I think It wouldn't cost much at all to make forward a pretty robust cost-effectiveness model for a CHW which rolls out a wide range of interventions. (I think Living Goods +- others might well have decent models already here?). I think you could even build this yourself? Some of the package would be easy to do because data is there (malaria, diarrhoea, pneumonia treatment, family planning antenatal care), while screenings and referrals are much harder to quantify and might have to be left out of the analysis pending better data.

I agree the bet argument is pretty good if the goal is government adoption. Regardless of the nuance of cost-effectiveness, CHWs will always be more cost-effective than most health things govt. could do and it would likely displace less cost-effective things. Unfortunately I don't think EA funders have seriously considered scale through govt. something worth chasing as a bet, but I really like the idea.

Thanks @Madeleine Ballard. As you know I'm a big fan of the Pro-CHW movement and all the work you do. I agree facility care can't reach everyone. I agree we need pro CHWs across Africa and like you say CHWs are often the main people doing the work facilitating EA funded vertical programs like mosquito nets and deworming. Sometimes this platform isn't recognised, and the cost borne by governments and donors who support these CHWs isn't appreciated properly in calculations - which I think I've pointed out to GiveWell before.

I also agree that EA doesn't pay nearly enough attention to gradual government adoption of the most cost-effective interventions, and I hope we're going to see more discussion about this in the near future.

When looking at RP's report, this was what I said in context.

"I agree that CHWs are an essential, scalable part of many LMIC health systems. CHWs however are touted to be very cost-effective when they often aren’t. This statement made my eyes pop a little. “Many studies have found CHW programs to be cost-effective by various metrics. In Mozambique, the annual cost per beneficiary was just $47.12. 

That number $47.12 I quoted from the RP report is quite different from $10 you've quoted as the median cost per capita here and doesn't scream cost-effectiveness. The report which  your $10 number came from wasn't out at the time I wrote that response. $10 per beneficiary still seems high to me but it's hard to compare to other interventions (see below)

One thing I'm confused about, is that I don't think a cost as low as $0.59 per capita is possible for a salaried CHW. How did the researchers get to this number? Trying to be as conservative as possible - here's my sense check. The ceiling for the no. of people a CHW can cover on a door-to-door model is perhaps 800. A lower end CHW salary might be $40 a month. Then we add commodity and supervision costs (excluding digital/connectivity costs). I've estimated this at around 20 people tested for malaria, 10 given treatment for malaria and 5 each treated for pneumonia and diarrhoea. Often CHWs would treat more people than this, which would mean higher costs.

- Salary (12x40) = $480 
- Malaria tests 20 a month (20 x $0.5 x 12 months ) = $120
- Malaria treatment 10 a month (10 x $0.3 x 12) = $36
- Pneumonia treatment 5 a month (5 x $0.5 x 12) = $30
- Diarrhoea ORS + zinc 5 a month (5 x $0.5 x 12) = $30
- Supervision ? $60 a year

= $756 / 800 people = $0.95 per capita per year under pretty conservative assumptions. I might be missing something here though!

Also "Per capita cost" and "Cost per beneficiary" are useful metrics for governments to decide if CHWs are affordable/cheap enough, but not useful metrics for cost-effectiveness comparisons. What can these metrics be compared to? We can't compare this metric to any other intervention. I've considered we could use a metric such as "cost of universal primary care per-capita" which would combine the cost of CHWs + other services, which could then be compared to facility only based models + extra transport costs. This would be tricky though... 

For CHWs a more straightforward "Cost per DALYs averted" would be more useful to compare with other EA funded interventions, or a more thorough GiveWell style CEA. 

Also on this point "Any cost-effectiveness analysis that attributes the full fixed cost of that worker to one of those service lines is going to produce a misleading number." I don't think Givewell attributes the full fixed cost of CHWs in CEAs. They might apportion some of the cost, or perhaps none. I often have the opposite criticism that they don't always appreciate the underlying cost of government health workers used in mass-distribution campaigns. I could be wrong here though I haven't looked closely recently.

Again I'm a huge fan of CHWs and your work and I believe in your mission. I'm just challenging whether your numbers make sense, and I think we need more than per-capita/per-beneficiary numbers to compare CHWs to other cost-effective interventions under EA frameworks.

Hey @Albert Oyawa thanks for the contribution - on the forum here. These are decent points but this is clearly written by AI, and any AI use needs to be stated up-front here.

100% agree this would be the best solution. Unfortunately in almost all African Countries, perceived sovereignty and not "bowing the knee" to the West is put at far higher premium than things like drug approvals. This would be scoffed at accross the continent for this reason.

To be fair it's not like high income country's are doing great at getting their approvals sorted, partly for similar pride reasons.

@Vasco Grilo🔸 have a look at my latest reply to @Ben_West🔸 below. I think there is a worldview where it's important and "good" to know who wrote the words, and who we are interacting with. I think we might even start to see legislation and guidelines which demand disclosure of who wrote what. Before AI, it was just assumed that all our words were our own. There are exceptions in human norms to this like having a "ghost writer" but I think that's ethically wrong too.

Putting aside whether it's "good" or "bad" for something to be written by an AI, and putting aside the question of quality, at the very least given it's hard to detect if a human writes it or not, I think as a human readers should have the right to know who they are interacting with. Is it a human? Is it an AI? Is it a mix of both and how did they mix?

I think your perspective is reasonable here, it's just not what's important to me. Genuine unfiltered human interaction is important to me. Knowing that I'm talking with someone without an AI in between is important to me. If that's not important to you that's fine. This is important to me not only because I value true direct human interaction, but also (as a secondary problem) because I think AI writing is samey and boring. Maintaining a public writing space with true diversity, quirkiness and strong voices is part of what drives engagement and excitement.

When I see your name on something, I want it to be 100% your voice and your words like we are talking in a public space. Or at the very least I want you to tell me if it's not. If you're not concerned with that, then we have a fundamental almost axiomatic difference about what matters in a forum like this. I think that's part of the reason why there's a bit of a chasm between our views, and those who are happy with AI writing things. The quality of ideas and reasoning is only half of what matters for me. The other half is the discussion and interaction between us - the mingling of our minds. I'm not sure we can resolve this difference. If you genuinely don't mind who's "brain" words came from, and think that other's don't have the right to know that as well, that's reasonable but we may have fundamentally different beliefs.

A human or an AI could do good or bad research, I'm less concerned with that. Karma will sort that out. Karma can't answer the human interaction question above. We can discern from outside whether an argument is good or not. We can't discern from outside whose words they are - that's why we need the start-of-post disclosure at the very least (I would go further). An analogy might be if someone did a bunch of research for you and sent it to you, and then you used half of their words in your post. Ignoring the plagiarism element, that wouldn't be you talking with me it would be someone else which would be dishonest - unless you said "hey this article is half my research assistant's words and half mine).

 I think as a human I have the right to know who I'm interacting with.

Yep that word "moral" was the only dubiously EA coded looking one in your prompts to me. But like you say results seem to hold which is kind of wild...

I don't love the tone of this piece, but I heard him once recently on the Ezra Klein show and he did pretty poorly. He clearly hadn't prepared and acted defensive and incredulous at times. He does really need to work on his public appearances...