NOTE: This CEA was modeled based on our CEA for Fortify Health's iron fortification program. We have not yet adjusted the cell descriptions and notes in column A to refer specifically to Evidence Action's IFA supplementation program.
We have specifically highlighted inputs in red where we've made updates to our previous inputs to account for differences between Evidence Action's IFA supplementation program and Fortify Health's iron fortification program (explanations in "Notes" column).

See "Counterfactual burden" tab. This input includes upward adjustments to account for: 1) higher burden among children not receiving supplementation (since we model the benefits of Evidence Action's program as coming from an increase in the number of children receiving supplementation who would not have received it otherwise), and 2) higher burden in the 5 Indian states where EvAc is working relative to the national average.

See results for the effect of iron supplementation on iron deficiency among children 5-12 years on Table 3 of the Low et al. 2013 meta-analysis (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3832580/).

Adding 20% discount to account for weekly/biweekly vs daily supplementation and time children spend out of school/AWCs.

We roughly assume that trials included in Low et al. 2013 (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3832580/) measure the effect of moving from no supplementation to full coverage (i.e. that treatment groups received full coverage and control groups received no coverage). The discount here accounts for our best guess that some of the increase in coverage from IFA will come from going from partial coverage to full coverage, rather than no coverage to full coverage.

Based on Evidence Action's coverage survey results in Uttarakhand, Jharkhand, and Haryana, we calculate ~50% of children have received any IFA supplementation, and of those ~30% with partial coverage and ~70% with full coverage. Very roughly guessing that partial coverage provides about half the effect on dietary iron deficiency as full coverage, we calculate an ~85% discount. This is a very rough guess. See calculations here.

Calc. We assume the effect size reported in Low et al. 2013 for children 5-12 years is equivalent to the effect size for children 0-14 years.

Adjusted for program-specific age structure (6 months to 19 years)

See "Counterfactual burden" tab. This input includes upward adjustments to account for: 1) higher burden among children not receiving supplementation (since we model the benefits of Evidence Action's program as coming from an increase in the number of children receiving supplementation who would not have received it otherwise), and 2) higher burden in the 5 Indian states where EvAc is working relative to the national average.

Adding 40% discount to account for weekly/biweekly vs daily supplementation and time children spend out of school/AWCs, and children under 3 likely not benefitting from cognitive improvements.

Adjusted for program-specific age structure (6 months to 19 years)

See "Counterfactual burden" tab. This input includes upward adjustments to account for: 1) higher burden among children not receiving supplementation (which we assume to be the primary IFA program beneficiaries), and 2) higher burden in the 5 Indian states where EvAc is working.

Adding 20% discount to account for weekly/biweekly vs daily supplementation and time children spend out of school/AWCs.

This is a rough guess. It is intended to capture risk of wastage, quality of monitoring and evaluation, and confidence in funds being used for the intended purpose.

Based on our conversation with an expert in iron interventions, our understanding is that supplementation and fortification are relatively less effective against severe anemia. In India, we estimate roughly 20% of anemia burden is due to severe anemia (calculations are internal). Because we're unsure about how important this adjustment is, we apply a lower adjustment.

We've also tried modeling supplementation programs as reduction in anemia (as opposed to iron deficiency). We find results are slightly lower using this approach so apply a small downward adjustment.

In addition to iron, supplements also provide folic acid. We have not modeled additional benefits of folic acid but apply a small upward adjustment.

The target population only includes children in private schools; however, states might decide to scale the program to all private schools, depending on the result of the pilot (see associated grant page for more information). We’d guess this would increase the target population by ~30% (see "Source: Targeted children" tab). We guess this is fairly unlikely (<50% chance) so the positive adjustment is less than 30%.

Cost-effectiveness, multiples of cash transfers from GiveDirectly, after downside adjustments and inclusion/exclusion

Cost-effectiveness, multiples of cash transfers from GiveDirectly, after accounting for leverage/funging

Calc. We add a leverage/funging adjustment since this IFA is a technical assistance intervention.