I meant that, based on Sanjay’s info, that only 1 out of 800 people are a match for anyone, and they will spend around 1 hour to fill in forms, do the swap (which in itself creates costs etc.), this should be included in a full costing analysis of the cost per death averted figure and make it no longer competitive with other priority interventions.
Here’s another more recent related study that similarly suggests that bone marrow transplantation is quite expensive to treat even for advanced economy standards:
I’m not sure these costs (the $170,000+) should be included, or if they are, they may need to be weighed down significantly. We’re not looking at the cost-effectiveness of implementing such a program; we’re looking at the cost-effectiveness of an extra person signing up and possibly an extra donation (with small probability). If you want to include the $170,000+, you should ask what else would have been done with that money, because it’s not ours to use.
As I said that in my previous reply, I do agree with you that sometimes it can be very effective to improve a generally not very cost-effective intervention such as this. I also agree that one’s marginal time might not be very valuable (yet not zero). So I do agree that in theory this could be a cost-effective intervention.
However, I think the marginal cost per death averted figure above of spreading this meme were overstated for two reasons:
1. Crucially, on the marginal costs:
Assume that including cognitive overhead that it takes roughly an hour for an EA to do the swap (e.g. using the swap, mailing it etc.) and the marginal value of my time on Sunday afternoon when I would otherwise just do something pointless is $10 (this is conservative see http://globalprioritiesproject.org/wp-content/uploads/2015/03/NeutralHours.pdf ).
This buys us a 1 in 800, or 0.125% chance of being a match.
This means we need to convince 800 EAs to spend $10 to find a match. The cost is $8000. In other words, your own probability of saving a life is 1 in 800 but you have to pay $10 for it.
Technically, but less crucially, we also need to add that one EA will actually have to go through the donation process for this to have an effect. 20 hours * $10 = $200. If you’re unwilling to accept a 1 in 800 probability of going through the process, there’ll be no benefit. For one person that means that we need to add 1 in 800 probability of paying $200 (or 25 cents). So the total for one person comes to $10.25.
2. Also less crucially on the benefits side: As shown in the papers above the QALYs gained from might not add up to a “whole life saved”, which can be seen by their measure “per additional treatment success’ (“For patients at standard risk for disease, the treatment success rate was 80.3% for BMT recipients” [...] For patients with high-risk disease, the treatment success rate was 23.5% for BMT recipients”). Then, even after successful treatment, cancers often come back, which reduces the cost-effectiveness (“only 62% of patients survived the first year post-BMT, 98.5% of patients alive after 6 years survived at least another year. Almost 1⁄3 (31%) of the deaths in long-term survivors resulted from causes unrelated to transplantation or relapse. ” see https://www.ncbi.nlm.nih.gov/pubmed/16545726 ).
Worth noting (some of these are also mentioned in my other comment, with sources there):
1. If you’re male and 16-30 (a significant proportion of EAs), the chance of being asked to donate is more like 1 in 200.
2. I’d say the time cost of registering is closer to 15 minutes than an hour. Personally I filled in the form at work during my lunch break, and spent a few minutes doing the cheek swab at home.
3. It’s not clear that not doing a transplant would be cheaper for healthcare systems than doing the transplant. In general, complications for conditions which would require a transplant can be expensive to treat, and healthcare systems like the NHS would treat them. See the linked Guardian article in my other comment which suggests that transplants could actually be saving the NHS money vs treating the symptoms. Also see the other link that suggests UK transplants from UK donors are much cheaper than UK transplants from foreign donors (assuming these transplants would be happening anyway).
4. I do think health concerns to donors need to be taken into consideration. Most people (90%) can donate in a way similar to blood donation, but 10% undergo general anesthesia. This isn’t risk-free, though generally the criteria which filter for suitable donors upon registering exclude people who would be at high risk when undergoing general anesthesia.
If you’re chosen to donate, someone else is not, so they save time, although overall, more time is wasted if you don’t sign up, because the organization will have to deal with an extra potential donor. Roughly one more person will be signed up in expectation if you sign up, unless you think you have an important effect on others. You’re not wasting the time of anyone else who signed up if you sign up, since they were already going to pay that cost.
If you’re not chosen to donate, you’ll waste your own time, and the time of the organization.
Overall, though, if you’re EA-minded, your time is probably worth much more than the average person who signed up, and I think this should dominate time considerations.
I meant that, based on Sanjay’s info, that only 1 out of 800 people are a match for anyone, and they will spend around 1 hour to fill in forms, do the swap (which in itself creates costs etc.), this should be included in a full costing analysis of the cost per death averted figure and make it no longer competitive with other priority interventions.
Cost-effectiveness analysis looking at bone marrow transplantation often include all these costs (e.g. “The median cost of the first 6 months of care including donor identification, marrow collection, patient hospitalization for transplantation and all outpatient medications and readmissions through 6 months postmarrow infusion was $178,500” https://ashpublications.org/blood/article/92/11/4047/133908/The-Costs-and-Cost-Effectiveness-of-Unrelated ).
Here’s another more recent related study that similarly suggests that bone marrow transplantation is quite expensive to treat even for advanced economy standards:
https://www.ncbi.nlm.nih.gov/pubmed/20348004
Note that I haven’t looked into this very deeply, and this this just a hunch.
Also, again, I agree that sometimes, it can be effective to improve a generally not very effective intervention.
I’m not sure these costs (the $170,000+) should be included, or if they are, they may need to be weighed down significantly. We’re not looking at the cost-effectiveness of implementing such a program; we’re looking at the cost-effectiveness of an extra person signing up and possibly an extra donation (with small probability). If you want to include the $170,000+, you should ask what else would have been done with that money, because it’s not ours to use.
The organization gets it funding from the government and donations.
As I said that in my previous reply, I do agree with you that sometimes it can be very effective to improve a generally not very cost-effective intervention such as this. I also agree that one’s marginal time might not be very valuable (yet not zero). So I do agree that in theory this could be a cost-effective intervention.
However, I think the marginal cost per death averted figure above of spreading this meme were overstated for two reasons:
1. Crucially, on the marginal costs:
Assume that including cognitive overhead that it takes roughly an hour for an EA to do the swap (e.g. using the swap, mailing it etc.) and the marginal value of my time on Sunday afternoon when I would otherwise just do something pointless is $10 (this is conservative see http://globalprioritiesproject.org/wp-content/uploads/2015/03/NeutralHours.pdf ).
This buys us a 1 in 800, or 0.125% chance of being a match.
This means we need to convince 800 EAs to spend $10 to find a match. The cost is $8000. In other words, your own probability of saving a life is 1 in 800 but you have to pay $10 for it.
Technically, but less crucially, we also need to add that one EA will actually have to go through the donation process for this to have an effect. 20 hours * $10 = $200. If you’re unwilling to accept a 1 in 800 probability of going through the process, there’ll be no benefit. For one person that means that we need to add 1 in 800 probability of paying $200 (or 25 cents). So the total for one person comes to $10.25.
We’re also subsidizing expensive health care, spreading the meme that this is good, and perhaps in the process displacing more effective treatment, which can do net harm (see https://www.nice.org.uk/news/blog/carrying-nice-over-the-threshold ).
2. Also less crucially on the benefits side: As shown in the papers above the QALYs gained from might not add up to a “whole life saved”, which can be seen by their measure “per additional treatment success’ (“For patients at standard risk for disease, the treatment success rate was 80.3% for BMT recipients” [...] For patients with high-risk disease, the treatment success rate was 23.5% for BMT recipients”). Then, even after successful treatment, cancers often come back, which reduces the cost-effectiveness (“only 62% of patients survived the first year post-BMT, 98.5% of patients alive after 6 years survived at least another year. Almost 1⁄3 (31%) of the deaths in long-term survivors resulted from causes unrelated to transplantation or relapse. ” see https://www.ncbi.nlm.nih.gov/pubmed/16545726 ).
Worth noting (some of these are also mentioned in my other comment, with sources there):
1. If you’re male and 16-30 (a significant proportion of EAs), the chance of being asked to donate is more like 1 in 200.
2. I’d say the time cost of registering is closer to 15 minutes than an hour. Personally I filled in the form at work during my lunch break, and spent a few minutes doing the cheek swab at home.
3. It’s not clear that not doing a transplant would be cheaper for healthcare systems than doing the transplant. In general, complications for conditions which would require a transplant can be expensive to treat, and healthcare systems like the NHS would treat them. See the linked Guardian article in my other comment which suggests that transplants could actually be saving the NHS money vs treating the symptoms. Also see the other link that suggests UK transplants from UK donors are much cheaper than UK transplants from foreign donors (assuming these transplants would be happening anyway).
4. I do think health concerns to donors need to be taken into consideration. Most people (90%) can donate in a way similar to blood donation, but 10% undergo general anesthesia. This isn’t risk-free, though generally the criteria which filter for suitable donors upon registering exclude people who would be at high risk when undergoing general anesthesia.
If you’re chosen to donate, someone else is not, so they save time, although overall, more time is wasted if you don’t sign up, because the organization will have to deal with an extra potential donor. Roughly one more person will be signed up in expectation if you sign up, unless you think you have an important effect on others. You’re not wasting the time of anyone else who signed up if you sign up, since they were already going to pay that cost.
If you’re not chosen to donate, you’ll waste your own time, and the time of the organization.
Overall, though, if you’re EA-minded, your time is probably worth much more than the average person who signed up, and I think this should dominate time considerations.