454: Implementation of quantitative blood loss does not improve prediction of hemoglobin drop in deliveries with average blood loss

Abstract

Hemorrhage is a leading cause of morbidity and mortality. The National Partnership for Maternal Safety released a PPH bundle in 2015 to improve safety and reduce morbidity, and the use of blood products. The bundle recommends quantitative blood loss (QBL) assessment for all deliveries. QBL can be time and resource intensive. We sought to determine whether QBL, as a specific part of the bundle, more accurately predicts hemoglobin (hgb) drop than visually estimated blood loss (EBL) among all deliveries. This is a prospective observational study. Pre-intervention data (PRE) was collected on all deliveries between 10/15/13 to 12/15/13. Deliveries were included if EBL, admission hgb, and 12 hour postpartum hgb (12hrCBC) were available. In July 2015, we implemented a hemorrhage bundle, including QBL. Post intervention data (POST) was collected on all deliveries from 10/20/15 to 12/20/15. A 500cc blood loss was predicted to result in a 1g/dL hgb drop at 12hrCBC. Means and standard deviations were calculated. Student’s t test was used to compare means. 592 of 626 (95%) of PRE and 583 of 613 (95%) of POST deliveries were included. 278 of 583(48%) POST deliveries had QBL recorded. Average EBL in the PRE and POST groups was 500cc and 530cc respectively. Average QBL in the POST group was 640cc. In the PRE group, actual hgb drop differed from predicted based on EBL by an average of 0.6g/dL for all deliveries. In the POST group, the difference between actual hgb drop and predicted by QBL was also 0.6g/dL. When evaluating deliveries with an EBL >1000cc (45 in PRE and 41 in POST), QBL in POST group trended toward improved prediction of hgb drop when compared to EBL in PRE group (difference of 0.2g/dL vs. 0.5g/dL, p=0.17). When looking at all deliveries, QBL does not predict hgb drop at 12hrCBC more accurately than EBL. However, when evaluating only deliveries with EBL >1000cc, QBL trends towards improved accuracy. It may be prudent to perform QBL only on deliveries where EBL is greater than 1000cc in order to balance the benefit of accuracy and the consequences of a resource intensive process.