GiveWell Top Charities Explained: Helen Keller International

Link post

[This is the sec­ond post in a brief se­ries ex­plain­ing the cur­rent GiveWell top char­i­ties. You can get all the in­for­ma­tion in this post on GiveWell’s web­site, but my blog post is both shorter and less bor­ing.]

He­len Kel­ler In­ter­na­tional performs Vi­tamin A sup­ple­men­ta­tion. It is not only a brand new top char­ity this year, it is also the sec­ond new cause GiveWell recom­mended dona­tions in this year.

Vi­tamin A defi­ciency can cause stunt­ing, ane­mia, blind­ness, more se­vere in­fec­tions, and death. It is par­tic­u­larly harm­ful to preg­nant or lac­tat­ing peo­ple, in­fants, and chil­dren. Peo­ple in de­vel­op­ing coun­tries who don’t eat much meat and don’t eat Vi­tamin-A-for­tified food may suffer from defi­cien­cies. (Vi­tamin A is also a nu­tri­ent of con­cern for ve­g­ans and veg­e­tar­i­ans in the de­vel­oped world, but if you have an am­ple and di­verse diet of fruits and veg­eta­bles you prob­a­bly don’t have to worry about it.) Luck­ily, your liver can store Vi­tamin A for sev­eral months; if you take a very high dose of Vi­tamin A once ev­ery six months, you won’t suffer from Vi­tamin A defi­ciency. It is recom­mended that chil­dren be­tween 6 months and 5 years old get a sup­ple­ment once ev­ery six months.

He­len Kel­ler In­ter­na­tional pro­vides tech­ni­cal as­sis­tance, ad­vo­cacy, and fund­ing. Tech­ni­cal as­sis­tance in­cludes helping coun­tries mon­i­tor how many chil­dren are get­ting Vi­tamin A sup­ple­ments, run­ning cam­paigns that ed­u­cate par­ents about the im­por­tance of Vi­tamin A sup­ple­ments, train­ing health work­ers to give out sup­ple­ments, and helping gov­ern­ments figure out why their vi­tamin A sup­ple­men­ta­tion rates are so low and how they can fix them. Ad­vo­cacy in­volves con­vinc­ing gov­ern­ments that they should pri­ori­tize mass vi­tamin A sup­ple­men­ta­tion. He­len Kel­ler In­ter­na­tional also some­times pro­vides grants to gov­ern­ments to help them pay for vi­tamin A sup­ple­men­ta­tion pro­grams.

It is some­what un­clear whether vi­tamin A sup­ple­men­ta­tion ac­tu­ally works to re­duce mor­tal­ity rates. Nor­mally, this sort of thing is un­clear be­cause we don’t have any ev­i­dence about it. In the case of vi­tamin A sup­ple­men­ta­tion, we have two pieces of re­ally good ev­i­dence; they just point in op­po­site di­rec­tions.

The Cochrane Col­lab­o­ra­tion, whom you may re­mem­ber from the pre­vi­ous post as the peo­ple who aren’t sure whether floss­ing your teeth makes your teeth bet­ter, performed a meta-anal­y­sis that sug­gests that vi­tamin A sup­ple­men­ta­tion re­duces all-cause mor­tal­ity by 24%, with a 95% con­fi­dence in­ter­val rang­ing from 17% to 31%. The Cochrane Col­lab­o­ra­tion pri­mar­ily used stud­ies that were con­ducted in the 1980s and 1990s.

Not long af­ter the Cochrane Col­lab­o­ra­tion’s meta-anal­y­sis came out, we learned the re­sults of the DEVTA trial. The DEVTA trial is the sin­gle largest ran­dom­ized con­trol­led trial ever con­ducted, with one mil­lion chil­dren par­ti­ci­pat­ing. It es­ti­mated that Vi­tamin A sup­ple­men­ta­tion re­duces child mor­tal­ity rates by 4% and could not rule out the pos­si­bil­ity that it did not af­fect child mor­tal­ity rates at all. You may no­tice that 4% is in fact much much smaller than 24%.

What the fuck is go­ing on?

  • It might just be ran­dom chance. That’s pretty un­likely: the Cochrane Col­lab­o­ra­tions 95% con­fi­dence in­ter­val doesn’t over­lap at all with DEVTA’s.

  • DEVTA might not have treated as high a per­centage of the chil­dren in the study as claimed. DEVTA claims to have treated 86% of chil­dren, but some re­searchers are skep­ti­cal be­cause DEVTA was done very cheaply. DEVTA seems to be us­ing broadly rea­son­able strate­gies to get all chil­dren to take vi­tamin A sup­ple­ments and to figure out how many chil­dren ac­tu­ally took it, but their strate­gies aren’t very well-doc­u­mented and some­times they didn’t im­ple­ment them un­til halfway through the study. Nev­er­the­less, the per­centage of chil­dren treated in or­der to make ‘they didn’t treat enough kids’ plau­si­ble as an ex­pla­na­tion is so much lower than the per­centage of chil­dren claimed to have been treated that this is not a very plau­si­ble ex­pla­na­tion.

  • DEVTA might have treated a pop­u­la­tion with less se­vere or preva­lent vi­tamin A defi­ciency. How­ever, the rate of vi­tamin A defi­ciency, se­vere vi­tamin A defi­ciency, and com­pli­ca­tions re­lated to vi­tamin A defi­ciency is similar in DEVTA as it is in other stud­ies. It’s more likely they un­der­es­ti­mated vi­tamin A defi­ciency than that they over­es­ti­mated it.

  • DEVTA’s pop­u­la­tion might be healthier than other pop­u­la­tions. Vi­tamin A defi­ciency doesn’t gen­er­ally kill chil­dren di­rectly; it kills them in­di­rectly, by mak­ing them more sus­cep­ti­ble to in­fec­tions. If those deaths are be­ing pre­vented some other way (e.g. measles vac­cia­na­tions, oral re­hy­dra­tion treat­ment), treat­ing Vi­tamin A defi­ciency saves fewer chil­dren’s lives. DEVTA had a lower child mor­tal­ity rate than most of the stud­ies in the Cochrane re­view, and gen­er­ally stud­ies with a lower child mor­tal­ity rate show a smaller effect from vi­tamin A sup­ple­men­ta­tion. How­ever, it’s un­clear whether measles and di­ar­rhea– the two biggest kil­lers re­lated to Vi­tamin A– were less com­mon in DEVTA than in other stud­ies.

GiveWell thinks the most likely ex­pla­na­tion is the last one. That means that whether Vi­tamin A sup­ple­men­ta­tion is cost-effec­tive de­pends on not just how high the Vi­tamin A defi­ciency rates are but also how high the child mor­tal­ity rate is.

The coun­tries He­len Kel­ler In­ter­na­tional works in typ­i­cally have lower rates of vi­tamin A defi­ciency than in any study of the effects of vi­tamin A sup­ple­men­ta­tion: they work in coun­tries where 20% of preschool-aged chil­dren have vi­tamin A defi­ciency, com­pared to 59% in the Cochrane meta-anal­y­sis. Hel­len Kel­ler In­ter­na­tional works in coun­tries where 12 chil­dren out of ev­ery 1000 die ev­ery year; pre­vi­ous stud­ies have found an effect of vi­tamin A sup­ple­men­ta­tion if more than 10 chil­dren out of ev­ery 1000 die ev­ery year.

How­ever, there are a lot of limi­ta­tions of this es­ti­mate. He­len Kel­ler In­ter­na­tional typ­i­cally works in re­gions, rather than in whole coun­tries, which might have higher or lower child mor­tal­ity rates than the coun­try as a whole. It seems re­ally un­likely that vi­tamin A sup­ple­men­ta­tion doesn’t do any­thing be­low 10 chil­dren out of 1000 dy­ing each year and then sud­denly has a big effect as soon as you get to 10; it’s prob­a­bly a smoother effect that’s harder to an­a­lyze. “Child mor­tal­ity” is a statis­tic that in­cludes a lot of differ­ent things. It’s un­clear whether vi­tamin A sup­ple­men­ta­tion helps with all in­fec­tious dis­eases or a sub­set, such as measles and di­ar­rhea; it’s pretty clear that vi­tamin A sup­ple­men­ta­tion has no effect on some other causes of child mor­tal­ity, like car ac­ci­dents. If a re­gion has a high child mor­tal­ity rate be­cause there are a lot of car ac­ci­dents, vi­tamin A sup­ple­men­ta­tion might not do any­thing. More re­al­is­ti­cally, if vi­tamin A sup­ple­men­ta­tion re­duces deaths by caus­ing chil­dren to be less likely to die if they get measles, then if a re­gion has a low rate of measles, it won’t have a big effect from vi­tamin A sup­ple­men­ta­tion, even if it has a high child mor­tal­ity rate.

There are not likely to be any nega­tive long-term side effects from vi­tamin A sup­ple­men­ta­tion. In the short term, less than ten per­cent of chil­dren ex­pe­rience some sort of ad­verse side effect, such as headaches, nau­sea, vom­it­ing, ir­ri­ta­bil­ity, fever, or loose stools. Vi­tamin A sup­ple­men­ta­tion does not cause vi­tamin A over­dose or in­crease mor­tal­ity when given alongside an in­ac­ti­vated vac­cine.

He­len Kel­ler In­ter­na­tional gives a rel­a­tively high per­centage of tar­get chil­dren vi­tamin A sup­ple­ments (be­tween 46% and 81%, de­pend­ing on re­gion).

GiveWell be­lieves He­len Kel­ler In­ter­na­tional’s grants cause vi­tamin A sup­ple­men­ta­tion dis­tri­bu­tions that oth­er­wise would not oc­cur, but does not know whether its tech­ni­cal as­sis­tance helps coun­tries to give vi­tamin A to chil­dren who oth­er­wise wouldn’t re­ceive vi­tamin A. For this rea­son, GiveWell’s analy­ses in­clude only the effect of grant­mak­ing, not the effect from tech­ni­cal as­sis­tance.

He­len Kel­ler In­ter­na­tional needs $20.7 mil­lion over the next three years.

One big area of un­cer­tainty comes from the fact that He­len Kel­ler In­ter­na­tional has been in­ves­ti­gated in less de­tail than other top char­i­ties. In gen­eral, over time, GiveWell tends to be­come more un­cer­tain about char­i­ties, learn more about their limi­ta­tions, and have a higher cost-per-life-saved-equiv­a­lent num­ber at­tached to the char­ity.

Why might you donate to He­len Kel­ler In­ter­na­tional?

  • You want to save the lives of chil­dren un­der 6.

  • You want to donate to some­thing that definitely won’t cause sig­nifi­cant harm, even if it might not have an effect.

  • You’re op­ti­mistic about GiveWell’s fur­ther in­ves­ti­ga­tions find­ing that He­len Kel­ler In­ter­na­tional is as effec­tive as we thought, not less effec­tive.

  • You think He­len Kel­ler In­ter­na­tional’s tech­ni­cal as­sis­tance is more likely to be effec­tive than GiveWell thinks it is.

  • You’re not too con­cerned about low rates of vi­tamin A defi­ciency or about the un­cer­tain­ties as­so­ci­ated with child mor­tal­ity rates.

No comments.