Worked example · Model v0.1
Preventing malaria in young children in Chad
What could $1,000 in program spending prevent? We combine published cost and effect estimates with a simple model of acute malaria suffering.
Already funded. This estimate describes a program, not a verified opportunity for an additional donation.
Estimated acute-malaria USU averted
≈170per $1,000 in total program costs
malaria episodes prevented
direct malaria deaths prevented
All three outcomes use the same $1,000 cost basis. They are modeled expectations, not observed results. Gross acute-malaria benefit; side effects and long-term effects are unestimated.
Reading the result
A concrete use for a common unit
Under our central assumptions, $1,000 of total delivery costs corresponds to approximately 170 USU of acute malaria suffering prevented.
A USU uses the stipulated six-hour kidney-stone pain reference of 47.25 intensity-hours. It makes the scale explicit; it does not imply that the child experienced renal colic, or that different forms of suffering have been empirically calibrated to that reference.
The value can help readers compare future analyses using the same cost and suffering definitions. It captures only selected experienced suffering. Deaths prevented are shown separately because death itself and lost life years are not assigned USU.
The decimals describe an expected effect across a program. An estimate of 0.17 deaths prevented is not a fraction of a person observed to survive.
Follow the calculation
From program costs to illness avoided
Estimate the delivery funded by $1,000
$1.9298367 per monthly cycle × 4.666167 cycles = about $9.00 per child-season. That corresponds to 111.05 seasonal rounds per $1,000, using unrounded costs.
Apply the published disease and prevention inputs
Child-seasons × 0.45928 annual episodes × 70% seasonal share × 79.3695% expected reduction = 28.34 episodes prevented.
Child-seasons × 0.002932651202 annual direct mortality × 0.94 vaccine adjustment × 70% seasonal share × 79.3695% reduction = 0.1701 direct deaths prevented.
Assign one course to each prevented episode
Nonfatal episodes use four days at average intensity 3. Fatal episodes use one complete four-day course at intensity 5. We subtract fatal episodes from the ordinary group before adding the two results.
(28.3366 − 0.1701) × 6.0952 USU
+ 0.1701 × 10.1587 USU
= 173.41 USU, rounded to ≈170Shown inputs are rounded for reading; calculations use the full stored values. USU per course = days × 24 × intensity ÷ 47.25. Linear weighting, p = 1.
Inspect the source inputs and cost basis
These inputs are from the Chad column of GiveWell SMC CEA, 2025 version 3, Main CEA, retrieved 9 October 2026. The cost denominator includes all contributors, including domestic government costs. Dollars retain the source model’s basis and are not inflation-adjusted to 2026.
| Input | Value | Source cell |
|---|---|---|
| Full cost per monthly cycle | $1.9298367 | J28 |
| Mean cycles per season | 4.666167 | J29 |
| Annual incidence, under 5 | 0.45928 episodes per child-year | J74 |
| Share in the SMC season | 70% | J56 |
| Expected incidence reduction among recipients | 79.36949983% | J47 |
| Annual direct malaria mortality | 0.002932651202 deaths per child-year | J50 |
| Vaccine adjustment to mortality | −6% | J53 |
| Mortality effect / incidence effect | 1 | H48 |
The model pairs approximately $11.56 million in all-contributor costs with 1.283 million child-seasons for the grant-supported scope. The grant itself is about $10.43 million. Neither the grant amount alone nor the broader country budget is the denominator for this result.
Unresolved source difference: the grant narrative quotes $8.80 per round, while the linked model gives $9.00494 in full costs and $8.12470 in grantee/upstream costs. We use the linked model consistently. The reason for the discrepancy is not established.
The source uses under-5 incidence as a proxy for the target ages. It adjusts mortality for vaccine rollout but does not apply the same adjustment to incidence. We retain those choices and do not import GiveWell’s moral weights.
Explore the judgment
How much do the USU assumptions matter?
Duration and intensity are provisional choices. Change them below while holding program costs, episodes and deaths prevented fixed.
173 acute-malaria USU averted per $1,000
Central: nonfatal course, 4 days at intensity 3; fatal course, 4 days at intensity 5.
Whole-USU rounding here makes the alternatives easier to inspect. The headline uses ≈170 to avoid false precision.
| Scenario | Nonfatal course | Fatal course | USU / $1,000 |
|---|---|---|---|
| Lower assumptions | 2 days × intensity 2 | 2 days × intensity 3 | 58 |
| Central assumptions | 4 days × intensity 3 | 4 days × intensity 5 | 173 |
| Higher assumptions | 7 days × intensity 4 | 7 days × intensity 7 | 405 |
58–405 is a mapping-only sensitivity range. It is not a confidence interval or a full uncertainty range. Source disease burden, program effectiveness, costs and delivery can also differ substantially.
Why four days?
A Nigerian pediatric treatment trial reported mean fever clearance about 46–49 hours after treatment. Four days allows additional time before care and in early recovery; it is our assumed full course, not a measured Chad average.
A Malian hospital study found about 3.7 days of illness before admission in fatal severe febrile cases, with most deaths soon after admission. These were presumed severe malaria cases, not all confirmed malaria, and not a representative Chad community sample. This offers only weak context for the four-day fatal assumption.
Why intensities 3 and 5?
WHO’s symptom descriptions support meaningful acute distress and potentially more severe illness. They do not supply a numerical USU conversion. Intensities 3 and 5 are explicit judgments, not measured pediatric ratings.
Both are averages over the whole course, including relief, sleep and periods of reduced or absent awareness. They are not continuous peak pain. Severe survivors are folded into the single nonfatal average.
Scope and limits
What the estimate includes
Included
Acute malaria episodes prevented among the modeled young-child population, including complete fatal courses for direct malaria deaths prevented. Each avoided episode receives one course.
Death prevention remains visible
The mortality figure includes direct malaria deaths only. The source model’s indirect deaths and effects in older children are outside this pilot.
Unestimated
Drug side effects, long-term complications, caregiver distress, bereavement and other spillovers. Death itself and lost life years are not assigned USU.
This is a gross acute-malaria benefit, not net total suffering averted and not a guaranteed lower bound. Both the program model and the USU mapping can be substantially wrong.
Would the same calculation work for treatment?
Prevention can avert a whole illness episode. Treatment may shorten a course or replace a fatal course with a surviving course. A treatment analysis must estimate the difference between the experience with and without treatment; it cannot automatically credit the full prevented episode used here.
Funding status
A worked example, already funded
GiveWell’s June 2025 grant supports the modeled Chad expansion. This page does not establish the effect of another donation today.
For a provisional estimate of future support, see the AMF malaria-prevention analysis. It uses explicit extrapolation assumptions and separates AMF’s reported funding need from modeled impact. It does not identify a named open campaign.
Inspect and reuse
Sources and calculation files
Program facts come from published models. The USU duration and intensity assumptions are WSI judgments. They are kept separate in the downloads.
- GiveWell SMC CEA, 2025 version 3Retrieved 9 Oct 2026
- GiveWell Chad expansion grantPublished Sep 2025
- WHO malaria fact sheetUpdated 4 Dec 2025
- Pediatric treatment trial, NigeriaBMC Infectious Diseases, 2025
- Willcox et al., severe febrile illnessPLOS ONE, 2014
- AMF funding update24 Aug 2026; checked 9 Oct
- GiveWell net campaigns grantDec 2024 grant; checked 9 Oct
Compare malaria prevention programs
The DR Congo net example uses the same malaria illness courses. The comparison uses core program-model estimates before final grant-level adjustments. It counts young-child acute-malaria benefits and does not establish the return on a new donation.
Read the two-program comparison →