Accuracy of Visual Postpartum Blood Loss Estimation Before and After Educational Intervention in Simulated and Real Clinical Settings

Abstract

Background: Postpartum hemorrhage (PPH) remains a major maternal safety concern in healthcare systems. Visual estimation of blood loss is widely practiced due to its convenience and rapid application, however, it is prone to significant inaccuracy, potentially leading to delayed diagnosis and management. Simulation-based training may enhance healthcare personnel’s ability to estimate blood loss more accurately.

Objectives: To evaluate the accuracy of visual postpartum blood loss estimation in both simulated and real clinical settings and to examine the effect of an educational intervention on estimation accuracy among health care personnel at Somdet Phra Yuparat Wiang Sa Hospital.

Methods: This pre-experimental study employed a one-group pretest–posttest design. The simulated phase (May 1–7, 2025) involving 39 participants, including intern physicians and nurses responsible for intrapartum care, who assessed blood loss across 12 simulation stations. The real clinical phase (May 20–July 30, 2025) 13 labor room nurses and nurse anesthetists assessed blood loss in 60 women who delivered at the hospital, with each patient independently evaluated by two assessors. Calculated blood loss was derived from pre- and post-delivery hematocrit values. An educational intervention on blood loss estimation was then provided, followed by reassessment in both settings. Data were analyzed using descriptive and inferential statistics, with statistical significance set at p < 0.05.

Results: In the simulated setting, significant estimation errors were observed before the educational intervention, particularly underestimation in scenarios involving blood loss greater than 500 mL. After the intervention, estimation accuracy improved, with no significant difference between visually estimated and calculated blood loss in 7 of 12 simulation stations. Professional background significantly influenced estimation error (p = 0.011), whereas years of work experience did not affect accuracy. In the real clinical setting, mean calculated blood loss was 395.65 ± 348.65 mL, visual estimation yielded lower mean values by approximately 82–89 mL (20.8–22.5%). Although the difference was not statistically significant, a tendency toward underestimation was observed in higher blood loss cases.

Conclusion: Visual estimation of postpartum blood loss demonstrates limited accuracy in both simulated and clinical settings. Educational training improves accuracy but does not eliminate discrepancies. Combining staff training with quantitative measurement methods, particularly in high-risk patients, may improve maternal safety within health care systems.