373: An evaluation of maternal early warning systems to predict specific maternal morbidity patterns

Abstract

Vital sign scoring systems that alert providers to clinical deterioration prior to critical illness have been proposed as a means of reducing maternal risk. Unique parameters distinct from the general medical and surgical population are required given the physiology of pregnancy. Several maternal-specific early warning systems (MEWS) have been proposed with different vital sign parameters. This study compares the predictive ability of varying MEWS criteria for specific adverse outcomes. This retrospective cohort study analyzed all pregnant patients ≥18 years old at any gestational age admitted to the obstetric service of a large academic hospital system in 2018. Vital signs and clinical data were extracted from the electronic medical record. Three sets of MEWS criteria were evaluated (Table 1). Maternal morbidity was defined as hemorrhage, infection, cardiac dysfunction, and respiratory distress ascertained from administrative and clinical data. The test characteristics of each maternal early warning system were compared for any maternal morbidity as well as each of the four outcomes. 14,597 obstetric admissions along with 5.1 million vital sign measurements were included in the study. 2,451 patients experienced a morbidity event (16.8%). There were 980 cases of hemorrhage (6.7%), 1,337 of infection (9.2%), 362 of cardiac dysfunction (2.5%), and 275 of respiratory distress (1.9%). The sensitivities (15.3%-64.8%) and positive predictive values (22.3%-44.5%) of the three MEWS criteria differed significantly as did specificities (56.8%-96.1%) and negative predictive values (84.9%-88.9%) (Table 1). Test characteristics of MEWS criteria for individual outcomes also ranged broadly (Table 2). Test characteristics for MEWS criteria differ significantly with either relatively low sensitivity or specificity. Further refinement of the parameters is required to optimize care. Hospitals initiating an early warning system may be forced to choose between low sensitivity and high false positive rates.View Large Image Figure ViewerDownload Hi-res image Download (PPT)