The Red Cross claims that 1 pint saves “up to 3 lives”. I’m not sure what to make of that, given that it’s an upper bound and presented by a non-partial source.
If anyone can do better, I would be very interested in knowing the answer. I always try to give blood as often as possible under the assumption that I save at least one life each time, but a more robust figure would be nice.
Given you and wedrifid’s responses, I am now updating my estimate of number of lives saved significantly downwards. However, I am curious as to why it’s obvious to you that 3 lives is too high of a number on the margins.
Maybe I’m just being naive here, but in a case that straightforward and that possible for the average person to understand, where there’s nothing odd or unprestigious about the action and lots of people are doing it already, where, on the margins, an additional American life is saved each time another person donates blood, I have trouble believing that even a world this insane wouldn’t push blood donations a little harder.
I take it that you’re suggesting marginal analysis based on the standard correct classical causal decision theory (in which no one is responsible for saving a life by donating blood unless someone would have actually died had that donation not been made) out of either belated humility about the probability of an SIAI-originating decision theory being correct, or because you’re planning to actually convince someone and you don’t want to invoke Hofstadterian superrationality in place of the standard correct decision theory?
:)
My guess would be that at the margin, a blood donation saves less than 0.00001 lives. (Otherwise, compensation would be increased for the paid donors). But, if you want to use a TDT/UDT style analysis, here are some relevant statistics from the American Red Cross:
The number of blood donations collected in the U.S. in a year: 16 million (2006).
The number of patients who receive blood in the U.S. in a year: 5 million (2006).
Given these numbers, I would estimate that roughly 0.5 million (US) lives are saved (more accurately, extended) by blood products annually. If you adopt the assumption that all blood comes from voluntary, uncompensated donations, and divide those 0.5 million lives among the 16 million annual donations, you get one life saved for every 32 pints donated—not as much as jsteinhardt hoped, but still significant enough to earn a major warm-and-fuzzy.
I happen to administer a lot of blood to my patients, so let me answer some of the factual questions.
The way they calculate “up to 3 lives” is in the most trivial way: blood you donate is fractionated into red cells, plasma, and platelets. Each of those may go to a different recipient.
All blood administered to patients comes from voluntary, uncompensated donations. Plasma used in research studies may be compensated, but may not be transfused. This is the most important factor keeping our blood supply safe, and is far more effective than laboratory testing alone.
Given that blood banks need to keep a sufficient store of blood available of each type, rarer blood types are generally in greater need than, say, A After all, a larger proportion of blood of those types must be discarded. O blood is obviously highly useful in trauma situations, and is therefore in high demand as well.
The distribution of donors’ and recipients’ blood types should not be assumed to be equal: people with blood type A are significantly more likely to donate than people with blood type B. This exacerbates the discrepancies due to point 3.
The number of lives saved can be calculated in two ways:
a. the feel-good way. Every time a physician gives a unit of blood to a patient e does so believing it is a life-saving procedure. So if 3 units are given the patient’s life was saved 3 times in rapid succession. (You have to be willing to save a life multiple times, because that’s the analysis we’re using for the rest of this discussion: multiple mosquito nets saved the same kid’s life multiple times over his lifetime; that same kid was then saved by anti-diarrheal treatments; etc. The same analysis belongs here). Now, we subtract the number of patients who die, but that’s a small number.
So 26 million transfusions/16 million donations = 1.6 lives saved per donation.
b. the marginal way. Donations are currently sufficient for usage; we benefit in three ways from more donations. First, we can be slightly more profligate with trauma patients who have a low survival chance; this saves a minimal number of lives. Second, fresher blood is associated with better outcomes than older blood; the extent of this effect is unknown but is an area of current research interest. The calculation would have to look at the likelihood that your donation reduced the average shelf age of the blood being administered times the survival improvement from the fresher blood. Third, blood from multiparous women is associated with ARDS; an increase in donation would allow us to stop using it.
people with blood type A are significantly more likely to donate than people with blood type B
I’ve donated blood a few times and I’m type A+. Why is it that B’s are less likely to donate, or is that unknown? Are my donations likely to be marginally useless?
I have mostly donated blood in the past for signaling reasons, conversational high ground, and a vague desire to match the 15-gallon mark that my grandfather got his name in the paper for. There’s a plaque of the newspaper mention in my grandma’s house and I’ve been looking at it my whole life. Also I figure the Red Cross will let me know if I come down with one of the diseases they screen for, and it’s a free way to get my iron levels checked (attempting to donate blood was how I found out I was anemic in the first place). These reasons aren’t likely to evaporate if I find that I have been saving only tiny fractions of expected lives, but I would probably endure less inconvenience in order to donate for only these reasons as opposed to these reasons on top of lifesaving.
Your donations are not marginally useless! (unless you’ve been pregnant a couple times—in that case, consider stopping).
The reason for the discrepancies in donation rates between types A and B is both simple and complex: ethnicity. In the interests of safety (avoidance of Hepatitis C, HIV, etc) we’ve set up a system that subtly encourages certain types of donors and discourages others. The system is not racist per se, but it is most effective in obtaining donations from white, middle-aged, middle-class males.
Regarding signaling reasons: we are obviously very afraid of blood donated for signaling purposes. Accordingly, we do not allow people to donate to their relatives except under very unusual circumstances. Additionally, we give people an “out” by checking a box which tells the center to draw and discard their blood. That way people who fear they may be high-risk donors can get the social approval of donating without harming any patients.
Essentially this. The A/B/O blood groups represent the most relevant antigens in human blood. There are a host of others (Rh, Duffy, Kell, etc.) which typically create only minor problems in a transfusion and which can be ignored in an emergency. But a person who has been exposed to allogeneic blood via multiple transfusions or pregnancies becomes more likely to develop antibodies to some of these antigens. The donor’s antibodies or white cells can react to the person being transfused, causing lung damage.
There are a host of others (Rh, Duffy, Kell, etc.) which typically create only minor problems in a transfusion and which can be ignored in an emergency.
In the case of the Rhesus factor it should be noted that it is minor once and then only minor for males. Being thereafter unable to safely give birth to healthy Rh+ children is definitely not a minor consequence even if it is better than ‘probably going to die today’. (Unless, I suppose, you happen to some Rh+ antiserum lying around but no Rh- blood, which will usually avoid the future difficulties.)
I use “minor” differently than you do, to mean “unlikely to cause death”. Obviously cross-matched blood is always preferable for a variety of reasons (including possible infertility, in the case of young females).
I would avoid RhoGam in the case of a patient who needs a RBC transfusion, incidentally. It would be unlikely to be safe or effective.
I use “minor” differently than you do, to mean “unlikely to cause death”.
I would like to express that my approval of this phrase extends beyond the capacity of upvoting and into the capacity of a comment expressing approval.
we’ve set up a system that subtly encourages certain types of donors and discourages others
While we’ve got you here, can you explain why gay men cannot donate? This upsets a lot of gay people that I know.
I understand that it’s easier to catch STDs (not just HIV/AIDS) from a man than from a woman. But the current U.S. rule (A man cannot donate if he’s had sex with a man; a woman cannot donate if she’s had sex with a man who’s had sex with man.) is lopsided.
The even-handed rule that you cannot donate if you’ve had sex with a man would keep the supply safe without having to rely on people’s being able to trust their partners. But it would keep most women from donating, so maybe it’s not worth it. The even-handed rule that you cannot donate if you’ve had sex with man who’s had sex with a man would still keep out most gay men, but it would probably help to heal the rift.
The even-handed rule that you cannot donate if you’ve had sex with man who’s had sex with a man would still keep out most gay men
If a man is gay and sexually active, he’s almost certainly had sex with a man who’s had sex with a man, even if the men he has had sex with has only had sex with him. I don’t see how this phrasing of the rule would be an improvement.
My phrasing was unclear; make it “if you’ve had sex with man who’s previously had sex with a man (other than you)”. There wouldn’t be any point in forbidding me from donating (if I’m male) because the man that I’ve had sex with has had sex with me!
This change would include more people; it includes monogamous gay male couples who began their relationship as virgins (as well as some other people). Not many more, but it makes it clear that the blood collector is only willing to trust you and your partners, no further.
Frankly, the first even-handed rule (no sex with a man, period) makes more sense to me. Why should the blood collector trust that I know (if I’m a woman) whether all of the men that I’ve had sex with have had sex only with women? (No doubt many women are donating contrary to guidelines because they don’t know this about their partners.) But because this would cut the potential donor pool in half, the blood collector is basically forced to trust me about my partners too.
In fact, the blood collectors trust women to know the sexual history of their partners, but not men. They are not asking everybody the same questions.
Another possible solution, not even-handed, but more honest: Just don’t ask women anything about the subject.
The idea that a person can be trusted to know about their partners’ partners is preposterous; no other question (in the U.S.) asks the donor about other people’s behaviour, and for good reason. Instead of half-assedly trying to be even-handed about it, just admit what they’re doing: ruling out men who’ve had sex with men, because many of their partners will have had sex with other men, and so on back (in many cases) a long way; but accepting women who’ve had sex with men, because most of their partners won’t have had sex with men, stopping the transmission-from-men sequence.
I’m confident that they already accept blood from most women who’ve had sex with men who’ve had sex with men (because the women don’t know this about their partners), and they are surely aware of this (if I am correct) fact. So why are they asking questions of people who don’t actually know the answers?
Gay people will still be upset that they can’t donate, but I at least would be more willing to trust that the blood collectors are actually making an honest decision.
Mostly, my faith in the quality of the blood supply derives from what testing they’re doing to the blood, not from what unenforceable policies they’re suggesting to the donors.
I’d actually be surprised if the latter significantly affected the quality of the blood.
Mostly, I think the problem they are a solution for is maintaining public confidence in the blood supply. Which I acknowledge is an important problem. And it may well be that being perceived as excluding gay men and their partners is a better solution to that problem than anything else they might do; I don’t know.
That said, if I’m wrong and these policies really do solve a problem related to the blood supply, yet another possible solution is: don’t allow people who have had unprotected sex to donate.
Or, if that’s too big a chunk of your potential donor base, make it people who have had unprotected sex outside of a monogamous relationship.
The original rule bars ‘a man who has had sex with a man’ - X—and then any women who’ve had sex with X. It’s a logical phrasing but unfortunately X maps exactly onto “gay man”, so it feels like gay men are being specifically targeted. The rephrasing mollifies that sense of targeting without, as far as I can tell, changing the included or excluded people.
The original phrase is even-handed, however. If you overspecified an even-handed rule and said “1) You cannot donate if you’re a man who has had sex with a man who has had sex with a man, and 2) you cannot donate if you’re a woman who has had sex with a man who has had sex with a man”—ie, prevent “man who has had sex with a man” from coming into sexual contact with any donor—you could reduce 1) down to “man who has had sex with a man” (it logically implies three, four, and so on iterations). This, therefore, reduces down to the actual rule they have in place.
Every time a physician gives a unit of blood to a patient e does so believing it is a life-saving procedure. So if 3 units are given the patient’s life was saved 3 times in rapid succession. (You have to be willing to save a life multiple times, because that’s the analysis we’re using for the rest of this discussion: multiple mosquito nets saved the same kid’s life multiple times over his lifetime; that same kid was then saved by anti-diarrheal treatments; etc. The same analysis belongs here)
There are not many times I see a line of reasoning and have to reject it at every single step. Apart from being conceptually absurd the very thought is morally objectionable. It totally devalues the value of ‘saving a life’ to the point of utter meaningless. How could that ever make someone ‘feel-good’?
It totally devalues the value of ‘saving a life’ to the point of utter meaningless.
Which part? I thought that started silly (it’s explaining the logic behind a non-profit’s puffery, did you expect it to be rigorous?) but then got better. The idea of “saving a life” is pretty meaningless when you poke at it- it’s all just lifespan extension. And so the idea that each emergency treatment extends lifespans by the ‘natural span of a life’ is silly. If someone would die if they don’t receive a unit of blood at 50 separate occasions on their life, should each transfusion get the full moral weight of saving a life? If so, we just gave this person 50 lives. If not, then we need to abandon the language of “saving a life” and talk about “extending a lifespan” (because we can say those units of blood each added a year to the person’s life, for example).
I didn’t realize that donations were sufficient for usage. Is this barely maintained by calling people when blood supplies are low, or does blood regularly get thrown out, or is there some other reason that supplies closely match need?
A combination of the above. We have a core group of donors who can be called in emergency situations, we increase the intensity of blood drives when supplies are low, we reduce marginally-beneficial uses of blood when supplies become low, and we are better able to discard the oldest least-effective blood whenever supplies increase.
We are likely to face challenges in meeting future need. The cohort that most regularly donates blood is aging...
I’m assuming you are from the US, do you think the same is true for other countries? Also which demographic are you referring too?
I’m more interested in the cohort that isn’t aging. What is their secret? A new and improved Calorie Restricted diet? Perhaps that explains their inability to generate sufficient excess blood for donation.
Only vaguely relatedly: if you have pointers to (or are willing to synthesize) a reliable calculation of expected lives-saved/deaths-caused by maintaining or discarding the existing Red Cross policies about who is “allowed” to donate blood, especially the relatively controversial ban on male donors with homosexual acts in their sexual history, I would be interested.
Full disclosure: I do have a personal/emotional stake in this question, but I really really don’t want to set off a political/ethical conversation about it. I’m asking it here because, as with a lot of politically charged topics, the arguments I’ve found on both sides are mostly a case of framing the question so as to give the answer one wants to give, rather than so as to answer the question that was asked, and I’m looking for a more objective analysis.
Giving blood is important to me. It is so important that I have chosen not to pursue relationships with other men in order than I can continue to give blood without lying to do so. I expect that sooner or later, I will choose otherwise, and a sexual relationship will be important enough to me to sacrifice my ability to ever give blood again, and this distresses me.
I can accept that the risks of HIV may be high enough to make this a reasonable choice on the part of United Blood Services / Red Cross. However, I would like to be quite sure that this is the case, or to be told that my blood isn’t as important as I previously though it was. I was previously giving blood on the impression that each donation saves around a twentieth of a life; this thread doesn’t change that estimate enough for me to feel like I can stop donating in good conscience.
Giving blood is important to me. It is so important that I have chosen not to pursue relationships with other men in order than I can continue to give blood without lying to do so.
On the margins, I expect that each marginal pint of blood saves only a very small fraction of a life. As several readers pointed out, this doesn’t mean that we should ordinarily be calculating on the margins, since it’s not like you can use a pint of blood for something else instead; in terms of moral credit, you should think of yourself as part of a reference class of people who all choose to donate blood for around the same reasons, and who all get an equal share of the lives saved.
However, the Red Cross has already decided that they’re willing to X out the entire homosexual community, and I would expect the reference class of those who refrain from sexual activity in order to continue donating blood to be small, and I would guess that if this entire reference class refrained from donating blood, not a single additional life might be lost.
Modern-day hospitals are not, so far as I know, blood-limited. They need a routine flow of blood in order to routinely save lives. They do not need more blood to save more lives. That’s the impression I got, anyway; some quick Googling even said that they usually have enough blood to just use O-negative instead of matching types.
I hate to say this, but I think you’re making the wrong sacrifices here. I estimate a very high information value for further investigation on your part; I would expect it to show that you were safe to stop donating blood and resume sexual activity without costing anyone one-twentieth of a life. If you’re really feeling guilty or worried, resume sexual activity and send a donation to the Singularity Institute as a carbon offset. If you can speed up a positive Singularity by one minute that works out to around 100 lives, never mind increasing the probability.
I think I was accidentally misleading by failing to add that I’m bisexual. Not giving blood reduces my pool of potential romantic partners by roughly 10%, and doesn’t prevent me from having fulfilling relationships. I don’t think I would abstain from sex in order to give blood even if I knew I could save a life with each donation. Even if that’s an incredibly selfish decision, I’m just not that good a person.
Regardless, the support of everyone who replied is very much appreciated.
...technically, doesn’t speeding up a negative singularity also save lives—the lives of those who would otherwise have been born and then killed but were instead never born and therefore couldn’t be killed? In fact, I think speeding up a negative singularity actually “saves” more lives than speeding up a positive one using this calculation—a quick Google search indicates ~250 people are born every minute and ~100 people die every minute.
In a fairly meaningful sense, no life has ever been saved before. Nobody has actually been prevented from dying yet. A positive singularity could change that.
I believe you can make an easier calculation: change the denominator from lives to units of blood.
How much effort/money/social capital would it take you to convince one more person to donate one more unit? [ignore the cost to that person, as it’s likely zero or slightly beneficial]. Calculate the effort it therefore would take you to replace yourself as a donor while keeping the blood supply constant; this should serve as an upper bound for the self-sacrifice you should make in terms of sexual restraint.
You make an excellent point. I clarified that the sexual restraint required is not as great as it may seem, but convincing other people to donate regularly (I have done so at least twice in my life) is still much less of a sacrifice.
(nods) For me, it’s not a pragmatic question of whether I donate or not: after ~20 years in a mutually monogamous relationship, I am confident that my donating blood reduces the percentage of infected blood in the supply, regardless of my gender, and that’s the metric that matters.
But I spent some time trying to make sense of the arguments pro and con, a few years back, and mostly came to the conclusion that I didn’t trust anyone’s arguments.
It is certainly true that if you divide the community of potential donors into two groups, and the frequency of blood-born pathogens is higher in group A than group B, and your filtering mechanisms aren’t 100% reliable, then the blood supply is N% safer if you remove group A from potential donors.
It is equally certainly true that you can do that division in thousands of different ways, and each way of doing that division gets you a different N.
I was hoping to find a comparison of estimated Ns for different plausible policies, and perhaps a recommendation for the best policy.
What I found instead was that defenders of the existing policy were making the first argument and saying “See? The policy makes the blood supply N% safer! We have to keep doing it, to do otherwise would be unsafe!” while at the same time disregarding questions about how large N actually was (i.e.., how many lives were actually at stake? 1000? .001? Somewhere in between?) and whether a different policy might get you a much larger N, while opponents of the policy were disregarding the first argument altogether.
But I spent some time trying to make sense of the arguments pro and con, a few years back, and mostly came to the conclusion that I didn’t trust anyone’s arguments.
My conclusion is somewhat related. I have no particularly good reason to believe that I am better able to establish blood donation and usage policy than the Red Cross or the medical practitioners. I just give them my blood and they can use it or not as they see fit. I’d do it just for the health benefits anyway.
For my own part, I appreciate that the Red Cross (and etc.) is trying to satisfy multiple constraints, only one of which is the actual safety of their blood supply, and I don’t object to that. But the constraints that apply to them in articulating a policy don’t necessarily apply to me in donating blood.
On the other hand you have constraints that they do not have, not least of which is the lack of scaling benefits for your research and decision making efforts.
We are left with an optimal approach of considering what we know of our own blood that the collection agency does not (or is forbidden from discriminating on). We can approximate whether this knowledge would make the blood more suitable or less. Only if ‘less’ do we need worry about how significant that extra knowledge is.
I don’t do it, mostly because I’m so irritated by the policy that I’ve worked my way into a completely counterproductive “F—k it, then, donate your own f—king blood, see if I care” kind of sulk about it. I’m not proud of this, but there it is.
Yes, I condone it… indeed, I endorse it… in situations very much like mine.
They aren’t assessing that risk in a logical fashion. If they were, they would have similar restrictions on donation by ethnic group. (It is possible that the Red Cross would like to do that also but knows that it is political unfeasible.)
All blood administered to patients comes from voluntary, uncompensated donations. Plasma used in research studies may be compensated, but may not be transfused. This is the most important factor keeping our blood supply safe, and is far more effective than laboratory testing alone.
This article on the ethics and pragmatics of blood source—compensated vs uncompensated—was fascinating, IMO. Though it may be somewhat out-of-date.
This is Holden Karnofsky, the co-Executive Director of GiveWell, which is referenced in the top-level article and elsewhere on this thread.
I think there is an important difference between discussing the marginal impact of a blood donation and the marginal impact of a vote. When it comes to blood donations, it is possible for everyone to simultaneously follow the rule: “Give blood only when the supply of donations is low enough that an additional donation would have high expected impact”, with a reasonable outcome. It is not possible for everyone to behave this way in elections: no voter is able to consider the existing distribution of votes before casting their own.
I am only casually familiar with TDT/UDT, but it seems to me that that “Give blood only when the supply of donations is low enough that an additional donation would have high expected impact” should get about the same amount of credit under TDT/UDT as giving blood, and thus the extra impact of actually giving blood (as opposed to following that rule) is small regardless of what decision theory one is using.
I bring this up because the discussion of marginal blood donations is parallel to analysis GiveWell often does of the marginal impact of donations. We do everything we can to understand the marginal (not average) impact of a donation and recommend organizations on this basis, and we believe this is a very important and unique element of what we offer (more on this issue). We try to push donors to underfunded charities and away from overfunded ones, and I do not think the validity of this depends on any controversial (even controversial-within-Less-Wrong) view on decision theory, though I am open to arguments that it does.
Completely agree with your general point on marginal analysis (although I’m a TDT skeptic), and am a fan of GiveWell, but this is trivially wrong:
It is not possible for everyone to behave this way in elections: no voter is able to consider the existing distribution of votes before casting their own.
This seems to assume away information about the size of the electorate as well as any predictive power about the outcome. Surely the marginal benefit of a Presidential vote in a small swing state is massively higher than in a large solidly Democratic state, for example. And in addition to historical results, there is polling data in advance of the election to improve predictions.
Besides this being theoretically true, we can see it empirically from the spending patterns of both Presidential campaigns and political parties on Congressional races. They allocate money to the states / races where they believe it will do the most marginal good, which is often a very inequal distribution. Thus they do, in fact “consider the existing distribution of votes before casting” their advertising dollars.
Patrissimo, fair enough. I was thinking that voters can’t vote with the same degree of knowledge of the existing situation that they can have with blood donations. Arguments over TDT certainly seem more relevant to voting than to blood donations. But you are right that voters have lots of relevant information about the likely distribution of votes that can be productively factored into their decisions regardless of the TDT debate. Glad to hear you’re a fan of GiveWell.
0.00001 sounds low to me. Given that hospitals aren’t normally blood-limited, there are always fluctuations around the average, and I’d be surprised if becoming blood-limited happens less than 1 day in 10^5. Two trauma cases can be enough to create a local crisis
My actions and your actions aren’t perfectly correlated, because we’re somewhat different.
No matter how you handle this, it seems to suggest that my donation would acausally affect some fraction of other people’s donations. So it might count as, e.g., +/- 2 million, which is still more-or-less marginal, since the costs are multiplied as well.
Maybe more than that?
It’s stil a far cry from just dividing 5 million/16 million.
Edit: Isn’t utility maximized if the abstract computations “What humans do” and “The thing with greatest marginal benefit” equalized, though? If utility is convex, yes. So there should be some other rule, like, if you’re at a bad equilibrium, act so as to break it. I am unsure how this works.
If you adopt the fiction that all blood comes from voluntary, uncompensated donations
I know that plasma and such are compensated, but where is blood paid for? Places where it’s cheaper than transporting it from areas that have surplus volunteer efforts? Or they don’t publicize that compensation is available because that would shrink the volunteer base?
However, I am curious as to why it’s obvious to you that 3 lives is too high of a number on the margins.
Around 15 million pints of whole blood are donated per year in the US. At 3 lives per pint that comes out to 45 million. We can also assume that if lives per pint is 3 at the margin then the more efficient cases it will be even more than that. The population of the US just isn’t high enough to account for that.
Oh, then there there is the fact that a lot of cases use a whole heap more than one pint of blood. (For example.)
Dead babies or children are a bad metric precisely because of this reason.
Years in good physical and mental health seem a better way to measure what people are going for.
A donation of blood saves less than one life in my estimates, but it improves quality of life and adds in my opinion a few years of healthy happy life.
It seems odd the more people of a certain blood type would sustain injuries requiring blood, or that people of a certain blood type would care more about blood donation
So if the people who heard about blood banks and were interested, and the people who needed them are nearly randomly distributed in the population, I would expect the demand vs supply of each type to average out to the same figure.
However, I have heard blood banks ask specifically for people with rare blood types to donate, so it would appear that this theory is wrong. Alternatively, there is an equal shortage of all types, and someone in marketing thought that the specification would attract more people. (Even though if I was going to use the dark arts to make a group more likely to come, I would target the largest one)
It seems odd the more people of a certain blood type would sustain injuries requiring blood, or that people of a certain blood type would care more about blood donation.
For the same reason donating organs if one is mixed race is worth more (than organ donations by individuals of predominantly monoracial ancestry) because of complications in compatibility for the demographic.
The reason for this is the compatibility of the blood types, for example O-negative-blood can be donated to everyone and is therefore used in emergencies where the blood type of the recipient is not known.
I know pretty much nothing about the mechanics of saving and donating blood, but I’d expect logistical effects to push demand for rare types up even if the actual need for them is proportional to their rarity.
The Red Cross claims that 1 pint saves “up to 3 lives”. I’m not sure what to make of that, given that it’s an upper bound and presented by a non-partial source.
I presume it means that the first pint of blood donated, if allocated with efficient triage, could save three lives. At the margin I assume the figure is a small fraction of a life per pint donate.
The Red Cross claims that 1 pint saves “up to 3 lives”. I’m not sure what to make of that, given that it’s an upper bound and presented by a non-partial source.
If anyone can do better, I would be very interested in knowing the answer. I always try to give blood as often as possible under the assumption that I save at least one life each time, but a more robust figure would be nice.
That can’t possibly be right, not on the margins.
Given you and wedrifid’s responses, I am now updating my estimate of number of lives saved significantly downwards. However, I am curious as to why it’s obvious to you that 3 lives is too high of a number on the margins.
Maybe I’m just being naive here, but in a case that straightforward and that possible for the average person to understand, where there’s nothing odd or unprestigious about the action and lots of people are doing it already, where, on the margins, an additional American life is saved each time another person donates blood, I have trouble believing that even a world this insane wouldn’t push blood donations a little harder.
I take it that you’re suggesting marginal analysis based on the standard correct classical causal decision theory (in which no one is responsible for saving a life by donating blood unless someone would have actually died had that donation not been made) out of either belated humility about the probability of an SIAI-originating decision theory being correct, or because you’re planning to actually convince someone and you don’t want to invoke Hofstadterian superrationality in place of the standard correct decision theory?
:)
My guess would be that at the margin, a blood donation saves less than 0.00001 lives. (Otherwise, compensation would be increased for the paid donors). But, if you want to use a TDT/UDT style analysis, here are some relevant statistics from the American Red Cross:
The number of blood donations collected in the U.S. in a year: 16 million (2006).
The number of patients who receive blood in the U.S. in a year: 5 million (2006).
Given these numbers, I would estimate that roughly 0.5 million (US) lives are saved (more accurately, extended) by blood products annually. If you adopt the assumption that all blood comes from voluntary, uncompensated donations, and divide those 0.5 million lives among the 16 million annual donations, you get one life saved for every 32 pints donated—not as much as jsteinhardt hoped, but still significant enough to earn a major warm-and-fuzzy.
I happen to administer a lot of blood to my patients, so let me answer some of the factual questions.
The way they calculate “up to 3 lives” is in the most trivial way: blood you donate is fractionated into red cells, plasma, and platelets. Each of those may go to a different recipient.
All blood administered to patients comes from voluntary, uncompensated donations. Plasma used in research studies may be compensated, but may not be transfused. This is the most important factor keeping our blood supply safe, and is far more effective than laboratory testing alone.
Given that blood banks need to keep a sufficient store of blood available of each type, rarer blood types are generally in greater need than, say, A After all, a larger proportion of blood of those types must be discarded. O blood is obviously highly useful in trauma situations, and is therefore in high demand as well.
The distribution of donors’ and recipients’ blood types should not be assumed to be equal: people with blood type A are significantly more likely to donate than people with blood type B. This exacerbates the discrepancies due to point 3.
The number of lives saved can be calculated in two ways:
a. the feel-good way. Every time a physician gives a unit of blood to a patient e does so believing it is a life-saving procedure. So if 3 units are given the patient’s life was saved 3 times in rapid succession. (You have to be willing to save a life multiple times, because that’s the analysis we’re using for the rest of this discussion: multiple mosquito nets saved the same kid’s life multiple times over his lifetime; that same kid was then saved by anti-diarrheal treatments; etc. The same analysis belongs here). Now, we subtract the number of patients who die, but that’s a small number. So 26 million transfusions/16 million donations = 1.6 lives saved per donation.
b. the marginal way. Donations are currently sufficient for usage; we benefit in three ways from more donations. First, we can be slightly more profligate with trauma patients who have a low survival chance; this saves a minimal number of lives. Second, fresher blood is associated with better outcomes than older blood; the extent of this effect is unknown but is an area of current research interest. The calculation would have to look at the likelihood that your donation reduced the average shelf age of the blood being administered times the survival improvement from the fresher blood. Third, blood from multiparous women is associated with ARDS; an increase in donation would allow us to stop using it.
I’ve donated blood a few times and I’m type A+. Why is it that B’s are less likely to donate, or is that unknown? Are my donations likely to be marginally useless?
I have mostly donated blood in the past for signaling reasons, conversational high ground, and a vague desire to match the 15-gallon mark that my grandfather got his name in the paper for. There’s a plaque of the newspaper mention in my grandma’s house and I’ve been looking at it my whole life. Also I figure the Red Cross will let me know if I come down with one of the diseases they screen for, and it’s a free way to get my iron levels checked (attempting to donate blood was how I found out I was anemic in the first place). These reasons aren’t likely to evaporate if I find that I have been saving only tiny fractions of expected lives, but I would probably endure less inconvenience in order to donate for only these reasons as opposed to these reasons on top of lifesaving.
Your donations are not marginally useless! (unless you’ve been pregnant a couple times—in that case, consider stopping).
The reason for the discrepancies in donation rates between types A and B is both simple and complex: ethnicity. In the interests of safety (avoidance of Hepatitis C, HIV, etc) we’ve set up a system that subtly encourages certain types of donors and discourages others. The system is not racist per se, but it is most effective in obtaining donations from white, middle-aged, middle-class males.
Regarding signaling reasons: we are obviously very afraid of blood donated for signaling purposes. Accordingly, we do not allow people to donate to their relatives except under very unusual circumstances. Additionally, we give people an “out” by checking a box which tells the center to draw and discard their blood. That way people who fear they may be high-risk donors can get the social approval of donating without harming any patients.
I’ve never been pregnant, but what is it about multiple pregnancies that renders the blood non-preferred?
Obvious guess: Your blood then contains antibodies to the blood type of your babies.
Essentially this. The A/B/O blood groups represent the most relevant antigens in human blood. There are a host of others (Rh, Duffy, Kell, etc.) which typically create only minor problems in a transfusion and which can be ignored in an emergency. But a person who has been exposed to allogeneic blood via multiple transfusions or pregnancies becomes more likely to develop antibodies to some of these antigens. The donor’s antibodies or white cells can react to the person being transfused, causing lung damage.
In the case of the Rhesus factor it should be noted that it is minor once and then only minor for males. Being thereafter unable to safely give birth to healthy Rh+ children is definitely not a minor consequence even if it is better than ‘probably going to die today’. (Unless, I suppose, you happen to some Rh+ antiserum lying around but no Rh- blood, which will usually avoid the future difficulties.)
I use “minor” differently than you do, to mean “unlikely to cause death”. Obviously cross-matched blood is always preferable for a variety of reasons (including possible infertility, in the case of young females).
I would avoid RhoGam in the case of a patient who needs a RBC transfusion, incidentally. It would be unlikely to be safe or effective.
I would like to express that my approval of this phrase extends beyond the capacity of upvoting and into the capacity of a comment expressing approval.
And having both your arms removed is “Just a flesh wound!”.
Oops, I see that this has already been asked.
While we’ve got you here, can you explain why gay men cannot donate? This upsets a lot of gay people that I know.
I understand that it’s easier to catch STDs (not just HIV/AIDS) from a man than from a woman. But the current U.S. rule (A man cannot donate if he’s had sex with a man; a woman cannot donate if she’s had sex with a man who’s had sex with man.) is lopsided.
The even-handed rule that you cannot donate if you’ve had sex with a man would keep the supply safe without having to rely on people’s being able to trust their partners. But it would keep most women from donating, so maybe it’s not worth it. The even-handed rule that you cannot donate if you’ve had sex with man who’s had sex with a man would still keep out most gay men, but it would probably help to heal the rift.
If a man is gay and sexually active, he’s almost certainly had sex with a man who’s had sex with a man, even if the men he has had sex with has only had sex with him. I don’t see how this phrasing of the rule would be an improvement.
My phrasing was unclear; make it “if you’ve had sex with man who’s previously had sex with a man (other than you)”. There wouldn’t be any point in forbidding me from donating (if I’m male) because the man that I’ve had sex with has had sex with me!
This change would include more people; it includes monogamous gay male couples who began their relationship as virgins (as well as some other people). Not many more, but it makes it clear that the blood collector is only willing to trust you and your partners, no further.
Frankly, the first even-handed rule (no sex with a man, period) makes more sense to me. Why should the blood collector trust that I know (if I’m a woman) whether all of the men that I’ve had sex with have had sex only with women? (No doubt many women are donating contrary to guidelines because they don’t know this about their partners.) But because this would cut the potential donor pool in half, the blood collector is basically forced to trust me about my partners too.
In fact, the blood collectors trust women to know the sexual history of their partners, but not men. They are not asking everybody the same questions.
Another possible solution, not even-handed, but more honest: Just don’t ask women anything about the subject.
The idea that a person can be trusted to know about their partners’ partners is preposterous; no other question (in the U.S.) asks the donor about other people’s behaviour, and for good reason. Instead of half-assedly trying to be even-handed about it, just admit what they’re doing: ruling out men who’ve had sex with men, because many of their partners will have had sex with other men, and so on back (in many cases) a long way; but accepting women who’ve had sex with men, because most of their partners won’t have had sex with men, stopping the transmission-from-men sequence.
I’m confident that they already accept blood from most women who’ve had sex with men who’ve had sex with men (because the women don’t know this about their partners), and they are surely aware of this (if I am correct) fact. So why are they asking questions of people who don’t actually know the answers?
Gay people will still be upset that they can’t donate, but I at least would be more willing to trust that the blood collectors are actually making an honest decision.
Mostly, my faith in the quality of the blood supply derives from what testing they’re doing to the blood, not from what unenforceable policies they’re suggesting to the donors.
I’d actually be surprised if the latter significantly affected the quality of the blood.
Mostly, I think the problem they are a solution for is maintaining public confidence in the blood supply. Which I acknowledge is an important problem. And it may well be that being perceived as excluding gay men and their partners is a better solution to that problem than anything else they might do; I don’t know.
That said, if I’m wrong and these policies really do solve a problem related to the blood supply, yet another possible solution is: don’t allow people who have had unprotected sex to donate.
Or, if that’s too big a chunk of your potential donor base, make it people who have had unprotected sex outside of a monogamous relationship.
The original rule bars ‘a man who has had sex with a man’ - X—and then any women who’ve had sex with X. It’s a logical phrasing but unfortunately X maps exactly onto “gay man”, so it feels like gay men are being specifically targeted. The rephrasing mollifies that sense of targeting without, as far as I can tell, changing the included or excluded people.
The original phrase is even-handed, however. If you overspecified an even-handed rule and said “1) You cannot donate if you’re a man who has had sex with a man who has had sex with a man, and 2) you cannot donate if you’re a woman who has had sex with a man who has had sex with a man”—ie, prevent “man who has had sex with a man” from coming into sexual contact with any donor—you could reduce 1) down to “man who has had sex with a man” (it logically implies three, four, and so on iterations). This, therefore, reduces down to the actual rule they have in place.
There are not many times I see a line of reasoning and have to reject it at every single step. Apart from being conceptually absurd the very thought is morally objectionable. It totally devalues the value of ‘saving a life’ to the point of utter meaningless. How could that ever make someone ‘feel-good’?
Which part? I thought that started silly (it’s explaining the logic behind a non-profit’s puffery, did you expect it to be rigorous?) but then got better. The idea of “saving a life” is pretty meaningless when you poke at it- it’s all just lifespan extension. And so the idea that each emergency treatment extends lifespans by the ‘natural span of a life’ is silly. If someone would die if they don’t receive a unit of blood at 50 separate occasions on their life, should each transfusion get the full moral weight of saving a life? If so, we just gave this person 50 lives. If not, then we need to abandon the language of “saving a life” and talk about “extending a lifespan” (because we can say those units of blood each added a year to the person’s life, for example).
Thanks, this is exceptionally informative.
I didn’t realize that donations were sufficient for usage. Is this barely maintained by calling people when blood supplies are low, or does blood regularly get thrown out, or is there some other reason that supplies closely match need?
A combination of the above. We have a core group of donors who can be called in emergency situations, we increase the intensity of blood drives when supplies are low, we reduce marginally-beneficial uses of blood when supplies become low, and we are better able to discard the oldest least-effective blood whenever supplies increase.
We are likely to face challenges in meeting future need. The cohort that most regularly donates blood is aging...
I’m assuming you are from the US, do you think the same is true for other countries? Also which demographic are you referring too?
I’m more interested in the cohort that isn’t aging. What is their secret? A new and improved Calorie Restricted diet? Perhaps that explains their inability to generate sufficient excess blood for donation.
They could also be brain uploads, which would also explain the inability.
Good point. I wonder if they would consider donating CPU time instead!
Thanks for data!
Only vaguely relatedly: if you have pointers to (or are willing to synthesize) a reliable calculation of expected lives-saved/deaths-caused by maintaining or discarding the existing Red Cross policies about who is “allowed” to donate blood, especially the relatively controversial ban on male donors with homosexual acts in their sexual history, I would be interested.
Full disclosure: I do have a personal/emotional stake in this question, but I really really don’t want to set off a political/ethical conversation about it. I’m asking it here because, as with a lot of politically charged topics, the arguments I’ve found on both sides are mostly a case of framing the question so as to give the answer one wants to give, rather than so as to answer the question that was asked, and I’m looking for a more objective analysis.
I also wanted to ask this question.
Giving blood is important to me. It is so important that I have chosen not to pursue relationships with other men in order than I can continue to give blood without lying to do so. I expect that sooner or later, I will choose otherwise, and a sexual relationship will be important enough to me to sacrifice my ability to ever give blood again, and this distresses me.
I can accept that the risks of HIV may be high enough to make this a reasonable choice on the part of United Blood Services / Red Cross. However, I would like to be quite sure that this is the case, or to be told that my blood isn’t as important as I previously though it was. I was previously giving blood on the impression that each donation saves around a twentieth of a life; this thread doesn’t change that estimate enough for me to feel like I can stop donating in good conscience.
On the margins, I expect that each marginal pint of blood saves only a very small fraction of a life. As several readers pointed out, this doesn’t mean that we should ordinarily be calculating on the margins, since it’s not like you can use a pint of blood for something else instead; in terms of moral credit, you should think of yourself as part of a reference class of people who all choose to donate blood for around the same reasons, and who all get an equal share of the lives saved.
However, the Red Cross has already decided that they’re willing to X out the entire homosexual community, and I would expect the reference class of those who refrain from sexual activity in order to continue donating blood to be small, and I would guess that if this entire reference class refrained from donating blood, not a single additional life might be lost.
Modern-day hospitals are not, so far as I know, blood-limited. They need a routine flow of blood in order to routinely save lives. They do not need more blood to save more lives. That’s the impression I got, anyway; some quick Googling even said that they usually have enough blood to just use O-negative instead of matching types.
I hate to say this, but I think you’re making the wrong sacrifices here. I estimate a very high information value for further investigation on your part; I would expect it to show that you were safe to stop donating blood and resume sexual activity without costing anyone one-twentieth of a life. If you’re really feeling guilty or worried, resume sexual activity and send a donation to the Singularity Institute as a carbon offset. If you can speed up a positive Singularity by one minute that works out to around 100 lives, never mind increasing the probability.
I think I was accidentally misleading by failing to add that I’m bisexual. Not giving blood reduces my pool of potential romantic partners by roughly 10%, and doesn’t prevent me from having fulfilling relationships. I don’t think I would abstain from sex in order to give blood even if I knew I could save a life with each donation. Even if that’s an incredibly selfish decision, I’m just not that good a person.
Regardless, the support of everyone who replied is very much appreciated.
...technically, doesn’t speeding up a negative singularity also save lives—the lives of those who would otherwise have been born and then killed but were instead never born and therefore couldn’t be killed? In fact, I think speeding up a negative singularity actually “saves” more lives than speeding up a positive one using this calculation—a quick Google search indicates ~250 people are born every minute and ~100 people die every minute.
Replace “save lives” with “extend lifespans.” All the math will suddenly start working out better.
Agreed, I retrospect I should have phrased the original question in terms of QALYs or some similar metric.
In a fairly meaningful sense, no life has ever been saved before. Nobody has actually been prevented from dying yet. A positive singularity could change that.
I believe you can make an easier calculation: change the denominator from lives to units of blood. How much effort/money/social capital would it take you to convince one more person to donate one more unit? [ignore the cost to that person, as it’s likely zero or slightly beneficial]. Calculate the effort it therefore would take you to replace yourself as a donor while keeping the blood supply constant; this should serve as an upper bound for the self-sacrifice you should make in terms of sexual restraint.
You make an excellent point. I clarified that the sexual restraint required is not as great as it may seem, but convincing other people to donate regularly (I have done so at least twice in my life) is still much less of a sacrifice.
(nods) For me, it’s not a pragmatic question of whether I donate or not: after ~20 years in a mutually monogamous relationship, I am confident that my donating blood reduces the percentage of infected blood in the supply, regardless of my gender, and that’s the metric that matters.
But I spent some time trying to make sense of the arguments pro and con, a few years back, and mostly came to the conclusion that I didn’t trust anyone’s arguments.
It is certainly true that if you divide the community of potential donors into two groups, and the frequency of blood-born pathogens is higher in group A than group B, and your filtering mechanisms aren’t 100% reliable, then the blood supply is N% safer if you remove group A from potential donors.
It is equally certainly true that you can do that division in thousands of different ways, and each way of doing that division gets you a different N.
I was hoping to find a comparison of estimated Ns for different plausible policies, and perhaps a recommendation for the best policy.
What I found instead was that defenders of the existing policy were making the first argument and saying “See? The policy makes the blood supply N% safer! We have to keep doing it, to do otherwise would be unsafe!” while at the same time disregarding questions about how large N actually was (i.e.., how many lives were actually at stake? 1000? .001? Somewhere in between?) and whether a different policy might get you a much larger N, while opponents of the policy were disregarding the first argument altogether.
My conclusion is somewhat related. I have no particularly good reason to believe that I am better able to establish blood donation and usage policy than the Red Cross or the medical practitioners. I just give them my blood and they can use it or not as they see fit. I’d do it just for the health benefits anyway.
For my own part, I appreciate that the Red Cross (and etc.) is trying to satisfy multiple constraints, only one of which is the actual safety of their blood supply, and I don’t object to that. But the constraints that apply to them in articulating a policy don’t necessarily apply to me in donating blood.
On the other hand you have constraints that they do not have, not least of which is the lack of scaling benefits for your research and decision making efforts.
We are left with an optimal approach of considering what we know of our own blood that the collection agency does not (or is forbidden from discriminating on). We can approximate whether this knowledge would make the blood more suitable or less. Only if ‘less’ do we need worry about how significant that extra knowledge is.
We also need to worry if the answer is ‘more’ and because of that we decide to lie on the answer form so that we can donate.
I kind of get the impression that TheOtherDave is doing that, or at least would condone it under circumstances very much like his.
I don’t do it, mostly because I’m so irritated by the policy that I’ve worked my way into a completely counterproductive “F—k it, then, donate your own f—king blood, see if I care” kind of sulk about it. I’m not proud of this, but there it is.
Yes, I condone it… indeed, I endorse it… in situations very much like mine.
They aren’t assessing that risk in a logical fashion. If they were, they would have similar restrictions on donation by ethnic group. (It is possible that the Red Cross would like to do that also but knows that it is political unfeasible.)
Will Saletan has an article on this.
Thx.
This article on the ethics and pragmatics of blood source—compensated vs uncompensated—was fascinating, IMO. Though it may be somewhat out-of-date.
This is Holden Karnofsky, the co-Executive Director of GiveWell, which is referenced in the top-level article and elsewhere on this thread.
I think there is an important difference between discussing the marginal impact of a blood donation and the marginal impact of a vote. When it comes to blood donations, it is possible for everyone to simultaneously follow the rule: “Give blood only when the supply of donations is low enough that an additional donation would have high expected impact”, with a reasonable outcome. It is not possible for everyone to behave this way in elections: no voter is able to consider the existing distribution of votes before casting their own.
I am only casually familiar with TDT/UDT, but it seems to me that that “Give blood only when the supply of donations is low enough that an additional donation would have high expected impact” should get about the same amount of credit under TDT/UDT as giving blood, and thus the extra impact of actually giving blood (as opposed to following that rule) is small regardless of what decision theory one is using.
I bring this up because the discussion of marginal blood donations is parallel to analysis GiveWell often does of the marginal impact of donations. We do everything we can to understand the marginal (not average) impact of a donation and recommend organizations on this basis, and we believe this is a very important and unique element of what we offer (more on this issue). We try to push donors to underfunded charities and away from overfunded ones, and I do not think the validity of this depends on any controversial (even controversial-within-Less-Wrong) view on decision theory, though I am open to arguments that it does.
Completely agree with your general point on marginal analysis (although I’m a TDT skeptic), and am a fan of GiveWell, but this is trivially wrong:
This seems to assume away information about the size of the electorate as well as any predictive power about the outcome. Surely the marginal benefit of a Presidential vote in a small swing state is massively higher than in a large solidly Democratic state, for example. And in addition to historical results, there is polling data in advance of the election to improve predictions.
Besides this being theoretically true, we can see it empirically from the spending patterns of both Presidential campaigns and political parties on Congressional races. They allocate money to the states / races where they believe it will do the most marginal good, which is often a very inequal distribution. Thus they do, in fact “consider the existing distribution of votes before casting” their advertising dollars.
Patrissimo, fair enough. I was thinking that voters can’t vote with the same degree of knowledge of the existing situation that they can have with blood donations. Arguments over TDT certainly seem more relevant to voting than to blood donations. But you are right that voters have lots of relevant information about the likely distribution of votes that can be productively factored into their decisions regardless of the TDT debate. Glad to hear you’re a fan of GiveWell.
0.00001 sounds low to me. Given that hospitals aren’t normally blood-limited, there are always fluctuations around the average, and I’d be surprised if becoming blood-limited happens less than 1 day in 10^5. Two trauma cases can be enough to create a local crisis
My actions and your actions aren’t perfectly correlated, because we’re somewhat different.
No matter how you handle this, it seems to suggest that my donation would acausally affect some fraction of other people’s donations. So it might count as, e.g., +/- 2 million, which is still more-or-less marginal, since the costs are multiplied as well.
Maybe more than that?
It’s stil a far cry from just dividing 5 million/16 million.
Edit: Isn’t utility maximized if the abstract computations “What humans do” and “The thing with greatest marginal benefit” equalized, though? If utility is convex, yes. So there should be some other rule, like, if you’re at a bad equilibrium, act so as to break it. I am unsure how this works.
I know that plasma and such are compensated, but where is blood paid for? Places where it’s cheaper than transporting it from areas that have surplus volunteer efforts? Or they don’t publicize that compensation is available because that would shrink the volunteer base?
Around 15 million pints of whole blood are donated per year in the US. At 3 lives per pint that comes out to 45 million. We can also assume that if lives per pint is 3 at the margin then the more efficient cases it will be even more than that. The population of the US just isn’t high enough to account for that.
Oh, then there there is the fact that a lot of cases use a whole heap more than one pint of blood. (For example.)
Dead babies or children are a bad metric precisely because of this reason.
Years in good physical and mental health seem a better way to measure what people are going for.
A donation of blood saves less than one life in my estimates, but it improves quality of life and adds in my opinion a few years of healthy happy life.
I wonder if getting too focused on the best (or worst) case scenario is a named logical error.
I’m also not sure whether giving a rare blood type is likely to save more lives than giving a more common blood type.
It seems odd the more people of a certain blood type would sustain injuries requiring blood, or that people of a certain blood type would care more about blood donation So if the people who heard about blood banks and were interested, and the people who needed them are nearly randomly distributed in the population, I would expect the demand vs supply of each type to average out to the same figure.
However, I have heard blood banks ask specifically for people with rare blood types to donate, so it would appear that this theory is wrong. Alternatively, there is an equal shortage of all types, and someone in marketing thought that the specification would attract more people. (Even though if I was going to use the dark arts to make a group more likely to come, I would target the largest one)
Blood types vary by ethnicity, SES varies by ethnicity, injuries and donations vary by SES.
For the same reason donating organs if one is mixed race is worth more (than organ donations by individuals of predominantly monoracial ancestry) because of complications in compatibility for the demographic.
The reason for this is the compatibility of the blood types, for example O-negative-blood can be donated to everyone and is therefore used in emergencies where the blood type of the recipient is not known.
I know pretty much nothing about the mechanics of saving and donating blood, but I’d expect logistical effects to push demand for rare types up even if the actual need for them is proportional to their rarity.
I presume it means that the first pint of blood donated, if allocated with efficient triage, could save three lives. At the margin I assume the figure is a small fraction of a life per pint donate.
Thanks to both of you. Yes, what I’m curious about is essentially the figure at the margin,