Postpartum hemorrhage: measure blood loss and treat early

In a busy maternity ward, a woman has just had her second child in an uncomplicated vaginal delivery. The placenta came out complete, the baby is in her arms and the team is already preparing for the next admission. Twenty minutes later, the nursing technician changes the bed pad and comments that she "bled a little more than usual". The blood pressure is still good, the woman is talking and says she is just tired. Nobody weighed anything, nobody measured anything, and the general impression is that everything is under control. The scene is fictional and was written for teaching. This text uses that picture as a setting for clinical reasoning: why the eye is fooled in postpartum hemorrhage, how to measure blood loss objectively and why treatment works better when it is triggered as a package, all at once, and not as a sequence of attempts.

NEXO is a daily medical diagnosis game: one clinical case a day, up to six progressive clues, and ICD-10 proximity feedback across five levels.

The text is not a treatment protocol. Drugs, doses, routes and escalation pathways follow national guidelines and each service's training. The interest here is the reasoning: why late detection is the most common weak point and why a simple step, such as placing a graduated collector under the woman, changes the decision. If you play the case of the day on NEXO, you will recognize the move: the clue that looks small, once measured, changes the hypothesis and the urgency.

Key points

• Visual estimation of blood loss usually underestimates bleeding, and the error grows the greater the loss.
• The World Health Organization recommends objectively measuring blood loss in every birth.
• The treatment package brings together uterine massage, a uterotonic, tranexamic acid, intravenous fluids, examination and escalation, started together.
• Tranexamic acid works best when given as early as possible, ideally within the first hours.
• Delays in recognizing, deciding and acting are the most frequent pattern when hemorrhage ends badly.

Why postpartum hemorrhage is a reasoning setting

Postpartum hemorrhage is traditionally defined as blood loss of half a liter or more in the first hours after birth. It remains one of the leading causes of maternal death worldwide, and a good part of those deaths is considered preventable. The difficulty is rarely in knowing what to do. It is in realizing in time that something must be done.

Coomarasamy and colleagues described, in a recent review, postpartum hemorrhage as a succession of possible delays: in recognizing the bleeding, in calling for help, in starting treatment, in escalating, in transferring and in transfusing (Coomarasamy and colleagues, 2026). Each delay looks small on its own. Added together, they turn a treatable complication into an emergency.

What the scene teaches

The scene has three weak clues. The first is the phrase "bled a little more than usual", which is a visual estimate. The second is the still normal blood pressure, which reassures the team but does not rule out a significant loss, because many young, healthy women compensate for a long time. The third is the tiredness, which may be just the effort of childbirth or the start of hypovolemia. None of the three is conclusive. What is missing is a number.

The post on differential diagnosis shows how weak clues, added together, change the order of the hypotheses. Here the logic is the same, with one difference: the most important clue can be measured.

The eye is fooled: the problem of visual estimation

Estimating blood loss by looking at sheets, compresses and the floor of the room is the oldest practice and still very common. It is also unreliable. Natrella and colleagues compared visual estimation with measurement and summed up the finding in one sentence: the more that is lost, the more that goes unseen (Natrella and colleagues, 2018). Small losses tend to be overestimated, and large losses underestimated, precisely when the error costs more.

The Cochrane review by Yunas and colleagues assessed the accuracy of diagnostic methods for postpartum hemorrhage and concluded that visual estimation misses about half of diagnoses (Yunas and colleagues, 2025). The same review found high sensitivity and specificity when a calibrated collector was used together with observation of vital signs.

Why the error is systematic

Blood spreads over different surfaces, mixes with amniotic fluid and urine, is retained in pads and clots, and runs out of the field of view. Those in the room are also busy with the baby, the placenta and conversation. In addition, the team tends to anchor on the first impression of an uneventful birth. The post on examples of heuristics in clinical reasoning shows how shortcuts that are useful day to day become traps when the situation changes.

An earlier Cochrane review, by Diaz, Abalos and Carroli, found insufficient evidence to prefer one estimation method over another for clinical outcomes (Diaz and colleagues, 2018). The conclusion seemed frustrating at the time. It changed with the E-MOTIVE trial, which tested measurement not as an end in itself but as a trigger for treatment.

Measure instead of estimate

Objective measurement can be done in two main ways. The volumetric method uses a graduated plastic collector placed under the buttocks right after birth, which collects the blood and allows the volume to be read directly. The gravimetric method weighs pads and drapes and subtracts the dry weight. Both require routine and training, but neither depends on expensive technology.

The 2023 World Health Organization recommendations for the diagnosis and treatment of postpartum hemorrhage call for routine objective measurement of blood loss in all births, not only when someone already suspects excessive bleeding (World Health Organization, 2023). The logic is simple: if measurement only starts when someone notices the problem, it inherits the delay of visual estimation.

The role of vital signs

Blood pressure, heart rate and the shock index, which is the ratio of heart rate to systolic pressure, help assess the impact of the loss. But they do not replace measurement. In the review by Yunas and colleagues, a high shock index had low sensitivity as a standalone criterion (Yunas and colleagues, 2025). In other words, when the shock index rises, the loss is usually already significant. Waiting for it is waiting too long.

That is why the recommendation combines measurement and warning signs. The bundle should be triggered when the measured loss reaches half a liter or when it reaches about three hundred milliliters accompanied by abnormal vital signs or other warning signs (World Health Organization, 2023).

What makes measurement feasible in routine care

Measurement only works if the collector is placed before the bleeding starts, that is, right after the baby is born, and if someone is responsible for checking the volume at agreed times. Without this routine, the collector becomes just another object on the bed. It is also necessary to decide what to do with amniotic fluid and urine, which can inflate the reading at the start, and to agree on when the collector will be removed.

Another condition is that the number has consequences. If the technique announces that the threshold has been reached and no one reacts, the team quickly learns that measuring changes nothing. That is why measurement and the bundle were tested together: one without the other loses much of the effect.

The table: visual estimation and objective measurement

The comparison below summarizes what changes when the team starts measuring.

Aspect	Visual estimation	Objective measurement
How it is done	Impression of whoever looks at sheets, pads and floor	Graduated collector under the woman or weighing of pads
Typical error	Underestimates large losses, overestimates small losses	Smaller, as long as the collector is placed right after birth
Timing of diagnosis	Tends to be late, when signs of shock appear	Earlier, with a clear threshold for action
Trigger for treatment	Depends on individual perception	Defined volume, with or without warning signs
Record	Vague, hard to compare	A number the whole team can read
Use in audit	Of little use for learning	It lets you review how long each step took

The table does not say that clinical signs stop mattering. It says that the number changes the conversation between the people in the room. A technique that says "the collector already marks almost half a liter" is heard differently from one that says "I think there was a lot of bleeding".

The E-MOTIVE trial and the idea of a bundle

The E-MOTIVE trial, published by Gallos and colleagues, randomized hospitals in four African countries to use or not use a two-part strategy (Gallos and colleagues, 2023). The first was early detection with a calibrated collector. The second was a treatment bundle triggered all at once: uterine massage, uterotonic, tranexamic acid, intravenous fluids, examination of the genital tract and escalation when needed.

The results were strong. The composite outcome of severe hemorrhage, laparotomy or maternal death from bleeding fell substantially in the intervention group. Detection of hemorrhage went from about half of cases to more than nine in ten, and the bundle came to be applied in the great majority of cases, against a minority in the control group. The trial did not test a new drug. It tested a different way of organizing what was already known.

Why a bundle and not a sequence

Traditional practice used to work in steps: massage, wait; uterotonic, wait; another uterotonic, wait. Each wait seems prudent, but it consumes minutes of active bleeding. The bundle reverses the logic. The first measures start together, and reassessment comes afterward, with the decision to escalate if the bleeding does not stop.

The 2023 World Health Organization recommendation incorporated this idea: the bundle should be started as soon as hemorrhage is diagnosed, preferably within fifteen minutes, and those who do not respond should move on to additional measures (World Health Organization, 2023). The post on clinical decision-making discusses why deciding early, with clear criteria, is usually better than waiting for certainty.

What the trial does not answer

Like every trial, E-MOTIVE has limits. It was done in vaginal births, in hospitals with some level of infrastructure, and with implementation support that not every service will have. It also does not separate the effect of each component, because the goal was precisely to test the whole. Even so, the result is consistent with what was known about timing: treating early, and with several measures at the same time, makes a difference when bleeding is active. Services in other contexts can adapt the model, keeping the two central principles: measure and act together.

The components of the bundle

Each component of the bundle addresses a possible cause or a consequence of the bleeding. The order below is for teaching; in practice, several people do several things at the same time.

Uterine massage. Atony is the most common cause of postpartum hemorrhage. Massaging the uterine fundus helps the uterus contract while the drugs start to act.

Uterotonic. Oxytocin remains the most studied first-line drug. The network review by Parry Smith and colleagues on first-line treatment concluded that oxytocin is probably more effective than misoprostol, with fewer adverse effects, and that adding misoprostol to oxytocin makes little or no difference (Parry Smith and colleagues, 2020).

Tranexamic acid. It is an antifibrinolytic that enters the bundle from the start, discussed in the next section.

Intravenous fluids. Large-bore venous access and volume replacement help maintain perfusion while the bleeding is controlled. Blood collection for typing and tests is usually done at the same moment.

Examination. Reviewing the uterus, placenta and genital tract looks for causes that drugs do not resolve, such as lacerations, retained placental tissue or coagulation changes. The classic reasoning about causes, summarized as tone, trauma, tissue and thrombin, helps keep any of them from being forgotten.

Escalation. If the bleeding continues, the team calls for experienced help, considers other uterotonics, balloon tamponade, surgical procedures, transfusion and transfer. The bundle is not the end of treatment; it is the organized beginning.

Who does what

In active bleeding, the question "who is going to do this?" costs time. Teams that train with defined roles, such as who does compressions, who prepares the drugs, who handles access and who records, tend to act faster. The post on interprofessional clinical reasoning shows how shared decisions depend on a common language between professions.

Tranexamic acid: the clock matters

The WOMAN trial randomized women with postpartum hemorrhage to receive tranexamic acid or placebo, in addition to usual care (WOMAN Trial Collaborators, 2017). Tranexamic acid reduced death from bleeding, with a clearer benefit when given in the first three hours after delivery. It did not reduce the need for hysterectomy and did not increase thromboembolic events.

The finding about timing is what matters most to reasoning. A drug whose benefit decreases with delay cannot be kept for when "nothing else worked". It needs to come in early, along with the other measures. That is why the 2023 World Health Organization recommendation includes tranexamic acid in the initial package, and not at a later step (World Health Organization, 2023).

A common reasoning error

There is a temptation to save the "strongest" drug for later, as if using it early were excessive. This reasoning makes sense for interventions with high risks, but not for a safe measure whose benefit depends on the clock. The post on diagnostic uncertainty discusses how to weigh the cost of acting early against the cost of waiting.

Prevention: what happens before the bleeding

The best treatment for postpartum hemorrhage starts before it. Using a prophylactic uterotonic right after birth is one of the most effective preventive measures. The CHAMPION trial, by Widmer and colleagues, showed that heat-stable carbetocin was not inferior to oxytocin in preventing loss of half a liter or more (Widmer and colleagues, 2018). The detail matters in places where keeping oxytocin refrigerated is difficult, because a degraded drug protects no one.

The Cochrane network review by Gallos and colleagues on prevention compared several uterotonics and found that combinations such as ergometrine with oxytocin and misoprostol with oxytocin may be more effective than oxytocin alone, but with more adverse effects, while carbetocin had an adverse effect profile similar to oxytocin (Gallos and colleagues, 2025). The choice depends on the context, the resources and each woman's contraindications.

Risk factors are not enough

Multiple pregnancy, prolonged labor, hemorrhage in a previous delivery and anemia increase the risk. But many hemorrhages happen in women with no known risk factor, like the one in the scene. That is why prevention and measurement are universal. Classifying someone as "low risk" should not reduce vigilance. The post on base rate in diagnosis shows how the frequency of a condition in the population should weigh more than the impression about an individual case.

The delays and how the team reduces them

The review by Coomarasamy and colleagues organizes the problem into successive delays, and each of them has a practical antidote (Coomarasamy and colleagues, 2026). The delay in recognizing decreases with objective measurement. The delay in acting decreases with the package and with defined roles. The delay in escalating decreases with clear criteria for when to call for help. The delay in transfusing decreases with workflows coordinated with the blood bank.

The 2023 World Health Organization recommendation also describes implementation strategies: audit and data feedback for the team, reference professionals who lead the change in each service and a postpartum hemorrhage cart that is always restocked (World Health Organization, 2023). None of them is sophisticated. All of them treat the problem as a team process, and not as a failure of whoever was on shift.

Learning from the cases that went wrong

When a hemorrhage ends in hysterectomy or death, it is easy to look back and say it "was obvious". The post on hindsight bias and outcome bias explains why that judgment is unfair and of little use. Structured review sessions, like those described in the post on morbidity and mortality sessions, work better when they ask how long each step took and what would have shortened that time.

Early closure and false reassurance

Croskerry described early closure as the tendency to accept a diagnosis before it has been verified (Croskerry, 2003). In postpartum hemorrhage, early closure often takes the form of "calm delivery, all fine". The first impression is good, and every new clue is read in its light: the normal blood pressure confirms it, the tiredness is expected, the soaked pad is "a little more than normal".

Objective measurement works as an interruption of this process. It does not depend on anyone changing their mind; it shows a number that forces a reconsideration. The post on diagnostic pause describes another way to interrupt early closure: asking, on purpose, what else could be happening.

Calibrating confidence

Experienced professionals often trust their own visual estimate a great deal, and that is understandable: they have seen many deliveries. But experience does not correct a systematic error of the human eye. The post on calibration of confidence shows how comparing the impression with the measured result is the fastest way to adjust one's own confidence.

How NEXO helps train this reasoning

In NEXO, the case of the day is fictional, with a real clinical presentation and up to six progressive clues. You propose a diagnosis by looking for the ICD-10 code and receive proximity feedback on five levels, which shows whether the guess landed on the exact ICD, the same category, the same group, the same chapter or a different chapter. This format speaks to the theme of this text: a clue that seems trivial can change the urgency and the course of the case.

To use the case of the day as training, try three habits. Before each guess, ask which clue could be measured instead of estimated. When the picture seems calm, ask what would make you change your mind. After the result, note at what point a number would have anticipated the decision.

Three exercises with cases

Swap impression for number. Faced with a vague clue, such as "bled a lot", ask which measurement would make that clue objective.

Think in a bundle. In emergency cases, list the actions that can start together instead of waiting for one at a time.

Count the delays. When reviewing a case, mark every point where the team waited and ask what would shorten that wait.

The post on learning based on clinical cases shows how cases help fix patterns that require a quick response.

The limits on using the game are described on the NEXO economy page. If you research the teaching of clinical reasoning, the page for researchers explains how to get in touch. Professionals who write or review cases find information under curation and under creators, and institutions can learn about the proposal on the partnerships page.

Frequently asked questions

Does visual estimation usually underestimate blood loss?

Yes. Studies show that the error grows the greater the loss, and the 2025 Cochrane review estimated that visual estimation misses about half of the diagnoses.

Should objective measurement be done in every delivery?

Yes. The 2023 World Health Organization recommendation calls for routine objective measurement, not only when excessive bleeding is suspected.

Should the treatment bundle start all at once?

Yes. Uterine massage, a uterotonic, tranexamic acid, intravenous fluids and examination start together, and escalation follows if the bleeding continues.

Does tranexamic acid work better when given early?

Yes. In the WOMAN trial, the benefit for death from bleeding was clearer in the first three hours after delivery.

Can women without risk factors also have postpartum hemorrhage?

Yes. Many hemorrhages happen with no known risk factor, which is why prevention and measurement apply to every delivery.

Does training the team help reduce delays?

Yes. Defined roles, a restocked cart, audit and data feedback are implementation strategies described by the World Health Organization.

Final checklist

• I placed the graduated collector right after birth.
• I recorded the measured loss, not just the impression.
• I watched blood pressure, heart rate and warning signs along with the volume.
• I triggered the bundle when the agreed threshold was reached.
• I started massage, a uterotonic, tranexamic acid, fluids and examination at the same time.
• I reviewed tone, trauma, tissue and thrombin.
• I defined who does what and called for help early.
• I escalated when the bleeding did not stop.
• I reviewed the case afterwards, counting how long each step took.

Want to practice the reasoning with progressive clues? Play the case of the day.

References

1. Coomarasamy A, Devall AJ, Bell S, et al. Diagnosis and treatment of postpartum haemorrhage: a race against time. Lancet. 2026. PMID 42285120.
2. Croskerry P. The importance of cognitive errors in diagnosis and strategies to minimize them. Acad Med. 2003. PMID 12915363.
3. Diaz V, Abalos E, Carroli G. Methods for blood loss estimation after vaginal birth. Cochrane Database Syst Rev. 2018. PMID 30211952.
4. Gallos I, Devall A, Martin J, et al. Randomized trial of early detection and treatment of postpartum hemorrhage. N Engl J Med. 2023. PMID 37158447.
5. Gallos ID, Yunas I, Devall AJ, et al. Uterotonic agents for preventing postpartum haemorrhage: a network meta-analysis. Cochrane Database Syst Rev. 2025. PMID 40237648.
6. Natrella M, Di Naro E, Loverro M, et al. The more you lose the more you miss: accuracy of postpartum blood loss visual estimation. A systematic review of the literature. J Matern Fetal Neonatal Med. 2018. PMID 28002983.
7. World Health Organization. WHO recommendations on the assessment of postpartum blood loss and use of a treatment bundle for postpartum haemorrhage. 2023. Document.
8. Parry Smith WR, Papadopoulou A, Thomas E, et al. Uterotonic agents for first-line treatment of postpartum haemorrhage: a network meta-analysis. Cochrane Database Syst Rev. 2020. PMID 33232518.
9. Widmer M, Piaggio G, Nguyen TMH, et al. Heat-stable carbetocin versus oxytocin to prevent hemorrhage after vaginal birth. N Engl J Med. 2018. PMID 29949473.
10. WOMAN Trial Collaborators. Effect of early tranexamic acid administration on mortality, hysterectomy, and other morbidities in women with post-partum haemorrhage (WOMAN): an international, randomised, double-blind, placebo-controlled trial. Lancet. 2017. PMID 28456509.
11. Yunas I, Gallos ID, Devall AJ, et al. Tests for diagnosis of postpartum haemorrhage at vaginal birth. Cochrane Database Syst Rev. 2025. PMID 39821088.