Charity Science Health (CSH) currently has a room for funding gap of approximately $495,000 USD over the next 2.5 years (or $247,500 over 1 year). Filling this gap would allow us to a) run our program and directly save lives and b) run an RCT on our program to evaluate its effects which would substantially increase the odds of us getting GiveWell recommended, thus indirectly saving more lives.
Getting GiveWell recommended would enable us to raise large amounts of funds, including outside of the EA community, for years to come, thus leveraging current donations to cause more in the future. Furthermore, success would inspire others in the EA movement to start GiveWell recommended charities and attract more talent to our organization.
Case for impact
Potential for GiveWell recommendation
While there are already six RCTs on our program in the developing world and dozens in the developed world, our intervention of sending SMS vaccine reminders is a behavior change intervention, so it is more likely to suffer from external validity programs. People are much more variable in behavior than biology, so medical interventions, like bed nets, that work in one area are very likely to work in another. However, due to cultural differences, reminders that work in Kenya could not work in India. This is on top of the fact that there are already massive issues with external validity among all development research because of variation in the details of implementation, different contexts, etc. That is why GiveWell would be much more likely to recommend us if we had a rigorously designed study on our particular intervention to establish its effects.
Funding to CSH has additional indirect effects, one of which is inspiring others. Already due to our preliminary success, another org (Fortify Health) has started and a third org is in the works. Getting a GiveWell recommendation would no doubt galvanize even more people to start down the path of charity entrepreneurship. This impact is arguably larger than even our direct impact of saving lives if multiple effective charities end up being founded due to this project.
Another indirect effect is the allocation of CSH staff time. There are many EAs working at CSH and right now their time is caught up in CSH almost entirely. If the RCT comes back negative, it would save many years of EA time for other high impact activities.
Lastly, if you are concerned about the flow through effects of poverty interventions on farm animals, CSH works in North India where meat consumption is very low. This can be a good option if you care about that concern but still wish to alleviate poverty.
There are currently six RCTs on our intervention in the developing world, and enough in the developed world to have two Cochrane reviews on it. You can see a summary of the evidence here and here.
GiveWell has published a preliminary cost-effectiveness analysis (CEA) that puts us in the realm of being as cost-effective as AMF. This is based on slightly out of date costs (we are now a bit more cost-effective) and projected effect size. The latter has largely remained unchanged other than that we work in lower vaccination rate locations than the national average. If we get more funding we will be able to experiment with cheaper ways of signing people up into our program and run an RCT to ascertain exactly how much we increase vaccination rates.
The GiveWell CEA was conservative about the effect size, predicting a 2-4% increase depending on current vaccination rate, whereas the previous studies found a 4% to 17.6% effect size. Nonetheless, according to this current estimate, we just need a moderate effect size to be as cost-effective as AMF, even if we cannot lower our costs with scale and new experiments.
For our internal calculations we use this same model with the same numbers but with the updated, more state-specific vaccination rates (~70% vs ~87%) and lower cost per enrollee ($0.50 vs $0.65). If you make those modifications it updates our cost-effectiveness estimate to about two times AMF’s current cost-effectiveness. This is our current best guess estimate.
Why is there a gap?
We have not had a gap in the last two years because our full room for more funding was filled before we could make it public. This year is different for two reasons:
In our first year and a half of running we have accomplished a lot. We’ve received two rounds of GiveWell incubation funding, hired a team on the ground in India, and become registered as a charity in India. After initially experimenting with different methods to acquire numbers, we found one that has allowed us to sign up 100,000 people in 4 months. This method of going door-to-door is very replicable and scalable; it just needs funds. If the RCT reveals that it is cost-effective, it could help millions of children.
We are also signing up hospitals, such that every person who gives birth there is enrolled. This is slower to scale because different bureaucracies have to be persuaded for each hospital. However, once they are signed up it costs less per person as the signup is passive. At the moment we have four hospitals signed up to our program, and with the social signalling of having other hospitals on board, others are joining faster and easier than previously.
We also generally have a good track record of getting things done at low cost. Joey and Katherine Savoie have successfully started and run Charity Science Outreach, raising nine counterfactual dollars for every dollar spent on fundraising, as well as testing several different fundraising methods cost-effectively. The Charity Entrepreneurship research that lead to the founding of CSH and Fortify Health was done on a total budget of $60,000. We also played a role in getting research cost-effectively done in the EA field of animal rights.
Budget and timeline
Here is our budget and timeline for the year. Broadly speaking, marginal funds will first fund basic operations (e.g. signing people up into our program, staff costs), then a small retrospective cohort study, then a full randomized controlled trial. The retrospective cohort study is a relatively inexpensive way to test the effect size, which will help narrow down the necessary sample size for the full RCT. Additionally, if it comes back with sufficiently precise null results, it may disprove our impact sooner, thus saving the money and effort of the full RCT. The marginal funds for the RCT will improve its sample size and rigour, which in turn will increase the probability of a GiveWell recommendation.
If you are interested in donating you can donate online. If you are making a donation of $1000 or more, please send an email to Peter Quin-Conroy, our operations officer, at email@example.com, so we can provide you with donation options that could significantly reduce third-party fees.
If you do not live in Canada, you can still donate to us. If you would like a tax benefit, send us an email and we can give you the best options for tax deductibility given your country.
1. Of note, if you are concerned about that, Deworm the World does a good portion of its work in India as well