In a nutshell
Food for the Hungry implements a
diverse set of programs aimed at improving health and reducing poverty across
20 countries in the developed and developing world. We named it a finalist on the strength of its Care Group model, a
CSHGP-funded project that we believe demonstrates a logical and cost-effective approach to saving lives. However, in trying to examine the organization as a whole, we are unable to gain the same confidence in a large enough portion of its activities, and therefore cannot confidently recommend the organization.
The details
We first discuss the Care Group model - what it is, the evidence that it works, and our estimate of "lives saved per dollar" - and then discuss our relatively limited understanding of the rest of the organization.
Care Group model
What do they do?
The Care Group model is a
CSHGP-funded program that FHI has implemented in Mozambique and Bolivia, and like many such programs, it takes a broad approach to improving health and quality of life (rather than focusing narrowly on a particular problem); includes detailed monitoring that gives us confidence in its effectiveness; and appears to save lives cost-effectively. In these ways, it is a good match with our
priorities for this cause.
The model centers around the use of FHI-employed "Promoters," who train groups of 10-12 "Leader Mothers" (Attachment B-3 Pg 5) in basic health practices, including:
These Leader Mothers are then expected to train 10-14 mothers each (Attachment B-3 Pg 5). In addition, FHI leverages Ministry of Health services to provide immunizations, deworming medication, and vitamin A supplementation to the children living in the region (Attachment B-6, slide 8).
Does it work?
FHI provided relatively strong evidence of effectiveness for its Care Group programs in Bolivia and Mozambique (particularly Mozambique):
- Survey data from Mozambique (Attachment B-6 Pg 13-17) shows significant declines in both diarrhea episodes and cases of stunted growth, and increases in the use of ORT and Vitamin A over the 1997-2000 period. It also describes a significant decline in actual child mortality, though the details of how this was measured are unclear to us (more on this in the following section).
- Additional survey data from Mozambique (Attachment B-3 Pg 7-10) indicates improvement in general health knowledge and practices, particularly nutritional practices.
- Survey data from Bolivia indicates improved access to (and practices regarding) nutrition and sanitation (Attachment B-2 Pg 150-151).
What do you get for your dollar?
Attachment B-6 Pg 18 states that between 1999 and 2004, the under-5 mortality rate in Mozambique Care Group areas fell from 107 to 41 deaths per 1000 live births; converted into an annual mortality number, this implies that annual mortality among under-5 children (the main targets of the program, as the above description shows) fell from 2.1% to 0.8%. We do not have a good deal of confidence in this statement, as we explain below, but for the moment we take it at face value in order to estimate the total number of lives saved per year.
We first estimate the total number of children reached by the program, based on the assumption that 1457 Leader Mothers were trained (as Attachment B-6 Pg 25 states), that each trained 10-14 more mothers (Attachment B-3 Pg 5), and that each of the mothers reached had 2-4 children in the household - implying that a total of around 30,000-80,000 children were reached. If the annual mortality rate among these children fell from 2.1% to 0.8%, that's 400-1000 lives saved per year, at a total cost of $500,000 per year (Attachment A-1 Pg 3) - implying
$500 to $1,250 per life saved.
However, we know very little about how FHI collected this data in general or how it assessed or estimated mortality rates in particular. The only information we have comes from Attachment B-6 Pg 10, which describes whom they surveyed and where. We aren't sure whether the decline in mortality is based on direct observation (i.e., counting the number of deaths reported), projections based on survey responses, or theoretical models of lives saved based on the survey data described above (i.e., how many lives would be theoretically saved if better nutritional practices were observed). Without information about this number, we have very low confidence in our estimate of cost-effectiveness, although it is in the ballpark of
USAID's estimate for similar programs.
Other programs
In aggregate, we know relatively little about FHI's activities. The following table summarizes the projects that we do have documentation on:
| Location | Description | Duration | Budget | Source |
| Bolivia | Child survival program (see above) | 2002-2006 | $23m | Attachment B-2 |
| Mozambique | Child survival program (see above) | 2006-2010 | $3.3m | Attachment B-3 |
| Mozambique | Improving farmers' incomes through training in better farming technologies and marketing their products; child survival program | 2005 | ? | Attachment B-7 |
| Sudan | Building schools and promoting equality in education for girls | ? | ? | Attachment B-5 |
| Ethopia, Mozambique, Haiti, Nigeria | Abstinence promotion | ? | $2.2m | Attachment B-1 |
| Uganda | Construction of latrines and water pumps; agricultural training | ? | $559k | Attachment B-16 |
| Sudan | Distributing seeds and training farmers to increase food security | ? | $937k | Attachment B-9 |
| Bangladesh | Microfinance (savings and loans); literacy training | 2006 | ? | Attachment B-8 |
| Indonesia | Tsunami relief: shipping supplies, distributing bednets, etc. | 2005 | $973k | Attachment B-13 |
| Dominican Republic | Latrine construction | - | $6k | Attachment B-4 |
| Rwanda | Proposal to improve farming practices through terracing | - | $874k | Attachment A-3 |
The concerns we have are:
- We have very little information about the impact of these programs on people's lives, with the exception of the child survival programs discussed above. The sexual abstinence programs include reports of people trained (or reached via advertising), but no information on whether these programs had any effect on sexual behavior (surveys have been designed but not yet implemented). The other programs vary in how possible it is to measure their impact, but across the board, we have nothing that would help us gain confidence in (a) whether lives have been saved (b) how cost-effectively lives have been saved.
- These programs represent only a tiny fraction of FHI's activities. According to the most recent financial audit (Attachment C-1 Pg 3-4), FHI had cash expenses of $56 million in 2008; the programs above account for at most 10% of that.
Conclusion
Ultimately, we cannot confidently recommend FHI because we have too little information about the organization as a whole. We have neither comprehensive evidence on outcomes, nor an overall view of the organization's strategy; while there are some programs we are confident in and others we are less confident in, we have no sense of what to expect from this organization if and when it brings in more donations.
Attachments
A. Application and response
- Attachment B-1: Abstinence Promotion evaluation
- Attachment B-2: Community health, Bolivia
- Attachment B-3: Child health, Mozambique
- Attachment B-4: Latrine building, Dominican Republic
- Attachment B-5: Education, Sudan
- Attachment B-6: Reduction of child mortality in Mozambique, presentation
- Attachment B-7: Economic development and child health, Mozambique, 2005
- Attachment B-8: Economic development, Bangladesh
- Attachment B-9: Food security, Sudan
- Attachment B-10: Government of Rwanda, poverty policy paper
- Attachment B-11: Overview of monitoring tools that can be used by aid agencies
- Attachment B-12: Results from Mozambique and Bolivia community health programs
- Attachment B-13: Tsunami relief, Indonesia
- Attachment B-14: Uganda food and water proposal doc 1
- Attachment B-15: Uganda food and water proposal doc 2
- Attachment B-16: Uganda food and water proposal doc 3
- Attachment B-17: World Vision terracing project evaluation