Evidence Action — Syphilis Screening and Treatment in Pregnancy in Zambia Top-up grant (July 2025)

Note: This page summarizes the rationale behind a GiveWell grant to Evidence Action. Evidence Action staff reviewed this page prior to publication.

Published: March 2026

Background

In July 2022, GiveWell recommended a grant of approximately $15 million to Evidence Action to provide technical assistance to the governments of Zambia and Cameroon to support the scale up of syphilis testing and treatment in pregnancy. This grant funds Evidence Action’s efforts from 2022 to 2028 to support the governments in switching from HIV rapid tests to dual HIV/syphilis rapid tests and increasing syphilis treatment rates in routine antenatal care.

In March 2025, GiveWell recommended the following additional grants to support the program in Zambia:

• $142,515 to cover additional training that was left unfunded during the foreign aid freeze period in early 2025;
• $904,020 as a contingency amount to allow Evidence Action to procure HIV/syphilis dual tests if gaps are left from delays or funding cuts since PEPFAR had supported dual test procurement.

In July 2025, GiveWell recommended $665,654 to cover budget overruns primarily due to higher-than-expected program implementation costs. This page outlines our reasoning for recommending this top-up grant.

What we think this grant will do

Evidence Action requested an additional $665,654 to finish implementing the planned activities of their syphilis screening and treatment support program in Zambia. This top-up request represents a 9% increase to the total grant costs over the $7,046,249 initial grant we made for Zambia in July 2022.

This top-up will allow Evidence Action to complete the planned program activities, including:

• Working with the government to train providers to use the dual rapid test and provide treatment for syphilis, and to provide ongoing supportive supervision
• Working with the government to develop tools for training and supportive supervision of health facilities
• Supporting Zambia with procurement of tests and penicillin for treatment of syphilis, where needed, as well as the management of the supply chain
• Monitoring program activities and the impact of the program on coverage of syphilis screening and treatment among pregnant people.

Budget

The total requested budget is $665,654. This represents an underspend of $1,122,521 in the first year of the program, and an anticipated overspend of $1,788,174 in years 2–5. There was cost variation across different categories, but Evidence Action attributes1 the higher budget mainly to:

• Higher-than-budgeted program implementation costs,2 including:
  • Costs to work with another NGO as a sub-grantee3
  • Higher training costs (and per diems for district trainers): The Zambian Ministry of Health requested two and half days of training per health facility rather than one day, two trainers sent to each facility rather than only one, and a higher degree of supervision.4 Our understanding is that these adjustments were a prerequisite to run the program.
  • Higher stakeholder meeting costs.5
  • Higher tool printing costs.6
• A vehicle purchase: Evidence Action originally expected to rent or lease vehicles, but found through a cost-analysis that purchasing a vehicle was more efficient over the long term.7
• Support contracts: This includes mandatory services overlooked in the original budget such as a legal firm, tax consultant, and audit firm.8
• Higher monitoring and evaluation (M&E) costs: Evidence Action originally benchmarked these costs on the cost for M&E in Liberia's program, but Zambia ended up being a higher cost market for surveys.9
• Higher office rent costs.10
• An increase in Evidence Action's indirect cost (IDC) rate.11

These cost increases are somewhat offset by much lower-than-expected costs for personnel.12

Why we made this grant

The program is going well so far, and we think the program is still highly cost-effective even with additional costs.

Survey data we've received so far suggests higher rates of syphilis screening and a higher syphilis treatment rate under the program than we expected at the time of making the initial grant.13 Syphilis prevalence was also higher than expected (3.9% active syphilis rate compared to 3.0% expected at the time of the grant14 ), so overall we think this grant is having more impact than we previously thought.

At the time we made the original grant for Zambia, our best guess was that the program was 30x as cost-effective as GiveWell’s benchmark.15 We did not do a full cost-effectiveness analysis update to approve this top up grant, but some rough calculations suggest that even accounting for the additional costs from this grant and the contingency grants we recommended earlier this year, the cost-effectiveness of this program could range from 11–51x. If we update our estimate for the survey findings, but assume that there will be no benefits after the grant period (given uncertainty in the global funding landscape), our guess is that the grant would be 25x.

The program offers a valuable learning opportunity that we would lose by not funding this.

Of the countries where we’ve supported this program, our understanding is that data systems are the most mature in Zambia, so we expect to have more confidence in the program M&E results here than in other locations. For example, we have more confidence in the Comprehensive Facility Survey estimates for Zambia relative to Liberia, because most facilities where data was collected by Evidence Action in Zambia also had data available in DHIS2, and estimates (between the DHIS2 and data collected from registers) were broadly aligned.

Therefore, we think that M&E results from this program in Zambia is our current best opportunity to learn about the impact of Evidence Action’s program, which could inform our future grant decisions. This learning value would be forgone if Evidence Action wasn't able to complete the program as intended.

We should be a responsible funding partner, particularly given the current global aid landscape.

In the past, GiveWell has encouraged Evidence Action to expand to new geographies and program areas. We acknowledge that this involves inherently higher risks of unforeseen challenges, delays, and costs. We think it is responsible for GiveWell to shoulder some heightened associated costs to this, especially for such a cost-effective program.

Additionally, we think that Evidence Action’s relationship with the Zambian Ministry of Health could be damaged if they weren’t able to complete their support as intended.

This is especially important given the current global funding landscape. Along with many other countries, Zambia was exposed to cuts in USAID and PEPFAR funding in 2025. If we did not make this additional grant and Evidence Action had to scale down the scope of training, it would risk compounding challenges with supporting HIV/syphilis testing and treatment in antenatal care at a time when Zambia is already facing contracting resources.

Our main reservations

Long term sustainability of the program is uncertain due to the changing global aid landscape.

Both HIV programming (with which syphilis programming is often integrated) and future dual test procurement are vulnerable to recent cuts to foreign aid. Evidence Action told us that dual test procurement is currently supported by the Global Fund, PEPFAR, and via domestic resources in Zambia. It remains unclear whether PEPFAR or the Global Fund will be able to continue procurement at the same level in the medium-term, so we currently have a lot of uncertainty about whether syphilis testing rates will remain high after Evidence Action’s support concludes.

We did not update our broader research on this intervention, and may have missed important considerations.

We have not furthered our research on syphilis interventions since the original 2022 grant to this program. It’s possible we should have made headway on our research agenda to strengthen our understanding of the cost-effectiveness of this intervention before investing more money into this program. However, even in that case, we think some of the qualitative considerations for making this grant would still hold.

Plans for follow up

• Check-ins for this grant will align with our check-ins for the initial grant.
• We plan to continue making progress on our syphilis research agenda.
• Because Zambia appears to have high quality antenatal health data and high syphilis prevalence, we plan to consider ways we can collect data on the impact of this program on end health outcomes, like congenital syphilis rates. We will begin discussing this with Evidence Action.

Internal forecasts

For this grant, we are recording the following forecasts:

Confidence	Prediction	By time	Resolution
80%	We will assess this work as being more than 10 times as cost-effective as our benchmark	By December 2027

Our process

• Evidence Action shared a budget and narrative document with us on May 2, 2025.
• We had two calls with them to discuss these.
• We reviewed the budget, and updated our cost-effectiveness analysis.

Sources

Document	Source
Evidence Action, 2025.05.02 Zambia top-up budget	Source
Evidence Action, Syphilis-Free Start Zambia: Y3-5 Top-Up Proposal	Source
Evidence Action, Zambia Comprehensive Facility Survey Results & 2024 Program Coverage Estimation, 2025	Source
GiveWell, Evidence Action — Syphilis Screening and Treatment in Pregnancy in Zambia and Cameroon (July 2022)	Source
GiveWell, Maternal syphilis cost-effectiveness analysis for Zambia and Cameroon, 2022	Source
GiveWell, Maternal syphilis cost-effectiveness analysis for Zambia and Cameroon, 2025	Source
GiveWell, Our Approach to Foreign Aid Cuts	Source

• 1See narrative doc here: Evidence Action, Syphilis-Free Start Zambia: Y3-5 Top-Up Proposal, 2025
• 2Evidence Action, Syphilis-Free Start Zambia: Y3-5 Top-Up Proposal, 2025, table 2, p. 5
• 3“Our initial budget did not include the sub-grant we gave to CHAI Zambia. In late 2022, we decided to move forward with the sub-grant to enable CHAI Zambia to extend their national-level technical support to the MoH. Throughout 2021-2022, CHAI had provided very specific and targeted support to the MoH labs team to coordinate the adoption of dual testing - this included acting as the secretariat for the biweekly dual test working group calls, supporting the registration of the various HIV/syphilis dual test brands, and supporting the investigation of false positives due to the Bioline dual test. We felt it was more efficient to continue to leverage the CHAI / labs teams engagement rather than replacing it. This allowed Evidence Action to focus on the overall strategic planning and management of the maternal syphilis program and prioritize its relationship with the National PMTCT team at the outset. In all, the partnership proved useful -- CHAI was able to maintain the dual test working group, helped manage the dual test brand transition, advocated together with our team to lift the 2023 hold on dual test procurement which had prevented PEPFAR from donating test kits, ensured adequate annual quantification of test kits, and worked with the Labs team to develop stronger tools for dual test usage at the health facility. The agreement ended, as planned, in December 2024. In total, we transferred $394,228 to CHAI.” Evidence Action, Syphilis-Free Start Zambia: Y3-5 Top-Up Proposal, 2025
• 4See the narrative report, p. 6-7: Evidence Action, Syphilis-Free Start Zambia: Y3-5 Top-Up Proposal, 2025
• 5“Our initial estimate for stakeholder meetings was an annual spend of ~$5,000 per year, based very roughly on what we spend in Liberia annually. In reality, the true cost of this has been triple that of our budgeted amount (~$15,000 per year).” Evidence Action, Syphilis-Free Start Zambia: Y3-5 Top-Up Proposal, 2025, p. 7
• 6“We initially estimated that we would spend a total of $55,000 across five years to print various program tools (job aids, registers, training manuals, etc), based very roughly on what we spent in Liberia. Our realized costs were approx. twice as high at $114,133 ($54,686 of which we’ve spent to-date and a further $59,447 that we have budgeted to spend going forward). We encountered higher printing costs for two main reasons: (1) we developed more tools in Zambia which required printing than we had in Liberia, most important was the participant’s reference manuals which are small hand-held booklets given to all trained healthcare workers to serve as reference guides in case they forget any aspects of the training; and, (2) we needed to print more total units of every tool because there were more trainers and more healthcare providers than we had initially estimated.” Evidence Action, Syphilis-Free Start Zambia: Y3-5 Top-Up Proposal, 2025, p. 7
• 7“Vehicle purchase & maintenance: Our initial budget did not include the purchase of a vehicle, nor
  the costs associated with the vehicle (fuel, repairs, etc). However, in mid-2023, we decided to
  purchase a vehicle, with GiveWell’s approval, following a costing analysis of various ground
  transportation options. Specifically, that analysis found that purchasing a vehicle would have an
  up-front cost, likely not to exceed $50k, and then we anticipate we'll spend ~$1,800 monthly toward
  fuel, a driver, insurance, etc, which is in comparison to ~$3,000 to ~$5,000 monthly toward car
  rentals, leases, etc if we were unable to buy a car.” Evidence Action, Syphilis-Free Start Zambia: Y3-5 Top-Up Proposal, 2025, p. 8
• 8“Support contracts: These include a legal firm, tax consultant, and audit firm. In our initial budget,
  we entirely neglected to account for these costs, which run ~$33,000 annually. Ultimately, these are
  mandatory expenditures to maintain our program office in Zambia and ensure we are compliant with
  local regulations and Evidence Action policies.” Evidence Action, Syphilis-Free Start Zambia: Y3-5 Top-Up Proposal, 2025, p. 8
• 9“Our original estimate for the cost of a comprehensive facility survey in Zambia was based on the cost of implementing the same survey in Liberia (~$80,000). We anticipated the sample sizes to be the same and both countries are known for high transport costs and so we took the Liberia survey costs as the best point estimate for Zambia. However, upon soliciting proposals from survey enumeration firms in early 2023, we realized we had significantly underestimated the cost of surveying in Zambia as the two most qualified survey firms, CIDRZ and IDinsight both quoted survey costs around ~$145,000. This was further exasperated by recent inflation, which has driven up the cost of fuel / transport (a significant cost driver in data collection)." Evidence Action, Syphilis-Free Start Zambia: Y3-5 Top-Up Proposal, 2025, p. 7-8
• 10"In our initial budget, we estimated that we would spend ~$1,500 per person per month toward office rent, based on what partners shared about the costs of using a co-working space. As our team size grew substantially relative to what we had initially predicted (10 staff vs. 6 staff), it was more cost-efficient to move into a formal office space vs. continuing to use a co-working space. Our current office costs approx. $4,000 per month, driven in part by inflation and an increase in our monthly office service charge due to frequent power outages driving an increase in generator costs." Evidence Action, Syphilis-Free Start Zambia: Y3-5 Top-Up Proposal, 2025
• 11"Starting January 1, 2025, Evidence Action’s indirect cost (IDC) rate has increased to 22% (note: it was 18% at the time of our original Zambia budget submission)." Evidence Action, Syphilis-Free Start Zambia: Y3-5 Top-Up Proposal, 2025
• 12“We are anticipating a significant under-spend in local personnel and fringe. We developed the budget before we had done formal salary benchmarking for Zambia and had instead based our salary assumptions on what two local partners had informally shared about their own pay scales. In the end, the salaries we assumed for budgeting were much higher than our final salary ranges and so we have found significant cost-savings on this budget line” Evidence Action, Syphilis-Free Start Zambia: Y3-5 Top-Up Proposal, 2025
• 13
  • Syphilis screening rate at full scale: Increased from 78% to 83% based on the comprehensive facility survey (CFS). See slide 23: Evidence Action, Zambia Comprehensive Facility Survey Results & 2024 Program Coverage Estimation, 2025. We used figures from the preferred DQA approach, but in this case, taking the average of the DQA and ANC register approach would give a similar result.
    • We adjusted the 90% coverage rate estimated by Evidence Action for the 92.5% of antenatal care (ANC) attendees that we expect their program will reach. (90% * 92.5% = 83%)
  • Syphilis treatment rate: Increased from 70% to 85% based on CFS. See slide 23: Evidence Action, Zambia Comprehensive Facility Survey Results & 2024 Program Coverage Estimation, 2025. We used the preferred DQA approach, but in this case, averaging would give a similar result. Note: Evidence Action reviewed these estimates, and in early 2026 provided an updated treatment rate estimate of 90% (based on the antenatal care register approach), as they surfaced some concerns about the DQA approach in this context. We have left our modeled estimate unchanged as it was at the time we made this grant, but note that we would expect cost-effectiveness to be slightly higher than our model suggests.

• 14Active syphilis prevalence: Increased from 3.0% to 3.9% based on higher syphilis prevalence found in the comprehensive facility survey (CFS). See slide 23: Evidence Action, Zambia Comprehensive Facility Survey Results & 2024 Program Coverage Estimation, 2025. We averaged the DQA and ANC register estimation approaches, which indicated that the total prevalence of active and latent maternal syphilis among women attending antenatal care was 7.3%. Assuming that around 54% of reactive treponemal tests are "probable active syphilis," this implies an active syphilis rate of 3.9%. See our calculations here.
• 15Prior to November 2025, we used a benchmark that we described as multiples of “cash,” based on an outdated assessment of the impact of GiveDirectly's unconditional cash transfers program. While we’ve changed how we calculate our benchmark to no longer be directly tied to GiveDirectly's program, the value of that benchmark remains similar, which means our cost-effectiveness estimates remain comparable across the old and new approaches. The benchmark referred to in this grant page is the pre-November 2025 benchmark.