience sample of critical care nurses (n=32) from a nonprofit acute care facility (881 licensed beds) were recruited for reliability testing of the Covenant Health MEWS form. At time one, testing will be conducted in a simulation laboratory adjacent to the hospital using four low-fidelity mock hospital simulation scenarios developed from clinical data from hospitalized patients who died from sepsis.  Each scenario represented one of four levels of clinical concern based on a total calculated MEWS score. Physiologic assessment findings were written out in the simulated rooms for research subjects to assign a score directly onto the MEWS instrument. Each assigned score was added together to compute a total MEWS score and recorded on the tool. Total scores were translated by test subjects into assignment of a MEWS color and associated algorithm.  Higher scores represent worsening patient conditions. The derived color recording and steps to alert other clinicians of concern for declining condition were documented on the paper tool.  After an approximate three-week interval, subjects will repeat the procedure at time two.  The ordinal level paired scores at time one and time two will analyzed using Spearman’s Rho to compute test-retest reliability of the MEWS tool for use in this sample.

Results: The results are pending testing. 

Conclusion: The proposed study will establish reliability and validity of an EWSS tool modified for institutional use. Publishing of research findings add to nursing body of knowledge available to guide institutional modification and adoption of EWSS clinical tools. With a paucity of published reliability testing and established validity of MEWS instruments, this research attempts to address identified research gaps. Simulated scenarios can help establish reliability of a MEWS tool through preplanning confounding variable control measures.