Retrospective Review of Postpartum Hemorrhage Incidence, Risk Factors, and Maternal Morbidity

BACKGROUND: Postpartum hemorrhage (PPH) remains one of the most common causes of maternal mortality worldwide. In the US, PPH accounts for 2-3% of yearly maternal deaths and is also associated with serious morbidity including the need for blood transfusion, ICU admission, and hysterectomy. In the recent years, PPH risk assessment tools have been widely implemented in an attempt to identify woman at risk for hemorrhage and preemptively mobilize resources. The most commonly used trinary assessment tool from the California Maternal Quality Care Collaborative (CMQCC) assigns women to a low, medium or high-risk category; however, its utility has only been validated in relation to need for blood transfusion, which can be subjective. LOCAL PROBLEM: The American College of Obstetricians and Gynecologists (ACOG) recently released maternal safety bundles to reduce obstetric complications and rates of maternal morbidity and mortality. At UT Southwestern's Clements University Hospital (CUH), obstetricians and gynecologists recognized that a three-category approach to PPH risk stratification may hinder efficient PPH diagnosis and management. The Maternal-Fetal Medicine team developed a quality improvement project to identify hemorrhage risk factors and implement a PPH risk score in the labor and delivery unit. METHODS: Retrospective cohort review of all deliveries at CUH between January 1, 2018 and December 31, 2018 was conducted. Data on all antepartum and intrapartum hemorrhage risk factors using the trinary risk score developed by the CMQCC was analyzed and used to assign a risk category low, medium or high. A validated formula, utilizing maternal height, weight, ante- and post-partum hematocrits, was used to calculate each patient's blood loss. Calculated blood loss, need for intervention (uterotonic administration) and maternal morbidities (need for blood transfusion, intensive care unit (ICU) admission, and/or hysterectomy) were correlated to PPH risk categorization as well as to a developed numeric risk score. PPH was defined as blood loss exceeding 1,000 mL. Data was analyzed using standard methods of rates and proportions, Chi-square, and Wilcoxon rank-sum tests with p<0.05 considered significant. Quality improvement tools, including PDSA cycles, process maps, a SIPOC diagram, and FMEA were developed. INTERVENTIONS: A modified PPH risk assessment score that assigned a point value of 1 or 2 to hemorrhage risk factors based on their perceived potential to lead to PPH was developed. RESULTS: Of the 1855 deliveries, the median calculated blood loss was 879 mL. The overall PPH rate was 25.9%. The rates according to PPH risk groups were 19.5%, 36.6%, and 27.9%, respectively. The rate of PPH was significantly lower in the low risk group (p<0.001) but did not differ between the medium and high-risk groups (p= 0.11). The median blood loss was lowest in the low risk group (p<0.001). The transfusion rate correlated with risk stratification, with rates of 0.9%, 2.7% and 7% in the low, medium and high-risk groups, respectively. Overall, 178 women (9.5%) were treated with uterotonics, 38 (2.0%) required transfusion, 7 (0.4%) needed ICU admission and 12 (0.6%) underwent hysterectomy. Relative to low and medium-risk stratifications, women in the high-risk group were 2.3 times more likely to require uterotonic administration. Women in the low-risk group were 83% less likely to experience transfusion, ICU admission or hysterectomy compared to medium and high-risk women. Conversely, women in the high-risk group had a 4.1 fold increase in these same morbidities. CONCLUSION: Relative to the low and medium-risk stratifications, women classified as high risk for hemorrhage are indeed more likely to require uterotonic administration and also suffer disproportionately higher maternal morbidity. The trinary risk stratification tool commonly used to predict PPH distinguishes women at lower risk for hemorrhage compared to the general population. However, determining which women are most likely to hemorrhage at delivery evades prediction using current risk assessment tools. Though hypothesized that a numeric scoring system would better predict PPH and morbidity than the currently used trinary risk assessment, this was not substantiated by our data. More work needs to be conducted to understand which of the identified risk factors is most highly associated with hemorrhage, particularly in the low and medium risk groups, since women in these groups make up the majority of the obstetric population.