Abstract 16075: Predicting Clinical Deterioration on the Wards: Subjective vs Objective Metrics of Patient Acuity
BACKGROUND: The Modified Early Warning Score (MEWS), derived from vital signs, and the Patient Acuity Rating (PAR), a seven-point clinical judgment-based score, have both been independently validated for identifying ward patients at risk for cardiac arrest and intensive care unit (ICU) transfer. However, the two scores have never been compared to one another.
METHODS: We conducted a prospective cohort study of inpatients at an academic medical center between May 2010 and March 2011. Daily PAR scores obtained from housestaff physicians at the end of each workday were compared to the corresponding MEWS, calculated from the vital signs closest to but preceding the PAR documentation time. The primary outcome was cardiac arrest or ICU transfer in the next 24 hours. Test statistics were compared for MEWS, PAR, and a combined PAR+MEWS, using the cutpoint for each with an approximate specificity of 85%.
RESULTS: Data were collected and analyzed from 2974 inpatient-days occuring in 713 patients. Three patients suffered a cardiac arrest and 33 were transferred to the ICU, with a mean MEWS of 2.4 (vs 1.7 for controls, p<0.001), PAR of 4.0 (vs 2.9, p<0.001), and PAR+MEWS of 6.4 (vs 4.6, p<0.001). A MEWS ≥3 (33% sensitivity) identified 12 total and 7 unique deteriorations, while a PAR≥5 (39% sensitivity) identified 14 total and 9 unique deteriorations (Figure). A PAR+MEWS≥7 (50% sensitivity) performed best, detecting 18 events total. There was a trend toward increasing area under the receiver characteristics curve from MEWS (0.63, 95% CI: 0.53-0.73) to PAR (0.68, 95% CI: 0.59- 0.77) to PAR+MEWS (0.71, 95% CI: 0.61-0.80).
CONCLUSIONS: The PAR and MEWS each identify unique patients missed by the other system. A combination score of both physiologic parameters and clinical judgment may be superior for predicting clinical deterioration.
- © 2011 by American Heart Association, Inc.