Combining the National Early Warning Score with an early warning score based on common laboratory test results better discriminates patients at risk of hospital mortality

We hypothesised that combining an early warning score (EWS) based exclusively on laboratory tests (LDTEWS) with the National Early Warning Score (NEWS) would improve the discrimination of hospital mortality compared with each system individually. We used a combined electronic database of haematology, biochemistry and vital signs measurements collected routinely soon after admission for 88695 adult admissions for whom the admission specialty was Medicine. The data were divided into 23 sets (Q1-Q23), each corresponding to three months. LDTEWS was generated using decision tree analysis of haematology and biochemistry results for episodes from set Q1. LDTEWS (to haematology and biochemistry results) and NEWS (to vital signs measurements) were then applied in 22 discrete data sets each of three months long (Q2, Q3......Q23) (range of n = 3580 to 4186). A combined EWS was determined for each episode by summing, for each episode, the EWS values for NEWS and LDTEWS. The abilities of NEWS, LDTEWS and the combined EWS to discriminate in-hospital death were assessed using the area under the receiver-operating characteristic (AUROC) curve. The area under the receiver-operating characteristic curve values (95% CI), with in-hospital death as the outcome for the validation sets Q2-Q23: • for LDTEWS, ranged from 0.743 (0.718 to 0.768) (Q10) to 0.799 (0.773 to 0.825) (Q9) • for NEWS, ranged from 0.704 (0.675 to 0.733) (Q7) to 0.759 (0.732 to 0.786) (Q8) • for the combined EWS, ranged from 0.799 (0.776 to 0.822) (Q10) to 0.836 (0.813 to 0.861) (Q21). This study provides evidence that a combined EWS using commonly measured laboratory tests and vital signs better discriminates in-hospital mortality than using either an EWS based on laboratory data or vital signs alone. We hypothesise that, with appropriate modification, the combined EWS could be used for identification of patients at high risk of death in the short term (for example, 24 hour mortality).