Visual Blood Loss Estimation Accuracy: Directions

for Future Research Based on a Systematic

Literature Review
Rachel Phillips, Marc Friberg, Mattias Lantz Cronqvist, Carl-Oscar Jonson and Erik
Prytz

The self-archived postprint version of this journal article is available at Linköping

University Institutional Repository (DiVA):
http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-173322

N.B.: When citing this work, cite the original publication.

Phillips, R., Friberg, M., Lantz Cronqvist, M., Jonson, C., Prytz, E., (2020), Visual Blood Loss
Estimation Accuracy: Directions for Future Research Based on a Systematic Literature Review,
Human Factors and Ergonomics Society Annual Meeting Proceedings, 64(1), 1411-1415.
https://doi.org/10.1177/1071181320641337

Original publication available at:

https://doi.org/10.1177/1071181320641337
Copyright: Human Factors and Ergonomics Society
http://www.hfes.org/
 Visual Blood Loss Estimation Accuracy: Directions for Future
Research Based on a Systematic Literature Review
Rachel Phillips1, Marc Friberg2,3, Mattias Lantz Cronqvist2, Carl-Oscar Jonson3, and Erik Prytz2,3
1
Department of Psychology, Old Dominion University.
2
Department of Computer and Information Science, Linköping University.
3
Center for Disaster Medicine and Traumatology, and Department of Biomedical and Clinical Sciences,
Linköping University.

Visual blood loss estimation occurs in a variety of medical contexts and may impact everything from
interventions by immediate responders to the likelihood of receiving blood transfusions in a hospital
setting. However, research suggests that visual blood loss estimation is inaccurate for laypeople and
medical professionals. The aim of the current study was to conduct a systematic literature review to
determine the current state of knowledge on visual blood loss estimation accuracy and identify directions
for future research. A structured search resulted in 1799 titles that were subsequently screened. A total of
72 articles were coded for comparison. Based on the evaluation, several gaps were identified, most notably
related to factors of the situation that may influence estimation accuracy such as blood flow and
victim/patient gender. Directions for future research are proposed based on identified gaps.

INTRODUCTION equipment. Additionally, in some situations, there may be less

than five minutes to stop a life-threatening bleeding (Tjardes
Blood loss resulting from trauma, childbirth, or a surgical & Luecking, 2018), rendering secondary blood loss estimation
procedure can be life-threatening for the patient depending on techniques impractical. Thus, visual blood loss estimation is
the severity of the bleeding. Understanding how much blood a likely to continue to have a major impact on immediate
trauma victim or patient has lost, and is currently losing, is intervention decisions and associated health-related outcomes.
therefore of great clinical importance. One simple and often- From both a decision-making and educational perspective
employed first method of estimating blood loss is to examine then, it is necessary to understand not only when estimation
the source of the bleeding and the amount of blood present at errors occur, but why, and the factors that may influence
the scene. For example, the Prehospital Trauma Life Support estimation accuracy. To this end, we conducted a literature
(PHTLS) course teaches students to identify “significant review with the intention of determining the current state of
external hemorrhage” and “rapidly scan the patient for any knowledge on visual blood loss estimation accuracy and
sign of severe bleeding” (National Association of Emergency identifying directions for future research.
Medical Technicians, NAEMT, 2020). Visual blood loss
estimation therefore occurs in a variety of medical contexts METHOD
and may impact everything from interventions initiated by
immediate responders at an accident scene to the likelihood of The first step in the systematic literature review process
receiving blood transfusions in a hospital setting. was an exploratory database search using combinations of
However, a plethora of research suggests that visual blood relevant search terms such as “estimating blood loss” or
loss estimation is inaccurate and unreliable (e.g., Adkins et al., “bleeding estimation” and keywords such as “training”,
2014; Beer et al., 2005; Bose et al., 2006; Glover, 2003; “emergency”, “accidents”, “rate”, “flow”, “pool”, and other
Legendre et al., 2016; Roston et al., 2012). Blood loss can be terms deemed relevant to the topic. These searches were
grossly over- or underestimated by both laypeople and conducted in 33 academic databases during 2018. Articles
medical professionals. As this is a known problem, prehospital broadly related to the topic of human ability to visually assess
care providers and medical professionals complement visual blood loss were included in the initial pilot data set of
blood loss estimation with additional blood loss indicators approximately 145 articles. These were subject to a full-text
such as vital signs and other clinical signs that the patient's review and summarized. A structured search was conducted in
body is trying to compensate. However, vital signs may not be February 2020 based on this data set, where six major
an accurate indicator of blood loss either (Little et al., 1995) academic databases (Scopus, PubMed, JSTOR, SAGE,
and there is often a time lag between blood loss and clinical PsycInfo, and ScienceDirect) were searched using the
sign manifestation. Even more accurate blood loss estimation following search string: “("visual estimation" OR "visual
techniques are labor-intensive, time-consuming, and require estimate") AND ("blood loss" OR "bleed" OR "bleeding" OR
specialized equipment, such as the gravimetric method "hemorrhage")”. The results were screened first based on title,
(weighing the blood containing items) or app-based programs then abstract (which included duplication removal), and
that require taking pictures of blood containing items. finally full text. The results of this search and the selection
Therefore, obtaining additional measures to determine process is outlined in Figure 1.
blood loss may not be feasible or available. For example, The primary inclusion criterion was that the article should
untrained laypersons lack the knowledge to use vital signs in try to, in some way, test people’s ability to make visual blood
their evaluation of injury, even if they had access to loss estimates. Publications in languages other than English
were excluded, as were abstract-only sources and posters due 1990 and 1999, 23 published 2000-2009, and 39 published
to the limited information. Also excluded were articles that between 2010 and 2019. Thus, 86.11% of the articles included
primarily focused on the development of new technical in our review were published in the last 20 years. The studies
methods of assessing blood loss and only used visual estimates reported were conducted in a variety of locations: the United
as a baseline comparison for these technologies. The final States of America accounted for 33.33% of the publications,
dataset included 72 articles. followed by the United Kingdom (16.67%), Australia (8.33%),
France (6.94%), and Thailand (5.56%). There were two
articles from Israel, Hong Kong, Nigeria, Sweden, and India
(2.78% each). Finally, Colombia, Singapore, Sri Lanka,
Germany, Ireland, Turkey, Iran, South Africa, and Malaysia
were each represented by a single article (1.39% per location).

Study populations

The populations studied in the articles were primarily

clinical professionals (94.44% of the articles), with an
additional 4.17% incorporating medical and non-medical
participants, and 1.39% exclusively examining laypersons.
The most studied specialty was obstetrics (51.39%), followed
by various sub-specialties of surgery (26.39%), emergency
medicine (15.28%), and those classified as miscellaneous or
unspecified (6.94%).

Stimuli viewing conditions

The way in which the blood loss was presented to the

person, or persons, making the visual estimate differed among
The articles in the final dataset were coded into a the articles. Of the 72 articles included in the analysis, 36.11%
structured table for further analysis. The coding process was (26) reported studies of visual blood loss estimation associated
initially performed individually by four authors on subsets of with live medical events, specifically surgery (46.15%) and
the data set. The coding schemes were then compared, and a childbirth (53.85%). Typically, these estimates were made by
finalized coding scheme was developed based on unanimous clinicians present during the event and later compared to some
agreement. All articles were re-coded using this coding secondary method of blood-loss estimation for an accuracy
scheme. The resultant data was used to calculate the frequency assessment. Thirty-six articles (50%) described some form of
of characteristics or variables in the literature, and to identify experiment that required participants to view stimuli in person.
current knowledge gaps for future research. References in the In these, the amount of blood used to create the stimuli was
reference list that were also included in the analysis are most often known. Finally, 10 of the 72 included articles
marked with an asterisk. A complete list of references used in (13.89%) reported studies that used photographs only (60%), a
the analysis is available from the corresponding author by combination of photographs and videos (20%), or a
request. combination of photographs and live stations (20%).

RESULTS Blood stimuli

An initial assessment of the literature was made to Studies conducted during surgery and childbirth involved
determine if a meta-review of statistical effects would be patient blood mixed with whatever other fluids were present in
possible. Due to the diversity of the methods employed, the the scene. Because the studies involved live patients, the
frequent lack of experimental control, and scarce reporting of amount of blood was uncontrolled and had to be calculated for
statistical results it was concluded that a meta-review was not comparison against the provided estimations. The amount of
possible at this time. The literature was instead narratively blood lost was calculated using 11 different techniques
summarized according to a number of relevant themes. The including the gravimetric method, the alkaline-haematin
themes were derived from frequently occurring topics in the method, and calculations based on physiological indicators.
literature, as well as patterns derived during coding. Within the 36 articles describing studies in which
participants viewed stimuli or scenarios in person, 50% used
Publication data fake blood (33.33% purchased, 44.44% homemade, and
22.22% not specified), 30.56% used expired human blood,
A total of 72 articles were included in the analysis. 8.33% used human blood (whole, reconstituted, or packed red
Articles in the final data set were published between the years cells) but did not specify if it was expired, and 11.11% used
of 1967 and 2019. By decade, the 1960s, 70s, and 80s had one other materials (unspecified, water, chicken, or porcine blood).
article each. There were seven articles published between All studies contained at least two separate volumes of blood,
though the volumes themselves varied greatly, from as little as wards, 30% included at least one nonporous surface like the
.25mL (Simpson et al., 2001) to 4000mL (Roston et al., 2012). floor or mannequin, and 10% incorporated a birthing pool.
Within the ten articles describing studies using
photographs, a combination of photographs and videos, or a Patients or victims
combination of photographs and live stations, 30% used
homemade fake blood and 60% used human blood (fresh for For the 53.85% of studies involving live childbirth, all
one, expired for the rest). There were at least five blood patients were female. For the 46.15% of studies involving
volume conditions in 90% of the articles, the remaining 10% surgery, 58.33% used male and female patients, 25% did not
only had one (Harrison & Cooke, 1999). Volumes ranged specify the gender of the patients, and 16.67% included only
from as little as 5mL (de la Peña Silva et al., 2014) to as much one gender due to the nature of the procedure. For the articles
as 1500mL (Parayre et al., 2015; Pranal et al., 2018). reporting studies based on viewing stations or simulations, a
mannequin or person was present in at least one condition for
Blood flow 22.22% of these. In terms of patient gender, 41.67% of the
studies did not include gender information. In 47.22% of the
Although blood flow was undoubtably present, the rate of articles, the “patient” gender was either specified as female or
blood flow was not mentioned as a consideration during implied to be female due to the nature of the apparatus (such
estimation or measurement in any of the 26 studies conducted as birthing pads). One scenario (2.78%) included both
during live medical events. Blood flow was only mentioned in genders, and three studies (8.33%) included either a person or
one of the live stations (Maslovitz et al., 2008). However, mannequin but the gender of these “patients” was not
despite indicating that blood flow amount and intensity could identified within the article. For the photograph or video-based
be controlled by remote, the rate of flow was not reported or research, 20% of the articles did not report providing gender
further discussed. For the studies using photographs or videos, information, 70% implied female gender, and 10% included
blood flow was only present in the video conditions, and male actors for some of the conditions.
although present, the rate of flow was not specified or
systematically manipulated. Volume estimation units and classification of accuracy

Surfaces and containers For studies involving real cases, participants were
individuals with medical training. Thus, though 46.15% did
In addition to the different viewing conditions used (e.g. not specify the unit of measurement, they were likely
live stations versus photographs) the studies also differed in measured in mL like the 53.85% of articles that did specify the
terms of how the blood was present in the scene. Studies unit of measurement. In terms of accuracy, 84.61% considered
conducted during surgeries or childbirth often included blood any deviation from the measured amount to be inaccurate, and
spilled on the surfaces associated with the respective half of them conducted statistical comparisons. Two studies
procedures. This included a combination of absorbent (7.69%) defined accuracy as within 20% of the measured
materials (such as pads, gauze, and clothing) as well as volume, and 7.69% defined it as being within 400 or 500 ml.
collection devices (such as bed pans, kidney dishes, and For studies involving stations viewed in person, all but
canisters), and impervious surfaces (such as the table or four used people with medical training as participants.
hospital floor). Thus, visual blood loss estimations made in Therefore, although 58.33% of the 36 articles did not
these conditions often required combining information from a specifically state the unit of measurement participants
variety of visual stimuli. estimated in, these estimations were likely made in mL.
For the studies conducted in person using stations or Twelve articles (33.33%) specified that estimations were made
simulations, there was also often a combination of conditions. in mL, two (5.56%) indicated that they let participants choose,
Many studies incorporated conditions that involved a variety and one (2.78%) specified cubic centimeters as the unit of
of stimuli including things like medical sponges, tapes, reporting. Of the four articles that reported patient estimations,
gauzes, pads, swabs, sheets, canisters, cups, drapes, kidney one did not specify the unit used during estimation, one
dishes, and/or bed pans. Additionally, 33.33% of the studies specified mL, and the remaining two allowed participants to
included blood on impervious flooring, 5.56% included blood choose. Interestingly, the classification of accuracy also varied
on concrete, and 8.33% included blood on carpet for at least between studies. Fully 69.44% of the studies considered any
one condition. Blood on clothing was included for at least one deviation from the actual amount to be inaccurate (of these 25
condition in 19.44% of the articles and blood in the toilet or on articles, 28% evaluated the statistical difference between
a sink was present for at least a condition in 11.11%. estimated and actual blood volumes). The remaining 11
For the photograph or video-based research, blood was articles defined accuracy as within 50% (2.78%), within 20%
pictured in or on a variety of surfaces as well. Two of the (19.44%), within 10% (2.78%), within 5% (2.78%), or within
experiments (20%) used non-medical stimuli including 50 mL (2.78%) of the actual volume.
clothing, nonabsorbent surfaces of different colors, carpet of The photograph or video-based studies all involved
different colors, tarmac, patient hair, and grass. Four (40%) participants with medical training. Of the 10 articles included
used only absorbent medical materials including sponges, in our evaluation, 30% specified that the volumes were
gauze pads, and dressings and pads typically used in maternity reported in mL, 20% chose from pre-defined options in mL,
and the remaining 50% provided estimations in an unspecified
format. In terms of accuracy, any deviation from the actual one of the criteria for determining if a trauma requires a
volume was considered inaccurate for 70% of the articles, one tourniquet (Pons & Jacobs, 2017; Goolsby et al., 2018). Yet,
of which also evaluated the differences statistically. in the studies included in our evaluation, blood flow was only
Estimating within 20% of the actual value was considered mentioned in three, as part of the procedure. During an active
accurate for 20% of the articles. The remaining study bleeding event, blood flow may be a prioritized part of the
classified choosing the correct amount (from the provided dynamic estimation process that needs to occur to determine
options) or higher as accurate. the most appropriate first aid action. Unfortunately, at this
time, the interaction between estimated blood flow and visual
DISCUSSION blood loss estimation is unclear. If the speed of blood flow
alters the perceptions of the amount of blood (or vice versa),
Overall, it is clear that the methods used to investigate then educational interventions may be necessary to reduce any
visual estimates of blood loss are varied. Over a third of the deleterious effects associated with this interaction.
studies were conducted during real surgeries or birth. This is Another potentially important variable that has not been
beneficial as it increases the potential generalizability of systematically studied to date is victim or patient gender.
results. It also indicates that the method itself is considered to Patient gender was either not present or not specified in over a
be of clinical relevance, and that there are problems related to third of the articles included in our evaluation. Additionally,
the method in natural settings. However, the findings may be no article included patient gender as a between-subjects
hospital or situation specific and it is complicated to account variable when evaluating estimation accuracy. Thus, the extent
for potential confounds inherent to the environment. to which visual blood loss estimation accuracy is impacted by
Furthermore, the nature of live surgical or birthing events victim or patient gender is unclear. Previous research has
limits the ability to systematically examine factors that may be found gender related differences in prehospital care (Koval et
contributing to the over or underestimation of blood loss. al., 2006; Wahlin et al., 2016) and the likelihood of surviving
Over half of the studies were conducted using more great vessel trauma (Chen et al., 2019). If victim gender
controlled conditions where such factors could be impacts visual blood loss estimation accuracy, then it will
systematically investigated. However, these studies were not likely also impact the evaluations of injury severity and,
without their limitations. A general shortcoming is the lack of therefore, could influence prehospital care-related decisions
theoretically guided exploration. Not a single study found in and subsequent treatment. Such bias, should it exist, would
the current literature review relied on a theory of human need to be addressed to ensure all genders were receiving
perception or cognition in any form. Instead, the experimental equitable care at every stage post injury.
designs were often based on stimuli of clinical relevance to a Finally, what was considered an “accurate” estimate
particular subspecialty of medical professional (e.g., the use of varied across studies and the use of statistical analyses related
blood on delivery pads within obstetrics or blood on the scalp to accuracy were limited. Most articles conclude that visual
for trauma). As such, there has not been a concerted effort to estimates are “inaccurate” to some degree (e.g., Adkins et al.,
systematically compare, for instance, estimation accuracy 2014; Bose et al., 2006), generally an underestimate (e.g.,
across stimuli or surface type (with two exceptions: Tall et al., Beer et al., 2005; Glover, 2003; Roston et al., 2012), or over
2003, and Williams & Boyle, 2006). This is problematic as the or underestimate depending on volume (e.g., Legendre et al.,
size of a blood pool can be impacted by characteristics of the 2016). This must, however, be interpreted in relation to the
surface on which blood loss occurred (Adair & Gallardo, definitions of accuracy and experimental setups used in these
1999; Feng et al., 2018; Kreutziger et al., 2014) and studies. Though many articles marked any deviation from the
estimations may be influenced by characteristics of the stimuli true amount as inaccurate, others used either relative errors
themselves, such as shape, size, or color. Several studies (such as 5, 10, 20, or 50%) or absolute errors (such as 50 mL
varied both volume and medium of the stimuli simultaneously, or 500 mL). This makes it difficult to assess the quality of the
e.g. spilling 500 mL of blood on a cot, 25 mL on a 10-pack of results, particularly given the lack of experimental control in
4x4-inch gauze, 100 mL on a T-shirt, and 150 mL poured into about half the studies and the presence of confounds (e.g.,
a “commode filled with water” (Asburn et al., 2012). varying both volume and medium at the same time), in many
Buckland and Homer (2007) used, among others, 1000 mL of the studies using an experimental setup.
poured on sheets, 600 mL in a kidney dish (which had volume
markings), and 350 mL in a bedpan and concluded that visual CONCLUSIONS
estimates in containers were more accurate. The few studies
based on photographs or videos contained similar confounds The literature on visual blood loss estimations is mostly
that were not addressed and many failed to include based on studies within clinical populations, with about a third
standardized lighting or camera angles. Ultimately, this means featuring some form of live medical procedure and about two
that the extent to which the surface, blood stimulus, viewing thirds some form of experimental setup. The experimental
angle, clothing, background contrast, or container shape setups seem to have been mainly derived from clinical cases
impact blood loss estimation accuracy is not known. rather than guided by theory or an attempt to systematically
Another notable gap in the current knowledge is how the vary potentially relevant variables that could impact blood loss
flow of blood impacts visual blood loss estimations. Blood estimation accuracy. The literature review therefore indicates
flow is one of the components used to assess the severity of a need for rigorous studies of accuracy and identifying sources
surgical bleeding (Lewis et al., 2017; Spotnitz et al., 2018) and of errors for visual blood loss estimation.
 In terms of variables, blood flow, surface or container European Journal of Emergency Medicine, 21(5), 360-363. doi:
10.1097/MEJ.0000000000000092
types, laypeople populations, patient gender, and the effect of *Legendre, G., Richard, M., Brun, S., Chancerel, M., Matuszewski, S., &
specific training interventions are some of the more relevant in Sentilhes, L. (2016). Evaluation by obstetric care providers of
need of further investigation. Blood loss volume is typically simulated postpartum blood loss using a collector bag: A French
the main independent variable in past research, but the relation prospective study. The Journal of Maternal-Fetal & Neonatal
Medicine, 29(21), 3575-3581.
of volume and accuracy is still poorly understood. This is in https://doi.org/10.3109/14767058.2016.1139569
part because of the wide variety of accuracy definitions used, Lewis, K. M., Li, Q., Jones, D. S., Corrales, J. D., Du, H., Spiess, P. E.,
and in part because of the lack of systematic experimental Menzo, E. L., & DeAnda, A. (2017). Development and validation of an
investigation. intraoperative bleeding severity scale for use in clinical studies of
hemostatic agents. Surgery, 161, 771-781.
Inadequate understanding of errors at the evaluation stage
http://dx.doi.org/10.1016/j.surg.2016.09.022
may result in the development of faulty intervention and Little, R. A., Kirkman, E., Driscoll, P., Hanson, J., & Mackway-Jones, K.
training programs for laypersons or students early in their (1995). Preventable deaths after injury: Why are the traditional vital
medical careers. Further experimental research would enable signs poor indicators of blood loss? Journal of Accident & Emergency
Medicine, 12(1), 1-14.
the development of training interventions, guides, and *Maslovitz, S., Barkai, G., Lessing, J. B., Ziv, A., & Many, A. (2008).
technology that could increase the accuracy of estimations and Improved accuracy of postpartum blood loss estimation as assessed by
have a clinical impact on improving medical care both within simulation. Acta Obstetricia et Gynecologica Scandinavica, 87(9),
and outside of the hospital. 929–934. https://doi.org/10.1080/00016340802317794
National Association of Emergency Medical Technicians (NAEMT), Pre-
Hospital Trauma Life Support Committee, & American College of
ACKNOWLEDGMENTS Surgeons Committee on Trauma. (2020). PHTLS: Prehospital Trauma
Life Support. Jones & Bartlett Publishers.
This study was financed in part by the Swedish Civil *Parayre, I., Rivière, O., Debost-Legrand, A., Lémery, D., & Vendittelli, F.
(2015). Reliability of student midwives’ visual estimate of blood loss in
Contingencies Agency. the immediate postpartum period: A cross-sectional study.
International Journal of Nursing Studies, 52(12), 1798–1803.
REFERENCES http://dx.doi.org/10.1016/j.ijnurstu.2015.06.015
*de la Peña Silva, A. J., Delgado, R. P., Barreto, I. Y., & Martínez, M. D. L.
Adair, T. W., & Gallardo, A. C. (1999). Considering the target surface in P. (2014). Is visual estimation useful in determining the extent of
bloodstain pattern analysis: An unusual case of blood pooling. Journal perioperative haemorrhage? A study of correlation among anaesthetists
of Forensic Identification, 49(5), 485-493. of intermediate and high complexity hospitals in Cartagena, Colombia.
*Adkins, A. R., Lee, D., Woody, D. J., & White, W. A. (2014). Accuracy of Colombian Journal of Anesthesiology, 42(4), 247-254.
blood loss estimations among anesthesia providers. AANA Journal, Pons, P. T., & Jacobs, L. (2017). Stop the Bleed. Save a life: What everyone
82(4), 300-306. should know to stop bleeding after an injury. Heartford Consensus.
*Ashburn, J. C., Harrison, T., Ham, J. J., & Strote, J. (2012). Emergency Bleedingcontrol.org
physician estimation of blood loss. Western Journal of Emergency *Pranal, M., Guttmann, A., Ouchchane, L., Parayre, I., Rivière, O., Leroux,
Medicine, 13(4), 376-379. doi: 10.5811/westjem.2011.9.6669 S., ... & Vendittelli, F. (2018). Do estimates of blood loss differ
*Beer, H. L., Duvvi, S., Webb, C. J., & Tandon, S. (2005). Blood loss between student midwives and midwives? A multicenter cross-
estimation in epistaxis scenarios. Journal of Laryngology and Otology, sectional study. Midwifery, 59, 17-22.
119(1), 16–18. https://doi.org/10.1258/0022215053222752 https://doi.org/10.1016/j.midw.2017.12.017
*Bose, P., Regan, F., & Paterson-Brown, S. (2006). Improving the accuracy of *Roston, A. B., Roston, A. L., & Patel, A. (2012). Blood Loss: Accuracy of
estimated blood loss at obstetric haemorrhage using clinical visual estimation. In S. Arulkumaran, M. Karoshi, L. G. Keith, A. B.
reconstructions. BJOG: An International Journal of Obstetrics and Lalonde & C. B-Lynch (Eds.) A Comprehensive Textbook of
Gynaecology, 113(8), 919–924. doi: 10.1111/j.1471- Postpartum Hemorrhage: An essential Clinical Reference for Effective
0528.2006.01018.x Management 2nd Edition (pp. 71-72). Sapiens Publishing.
*Buckland, S. S., & Homer, C. S. (2007). Estimating blood loss after birth: *Simpson, R. R., Kennedy, M. L., Chew, S. B., & Lubowski, D. Z. (2001). Do
Using simulated clinical examples. Women and Birth, 20(2), 85-88. patients assess rectal bleeding accurately? ANZ Journal of Surgery,
doi:10.1016/j.wombi.2007.01.001 71(11), 650–651. https://doi.org/10.1046/j.1445-1433.2001.02232.x
Chen, S., Huang, J., Tee, Y., Chen, S., Wang, S., Fu, C., Hsieh, C., & Liao, C. Spotnitz, W. D., Zielske, D., Centis, V., Hoffman, R., Gillen, D. L.,
(2019). Contemporary management and prognosis of great vessels Wittmann, C., ... & McAfee, P. C. (2018). The SPOT GRADE: A new
trauma. Injury, 50, 1202-1207. method for reproducibly quantifying surgical wound bleeding. Spine,
https://doi.org/10.1016/j.injury.2019.03.054 43(11), E664-E671. doi: 10.1097/BRS.0000000000002447
Feng, C., Michielsen, S., & Attinger, D. (2018). Impact of carpet construction *Tall, G., Wise, D., Grove, P., & Wilkinson, C. (2003). The accuracy of
on fluid penetration: The case of blood. Forensic Science International, external blood loss estimation by ambulance and hospital personnel.
284, 184-193. https://doi.org/10.1016/j.forsciint.2018.01.009 Emergency Medicine, 15(4), 318-321.
*Glover, P. (2003). Blood loss at delivery: How accurate is your estimation? Tjardes, T., & Luecking, M. (2018). The platinum 5 min in TCCC: Analysis
Australian Journal of Midwifery, 16(2), 21–24. of junctional and extremity hemorrhage scenarios with a mathematical
Goolsby, C., Jacobs, L., Hunt, R. C., Goralnick, E., Singletary, E. M., Levy, model. Military Medicine, 183(5-6), e207-e215. doi:
M. J., Goodloe, J. M., Epstein, J. L., Strauss-Riggs, K., Seitz, S. R., 10.1093/milmed/usx016
Krohmer, J. R., Nemeth, I., Rowe, D. W., Bradley, R. N., Gestring, M. Wahlin, R. R., Ponzer, S., Lӧvbrand, H., Skrivfars, M., Lossius, H. M., &
L., & Kirsch, T. D. (2018). Stop the Bleed Education Consortium: Castrén, M. (2016). Do male and female trauma patients receive the
Education program content and delivery recommendations. Journal of same prehospital care? An observational follow-up study. BMC
Trauma and Acute Care Surgery, 84(1), 205–210. Emergency Medicine, 16(6), 1-9. doi: 10.1186/s12873-016-0070-9
*Harrison, J., & Cooke, M. W. (1999). Visual assessment of blood loss by *Williams, B., & Boyle, M. (2007). Estimation of external blood loss by
accident and emergency staff. Journal of Accident & Emergency paramedics: Is there any point? Prehospital and Disaster Medicine,
Medicine, 16(5), 390. https://doi.org/10.1136/emj.16.5.390-a 22(6), 502-506. https://doi.org/10.1017/S1049023X0000532X
Koval, K. J., Tingey, C. W., & Spratt, K. F. (2006). Are patients being
transferred to level-1 trauma centers for reasons other than medical
necessity? The Journal of Bone and Joint Surgery, 88(10), 2124-2132.
Kreutziger, J., Haim, A., Jonsson, K., Wenzel, V., Stark, M., & Nussbaumer,
W. (2014). Variation in size of blood puddles on different surfaces.