This is a useful discussion, and highlights the continuing problem around the conflation of socio-economic inequity with racial inequity. Both are real, but they aren't synonymous, and impact and overlap one another in complicated ways. As noted in the piece, the framework proposed by the National Academies would tackle this in a rational, evidence-based approach. But this won't alleviate the consistent pull toward driving all our efforts thru a racially-tinted lens prior to implementation by government. Here's hoping that as a nation we can address this equitably, efficiently, and without our typical political rancour.
Thank you for an important discussion and a much-needed perspective. I would like to suggest two guiding principles that would lead to similar conclusions but might (!) be an easier point of agreement to start from.
A. Make it voluntary, at least until all those who want it have gotten it. B. Save as many life-years as possible.
Principle B may need defending. But first, I think most would agree with "saving lives." Obviously, that means first of all protecting health-care workers (#2 above). It also means protecting those most likely to get sick (#1 above) and most likely to transmit it (#3 above). Furthermore, it says to ignore ethnicity, income, etc. as direct decision variables -- the main point of this article. (Of course, as pointed out, protecting those most in danger will correlate with ethnicity and income.)
So principle B seems to be the underlying principle of the article, and I think it would help to articulate that, because it seems so easy to agree on, with one exception ...
Older folks are far more at risk once they catch COVID than young people, and this is taken into account by the National Academies proposal mentioned above. But it should be noted that this will tend to overwhelm many other factors. The CDC says someone who is 85+ is 630 times as likely to die as someone who is 18--29. (Full disclosure: I'm only about 80 times more likely...)
The counter-argument to simply using the risk-of-death factor (e.g. 630 times) is that it's more tragic when a 25-year-old dies than when a 90-year-old dies, because many more years of life are lost. I agree with that logic. It would cause bickering to try to fine-tune life expectancy estimates, but the risk to older folks could easily be adjusted using social-security life-expectancy tables. And that's what I have in mind for principle B when I say "life-years."
thank you. It's a difficult calculus, as you note. Saving life years (which deceased kidney allocation takes into account - a different issue in a number of ways) in the case of COVID would mean targeting the younger yet they are the least likely to develop bad symptoms if infected...so maybe a wash.
You’re quite welcome. Being a numbers guy, I had already discussed this problem four or five months ago when my friends were working on a related op-ed that they published in Nature. So I had mentally checked that your concern was not a worry. But I just now ran the numbers, and prioritizing Life-Years save would not come anywhere close to reversing the old-age priority based on lives saved.
For example, ninety-year-olds would still get a priority of 48 (instead of 630), while 24-years-olds would have a priority of 1 in both systems. (Higher means getting the vaccine sooner.)
As I mentioned, the problem that may be perceived when using "Lives Saved," is that the age effect “will tend to overwhelm many other factors.” So I just checked the effects of the two systems when taking into account “Rich” and “Poor” (proxy labels for any factor that causes more deaths for the “Poor” group). I assumed 2.1 times as many deaths because that’s the CDC ratio for Blacks vs Whites.
The result is that when using “Lives Saved,” a 90-year-old “Rich” person would get priority over all “Poor” people younger than about 82. So age would pretty much overwhelm the Rich-Poor factor.
With “Life-Years Saved,” a 90-year-old “Rich” person would get priority over all “Poor” people younger than about 68. So the compensatory effect of taking account of “poverty” is allowed to play a larger role with the “Life-Years” approach.
To me, it seems fair that the 68-year-old poor person gets this small increase in priority relative to the aged because the 68-year-old has not yet had the benefit of some very valuable life years that the 90-year-old has already received. And that’s exactly what the Life-Year approach is taking into account.
The difference between the two approaches weakens at younger ages, and I think that too is sensible. If you are interested in such details and would like any assistance with calculations, just let me know. -Steve
Zip code is an intriguing idea
This is a useful discussion, and highlights the continuing problem around the conflation of socio-economic inequity with racial inequity. Both are real, but they aren't synonymous, and impact and overlap one another in complicated ways. As noted in the piece, the framework proposed by the National Academies would tackle this in a rational, evidence-based approach. But this won't alleviate the consistent pull toward driving all our efforts thru a racially-tinted lens prior to implementation by government. Here's hoping that as a nation we can address this equitably, efficiently, and without our typical political rancour.
How about doing it by prevalence in particular zip codes? The ones that have the most deaths or hospitalizations or other non-racial metric
Thank you for an important discussion and a much-needed perspective. I would like to suggest two guiding principles that would lead to similar conclusions but might (!) be an easier point of agreement to start from.
A. Make it voluntary, at least until all those who want it have gotten it. B. Save as many life-years as possible.
Principle B may need defending. But first, I think most would agree with "saving lives." Obviously, that means first of all protecting health-care workers (#2 above). It also means protecting those most likely to get sick (#1 above) and most likely to transmit it (#3 above). Furthermore, it says to ignore ethnicity, income, etc. as direct decision variables -- the main point of this article. (Of course, as pointed out, protecting those most in danger will correlate with ethnicity and income.)
So principle B seems to be the underlying principle of the article, and I think it would help to articulate that, because it seems so easy to agree on, with one exception ...
Older folks are far more at risk once they catch COVID than young people, and this is taken into account by the National Academies proposal mentioned above. But it should be noted that this will tend to overwhelm many other factors. The CDC says someone who is 85+ is 630 times as likely to die as someone who is 18--29. (Full disclosure: I'm only about 80 times more likely...)
The counter-argument to simply using the risk-of-death factor (e.g. 630 times) is that it's more tragic when a 25-year-old dies than when a 90-year-old dies, because many more years of life are lost. I agree with that logic. It would cause bickering to try to fine-tune life expectancy estimates, but the risk to older folks could easily be adjusted using social-security life-expectancy tables. And that's what I have in mind for principle B when I say "life-years."
In any case, I think it makes discussions more sensible if they can be based on simple principles, at least when those are easier to agree on than all the ramifications. I hope that could help in this case. https://www.cdc.gov/coronavirus/2019-ncov/covid-data/investigations-discovery/hospitalization-death-by-age.html
thank you. It's a difficult calculus, as you note. Saving life years (which deceased kidney allocation takes into account - a different issue in a number of ways) in the case of COVID would mean targeting the younger yet they are the least likely to develop bad symptoms if infected...so maybe a wash.
Dear Sally,
You’re quite welcome. Being a numbers guy, I had already discussed this problem four or five months ago when my friends were working on a related op-ed that they published in Nature. So I had mentally checked that your concern was not a worry. But I just now ran the numbers, and prioritizing Life-Years save would not come anywhere close to reversing the old-age priority based on lives saved.
For example, ninety-year-olds would still get a priority of 48 (instead of 630), while 24-years-olds would have a priority of 1 in both systems. (Higher means getting the vaccine sooner.)
As I mentioned, the problem that may be perceived when using "Lives Saved," is that the age effect “will tend to overwhelm many other factors.” So I just checked the effects of the two systems when taking into account “Rich” and “Poor” (proxy labels for any factor that causes more deaths for the “Poor” group). I assumed 2.1 times as many deaths because that’s the CDC ratio for Blacks vs Whites.
The result is that when using “Lives Saved,” a 90-year-old “Rich” person would get priority over all “Poor” people younger than about 82. So age would pretty much overwhelm the Rich-Poor factor.
With “Life-Years Saved,” a 90-year-old “Rich” person would get priority over all “Poor” people younger than about 68. So the compensatory effect of taking account of “poverty” is allowed to play a larger role with the “Life-Years” approach.
To me, it seems fair that the 68-year-old poor person gets this small increase in priority relative to the aged because the 68-year-old has not yet had the benefit of some very valuable life years that the 90-year-old has already received. And that’s exactly what the Life-Year approach is taking into account.
The difference between the two approaches weakens at younger ages, and I think that too is sensible. If you are interested in such details and would like any assistance with calculations, just let me know. -Steve
appreciate all your careful thought on this, Steven!