Secondary Logo

Journal Logo

Original Research

Estimating Blood Loss: Can Teaching Significantly Improve Visual Estimation?

Dildy, Gary A. III MD; Paine, Archana R. MD; George, Natalie C. RN, BSN; Velasco, Cruz PhD

Author Information
doi: 10.1097/01.AOG.0000137873.07820.34
  • Free

Hemorrhage remains one of the leading causes of maternal mortality in the United States1 and in other industrialized nations.2,3 In developing nations, where the vast majority of maternal deaths occur, the problem is exponentially greater.4,5 The World Health Organization estimated there were over 500,000 maternal deaths worldwide in the year 1990, one quarter due to hemorrhage.6 The most recent triennial Report on Confidential Enquiries into Maternal Deaths in the United Kingdom cites hemorrhage as the most frequent cause of substandard care associated with maternal death.7

Underestimation of peripartum blood loss and delayed blood component therapy seem to be common factors in many cases of avoidable hemorrhage-related maternal mortality. Inaccurate blood loss assessment can result in significant adverse sequelae; overestimation can result in unnecessary transfusion, and perhaps more important, underestimation in delayed treatment. Delayed diagnosis and treatment of hemorrhage may lead to hypovolemic shock, cardiopulmonary arrest, and death.8

The purpose of our study was to assess blood absorption characteristics of commonly used surgical materials, create a brief didactic presentation to educate clinicians about estimating blood loss, and investigate the effectiveness of the didactic session to improve estimating blood loss in mock clinical settings. We also evaluated the influence of clinical experience on estimation of blood loss, before and after the didactic session.


The protocol was approved by the Louisiana State University Health Sciences Center Institutional Review Board. Clinicians and clinicians-in-training (medical students, residents, fellows, and faculty) in all university clinical departments were invited to participate in this study. This study was conducted during a weekly obstetrics and gynecology grand rounds. All participation in the study was voluntary and there were no specific inclusion or exclusion criteria.

The regional blood bank provided reconstituted whole blood (hematocrit 35–40%) without platelets for this study. We applied precisely measured volumes of blood in increasing increments to the following surgical materials: laparotomy pads, surgical sponges, and perineum pads. All materials were photographed and weighed (Royal eX5 Postal Scale, Bridgewater, NJ) before and after blood application. Scale resolution was 1 gram. All measurements of pads and sponges were performed dry and wet (saturated with normal saline and hand wrung).

We created a 20-minute PowerPoint (Microsoft Corporation, Redmond, WA) slide presentation as the didactic tool to teach estimated blood loss. The slides illustrated the following 3 teaching tools for visually assessing blood volume.

First, mathematical formulas to calculate the volume of simple objects were described. The volume of a box (rectangular parallelepiped) is calculated as length × width × height. For a box with dimensions of 10 × 8 × 4 cm, the volume is 320 mL. The volume of a sphere is calculated as (4π/3) radius3 = (π/6) diameter3 = 0.52 (diameter3) ≅ (diameter3)/2. For a sphere with a diameter of 10 cm, using the latter simplest equation, the volume would be approximately 500 mL (actual volume is 523 mL).

Second, the volumes of familiar objects were demonstrated to the audience. The metric volume of a common 12-ounce soda can is 355 mL. The metric volume of a standard 1-L intravenous fluid bag is of course 1,000 mL. It was suggested to correlate blood collected in drapes, such as transparent plastic under-buttocks drapes at vaginal delivery and laparotomy side collection drapes used at cesarean delivery, with these familiar objects of known volume.

Third, we presented some simple rules of thumb regarding visual estimation of blood volume contained in common surgical materials. A standard dry 18-in. × 18-in. laparotomy sponge (AmeriNet Choice Lap Sponge AC450, Amerinet, St. Louis, MO) containing 25 mL, 50 mL, 75 mL, and 100 mL of blood will appear 50% saturated, 75% saturated, 100% saturated, and 100% saturated with excess blood dripping, respectively (Fig. 1). A dry 12-ply 4-in. × 4-in. surgical sponge (#3243, Dynarex Corporation, Orangeburg, NY) containing 5 mL of blood will appear completely saturated. For example, 5 completely soaked 12-ply surgical 4 × 4 sponges will contain approximately 25 mL of blood. A dry 3-in. × 11-in. extra-large perineum pad (OBNS-411, Securely Yours, La Verne, CA) containing 25 mL of blood will appear 50% saturated, and a completely saturated perineum pad will hold approximately 80 mL of blood.

Fig. 1.
Fig. 1.:
Blood absorption characteristics of a dry 18-in. × 18-in. laparotomy sponge. For this particular type, 25 mL of blood saturates about 50% of the surface area, 50 mL of blood saturates about 75% of the surface area, 75 mL of blood saturates the entire surface, and 100 mL of blood will saturate and drip from the sponge. If this were an actual clinical scenario, actual blood loss would be 250 mL.Dildy. Visual Estimation of Blood Loss. Obstet Gynecol 2004.

Study participants were given a brief introduction to the study. They were then instructed to view 7 simulated clinical scenarios and quantitatively estimate the blood volume in milliliters:

  1. Station 1: Under-buttocks drape with 350 mL of blood.
  2. Station 2: Two dry laparotomy sponges with 100 mL of blood.
  3. Station 3: Extra-large perineum pad with 25 mL of blood.
  4. Station 4: Twelve dry laparotomy sponges with 1,200 mL of blood.
  5. Station 5: Seven dry 12-ply 4 × 4 sponges with 35 mL of blood.
  6. Station 6: Six dry laparotomy sponges with 400 mL of blood.
  7. Station 7: Extra-large perineum pad saturated with 80 mL of blood.

Then the 20-minute didactic session (see above) was presented to the participants. The participants reevaluated the same 7 stations and the study was concluded. The total time of introduction, pretest, didactic session, and posttest was approximately 80 minutes.

Percent error of estimated blood loss was calculated as [(estimated blood loss − measured blood loss)/ measured blood loss]100 for each measurement. Because data showed lack of symmetry, lack of normality (with Shapiro-Wilk test), and apparent heteroscedasticity, summary data are expressed as medians. All tests were carried out in SAS 9.0 (SAS Institute, NC). We prospectively categorized percent error of estimated blood loss as overestimation (percent error more than +20%), accurate assessment (percent error between +20% and −20%), and underestimation (percent error less than −20%); these categories were empirical and not chosen based on clinical significance. Prelecture-postlecture comparisons of percent error of estimated blood loss were made with the Wilcoxon matched pairs signed rank test, the sign test, and Bowker’s test of symmetry. The effect of clinical experience on percent error estimated blood loss was assessed with the Kruskal-Wallis and Friedman’s tests. The effect of actual blood volume on the tendency to overestimate or underestimate volume was sought with the Jonckheere-Terpstra test. In all, 7 tests were carried out on the data set for a given scenario. In what follows, a Bonferroni adjustment has been taken into account to maintain an overall significance level of 0.05, therefore, any test yielding a P value smaller than .0071 (= .05/7) was declared significant.


Fifty-three participants completed the study (Table 1). All were medical students or physicians, with the exception of 1 obstetric nurse. All were obstetric clinicians, with the exception of the medical students and 1 anesthesiologist. Only 1 participant (the obstetric nurse) responded affirmatively to the question, “Have you ever been formally instructed how to estimate the amount of blood loss in a patient?” Clinical experience was defined as the number of years since graduation from medical or nursing school.

Table 1
Table 1:
Clinical Experience Was Defined as the Number of Years Since Graduation From Medical or Nursing School at the Time of Study

In all but 2 of the 7 clinical stations, there was a significant improvement in reduction of percent error of estimated blood loss (Table 2). Before the lecture, there was a trend (Jonckheere-Terpstra test) in overestimating blood loss at the lower volumes and underestimating blood loss at the higher volumes, but there was no trend after the lecture. In general, the educational program effected both a reduction in overestimation and underestimation of blood loss in all scenarios. As seen in Table 2, an under-buttocks drape containing 350 mL of measured blood loss (Station 1) was estimated to contain 250 mL before and 355 mL after the lecture, with a statistically significant (P < .001) improvement of percent error in estimated blood loss from −29% to 1% after the lecture. As illustrated in Figure 2, 12 laparotomy sponges with measured blood loss of 1,200 mL (Station 4) were underestimated by at least 20% in 33/53 cases before the lecture and only 1/53 cases after the lecture (P < .001).

Table 2
Table 2:
Measured Blood Loss, Estimated Blood Loss, Percent Error Estimated Blood Loss, and Difference in Percent Error Estimated Blood Loss for Each of the 7 Clinical Stations Before and After Didactic Session
Fig. 2.
Fig. 2.:
Estimated blood loss before and after didactic session for each of the 7 mock clinical stations. Dark bars represent overestimation (percent error +20% or more), white bars represent accurate assessment (percent error between +20% and −20%), and hatched bars represent underestimation (percent error −20% or less) of measured blood loss. Statistically significant (P < .001, McNemar’s test and Bowker’s test of symmetry) improvements in estimated blood loss were observed in all Prelecture and Postlecture pairs.Dildy. Visual Estimation of Blood Loss. Obstet Gynecol 2004.

Years of clinical experience did not influence the ability to assess blood loss, either before or after the lecture. Using the Friedman’s test on the difference in percent error of estimated blood loss by years of training, the smallest P value (0.061) was not statistically significant.


Recent reports from the World Health Organization6,9 estimate at least a half million maternal deaths worldwide per year, with a significant proportion attributed to severe bleeding. In the United States and other industrialized nations, the absolute number and relative rate of maternal mortality are substantially lower, nevertheless approximately one third of maternal deaths are due to hemorrhage.1 Recent triennial reports of the Confidential Enquiries into Maternal Deaths10–12 conclude that a significant proportion of maternal deaths due to hemorrhage in the United Kingdom may be attributed to delayed diagnosis and replacement of blood products. In 1967 Brant13 stated that “underappreciated and underestimated blood loss is a much more important factor in many maternal deaths than available statistics show.”

Postpartum hemorrhage has traditionally been defined as an estimated blood loss exceeding 500 mL. Using the colorimetric technique of measured blood loss in 580 obstetric cases, Wallace14 found that measured blood loss exceeded 500 mL in 6% of vaginal deliveries without episiotomy, 16% of vaginal deliveries with episiotomy, 30% of forceps deliveries, and 80% of cesarean deliveries. The classic work of Pritchard15 using radiolabeled red blood cells and Ueland16 using radiolabeled human serum albumin to directly measure blood loss showed average measured blood loss to be notably higher for vaginal delivery (at least 500 mL) and cesarean delivery (at least 1,000 mL).

During child birth, physiologic blood loss is usually well tolerated; however, shock may develop when blood loss reaches 15–25% of blood volume, usually about 1,000–1,500 mL.17 Often healthy young pregnant women remain stable during hemorrhage and show signs of cardiovascular compromise only after a serious degree of blood loss, at which point decompensation occurs rapidly.7

Blood loss can be measured by a variety of methods, each of which are cumbersome or impractical in general clinical practice. Gatch and Little18 were the first to study blood loss in general surgery operations using the colorimetric technique, which requires that hemoglobin be washed from surgical materials in a blender and measured in a colorimeter. The gravimetric method requires weighing materials such as laparotomy sponges on a scale and subtracting known weights of the surgical materials from blood.19 Blood volume estimation using dye-dilution or radioisotope dilution techniques are more technically difficult, requiring special equipment and serial measurements.16,20 More recently, computer-based mathematical modeling has been attempted to estimate blood loss during surgical procedures.21

In everyday practice, blood loss is usually estimated by subjective visual quantitation, which is generally based upon prior clinical experience. A number of studies have shown that visually estimated blood loss is fraught with error. In 1936, Pastore22 reported a study of 574 consecutive vaginal deliveries comparing estimated blood loss with measured blood loss and found that estimates are erroneous and misleading. Using the colorimetric technique of measured blood loss in 580 obstetric cases, Wallace14 found that blood loss estimates were underestimated by visual assessment for all modes of delivery. There is considerable evidence from studies using the colorimetric method that the higher the measured blood loss, the greater the underestimation by visual assessment.13,23,24 Razvi and colleagues24 found that when measured blood loss was less than 150 mL, the estimated blood loss was overestimated and when measured blood loss was more than 300 mL the estimated blood loss was underestimated; they proposed that estimated blood loss was clouded by conventional teaching that usual blood loss at delivery is 200–300 mL.

Several studies including ours have shown that the accuracy of estimated blood loss is not dependent upon age or clinical experience.25–27 Our survey of major surgical, obstetric, and gynecologic textbooks reveals no formal teaching of clinical assessment of blood loss. Only 1 of the 53 participants in our study had any formal training, thus we could not make statistical comparisons between those with and without prior formal training. Blood loss estimation is traditionally taught at the bedside; perhaps experienced clinicians have no more of a sound foundation for making these estimations than do their most junior counterparts.

Review of the published literature has identified only 2 studies which used known amounts of blood to teach blood loss assessment skills and test postdidactic improvements. Luegenbiehl28 found improvements in the tendency of nurses to overestimate blood loss on perineum pads after presentation of an educational program. Moscati and colleagues29 observed a modest improvement in emergency medical technicians’ estimation of blood spilled on flat surfaces after a limited training program.

Our method uses a combination of simple geometric formulas, known volumes of common objects, and known absorption characteristics of surgical materials to estimate blood loss in the delivery and operating room. We found that this teaching tool significantly reduced the error in blood loss estimation, for inexperienced as well as experienced clinicians. Of particular clinical importance was the reduction in underestimation of blood loss with greater degrees of measured blood loss. This reduction in underestimating heavy blood loss has the strongest potential to reduce hemorrhage-related morbidity and mortality.

Several limitations of our study deserve consideration. Our data does not allow us to estimate whether this information can be retained long-term, reduce medical errors, or positively affect clinical outcomes. Our scenarios may not mimic all blood loss circumstances. The blood absorption characteristics of various surgical materials need to be further defined; the appearance of blood-stained wet versus dry laparotomy sponges may differ, and the absorptive characteristics of one brand of laparotomy sponge may differ from another. Also, our method does not allow for consideration of concealed blood loss.

In summary, our data confirm previous studies that show blood loss is grossly overestimated at low volumes and underestimated at high volumes. Prior clinical experience does not seem to influence accuracy significantly. We have established a simple educational method to reduce error in estimating blood loss in the operating room and delivery room. Our method of blood loss assessment seems superior to traditional methods passed down from one generation of clinicians to the next. We believe that incorporation of this method into clinical training would be an effective way to improve estimated blood loss and recognize serious hemorrhage earlier, as long as a modest effort is made to inspect the surgical field during and after a procedure. Future research should determine whether this information can be retained long-term, reduce medical errors, or positively affect clinical outcomes. Assessment of blood loss, monitoring basic physiologic parameters such as blood pressure and pulse, acquisition of laboratory parameters such as hematocrit, platelet count, and clotting studies, and timely administration of blood component therapy remain essential in avoiding preventable morbidity and mortality due to obstetric and surgical hemorrhage.


1. Berg CJ, Atrash HK, Koonin LM, Tucker M. Pregnancy-related mortality in the United States, 1987–1990. Obstet Gynecol 1996;88:161–7.
2. Hogberg U, Innala E, Sandstrom A. Maternal mortality in Sweden, 1980–1988. Obstet Gynecol 1994;84:240–4.
3. Razum O, Jahn A, Blettner M, Reitmaier P. Trends in maternal mortality ratio among women of German and non-German nationality in West Germany, 1980–1996. Int J Epidemiol 1999;28:919–24.
4. Martey JO, Djan JO, Twum S, Browne EN, Opoku SA. Maternal mortality due to hemorrhage in Ghana. Int J Gynaecol Obstet 1993;42:237–41.
5. Kaye D, Mirembe F, Aziga F, Namulema B. Maternal mortality and associated near-misses among emergency intrapartum obstetric referrals in Mulago Hospital, Kampala, Uganda. East Afr Med J 2003;80:144–9.
6. World Health Organization. Revised 1990 estimates of maternal mortality: a new approach by WHO and UNICEF. Available at: Retrieved July 20, 2004.
7. de Swiet M. Maternal mortality: confidential enquiries into maternal deaths in the United Kingdom. Am J Obstet Gynecol 2000;182:760–6.
8. Dildy GA. Postpartum hemorrhage. Washington (DC): American College of Obstetricians and Gynecologists; 1998.
9. World Health Organization. Maternal mortality in 2000: estimates developed by WHO, UNICEF and UNFPA. Available at: Retrieved July 20, 2004.
10. Report on confidential enquiries into maternal deaths in the United Kingdom 1991–1993. London (United Kingdom): HMSO; 1996.
11. Why mothers die 1997-1999: the confidential enquiries into maternal deaths in the United Kingdom. London (United Kingdom): RCOG Press; 2001.
12. Why mothers die: report on confidential enquiries into maternal deaths in the United Kingdom 1994–96. London (United Kingdom): HMSO; 1998.
13. Brant HA. Precise estimation of postpartum haemorrhage: difficulties and importance. BMJ 1967;1:398–400.
14. Wallace G. Blood loss in obstetrics using a haemoglobin dilution technique. J Obstet Gynaecol Br Commonw 1967;74:64–7.
15. Pritchard JA. Changes in the blood volume during pregnancy and delivery. Anesthesiology 1965;26:393–9.
16. Ueland K. Maternal cardiovascular dynamics. VII. Intrapartum blood volume changes. Am J Obstet Gynecol 1976;126:671–7.
17. Bonnar J. Massive obstetric haemorrhage. Baillieres Best Pract Res Clin Obstet Gynaecol 2000;14:1–18.
18. Gatch WD, Little WD. Amount of blood lost during some of the more common operations. JAMA 1924;83:1075–6.
19. Buchman MI. Blood loss during gynecological operations. Am J Obstet Gynecol 1953;65:53–64.
20. Quinlivan WL, Brock JA. Blood volume changes and blood loss associated with labor. I. Correlation of changes in blood volume measured by I-131 albumin and Evans blue dye, with measured blood loss. Am J Obstet Gynecol 1970;106:843–9.
21. Brecher ME, Monk T, Goodnough LT. A standardized method for calculating blood loss. Transfusion 1997;37:1070–4.
22. Pastore JB. A study of the blood loss in the third stage of labor and the factors involved. Am J Obstet Gynecol 1936;31:78–92.
23. Duthie SJ, Ven D, Yung GL, Guang DZ, Chan SY, Ma HK. Discrepancy between laboratory determination and visual estimation of blood loss during normal delivery. Eur J Obstet Gynecol Reprod Biol 1991;38:119–24.
24. Razvi K, Chua S, Arulkumaran S, Ratnam SS. A comparison between visual estimation and laboratory determination of blood loss during the third stage of labour. Aust N Z J Obstet Gynaecol 1996;36:152–4.
25. Grant JM. Treating postpartum haemorrhage. Br J Obstet Gynaecol 1997;104:vii.
26. Patton K, Funk DL, McErlean M, Bartfield JM. Accuracy of estimation of external blood loss by EMS personnel. J Trauma 2001;50:914–6.
27. Meiser A, Casagranda O, Skipka G, Laubenthal H. [Quantification of blood loss. How precise is visual estimation and what does its accuracy depend on?]. Anaesthesist 2001;50:13–20.
28. Luegenbiehl DL. Improving visual estimation of blood volume on peripads. MCN Am J Matern Child Nurs 1997;22:294–8.
29. Moscati R, Billittier AJ, Marshall B, Fincher M, Jehle D, Braen GR. Blood loss estimation by out-of-hospital emergency care providers. Prehosp Emerg Care 1999;3:239–42.
© 2004 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.