Abstract
Objective
To describe postpartum hemorrhage (PPH) outcomes and evaluate the predictive power of a PPH risk tool in a community birth population.
Design
Secondary data analysis using data from the Midwives Alliance of North America (MANA) Statistics Project 4.0 database.
Setting
United States.
Participants
We analyzed data from the MANA database that pertained to women who had singleton term pregnancies with the fetus in vertex position and gave birth at home or in birth centers from 2011 to 2018 (N = 61,246).
Methods
We calculated descriptive statistics and risk of PPH using the Association of Women’s Health, Obstetric and Neonatal Nurses (AWHONN) Postpartum Hemorrhage Risk Assessment Tool. Using logistic regression, we calculated the area under the curve (AUC) three different ways to see how well it predicted actual PPH in our sample. We calculated (a) with the current AWHONN tool, (b) by adding nulliparous women to the medium/high-risk group and multiparous women without additional PPH risk factors to the low-risk group, and (c) by treating the modified AWHONN tool as cumulative PPH risk factors rather than simply as low versus medium/high risk. We performed postestimation tests for goodness of fit, AUC, and correct classification.
Results
In our sample, 6.4% of women (n = 3,899) experienced PPH. The AUC for PPH prediction using the AWHONN tool was 0.53. After adding nulliparity as a PPH risk factor and applying PPH risk scores, the AUC of the tool increased to 0.59.
Conclusion
The AWHONN PPH Risk Assessment Tool was poorly predictive for PPH in this sample. Incorporating additional PPH risk factors may improve accuracy of risk tools.