Planet Money

Saving lives with fewer dollars

November 27, 2025

Key Takeaways Copied to clipboard!

  • GiveWell, an organization focused on maximizing lives saved per dollar through rigorous research, faced the challenge of rapidly assessing a complex, multi-pronged health project in conflict-ridden Cameroon after USAID funding was cut. 
  • GiveWell's decision-making process is characterized by a ruthless focus on quantifiable metrics, such as 'life-saved equivalent,' which often conflicts with the messy, context-dependent data available in humanitarian crises. 
  • Despite the difficulty in obtaining perfect data due to instability in Cameroon, GiveWell ultimately approved a $1.9 million grant to ALIMA's project, recognizing the high impact and the necessity of learning from imperfect evidence in urgent situations. 

Segments

ALIMA’s Work in Cameroon
Copied to clipboard!
(00:00:22)
  • Key Takeaway: ALIMA provides essential, hands-on healthcare, including malnutrition screening using simple tools like colored tape, to nearly 400,000 people in Cameroon’s conflict-affected far north region.
  • Summary: ALIMA manages difficult logistics and builds trust to deliver basic healthcare, including monitoring pregnancies and training mothers to identify malnutrition early. Their Cameroon program was set to lose $1.9 million in funding from USAID. The loss of this funding immediately threatened the continuation of care for vulnerable populations, including malnourished children in hospitals.
GiveWell’s Research Mandate
Copied to clipboard!
(00:03:04)
  • Key Takeaway: GiveWell operates as a philanthropic group using ‘ruthless calculations’ to fund interventions that save or improve the most lives per dollar, often favoring specific, efficient, and neglected projects.
  • Summary: GiveWell’s methodology contrasts with large-scale government aid by focusing on maximizing cost-effectiveness, often using metrics like ’life-saved equivalent.’ This approach historically favored interventions like distributing mosquito nets, which are proven to be highly effective and cheap. The organization is now attempting to apply this rigorous, data-heavy process to a chaotic, multi-pronged humanitarian project previously funded by USAID.
Initial Data Gathering Meeting
Copied to clipboard!
(00:06:44)
  • Key Takeaway: GiveWell immediately recognized the Cameroon project presented a ‘steeper learning curve’ due to its conflict-zone context, contrasting with their usual preference for projects with ‘crispy, clear data.’
  • Summary: The GiveWell team began mapping out necessary information, prioritizing the Cameroon project because it was emblematic of the large, multi-pronged programs USAID supported. Historically, GiveWell has focused on interventions where randomized controlled trials can prove efficacy, a standard difficult to meet in this unstable region. The team needed to establish basic facts like local population size and baseline mortality rates.
Urgency and Operational Cuts
Copied to clipboard!
(00:10:56)
  • Key Takeaway: The Cameroon project was already scaling back operations significantly, reducing coverage from 14 to 4 health centers in one area and facing complete shutdown in another without immediate funding.
  • Summary: ALIMA’s program manager confirmed that services were not fully operational due to funding uncertainty, forcing staff reductions across multiple facilities. The loss of funding meant that if the project shut down, the local health system could not absorb the displaced patients. GiveWell immediately focused on understanding the urgency and gathering demographic data necessary for their mortality calculations.
Mortality Math Challenges
Copied to clipboard!
(00:12:45)
  • Key Takeaway: Calculating averted deaths in the Cameroon project involved complex and uncertain ‘child mortality math,’ specifically reconciling baseline mortality rates with the counterfactual scenario without ALIMA’s intervention.
  • Summary: The GiveWell team struggled with data quality, noting that mortality figures are hard to establish in a conflict zone with mobile, displaced populations who do not return for follow-up. They debated potential double-counting issues related to malnutrition complications like pneumonia. The team needed to determine how many children under five would die annually per 10,000 people with and without the project’s support.
On-the-Ground Research Limitations
Copied to clipboard!
(00:14:22)
  • Key Takeaway: GiveWell could not deploy independent researchers to ‘ground truth’ the Cameroon project’s data due to extreme danger in the area, forcing them to rely more heavily on the implementing organization’s figures.
  • Summary: Normally, GiveWell hires external firms to verify data points like staffing levels and population counts, but the area’s instability—where researchers had previously been killed—made this impossible. A comprehensive study would take a year, conflicting with the immediate need for funding. This lack of independent confirmation became a major sticking point for the research team.
Cost-Effectiveness Insights
Copied to clipboard!
(00:15:52)
  • Key Takeaway: GiveWell’s data-driven framework has previously revealed counterintuitive findings, such as an intervention preventing syphilis in pregnant women being thousands of times more cost-effective than one targeting rare maternal mortality.
  • Summary: This historical example illustrates how cold calculations can discipline compassion and override strong intuition regarding aid priorities. The syphilis testing program was funded because it saved significantly more lives per dollar than the maternal mortality intervention. This reinforces GiveWell’s commitment to evidence, even when the resulting decision (funding syphilis testing over direct maternal death prevention) feels difficult.
Data Language Mismatch
Copied to clipboard!
(00:17:48)
  • Key Takeaway: The GiveWell team and ALIMA struggled to reconcile standard humanitarian mortality metrics (deaths per 10,000 people per day) with standard demographic survey metrics (live births per year).
  • Summary: ALIMA staff explained that humanitarian aid often targets areas of conflict to maintain dignity for displaced populations, which differs from focusing solely on areas with the highest baseline mortality rates. The difference in how mortality data is measured across the industry created confusion for GiveWell’s quantitative models. ALIMA’s nutrition services were identified as a major draw that keeps mothers coming to facilities for other essential care.
Back-of-the-Envelope Modeling
Copied to clipboard!
(00:19:48)
  • Key Takeaway: Initial modeling suggested the Cameroon programs were significantly better than direct cash transfers at saving lives, though the multi-intervention nature of the project made attributing specific outcomes difficult.
  • Summary: GiveWell’s ‘Botex’ calculations compared the intervention’s impact against simply giving people money, finding the health services superior. The difficulty lay in isolating the impact of individual services (vaccines, nutrition, water sanitation) when staff provide bundled care during patient visits. Unquantifiable benefits, like training medical staff who then train others, also complicated the cost-effectiveness assessment.
Ground Truthing Validation
Copied to clipboard!
(00:30:09)
  • Key Takeaway: A blunt conversation with a local hospital director confirmed the catastrophic impact of ALIMA’s potential withdrawal, validating the high mortality figures GiveWell had been skeptical about.
  • Summary: The hospital director stated that if ALIMA pulled out, it ‘would be a catastrophe,’ confirming that ALIMA’s services were the primary driver for people seeking care. This external confirmation helped overcome GiveWell researchers’ reservations about relying solely on ALIMA’s internal data. This realization shifted the team’s perspective from doubting the data to accepting the high impact potential.
Grant Approval and Aftermath
Copied to clipboard!
(00:32:53)
  • Key Takeaway: GiveWell approved the full $1.9 million grant to cover the Cameroon project’s USAID hole for one year, but this success is an exception, as only 23 out of 140 identified programs received funding.
  • Summary: The approval allowed ALIMA workers to restart paused services like mental health and educational programming that had been cut. Madeleine noted that the reduction in overall aid funding creates a dangerous feedback loop where reduced visibility makes the need appear smaller to potential donors. This loss of information means that while need increases globally, the ability to fund effective aid decreases.