Targeting High-Risk Pregnancies: A Cost-Effectiveness Case for Low-Dose Aspirin

A century-old pill, given to the right people, could cost-effectively save many lives.

Summary

• Targeting low-cost prevention toward people at elevated risk may uncover highly cost-effective ways to address noncommunicable diseases. Hypertensive disorders in pregnancy could be a test case for this approach.
• Meta-analyses indicate that low-dose aspirin reduces perinatal death by ~20% when given to women at risk of preeclampsia, a pregnancy complication involving high blood pressure. Aspirin likely also provides other benefits to mothers and babies, and the World Health Organization recommends this treatment for women at moderate or high risk.
• Safety evidence is reassuring but incomplete. Serious harms appear uncommon, although small increases in specific maternal morbidities may occur.
• India illustrates a large evidence-to-practice gap. Aspirin is inexpensive and early antenatal care is common, yet limited studies suggest that many higher-risk women are not treated. Observational implementation studies suggest that integrating screening, treatment, and follow-up can substantially improve uptake.
• My basic model suggests that a program driving uptake of low-dose aspirin among high-risk pregnant women could meet GiveWell’s funding bar if its total cost, including screening and delivery, is roughly $32 or less per high-risk pregnancy treated. That target looks feasible when compared to available benchmarks.
• This intervention could be bundled with other antenatal care or hypertension treatment quality improvement programs to increase their impacts and spread out costs.
• This is a rapid, shallow investigation. I may have missed compelling negative evidence or reasons this will not work.

Background

Many low- and middle-income countries are undergoing a demographic transition, with longer life expectancy and lower fertility. At the same time, the burden of disease is shifting toward noncommunicable diseases (NCDs), such as heart disease and diabetes. This creates a growing need to find cost-effective ways to address NCDs. One promising strategy is to target low-cost preventive interventions toward people at elevated risk, concentrating resources where the expected benefit per person treated is greatest. This search is timely because GiveWell and Coefficient Giving have recently lowered their cost-effectiveness bars as they scale their giving, creating opportunities to support programs that might not previously have cleared their funding thresholds.

Pregnancy is a particularly promising setting for this approach. It creates a defined period when elevated risk can be identified, preventive treatment can be delivered, and an intervention could benefit both mother and baby. Here, I look into hypertensive disorders in pregnancy as a proof of concept that targeting preventive NCD care to high-risk populations can be highly cost-effective.

The evidence on aspirin and preeclampsia

Preeclampsia is a pregnancy complication in which high blood pressure from 20 weeks onward is accompanied by protein in the urine, signs that the mother’s organs are not functioning normally, or signs that the placenta is not functioning normally. Preeclampsia affects an estimated 3-8% of women who give birth worldwide and can cause serious illness in the mother, restrict the baby’s growth, or make an early delivery necessary. 

It also contributes to the global burden of perinatal death.^[1] Approximately 1.9 million babies are stillborn each year, while about 2.3 million babies born alive die during their first month of life. Though many factors contribute to this burden of perinatal death, the association with preeclampsia is substantial. One study across low- and middle-income countries found that preeclampsia was associated with nearly twice the odds of perinatal death.

Low-dose aspirin during pregnancy can reduce these risks. A systematic review of randomized trials among women at increased risk found that aspirin reduced preeclampsia by 15%, preterm birth by 20%, and perinatal death by 21%. A Cochrane review reached similar conclusions.^[2] Although much of this evidence comes from high-income settings, a randomized trial involving nearly 12,000 first-time mothers in six low- and middle-income countries also found that low-dose aspirin reduced preterm birth and perinatal death. 

The evidence on safety of low-dose aspirin in pregnancy is reassuring, but not conclusive. The main concerns are postpartum hemorrhage (heavy maternal bleeding after delivery) and placental abruption (placental separation from the uterus before delivery). Randomized trials have not demonstrated uniform effects on these complications, but they have low base rates and inconsistent reporting. The Cochrane review found a possible small increase in postpartum bleeding and could not rule out an increased risk of placental abruption. While serious harm appears uncommon, the evidence does not conclusively show that aspirin is risk-free.

The World Health Organization (WHO) recommends 75 mg of aspirin daily for women at moderate or high risk of developing preeclampsia. As the global normative health authority, WHO has weighed the evidence and issued a clear recommendation. It did so despite recognizing the remaining uncertainty about potential maternal harms. This indicates its judgment that, for women at moderate or high risk, the substantial expected benefits justify those risks.

A large evidence-to-practice gap in low- and middle-income countries

India offers one example of how this well-supported clinical recommendation can feasibly be implemented, but also hasn’t yet become widespread in preventive care. Community-based research confirms that preeclampsia and related hypertension affect a substantial number of pregnancies in India. Yet recent studies suggest that many pregnant women at increased risk do not take aspirin.

At one public hospital, a study showed that only about one in nine women classified as high risk started aspirin. A smaller study also found that aspirin was not routinely started for women whose early records indicated high risk. Notably, these studies were conducted in hospital settings: true rates of appropriate low-dose aspirin therapy in risky pregnancies are likely even lower in the general population. 

The health system already has many of the components needed to enable low-cost programs to increase uptake of this lifesaving intervention.

• Aspirin is an inexpensive, widely available generic medicine.
• Low-dose aspirin is most effective when started early in at-risk pregnancies. India already reaches many pregnant women during this window: according to a recent national health survey, 76% of mothers had an antenatal check-up during the first trimester for their most recent birth.
• WHO risk criteria rely mainly on pregnancy history and common medical conditions, although some more specialized testing (using ultrasound) is a component of some risk algorithms. 

The gap is not inevitable. In an intensive rural program in India, a large majority of women advised to take aspirin reported using it regularly after early screening, counseling, and community follow-up. Another multi-state implementation study successfully integrated risk screening, aspirin for women at high risk, closer monitoring, and delivery planning into routine antenatal screening. 

India’s experience therefore points to a surmountable implementation challenge: pregnant women need to be reached early, screened consistently, prescribed aspirin promptly, and take that aspirin throughout pregnancy.

Cost-effectiveness BOTEC

A hypothetical program

To enable cost-effectiveness modeling, I envision a program working in regions or facilities where antenatal care does not consistently identify and treat women at moderate or high risk of preeclampsia. This would integrate risk assessment and aspirin treatment into existing antenatal care programs instead of setting up a parallel program.

The precise method for identifying risky pregnancies is an important design uncertainty. The WHO provides clinical criteria based largely on pregnancy history, blood pressure, and common medical conditions, though more advanced diagnostics would be required to assess some WHO-defined risk factors. A program would need to select an appropriate, evidence-backed screening approach and determine whether more complex screening improves targeting enough to justify its cost.

The program would:

• Develop simple protocols for identifying eligible women during early antenatal visits.^[3] 
• Train and supervise antenatal care staff to assess risk, prescribe aspirin where appropriate, and counsel patients.
• Ensure the necessary equipment and a reliable aspirin supply.
• Support prompt dispensing, adherence, and refills through facility staff and/or community health workers.
• Track screening, treatment initiation, adherence, adverse events, and referrals through existing information systems where possible.

Together, these activities should increase the number of eligible women who are identified early, begin aspirin promptly, and continue taking it. The cost per woman treated would include all of these activities, including the cost of screening women who are not ultimately prescribed aspirin. It is not simply the cost of the aspirin.

“Break-even” cost-effectiveness

Because implementation costing would depend on many uncertain assumptions, I instead ask what fully loaded cost per high-risk pregnancy treated would meet GiveWell’s funding bar. The model combines reference data from GiveWell’s latest cost-effectiveness models with the evidence on aspirin and perinatal mortality. After applying a 50% discount for uncertainty, it indicates the program would need to cost approximately $32 or less per woman treated. It does not estimate other likely benefits, which could include reduced infant and maternal morbidity. ^[4]

We can benchmark these costs against different reference data.

• The Indian government spent ₹2,786 per resident on health in 2022-23, or roughly $29. Using an illustrative estimate that about 23% of pregnancies would be classified as higher risk, a nationwide program costing $32 per woman treated would add approximately $0.12 per resident per year. In other words, this is a tiny fraction of current health system expenditure.^[5] 
• Taimaka’s cost of treating a child with severe acute malnutrition in Nigeria is approximately $94.^[6] That intervention requires therapeutic food, medicines, and inpatient treatment of many cases. Adding aspirin prevention to existing outpatient or community-based antenatal care programs should cost substantially less.
• The Community Health Impact Coalition reports a median cost of $10 per patient per year across diverse community health worker programs. In a more directly relevant trial, community-based monitoring and initial management of pregnancy hypertension cost approximately $12 per pregnancy. The program did not improve outcomes overall, but it also did not include aspirin prophylaxis: it focused on detecting and managing hypertension in pregnancy.

While these examples do not prove cost-effectiveness, I think they do suggest low-dose aspirin prophylaxis programs could be within the range of GiveWell’s current funding bar.^[7] It would be very useful to have unit economic benchmarks for a variety of health system interventions to make a better assessment and rank this against other programs.

Recommendations

This was a time-constrained, shallow investigation, and I may have missed important adversarial evidence or practical constraints. Additional desk research and conversations with clinicians, researchers, policymakers, and implementing organizations could de-risk future work on this intervention. This should include fully evaluating more skeptical perspectives on external validity, maternal risks, and implementation in low-resource settings. 

At this stage, my claim is mainly that the intervention looks promising enough to merit closer scrutiny.^[8] A deeper investigation would:

• Assess the clinical evidence and its external validity.
  • Review the underlying trials and meta-analyses, including benefits and harms to mothers and babies.
  • Assess how well the evidence applies in lower-resource settings, particularly primary care facilities with limited emergency obstetric care capacity, where postpartum hemorrhage presents even greater risks to pregnant women.
• Establish where a meaningful, neglected treatment gap exists.
  • Estimate current screening, prescription, and adherence in different geographies, recognizing that published studies may be incomplete or outdated.
  • Map policies, guidelines, diagnostic equipment, aspirin supply, workforce constraints, existing programs, and the current standard of care.
• Clarify the intervention and extend the cost-effectiveness model.
  • Assess whether medical history and blood pressure are sufficient for risk assessment, or if specialized ultrasound tests or other diagnostics are needed.
  • Identify potential implementing organizations and implementation models.
  • Directly estimate program costs.
  • Build upon my simplified cost-effectiveness model to incorporate potential harms, benefits beyond lower perinatal mortality, program complexities, and fully loaded costs.
• Consider implementation tests. If this intervention still looks favorable after deeper desk research, field research or implementation pilots could test the biggest uncertainties. Such programs would measure screening, treatment initiation, aspirin adherence, and costs. Depending on program size and monitoring budget, they could also assess perinatal and maternal outcomes. 

Overall, I think the most likely barrier to impactful interventions in this space is the complexity of scaling up and monitoring programs to screen pregnant women, classify risk factors, appropriately prescribe aspirin, and ensure adherence while maintaining high fidelity. 

It may make sense to link this to broader efforts to improve antenatal care and treatment of hypertensive disorders in pregnancy. As long as such bundled programs focus on a clearly defined at-risk target population, it seems possible that shared costs and multiple health benefits would increase cost-effectiveness. There are precedents for such programs: Resolve to Save Lives is already doing closely related work to treat hypertension in pregnancy in India. Even if they have a more ambitious quality improvement agenda, such programs should track clear leading indicators with a robust evidence base, such as blood pressure control in hypertensive patients and low-dose aspirin uptake among high-risk pregnant women. 

Disclosures and Disclaimers

I initially looked into this as an area of intervention for HealthLearn. I think it is not quite “shovel ready” for HealthLearn, because I haven't yet found a simple, actionable national clinical protocol on low-dose aspirin for preeclampsia that I’m confident we could build training around.^[9] However, my initial research also convinced me that this is an interesting area and worth sharing more broadly.

I used AI to research this intervention, test my ideas, and draft this writeup. I closely edited the final product and I manually verified the linked claims. I’m grateful for the high-level feedback on drafts of this writeup I received from two researchers in the field. Any errors are mine.

This was a shallow investigation, so I have not extensively red-teamed this idea. I usually find important negative evidence or major concerns when digging deeper into any intervention that initially appears promising, and it is unusual for me to remain optimistic about an intervention after an investigation of this depth.

1. ^{^}

   Perinatal death generally means stillbirth or death during the first week after birth. 

2. ^{^}

   The Cochrane review found a slightly smaller reduction in mortality (15%), however it did not report perinatal mortality specifically.

3. ^{^}

   There are strong precedents for NGOs to support governments to successfully develop simple hypertension treatment protocols; this could be a natural extension of those efforts.

^{^}

This model is deliberately simple, and one could argue that it is oversimplified. I did build a separate model (not shown) to stress-test the absolute risk reduction numbers I used in this model.

^{^}

I recognize this does not show such a program could be delivered at low cost, it simply shows these costs are well within the range of current government expenditures.

^{^}

Marginal cost per child treated is likely much lower. This cost was retrieved from the Taimaka site in September, 2026. My understanding from the Taimaka team is these numbers may be updated in the near future (personal communication with Justin Graham).

^{^}

I do not have a strong sense, from my own work, whether a cost below $32 per at-risk patient treated is achievable. 

^{^}

In support of the potential of hypertensive disorders in pregnancy as a general area for intervention, Coefficient Giving recently made a grant for research on preeclampsia, though I could not confirm whether that grant encompasses interventions to promote aspirin uptake.

^{^}

Relevant high-level guidelines do exist in some countries, and I intend to explore this further in the future. For example, Uganda’s 2022 Essential Maternal and Newborn Clinical Care Guidelines explicitly recommend low-dose aspirin and identify risk criteria. Simple and short national protocols with accompanying training packages would strengthen suitability for HealthLearn's intervention.

Show all footnotes