SESSION TITLE: Outcomes/Quality Control Posters II
SESSION TYPE: Original Investigation Poster
PRESENTED ON: Wednesday, October 29, 2014 at 01:30 PM - 02:30 PM
PURPOSE: Red cell distribution (RDW), a measure of erythrocyte size variability, has been shown to be an independent predictor of all-cause mortality in critically ill patients. Early warning scores such as National Early Warning Score (NEWS) have been used to direct care of deteriorating patients on hospital wards. NEWS captures only changes in acute physiology, while RDW may represent chronic inflammation. The purpose of this study was to assess whether addition of RDW would improve the prognostic ability of the NEWS at the time of rapid response team (RRT) activation.
METHODS: This is an observational study of a random sample (N=105) of patients on the hospital wards whose physiological parameters met criteria for RRT activation. RDW and NEWS closest to the time of RRT activation were collected from the hospital’s electronic medical records. RDW was divided into these tertiles: Normal (<14%), moderately abnormal (14.1-15.7%), and highly abnormal (>15.7%). The primary outcomes were in-hospital death, transfer to higher level of care or transition to hospice. The predictive value of RDW, NEWS and combination of both were measured and compared using receiver operating characteristic (ROC) curves.
RESULTS: Out of the 105 RRT activations, 39 patients had at least one positive outcome. For the combined endpoints of in-hospital death, transition to higher level of care or hospice, addition of RDW to NEWS improved AU-ROC from 0.783 (95% CI, 0.691 to 0.874, p<0.05) to 0.796 (95% CI, 0.705-0.887, p<0.05). It also improved PPV from 0.45 to 0.54, NPV from 0.85 to 0.88. For the single outcome of death, the addition of RDW to NEWS improved AU-ROC from 0.772 (95% CI, 0.655-0.888, p<0.05) to 0.819 (95% CI, 0.721-0.916, p<0.05).
CONCLUSIONS: Addition of RDW appears to improve the prognostic performance of NEWS for deteriorating patients. This improvement is more significant when the outcome of interest is in-hospital mortality.
CLINICAL IMPLICATIONS: RDW is routinely reported as part of automated CBC, and can be a useful tool in predicting poor outcomes in patients at the time of RRT activation.
DISCLOSURE: The following authors have nothing to disclose: Assad Oskuei, Syed Amin, Daryl Connolly, Adiba Geeti, David Kaufman
No Product/Research Disclosure Information