Accuracy of Visually Estimated Blood Loss in the Operating Room: A Simulation Study

Technical Summary: Accuracy of Visually Estimated Blood Loss in the Operating Room

Problem Statement
Visually estimated blood loss (VEBL) remains the most ubiquitous method for assessing intraoperative hemorrhage, yet it is characterized by high subjectivity and inaccuracy. While prior literature has predominantly focused on the estimation errors of surgeons and anesthesiologists, data regarding the accuracy of broader operating room (OR) personnel—including nurses and anesthesia support staff—are limited. Furthermore, existing studies often lack analysis of how specific variables, such as provider role, department, and years of experience, influence estimation performance across diverse simulated scenarios involving different materials and fluid volumes.

Methodology
This simulation study, conducted at a single academic institution (Temple University), evaluated VEBL accuracy across a multidisciplinary cohort of 55 OR personnel. The participants included attending surgeons, surgical residents, attending anesthesiologists, anesthesia residents, Certified Registered Nurse Anesthetists (CRNAs), Student Nurse Anesthetists (SRNAs), and OR/PACU nurses.

The study utilized eight distinct stations depicting simulated blood loss scenarios using expired packed red blood cells (pRBCs) diluted with normal saline to match the mean hemoglobin (12.3 g/dL) and hematocrit (36.9%) of surgical patients at the institution. The stations presented blood on various surgical materials:

• Station 1: Sponge collection bag (300 mL).
• Station 2: Absorbent pad (150 mL).
• Station 3: Suction canister with irrigation fluid contamination (300 mL actual blood).
• Station 4: Chux pad (225 mL).
• Station 5: Ray-Tec sponge (50 mL).
• Station 6: Ray-Tec sponge (100 mL).
• Stations 7 & 8: 4×4 Lap sponges (20 mL each).

Participants independently estimated the blood volume in milliliters. A "correct" estimation was defined as a value within ±15% of the actual volume. Statistical analyses included chi-square tests for categorical outcomes (correct/over/underestimation) and ANOVA or t-tests for continuous outcomes (mean estimates), stratified by professional title, department, and years of experience (<10 vs. ≥10 years).

Key Results
The study found that VEBL accuracy was universally low across all stations, with correct estimation rates ranging from 3.6% to 27.3%. Overestimation was the predominant error pattern at most stations, though specific scenarios elicited systematic biases:

• Scenario-Specific Bias:

  • Irrigation Contamination (Station 3): This station showed the most extreme bias, with 96.4% of participants overestimating the 300 mL blood volume, likely due to the difficulty in distinguishing blood from the total fluid volume in the suction canister.
  • Large Surface Area (Station 4): Conversely, the 225 mL volume on a Chux pad was dominated by underestimation (70.9%), suggesting that blood distributed over a large area is perceived as less voluminous.
  • Low Volumes (Stations 7 & 8): Small volumes (20 mL) on white sponges were frequently overestimated (67.3% and 58.2% respectively), with correct rates of only 12.7% and 9.1%.

• Provider Title and Department Differences:

  • Attending Physicians: Generally provided estimates closest to actual volumes, particularly at low-volume stations (7 and 8). At Station 6, attending physicians did not overestimate any cases, whereas Registered Nurses (RNs) overestimated 78.6% of cases.
  • Nursing Staff: PACU nurses demonstrated the highest rates of overestimation at Stations 5 (92.3%) and 6 (84.6%). RNs and CRNAs/SRNAs consistently overestimated small volumes significantly more than attending physicians.
  • Residents: Residents showed the highest rate of underestimation at Station 4 (88.0%).
  • Anesthesia Team: Achieved the highest correct estimation rate at Station 6 (39.3%).

• Experience:

  • While more experienced providers (≥10 years) gave higher absolute estimates at low-volume stations compared to less experienced peers, the distribution of correct, over-, and underestimates did not differ significantly between experience groups at any station. This suggests that experience alone does not broadly improve accuracy, though it may refine specific perceptual skills.

Significance and Claims
The authors claim that this study extends the understanding of VEBL inaccuracies beyond surgeons and anesthesiologists to include a comprehensive range of OR personnel. The findings identify specific, targetable gaps in estimation performance:

1. Systematic Errors: The direction of error is not random but varies systematically based on volume, material type, and the presence of irrigation fluids.
2. Role-Specific Training Needs: The data suggests that training interventions should be prioritized for nursing staff and trainees, who demonstrated the greatest systematic biases.
3. Clinical Caution: The study argues against sole reliance on VEBL for high-stakes decisions, particularly when irrigation fluids are present or when assessing small volumes on light-colored materials, as these scenarios consistently lead to significant overestimation.
4. Educational Application: The authors posit that these reproducible error patterns can be directly incorporated into simulation-based curricula to improve estimation accuracy, rather than relying on the passive accumulation of clinical years.

The paper concludes that while VEBL remains an indispensable tool for real-time decision-making due to its simplicity, its limitations necessitate structured, scenario-specific training and a heightened awareness of its inherent biases.