Correlation between hemorrhage risk prediction score and severe maternal morbidity

Abstract

Obstetrical hemorrhage is a leading cause of severe maternal morbidity, a key indicator of a nation's healthcare delivery system and often associated with a high rate of preventability. Limited data suggest that a patient's hemorrhage risk score may be associated with risk for maternal morbidity such as severe hemorrhage, intensive care unit admission, or transfusion. Little is known regarding the relationship between hemorrhage risk score and nontransfusion-related morbidity. We sought to evaluate the association between a patient's California Maternal Quality Care Collaborative admission hemorrhage risk score and severe maternal morbidity. This was a retrospective cohort of delivery admissions from 2018 to 2019 in a single healthcare network. Admission risk scores were assigned to each patient using the California Maternal Quality Care Collaborative criteria. Rates of transfusion- and nontransfusion-associated severe maternal morbidity were compared across low-, medium-, and high-risk strata. We defined severe maternal morbidity as the presence of any International Classification of Diseases diagnosis or procedure codes outlined by the Centers for Disease Control and Prevention, need for intensive care unit admission, or prolonged postpartum hospital length of stay. A multivariable logistic regression was used to assess the association between hemorrhage risk score and severe maternal morbidity. In the overall cohort, severe maternal morbidity occurred in 2.4% (n=517) of all deliveries. Excluding cases requiring transfusion, 0.6% (n=131) of cases still had a severe maternal morbidity event. The incidence of severe maternal morbidity was 1.6% (n=264) in patients categorized as low risk for hemorrhage compared with 2.5% (n=118) and 13.6% (n=135) in patients who were categorized as medium or high risk for hemorrhage, respectively (P<.001). Patients classified as high risk had a significant association with both severe maternal morbidity (adjusted odds ratio, 8.8; 95% confidence interval, 7.0-11) and nontransfusion-associated severe maternal morbidity (adjusted odds ratio, 3.6; 95% confidence interval, 2.2-5.9). In addition to predicting the risk for obstetrical hemorrhage and transfusion, our findings indicate that the California Maternal Quality Care Collaborative admission hemorrhage risk tool predicts risk for transfusion- and nontransfusion-associated severe maternal morbidity. Our findings imply that despite awareness and the identification of patients at high risk for obstetrical hemorrhage on admission, significant hemorrhage-associated morbidity persisted. Our data indicate that the identification of risk alone may be insufficient to reduce morbidity and imply that further work is needed to investigate and implement new practices in response to a patient's score stratum.