The best places to donate for COVID-19

<NOTE: this piece is cross-posted to the SoGive blog>

TL;DR

We (SoGive) are re­ceiv­ing some queries about where to donate for donors want­ing to re­spond to COVID-19. Here we give our an­swers, us­ing our ap­proach in­fluenced by Effec­tive Altru­ism.

Here’s the short an­swer: If men­tion­ing just one or­gani­sa­tion, our an­swer is:

  • Johns Hop­kins Cen­ter for Health Se­cu­rity (CHS)

If it’s suit­able to provide a range of or­gani­sa­tions (in­clud­ing those which are more in­di­rectly rele­vant to COVID-19), our an­swers are:

  • Johns Hop­kins Cen­ter for Health Se­cu­rity (CHS)

  • Biose­cu­rity ini­ti­a­tive at Cen­tre for In­ter­na­tional Se­cu­rity and Cooperation

  • Univursa Health

  • Devel­op­ment Me­dia In­ter­na­tional (DMI)

  • CSER (Cen­tre for Study of Ex­is­ten­tial Risk)

Note that this work has been performed as quickly as pos­si­ble and our think­ing may evolve.

0. Exec Summary

This re­port briefly con­sid­ers rea­sons why COVID-19 may not be the best area to fund. Nonethe­less, we are con­scious that there are donors in­ter­ested in this area, so we have gen­er­ated our list us­ing the fol­low­ing method.

“We” refers to SoGive, an or­gani­sa­tion sup­port­ing donors to achieve high-im­pact giv­ing.

Method: We gen­er­ated a list of po­ten­tial dona­tion tar­gets by perform­ing a google search and reach­ing out to some of our con­tacts. This gen­er­ated around 20 pro­jects. We ex­cluded promis­ing COVID-19-rele­vant ideas which didn’t already have a shovel-ready or­gani­sa­tion ac­cept­ing funds for it, al­though there is an ap­pendix which ex­plores this area.

We then put the or­gani­sa­tions into an­a­lyt­i­cally rele­vant cat­e­gories:

  • Co-benefits: Th­ese or­gani­sa­tions have the prop­erty that they may ap­peal to donors in­ter­ested in COVID-19, and po­ten­tially have other benefits too

  • Ne­glected: we be­lieve that most COVID-19-re­lated work has plenty of fund­ing; any­thing which we con­sider to be ne­glected is in­cluded in this category

  • Not fund­ing-con­strained: for these ar­eas of work, there is a con­straint that is likely to bite be­fore our col­lec­tive abil­ity to fund (e.g. enough peo­ple with the right skills)

  • Crowd­ing out: We be­lieve that philan­thropy should only fund goods/​ser­vices when an­other stake­holder (e.g. gov­ern­ment or the pri­vate sec­tor) wouldn’t fund it. When donors at­tempt to fund such work, they risk crowd­ing out more suit­able fun­ders.

  • No plan: Some or­gani­sa­tions are re­ceiv­ing funds for work on COVID-19, but at the time of writ­ing do not ap­pear to have in­for­ma­tion on how the funds will be used

In our method­ol­ogy, the first two cat­e­gories were viewed more pos­i­tively than the fol­low­ing three.

Th­ese or­gani­sa­tions are listed in the table be­low.

At the mo­ment, where we have suffi­cient anal­y­sis on the or­gani­sa­tion (with pos­i­tive con­clu­sions!), and where it falls into a cat­e­gory which we as­sess pos­i­tively, we treat it as one of our recom­men­da­tions. The anal­y­sis may be anal­y­sis con­ducted by other or­gani­sa­tions (e.g. Founders Pledge) which we have then re­viewed).

Our recom­men­da­tions are ten­ta­tive, be­cause we have aimed to pro­duce this quickly, rather than perfectly. As more in­for­ma­tion arises, our opinions may change.



Note that “not recom­mended” does not nec­es­sar­ily mean that a dona­tion to that area is bad, but rather that we be­lieve that other pos­si­bil­ities likely out­perform.

Johns Hop­kins CHS is cho­sen as the sin­gle or­gani­sa­tion to recom­mend (if only one is cho­sen) be­cause of (a) the strong co-benefits and (b) the di­rect rele­vance to COVID-19.

1. Is COVID-19 a good area to fund?

The first ques­tion to ask is whether coro­n­avirus is the right area to sup­port if you want to donate. To as­sess this cause area us­ing the stan­dard three fac­tor frame­work of Im­pact-Tractabil­ity-Ne­glect­ed­ness, a cur­sory as­sess­ment sug­gests the fol­low­ing for fund­ing of coro­n­avirus-re­lated work as a whole:

Most ar­eas of COVID-19 work are crowded enough that one may ques­tion whether it’s the best area to donate to.

Our ex­pe­rience of the char­ity sec­tor at the time of the Global Fi­nan­cial Cri­sis of c2009 sug­gests that we ex­pect the fol­low­ing to hap­pen:

  • The need for all the var­i­ous (non-coro­n­avirus-re­lated) char­i­ta­ble pro­jects won’t go away; if any­thing the need will increase

  • Some char­i­ties will have lost money be­cause of the im­pact of the mar­kets on their investments

  • Many fun­ders will have lost money be­cause of volatility in the fi­nan­cial mar­kets, and will there­fore less able to pro­vid­ing funding

  • The re­main­ing fund­ing may be di­verted to coro­n­avirus-spe­cific work

We ten­ta­tively be­lieve that COVID-19 will be a pop­u­lar (or at least some­what pop­u­lar) dona­tion tar­get. Below is our as­sess­ment of how COVID-19 performs on our check­list of “propen­sity to be pop­u­lar with fun­ders”



As can be seen from this table, we pre­dict donor in­ter­est in COVID-19, but not to the same ex­tent as some other pop­u­lar causes, such as the tsunami ap­peal of c2005 or guide dogs or life­boats or home­less­ness, many of which would score similarly highly or more so on this chart.

How­ever we nonethe­less ex­pect that COVID-19 will nonethe­less at­tract sub­stan­tial donor in­ter­est. And given that fund­ing sources for other char­i­ta­ble ac­tivi­ties will likely con­tract, it is likely that the smart money should go to some­thing other than COVID-19.

In or­der to firm up our opinion on this, we would need to see a deeper anal­y­sis of the more effec­tive dona­tion op­por­tu­ni­ties high­lighted here. This would include

  1. Con­sid­er­a­tions around how much more fund­ing (if any) they would re­ceive as a re­sult of ele­vated in­ter­est in COVID-19, and what this means for the marginal use of the next dol­lar donated (es­pe­cially rele­vant for or­gani­sa­tions which have already been as­sessed and recom­mended prior to the COVID-19 out­break)

  2. Fur­ther mod­els of cost effec­tive­ness to fa­cil­i­tate com­par­i­sons with other dona­tion op­por­tu­ni­ties, to the ex­tent that this is pos­si­ble (es­pe­cially rele­vant for new pro­jects)

This doc­u­ment is geared to­wards those fun­ders who wish to sup­port COVID-19 de­spite these con­sid­er­a­tions.

2. Cat­e­gory 1: Co-benefits (Benefits for fu­ture pan­demics and other co-benefits)

<<< SoGive opinion on this cat­e­gory: pos­i­tive>>>

Or­gani­sa­tions in this cat­e­gory:

  • Johns Hop­kins Cen­tre for Health Security

    • a think tank at Johns Hop­kins Univer­sity which car­ries out re­search on biose­cu­rity and main­tains a COVID-19 re­source centre

  • Biose­cu­rity ini­ti­a­tive at Cen­tre for In­ter­na­tional Se­cu­rity and Cooperation

    • a re­search cen­tre at Stan­ford that car­ries out policy re­search and in­dus­try out­reach; has gen­eral rele­vance to pandemics

  • Cen­tre for the Study of Ex­is­ten­tial Risk (CSER)

    • a re­search cen­tre at the Univer­sity of Cam­bridge, in­tended to study pos­si­ble ex­tinc­tion-level threats, in­clud­ing pandemics

  • Tack­ling wet mar­kets (mul­ti­ple orgs)

    • The COVID-19 pan­demic started at a wet mar­ket in Wuhan

Or­gani­sa­tions which gen­er­ally tackle pan­demics are a good choice. Lots of work re­lat­ing to gen­eral pan­demics is also rele­vant to COVID-19 be­cause of the pos­si­bil­ity of a re­sur­gence or sec­ond wave of the pan­demic.

The or­gani­sa­tions listed here have the prop­erty that they may ap­peal to donors in­ter­ested in COVID-19, and po­ten­tially have other benefits too. For ex­am­ple the or­gani­sa­tions may also do other high-im­pact work re­lat­ing to ex­is­ten­tial risk, and dona­tions to these or­gani­sa­tions may funge with those high-im­pact ar­eas of work, as well as po­ten­tially fund­ing good work on pan­demics.

The ex­is­tence of co-benefits (e.g. rele­vance to fu­ture pan­demics or other sources of ex­is­ten­tial risk) is im­por­tant to our as­sess­ment of the work as pos­i­tive for a donor, given that most work which is spe­cific to COVID-19 is likely to already be highly crowded (rare ex­cep­tions noted in the sec­tion on ne­glected work).

Tack­ling wet mar­kets has a clear link with COVID-19, since this is thought to be where the COVID-19 epi­demic came from. The key out­stand­ing ques­tions are the ex­tent to which this re­mains an is­sue given the re­cent Chi­nese ban on wet mar­kets.

3. Cat­e­gory 2: Ne­glected work on coronavirus

<<< SoGive opinion on this cat­e­gory: pos­i­tive>>>

We strug­gle to find any work on coro­n­avirus which is ne­glected. The only cat­e­gory we know of is:

  • Some work on COVID-19 in the de­vel­op­ing world

Or­gani­sa­tions in this cat­e­gory:

  • Devel­op­ment Me­dia International

    • Pro­ject in Mozam­bique and Malawi (and pos­si­bly other coun­tries) to pro­mote hand­wash­ing and other rele­vant behaviours

  • Univursa Health

    • They are look­ing to build a di­ag­nos­tic tool us­ing al­gorithms and ma­chine learn­ing to rapidly iden­tify pos­si­ble out­breaks of COVID-19 within their bor­ders, where those coun­tries have limited ca­pac­ity to do lab­o­ra­tory-based test­ing. There is a par­tic­u­lar fo­cus on sub-Sa­haran Africa.

In our con­ver­sa­tions with Devel­op­ment Me­dia In­ter­na­tional (DMI) we un­der­stand that some COVID-19 re­lated work is ne­glected, es­pe­cially around things which are coun­try-spe­cific (e.g. vac­cine de­vel­op­ment should work on all peo­ple; de­vel­op­ing a ra­dio spot that works in Burk­ina Faso is spe­cific to Burk­ina Faso). At time of writ­ing we have not been able to fully ex­plore/​ver­ify this claim in gen­eral, and would wel­come fur­ther feed­back about the ex­tent to which this ap­plies.

We have in­cluded Univursa Health in the “ne­glected” cat­e­gory, but it could equally be in­cluded in the “co-benefits” cat­e­gory, since the mechanism could be used for de­tect­ing many other things.

4. Cat­e­gory 3: Not fund­ing-constrained

<<<SoGive opinion: ul­ti­mate im­pact un­clear, not recom­mended>>>

Ex­am­ples of the sorts of work that is not fund­ing-con­strained

  • Send­ing sup­plies such as per­sonal pro­tec­tive or hy­giene equip­ment to frontline workers

  • Health ed­u­ca­tion mes­sag­ing in the de­vel­oped world

  • Re­search on vac­cines and treatments

Or­gani­sa­tions in this category

For these items, there is a con­straint that is likely to bite be­fore our col­lec­tive abil­ity to fund. For ex­am­ple we will likely run out of abil­ity to man­u­fac­ture per­sonal pro­tec­tive equip­ment such as masks be­fore we run out of money to pay for them; we will hit the point of diminish­ing marginal re­turns on health ed­u­ca­tion mes­sag­ing in the de­vel­oped world be­fore de­vel­oped world gov­ern­ments run out of fund­ing for that mes­sag­ing; we will run out of skil­led peo­ple able to do high qual­ity re­search on vac­cines and treat­ments for coro­n­avirus (or to provide peer re­view on it) be­fore we hit a con­straint on money to fund that work. This ob­ser­va­tion re­lates fund­ing from all sources, not fund­ing just from an in­di­vi­d­ual donor, or even the sum of philan­thropic fund­ing.

So if you fund this sort of work, your money may well be no­tion­ally ear­marked to that work, how­ever this is likely to be illu­sory re­strict­ed­ness (oth­er­wise known as fung­ing). The true im­pact of your dona­tion is that the bod­ies who would in any case have funded that work now have more money to fund what­ever they would have funded at the mar­gin.

This may be no bad out­come at a time when gov­ern­ments and fund­ing bod­ies ev­ery­where are short of funds, how­ever it’s hard for us to provide an as­sess­ment on the value of this, es­pe­cially when some of those things (e.g. vac­cine de­vel­op­ment) are cur­rently be­ing funded mul­ti­lat­er­ally by lots of differ­ent fun­ders, so a full as­sess­ment of this ques­tion would need to con­sider who those fun­ders are, and how im­pact­ful their use of funds would be at the mar­gin.

Once the rele­vant stake­hold­ers and their fund­ing flows are clearly mapped, it may turn out that this fund­ing is equiv­a­lent to pay­ing ex­tra tax to the gov­ern­ment. Or it may be that ex­tra fund­ing is stored up which can then be used on fund­ing pan­demic vac­cine re­search over the com­ing years. Given that this form of re­search has been prone to cy­cles of be­ing un­der­funded and then over­funded, stor­ing up funds which will only be dis­bursed over sev­eral years may still be a pos­i­tive out­come, es­pe­cially for an area like bio risk, which is a fairly high-im­pact cause area.

5. Cat­e­gory 4: Tack­ling a prob­lem that risks crowd­ing out other players

<<<SoGive opinion: not recom­mended>>>

Ex­am­ples of types of work:

  • do­ing work which may soon be done by the gov­ern­ment, such as bailouts for restau­rant staff or freelancers

  • do­ing work which could be pro­vided by the pri­vate sec­tor, for ex­am­ple pro­vid­ing books to children

Or­gani­sa­tions in this cat­e­gory:

Donat­ing/​philan­thropy is a good idea for things that are ne­glected. If an­other body has in­cen­tives to provide cer­tain goods or ser­vices, then it’s bet­ter if they do it. Bodies such as gov­ern­ments or for-profit com­pa­nies are ex­am­ples of other such bod­ies.

As an ex­am­ple, donat­ing to a foun­da­tion which will help out free­lance work­ers risks be­ing un­helpful be­cause the gov­ern­ment may choose to provide a bailout pack­age any­way, in which case your dona­tion turned out to have min­i­mal value.

We have ten­ta­tively in­cluded food banks in the UK in this cat­e­gory. This is based on the ex­pec­ta­tion that the UK gov­ern­ment would have lit­tle in­cen­tive to al­low peo­ple to starve be­cause of the pan­demic, and that it would be difficult to ex­pect a vol­un­teer-run food bank sys­tem to func­tion effec­tively when mem­bers of the pub­lic are largely stay­ing at home.

While we con­sider it likely that the UK gov­ern­ment will take some ac­tion in this area, we are un­clear on when this will hap­pen, and how many peo­ple will suffer fi­nan­cial ruin be­fore then.

6. Cat­e­gory 5: No plan: Or­gani­sa­tions seek­ing fund­ing but pro­vid­ing no plan of ac­tion for funds

<<<SoGive opinion: not recom­mended>>>

Ex­am­ple dona­tion op­por­tu­ni­ties in this cat­e­gory:

We are aware of some or­gani­sa­tions who are cur­rently con­scious that peo­ple want to donate to sup­port the COVID-19 crisis. In light of this they are pro­vid­ing donors the op­por­tu­nity to donate to a fund to tackle this is­sue, but with­out spec­i­fy­ing how the money will be used.

Those of a sus­pi­cious dis­po­si­tion may fear mis­ap­pro­pri­a­tion. While this is pos­si­ble, we think it more likely that the or­gani­sa­tions in ques­tion sim­ply want to give them­selves more time to con­sider how best to re­spond to the crisis.

Nonethe­less, we do not feel in a po­si­tion to recom­mend a dona­tion op­por­tu­nity with­out any in­for­ma­tion about how the money will be used

Ap­pendix: not shovel-ready philan­thropy ideas

This re­view has fo­cused on shovel-ready pro­jects which can re­ceive funds now.

In this ap­pendix we list some other po­ten­tial ideas for things to fund. Th­ese things might not cur­rently ex­ist, but could be brought about by one large fun­der or a group of medium-sized fun­ders.

Th­ese ideas have not been care­fully as­sessed at time of writing

  • Prize to en­courage cre­ation and wide­spread adop­tion of con­tact trac­ing app

    • Con­tact trac­ing in­volves iden­ti­fy­ing peo­ple who have been in con­tact with those who have tested pos­i­tive for the in­fec­tion; it has been used suc­cess­fully in some jurisdictions

    • A key challenge is mak­ing this work in a way which re­spects data privacy

    • Suc­cess here re­quires both the cre­ation of the app and wide­spread adoption

    • Ar­gu­ments against this prize are that we know of two groups already who are work­ing on this

    • Ar­gu­ments in favour in­clude the fact that if the prize in­cludes an in­cen­tive for wide­spread adop­tion, this would sup­port out­comes which are not yet guaranteed

  • Prizes for fur­ther di­ag­nos­tic tools

    • We be­lieve that di­ag­no­sis would de­liver sub­stan­tial value; speci­fi­cally a mechanism which can help with di­ag­no­sis, albeit im­perfectly, but op­er­ate au­to­mat­i­cally and at scale is a gap

    • A ma­jor ar­gu­ment against this is crowd­ed­ness again. Billions of dol­lars are already be­ing poured into this sort of work.

  • Work on facemasks

    • Face­masks are rel­a­tively cheap and simple

    • They have at least some effec­tive­ness in helping con­trol the spread of disease

    • Gaps that we see in­clude un­der­stand­ing whether a sim­ple mask which doesn’t meet the re­quired stan­dards still has some effec­tive­ness, and could be used to alle­vi­ate the de­mand for masks that are needed in med­i­cal contexts

  • Broad-spec­trum antivirals

    • Broad-spec­trum an­tivirals are difficult: viruses are small, so tar­get­ing them is hard

    • A vac­cine or drug which would be (even par­tially) effec­tive against a wide ar­ray of viral pathogens would sub­stan­tially help to con­trol pan­demics such as the cur­rent COVID-19 pandemic

    • Re­search on this area is not to­tally ne­glected (see e.g. Martinez et al 2015 or Jones et al 2020), how­ever may be ne­glected rel­a­tive to the scale of the im­pact that could be achieved by such work

    • We are un­clear on the ex­tent to which re­search on this area has the po­ten­tial to be ap­pro­pri­ated for mal­i­cious purposes

    • This has the benefit that it would help against a sec­ond, mu­tated wave of COVID-19 pan­demic as well as fu­ture pandemics

Ap­pendix: Johns Hop­kins Cen­ter for Health Se­cu­rity (CHS)

The Johns Hop­kins Cen­ter for Health Se­cu­rity (CHS) is a think tank at Johns Hop­kins Univer­sity which car­ries out re­search on biose­cu­rity, and ad­vo­cates for im­proved policy in the US and in­ter­na­tion­ally.

Our recom­men­da­tion of CHS is based largely on the recom­men­da­tion by Founders Pledge, set out in their re­port on ex­is­ten­tial risk. Their recom­men­da­tion of CHS re­flects past suc­cesses such as its de­vel­op­ment of guidance for the US on a re­sponse to an­thrax, smal­l­pox, plague, and Ebola be­fore the US gov­ern­ment had any guidance or efforts on these is­sues.

The CHS is very ac­tive on COVID-19. For ex­am­ple, they are main­tain­ing a coro­n­avirus re­source cen­tre: https://​​coro­n­avirus.jhu.edu/​​

Ap­pendix: Univursa Health

Univursa Health is a new startup de­vel­op­ing a di­ag­nos­tics tool.

The con­cept was de­vel­oped by Freya Jeph­cott, a highly ex­pe­rienced field epi­demiol­o­gist and med­i­cal an­thro­pol­o­gist with field ex­pe­rience through her work with MSF and WHO and Emma Glen­non, a lead­ing in­fec­tious dis­eases mod­el­ler.

Univursa be­gan when Freya, hav­ing just re­turned from an out­break re­sponse in Ethiopia, gave a pre­sen­ta­tion on the challenges in­volved in iden­ti­fy­ing the causative agent of out­breaks, es­pe­cially in re­source-limited set­tings. Emma im­me­di­ately recog­nised that she could build math­e­mat­i­cal mod­els of the pro­cesses be­ing de­scribed and through them rapidly iden­tify the causative agents of out­breaks.

They have since then started work on a di­ag­nos­tic tool.

The tool can ei­ther:

  • Take in in­puts about symp­toms and pro­duce an es­ti­mated diagnosis

  • Or where a di­ag­no­sis isn’t pos­si­ble, at least in­di­cate how to pri­ori­tise which di­ag­nos­tic tools to use first

The tool has origi­nally been de­vel­oped for use on haem­or­rhagic con­di­tions such as ebola. It takes in a set of symp­toms called a “line list”.

The tool then uses an al­gorithm to turn this list of symp­toms into a di­ag­no­sis. In our con­ver­sa­tion with Univursa, we did not spend too much time on the al­gorithm, how­ever our un­der­stand­ing of the al­gorithm is sum­marised below

  • The al­gorithm takes in in­for­ma­tion about a num­ber of differ­ent symp­toms (say N symp­toms)

  • The val­ues for each symp­tom for a given case could be con­sid­ered as a po­si­tion within an N-di­men­sional space

  • The al­gorithm then con­sid­ers where in that space the read­ing lies, and calcu­lates the dis­tance be­tween the symp­tom-set con­sid­ered for that par­tic­u­lar in­put com­pared with the lo­ca­tions in N-di­men­sional space that are as­so­ci­ated with a par­tic­u­lar di­ag­no­sis, us­ing a straight­for­ward Eu­clidean dis­tance metric

  • The out­put is a prob­a­bil­ity as­so­ci­ated with each pos­si­ble di­ag­no­sis, from which an over­all di­ag­no­sis can be inferred

The tool has been tested in a the­o­ret­i­cal con­text. They have tested the tool on some past ex­am­ples of “cryp­tic cases” where di­ag­no­sis has been difficult and the tool has typ­i­cally been able to get it right within about 5 or so cases.

We are some­what scep­ti­cal of these ap­par­ently pos­i­tive find­ings. We worry that

  • Ap­par­ent suc­cess based on the cases that have been cho­sen to be the most cryp­tic may not nec­es­sar­ily gen­er­al­ise to work in the field

  • More gen­er­ally, the fact that the work has not been tested in the field means that some risk re­mains to the pro­ject’s success

Hav­ing said that, the tool is in­tended to be used in a re­source-con­strained set­ting, e.g. in a de­vel­op­ing world con­text where the most com­mon ap­proach may be to bring in for­eign ex­perts via or­gani­sa­tions like MSF or the WHO. Ex­perts who will now be in very short sup­ply. This means it has to beat a very low bar in or­der to out­perform.

And even if we don’t be­lieve that it’s likely to out­perform even this very low bar, the benefits are high enough that the op­por­tu­nity seems worth ex­plor­ing.

To as­sess this, we would like to quan­tify this in a model which con­sid­ers the scale of the benefit, as mea­sured in, e.g., lives saved that could be brought about as a re­sult of this work, com­pared with the cost of im­ple­ment­ing the work. This model has not been pro­duced at time of writ­ing.

The main tasks for Univursa to work on next are:

  • Adapt the tool to shift from haem­or­rhagic con­di­tions (like ebola) to res­pi­ra­tory (speci­fi­cally COVID-19)

  • Perform­ing field tests

We un­der­stand from Univursa that other similar tools are ei­ther data col­lec­tion tools to help with di­ag­no­sis, or anal­y­sis tools that could be used by some­one with a rel­a­tively so­phis­ti­cated grad school back­ground, and likely not ac­cessible for a typ­i­cal dis­trict com­mis­sioner in sub-Sa­haran Africa.

We have ex­plored with Univursa whether they have sought other sources of fund­ing. Thus far, Univursa have been in touch with some for-profit in­vestors, but are averse to this be­cause of the risk of mis­sion drift. We un­der­stand that they ap­plied to Wel­l­come and did not re­ceive fund­ing, with no feed­back pro­vided.

Over­all, we con­sider this a risky, but po­ten­tially high-im­pact piece of work, with po­ten­tially ap­pli­ca­bil­ity not just to the COVID-19 out­break to many other con­texts, and op­er­at­ing in a con­text which is rel­a­tively ne­glected.

Ap­pendix: Devel­op­ment Me­dia In­ter­na­tional (DMI)

DMI is a GiveWell stand­out char­ity, and has been in­de­pen­dently as­sessed by Im­pact Mat­ters, and re­ceived a pos­i­tive re­view from them.

We un­der­stand from our call with DMI on Wed 18th March 2020 that:

  • There are some sub-Sa­haran Afri­can coun­tries which have asked DMI to sup­port them in their pub­lic health mes­sag­ing, some of whom would not have that ca­pac­ity with­out DMI’s support

  • There is a short­age of fund­ing for pub­lic health mes­sag­ing work in sub-Sa­haran Africa

  • DMI has a fund­ing gap roughly of around a 6-figure sum in or­der to be able to perform this work

As­sum­ing this is all cor­rect, this would con­sti­tute a rare ex­am­ple of work which is:

  1. Rel­a­tively well-evidenced

  2. Neglected

  3. Directly rele­vant to COVID-19

In terms of the ev­i­dence, we have the fol­low­ing:

  • We have re­viewed the ex­ist­ing ev­i­dence of DMI’s work based on a ran­domised con­trol trial (RCT) performed in Burk­ina Faso which en­couraged ra­dio listen­ers to go to a health clinic if they were ill.

  • The RCT was in­suffi­ciently pow­ered to find an im­prove­ment in all-cause mor­tal­ity; this is be­cause of an im­prove­ment in baseline mor­tal­ity.

  • How­ever it did con­vinc­ingly find an im­prove­ment in vis­its to health centres

  • DMI has also shared with us a high-level sum­mary of the re­sults of past work re­lat­ing speci­fi­cally to hand­wash­ing (see chart be­low)

  • We have not at this stage been able to fully un­der­stand these re­sults; in par­tic­u­lar we do not un­der­stand the na­ture of the con­trol group, the pre­cise na­ture of the in­ter­ven­tion, or whether any work was done to as­sess the ul­ti­mate im­pact on di­ar­rhoea rates. We think that in or­der for this find­ing to be in­val­i­dated, it would be likely be be­cause un­der­ly­ing baseline hand­wash­ing rates were im­prov­ing in­de­pen­dently of DMI’s work; we know of no rea­son why this should hap­pen.

  • More gen­er­ally, we be­lieve that there is of­ten good rea­son to be cau­tious about as­sum­ing that a char­ity’s work is achiev­ing re­sults. How­ever in the case of DMI, we think it rea­son­able for our de­fault or “prior” as­sump­tion to be in­formed by the fact that ad­ver­tis­ing (prob­a­bly) works, and there­fore we should be more op­ti­mistic that DMI’s in­ter­ven­tion prob­a­bly works too.


We asked DMI whether they knew of ex­ist­ing mod­els to link a cer­tain in­crease in hand­wash­ing (say) with an im­pact on r or the re­pro­duc­tion num­ber. They did not know of any such mod­els, and we are also un­aware of any. This makes it harder for us to build a cost-effec­tive­ness model. At time of writ­ing we do not have a cost-effec­tive­ness model of DMI’s COVID-19-re­lated work.

A coun­ter­ar­gu­ment against work in the de­vel­op­ment world is around whether flat­ten­ing the curve is re­al­is­tic in that con­text. Flat­ten­ing the curve refers to con­trol­ling the spread of the pan­demic so that it be­comes man­age­able for the health sys­tem.

There have been ques­tion marks over whether it’s re­al­is­tic to con­trol the spread of the pan­demic enough to stay within the ca­pac­ity of the health sys­tems in the de­vel­oped world. It seems even more likely there­fore that the health sys­tems in the de­vel­op­ing world will be over­whelmed, and there­fore that any effort em­ployed here will have no effect.

This con­cern seems all the more rea­son­able given that DMI’s past record on hand­wash­ing sug­gests that they can bring about an im­prove­ment, but it’s un­clear whether they can bring about a dras­tic im­prove­ment (note that the con­text of a wide­spread, high-pro­file pan­demic may in­val­i­date our at­tempts to ex­trap­o­late from their past ex­pe­rience).

In dis­cus­sion with DMI it be­came clear that the benefits go be­yond lives saved from hand­wash­ing or (if ap­pli­ca­ble) so­cial iso­la­tion. It in­cludes, for ex­am­ple, tack­ling the risks of mis­in­for­ma­tion. For ex­am­ple in an en­vi­ron­ment with pre-ex­ist­ing eth­nic ten­sions, ru­mours may spread sug­gest­ing that the virus was brought about by (e.g.) an­other racial group, which may ex­ac­er­bate con­flict risk, at a time when the world can least cope with large re­fugee camps or strains on health sys­tems.

Our con­fi­dence in our ten­ta­tive recom­men­da­tion of DMI would be higher if we had the data and ca­pac­ity to put to­gether a model which re­flects these fac­tors.

Ap­pendix: Biose­cu­rity ini­ti­a­tive at Cen­tre for In­ter­na­tional Se­cu­rity and Cooperation

The Biose­cu­rity Ini­ti­a­tive at the Cen­ter for In­ter­na­tional Se­cu­rity and Co­op­er­a­tion (CISAC) is a re­search cen­tre at Stan­ford that car­ries out policy re­search and in­dus­try out­reach to re­duce the risk of nat­u­ral and en­g­ineered pandemics

Our recom­men­da­tion of the Biose­cu­rity Ini­ti­a­tive of CISAC is based largely on the recom­men­da­tion by Founders Pledge, set out in their re­port on ex­is­ten­tial risk. Their recom­men­da­tion re­flects the fact that their in-house biose­cu­rity ex­perts David Rel­man and Me­gan Palmer are both thought lead­ers in their field.

Our cur­rently cur­sory re­view of their work sug­gests that their work is not very fo­cused on COVID-19.

Ap­pendix: Fu­ture of Hu­man­ity In­sti­tute (FHI)

FHI is a mul­ti­dis­ci­plinary re­search in­sti­tute at Oxford Univer­sity study­ing big pic­ture ques­tions for hu­man civ­i­liza­tion.

FHI has been funded in the past by the Open Philan­thropy Pro­ject. We are fa­mil­iar with work performed by a num­ber of the thought lead­ers at FHI. We un­der­stand that FHI’s work mostly has lit­tle di­rect rele­vance to COVID-19, but some of it is rele­vant to pan­demics.

Prior to the COVID-19 out­break we re­viewed FHI and de­pri­ori­tised it be­cause it was our im­pres­sion that fund­ing was not a par­tic­u­lar con­straint for FHI. There­fore, based on our cur­rent im­pres­sions, we are keep­ing FHI in the “not fund­ing-con­strained” cat­e­gory. Other­wise, it would likely fit well in the “co-benefits” cat­e­gory, be­cause it does plenty of other work rele­vant to pan­demics, bio risk, and fund­ing di­rected to FHI may well funge with/​en­able other high-im­pact work.

Ap­pendix: CSER (Cen­tre for Study of Ex­is­ten­tial Risk)

The Cen­tre for the Study of Ex­is­ten­tial Risk (CSER) is a re­search cen­tre at the Univer­sity of Cam­bridge, in­tended to study pos­si­ble ex­tinc­tion-level threats

Our opinion that CSER is a high-im­pact dona­tion tar­get is sup­ported by our view that (a) tack­ling ex­is­ten­tial risk a high-im­pact cause area (b) hav­ing this re­source available in Cam­bridge is a solidly valuable thing. We un­der­stand that CSER’s work mostly has lit­tle di­rect rele­vance to COVID-19, but some of it is rele­vant to pan­demics and that they are look­ing to ex­pand this el­e­ment of their team.

We be­lieve that this may be a suit­able choice for fun­ders in­spired to sup­port pan­demics as a re­sult of the coro­n­avirus out­break.

Ap­pendix: Work on Wet markets

Tack­ling wet mar­kets has a clear link with COVID-19, since this is thought to be where the COVID-19 epi­demic came from. Fur­ther­more, mul­ti­ple epi­demics (avian flu and SARS as well as COVID-19) are thought to have origi­nated in wet mar­kets, be­cause keep­ing live an­i­mals there in­creases the risk of zoonotic trans­mis­sion. Tack­ling this is­sue has a num­ber of co-benefits, no­tably for an­i­mals, who, we un­der­stand, suffer con­sid­er­ably as a re­sult of these mar­kets. Put­ting this area of work into a cat­e­gory which we are pos­i­tive may sug­gest that we are pos­i­tive about fund­ing this area of work. How­ever fur­ther work would be needed be­fore we can con­clude this.

A rough model for the im­pact of work on wet mar­kets con­sid­ers:

  1. The amount cur­rently spent by gov­ern­ments and NGOs on this currently

  2. The scale of the im­pact of wet mar­kets in terms of epi­demic risk

  3. The scale of the im­pact of wet mar­kets in terms of an­i­mal welfare

  4. We would need a ful­ler re­view to know how much is cur­rently spent on this topic, how­ever roughly ex­trap­o­lat­ing from what we know of the land­scape, our rough guess is that the amount spent globally is fairly low, maybe c.$10mil­lion (rough or­der of mag­ni­tude)

  5. We know of three epi­demics thought to origi­nate from wet mar­kets in the last 20 years. A more care­ful model would re­quire more time than we have for this quick re­view, so let’s as­sume that the epi­demic risk is cap­tured by as­sum­ing that a pan­demic of the scale of COVID-19 will oc­cur once ev­ery hun­dred years as a re­sult of wet mar­kets. The eco­nomic im­pact of the coro­n­avirus pan­demic is thought to be at least $1 trillion (source: UN). This sug­gests that it would be worth­while to in­vest $10 billion of effort (i.e. c1000x more than the amount cur­rently ex­pended) to move from the his­toric level of risk to zero risk. (Note that, as set out be­low, we are prob­a­bly no longer at the his­toric level of risk any­way)

  6. To give some idea of the scale of suffer­ing, we un­der­stand that there are 20,000 wildlife farms in China and also that there were 20,000 wet mar­kets that China closed in Fe­bru­ary 2020. It is un­clear how many an­i­mals would be on a farm or would come through a wet mar­ket per year, how­ever if we as­sume that a mod­er­ately high es­ti­mate is 10,000 an­i­mals per farm, then this is c 200 mil­lion an­i­mals per farm. This is rel­a­tively small com­pared to the scale of in­dus­trial fac­tory farm­ing (e.g. c 60 billion fac­tory farmed chick­ens). How­ever, as­sum­ing that non-hu­man an­i­mals are given suffi­ciently high moral weight, this scale is still suffi­cient to add a ma­te­rial amount of ex­tra benefit to the work on wet mar­kets.

While the above ini­tial as­sess­ment sounds promis­ing, we would need to fur­ther con­sider the fol­low­ing factors

  1. whether most of the work that needs to be done has already been done, given that China has re­cently banned wet mar­kets and 80% of peo­ple in Beijing are op­posed to wildlife meat markets

  2. it is re­ported that wet mar­kets still per­sist, but to what ex­tent is this the case?

  3. might con­trol mea­sures may drive wet mar­kets un­der­ground in a way that might not be net beneficial

  4. might the ban spawn the cre­ation of new wet mar­kets in nearby countries

  5. is this area of work likely to be crowded as a re­sult of COVID-19

  6. a full model of the im­pact of work on wet mar­kets should con­sider what im­pact (if any) there will be on Chi­nese peo­ple’s nu­tri­tion as a re­sult of the ban, as well as the an­i­mal welfare im­pact and the risk of epi­demics.

We know of sev­eral or­gani­sa­tions which we un­der­stand to be work­ing on this:

  • Wildlife Con­ser­va­tion Society

  • Wildaid

  • WWF

  • EIA

  • Traffic

  • Flora and Fauna International

Ap­pendix: Long list of COVID-19 re­lated dona­tion opportunities

A list of options

  • Fu­ture of Hu­man­ity In­sti­tute (FHI)

    • a mul­ti­dis­ci­plinary re­search in­sti­tute at Oxford Univer­sity study­ing big pic­ture ques­tions, in­cludes a sec­tion on bio risks, in­clud­ing pandemics

  • Cen­tre for the Study of Ex­is­ten­tial Risk (CSER)

    • a re­search cen­tre at the Univer­sity of Cam­bridge, in­tended to study pos­si­ble ex­tinc­tion-level threats, in­clud­ing pandemics

  • Johns Hop­kins Cen­tre for Health Security

    • a think tank at Johns Hop­kins Univer­sity which car­ries out re­search on biose­cu­rity and main­tains a COVID-19 re­source centre

  • Biose­cu­rity ini­ti­a­tive at Cen­tre for In­ter­na­tional Se­cu­rity and Cooperation

    • a re­search cen­tre at Stan­ford that car­ries out policy re­search and in­dus­try out­reach; has gen­eral rele­vance to pandemics

  • Tack­ling wet mar­kets (mul­ti­ple orgs)

    • Wildlife Con­ser­va­tion Society

    • Wildaid

    • WWF

    • EIA

    • Traffic

    • Flora and Fauna International

  • WHO soli­dar­ity re­sponse fund: https://​​www.who.int/​​emer­gen­cies/​​dis­eases/​​novel-coro­n­avirus-2019/​​donate

    • Send es­sen­tial sup­plies such as per­sonal pro­tec­tive equip­ment to frontline health workers

    • En­able all coun­tries to track and de­tect the dis­ease by boost­ing lab­o­ra­tory ca­pac­ity through train­ing and equip­ment.

    • En­sure health work­ers and com­mu­ni­ties ev­ery­where have ac­cess to the lat­est sci­ence-based in­for­ma­tion to pro­tect them­selves, pre­vent in­fec­tion and care for those in need.

    • Ac­cel­er­ate efforts to fast-track the dis­cov­ery and de­vel­op­ment of life­sav­ing vac­cines, di­ag­nos­tics and treatments

  • Direct re­lief: https://​​www.di­rec­tre­lief.org/​​emer­gency/​​coro­n­avirus-out­break/​​

    • provide per­sonal pro­tec­tive equip­ment and es­sen­tial med­i­cal items to health work­ers re­spond­ing to coro­n­avirus (COVID-19)

  • Unicef: https://​​www.unicef.org.uk/​​donate/​​coro­n­avirus/​​

    • per­sonal pro­tec­tion for health work­ers; hy­giene ma­te­ri­als for thou­sands of schools, pub­lic health fa­cil­ities and other pub­lic places; and sup­port with health ed­u­ca­tion messaging

  • Global Giv­ing fund: https://​​www.global­giv­ing.org/​​pro­jects/​​coro­n­avirus-re­lief-fund/​​#menu

    • We found no in­for­ma­tion what­so­ever about how the money will be used

  • Cana­dian Red Cross: https://​​donate.red­cross.ca/​​page/​​55441/​​donate/​​1?lo­cale=en-CA

    • When we launch an ap­peal, dona­tions made in re­sponse to it are used to help in­di­vi­d­u­als and com­mu­ni­ties iden­ti­fied in the ap­peal re­cover from emer­gen­cies and disasters

  • CDC foun­da­tion: https://​​www.cd­cfoun­da­tion.org/​​

    • This was listed as a coro­n­avirus-rele­vant dona­tion op­por­tu­nity, how­ever we have not found any coro­n­avirus-rele­vant info on their webpage

  • South­ern Smoke Foun­da­tion https://​​south­ernsmoke.org/​​sup­port/​​

    • Help sup­port those in the food and drink in­dus­try in crisis

  • Ital­ian ICU gofundme cam­paign https://​​www.gofundme.com/​​f/​​coro­n­avirus-ter­apia-intensiva

    • Fund­ing ven­tila­tors and mon­i­tors etc for Ospedale San Raf­faele di Milano

  • First book: https://​​first­book.org/​​

    • First book aims to de­liver 6 mil­lion books to kids in light of school closures

  • Devel­op­ment Me­dia International

    • Pro­ject in Mozam­bique and Malawi (and pos­si­bly other coun­tries) to pro­mote hand­wash­ing and other rele­vant behaviours

  • Univursa health

    • They are look­ing to build a di­ag­nos­tic sup­port tool to help coun­tries to rapidly iden­tify pos­si­ble out­breaks of COVID-19 within their bor­ders, where those coun­tries have limited ca­pac­ity to do lab­o­ra­tory-based test­ing. There is a par­tic­u­lar fo­cus on sub-Sa­haran Africa.