Can NEWS predict hospital mortality in patients who left the medical ICU? A retrospective cohort study in resource-limited setting

INTRODUCTION: Patients discharged from the intensive care unit (ICU) have up to a 20.7% risk of in-hospital mortality due to several risk factors (e.g., age and length of stay). However, data remain limited regarding whether the National Early Warning Score (NEWS) can predict in-hospital mortality following ICU discharge in resource-limited setting. OBJECTIVE: To evaluate whether NEWS is independently related to in-hospital mortality after ICU discharge. MATERIALS AND METHODS: In our retrospective cohort study, conducted in a general medical ICU, patients had to be at least 18 years old, admitted to the medical ICU, and subsequently discharged from the ICU to a medical ward. Clinical factors predictive of in-hospital mortality after ICU discharge were computed using multivariable logistic regression analysis. RESULTS: Of the 633 patients who met the inclusion criteria, 89 (14.06%) died in hospital after ICU discharge. Four independent factors were associated with their mortality: age, length of ICU stay, use of a ventilator, and NEWS. NEWS had an adjusted odds ratio of 1.696 (95 % confidence interval [1.469–1.958]); a NEWS of 6 or more had a sensitivity of 46.07 % and a specificity of 89.52 %, whereas a NEWS of 3 or more had a sensitivity of 88.76 % and a specificity of 35.48 %. CONCLUSIONS: NEWS, especially of 6 or more, may be used as a component of risk stratification and enhanced monitoring after ICU discharge in resource-limited setting.