Why Does Cost-Effectiveness Matter In Global Poverty Alleviation?
The best charities are fifteen thousand times better than others. (Introductory, unlikely to be very interesting to experienced EAs.)
A study suggests that the average person thinks it costs about $488 to save a life in the developing world, and $360 if you donate to the most cost-effective charities. In fact, this is completely wrong. In fact, current research suggests that the most cost-effective charities can save a life for fifteen thousand times less money than other charities—and some charities may even save us money on net.
It's easy to make mistakes about how important cost-effectiveness is. People tend to think about cost-effectiveness as being mostly about how good charities are at budgeting. Maybe a more cost-effective charity can cut back on the air conditioning or fly United Airlines and squeeze out a couple more lives saved, but surely they can't be that much different. What people don't think about is that some health care is cheaper to provide than other health care.
Toby Ord's excellent The Moral Imperative Towards Cost-Effectiveness In Global Health reviews the literature, and I encourage you to read it if you're curious. But you can see the basic argument by looking at this chart:
A DALY, or disability-adjusted life year, is a way of measuring how good a treatment is. It accounts for how many years of life are saved, how sick a person is, and how much the treatment makes them better. Lots of national health services and public health officials use them to figure out which services they should fund and which don't offer enough bang for the buck. If you're interested in learning more about DALYs, this is a good introduction. But you don't need to know much about them to understand this blog post: just that they're a very carefully researched measure and that the more DALYs a treatment saves the better it is at increasing people's lifespans and making them less sick.
As you can see from this chart, surgical treatment for Kaposi's sarcoma (a common cancer in people with HIV) is so cost-ineffective that the bar doesn’t even show up. HIV education for high-risk groups is 1,400 times better: that is, if you give a thousand dollars to surgical treatment for Kaposi's sarcoma and a thousand dollars to HIV education for high-risk groups, you will prevent a thousand four hundred times as much HIV-related illness. That's a big difference!
Think about it this way: if you teach people about how to prevent HIV, they won't get HIV at all. HIV treatment is expensive: a year of antiretrovirals can cost hundreds of dollars, even though drug companies tend to lower prices for people in the developing world. If you prevent people from getting HIV in the first place, it costs less than treating them once they already have it. But it's important not to make assumptions about how the numbers will end up working out before you do research. Before we studied it, we might have thought that HIV education didn't work at all: maybe people already know how to prevent HIV and make bad decisions on impulse, so teaching them won't change their behavior. We can't know which way it would end up working out without testing the programs to see what happens.
The chart I showed you is only comparing programs that treat or prevent HIV. In reality, we can fund prevention and treatment for many many different illnesses. You know that some diseases are cheaper to treat than other diseases, even if they're equally severe. You can buy some antibiotics which can treat a life-threatening infection for twenty-five dollars, but chemotherapy costs many hundreds of thousands of dollars. The same thing is true in the developing world. Right now, if we pick the most effective programs, economists estimate that the best is about fifteen thousand times better than the worst.
But some programs are even better than that. In fact, the single health program Toby Ord thinks was the most cost-effective of all time is smallpox eradication. Smallpox eradication cost about $1.7 billion dollars in today's money, which is an eye-popping figure. But so far it has saved more than a hundred million lives. If smallpox were released from a lab tomorrow and began to wreck havoc on the globe, smallpox eradication would have cost a mere seventeen dollars per life saved. And of course that number is going to drop over time, because once a disease is eradicated it usually stays eradicated. What's more, smallpox eradication doesn’t even cost money on net: because we no longer have to pay for vaccination programs or medicines for smallpox, it pays for itself every few months.
So what's the takeaway, other than that we should eradicate more diseases? It's very important to pick the best possible charities. If you switch from donating to give people Kaposi’s sarcoma medications to the best of global health charities which currently exist, it's the same as going from donating one thousand dollars to donating a hundred and fifty thousand dollars-- without having to make any personal sacrifices.
Don't the last couple paragraphs about smallpox imply that we shouldn't actually be donating to stop malaria, but rather whichever disease we have the greatest chance of eradicating right now, even if its not actually directly saving very many l people? I'm sure the last few smallpox victims weren't very cost effective to treat, in terms of lives saved per dollar. But then if you accept that then you accept that future people count, and all of a sudden the math on existential risk dwarfs any disease we might eradicate.
How do you balance between treatment, prevention, research on the problem, and background research (more knowledge of biology, more math, better computers and computer programs)?