Journal des soins intensifs et critiques Libre accès

Abstrait

Accuracy of Modified Early Warning Scores for Predicting Mortality in Hospital: A Systematic Review and Meta-analysis

Sunee Suwanpasu and Youwanuch Sattayasomboon

Background: Patients at risk of rapid deterioration and critical illness often have preceding changes in their physiological parameters. Use of Modified Early Warning Score (MEWS) allows distinguishing points-of-care of patients who are at increased risk of serious adverse outcomes.

Objectives: To evaluate the prognostic accuracy of the MEWS to identify patients at risk of in-hospital death. Methods: The review process conformed to the recommendation form five steps to conduct a systematic review. Relevant studies from January 2000 to December 2015 were obtained from electronic databases. Standards for Reporting of Diagnostic accuracy (STARD) and Quality Assessment of Diagnostic Accuracy Studies instrument (QUADAS-2) were used to assess quality and bias of individual studies. MedCalc statistic software was used to merge the diagnostic accuracy data of recruited studies. The prognostic accuracy of predicting for in-hospital death was pooled on data from the diagnostic odds ratio and area under The ROC Curve (AUC) analysis.

Results: A total of 402 citations were identified yielding 16 studies for inclusion in this systematic review. Studies were statistically significant heterogeneous in terms of age and sample size. For predicting in-hospital death, high risk group of MEWS that get the threshold equal 4 or more and equal 5 or more had the Diagnostic odds ratio (DOR) of 14.278 (95% Confidence Interval [CI] 12.185 to 16.730, I2=56.59%]) and 3.28 (95%CI:2.489 to 4.323, I2=48.64%). On pooled AUC analysis, there was a trend for MEWS to estimate fair at the discriminative power of test. AUC of MEWS > 4 was 0.778 (95% CI : 0.715 to 0.841, I2=89.54%) and of MEWS > 5 was 0.646 (95% CI : 0.611 to 0.682, I2=49.69%).

Conclusions: The result showed a robust positive trend to predict in-hospital death. MEWS equal 4 or greater may be a favored threshold to alert to call for immediate appropriate action in hospitalized patients.

Avertissement: test