Stanford Algorithm Decided To Vaccinate Only Seven of Its Frontline COVID-19 Workers, Out of 5,000 Doses (theverge.com) 104
An algorithm determining which Stanford Medicine employees would receive its 5,000 initial doses of the COVID-19 vaccine included just seven medical residents / fellows on the list, according to a December 17th letter sent from Stanford Medicine's chief resident council. The Verge reports: Stanford Medicine leadership has since apologized and promised to re-evaluate the plan. "We take complete responsibility for the errors in the execution of our vaccine distribution plan," a Stanford Medicine spokesperson said in a statement to The Verge. "Our intent was to develop an ethical and equitable process for distribution of the vaccine. We apologize to our entire community, including our residents, fellows, and other frontline care providers, who have performed heroically during our pandemic response. We are immediately revising our plan to better sequence the distribution of the vaccine." The residents' letter also alleges that the error in the algorithm was found on Tuesday but that leadership opted not to make changes to the plan ahead of its December 17th release.
The initial plan led to demonstrations from medical staff in addition to the letter sent by the chief resident council. Here's how the algorithm reportedly worked, according to NPR: "According to an email sent by a chief resident to other residents, Stanford's leaders explained that an algorithm was used to assign its first allotment of the vaccine. The algorithm was said to have prioritized those health care workers at highest risk for COVID infections, along with factors like age and the location or unit where they work in the hospital. Residents apparently did not have an assigned location, and along with their typically young age, they were dropped low on the priority list."
The initial plan led to demonstrations from medical staff in addition to the letter sent by the chief resident council. Here's how the algorithm reportedly worked, according to NPR: "According to an email sent by a chief resident to other residents, Stanford's leaders explained that an algorithm was used to assign its first allotment of the vaccine. The algorithm was said to have prioritized those health care workers at highest risk for COVID infections, along with factors like age and the location or unit where they work in the hospital. Residents apparently did not have an assigned location, and along with their typically young age, they were dropped low on the priority list."
Summary contradicts title (Score:1)
And yeah, the average, median, and probably 80th percentile resident is under 30 and unlikely to suffer the worst of the covid outcomes. So yeah, it was probably the correct call to start with.
Re: (Score:3)
The Chief Resident Council's response [documentcloud.org] to the Administration's algorithm includes both a breakdown and their demands. The first paragraph states
There is still no articulated plan to vaccinate the remaining 1,300+ residents and fellows, including those on the front line directly treating
COVID-19 patients.
Re: (Score:3)
Re: (Score:2)
Older, sicker, and more at risk candidates were ahead of the young, healthy med students, and the healthy young med students think they take priority over an elderly orderly with diabetes and a heart condition that cleans up COVID patient rooms.
Re: (Score:2)
So they got 5,000 doses, they have 6,300 workers - that's why they prioritized the vaccination assignments. Absent 1,300 additional doses, who do the young, healthy, med students feel shouldn't get vaccinated so that the residents can be? The janitors? Orderlies? Elderly staff? Who?
They are just pissed because Stanford didn't get enough for everyone - who would the residents bump off the list for vaccines?
Re: Summary contradicts title (Score:2)
The correct call is those likely to spread it, not those likely to die from it.
Exactly! (Score:2)
Like in old folks homes. They need to vaccinate the cleaning people and staff, not the residents. The residents don't go anywhere. The maintenance people are probably the ones that spread it. Inoculate the other two and the other population should be fine.
Re: (Score:1)
Exactly wrong. Vaccinate those most as risk first. Going down the risk level scale until general availability.
Especially since it is unknown if the vaccine prevents spread vs just making it so you do not get sick, but could still be a carrier / spreader.
No, wrong assumptions (Score:3)
If you build a wall of vaccinated people around the high risk people (nursing/care home residents don't travel), then you don't need to vaccinate the high-risk people.
Use your brain, not your endorphins.
The staff at these places are the ones generally responsible for these infections, since by definition the residents are sheltering-in-place.
Vaccinate the ones that move, not the ones that don't. They're the ones exposing everyone.
Re: (Score:1)
If you build a wall of vaccinated people around the high risk people you have a corralled group who can't get away from a still infectious person. Which is no better than not vaccinating them at all. Vaccination doesn't create a magical Batfink style shield of steel that the virus just pings off.
You vaccinate the ones that can't move first to protect them from those that can.
Re: (Score:2)
5,000 doses, 6,300 workers - everybody isn't getting the vaccine in the first round, period. Young healthy med students, and a thousand other hospital workers didn't make the cut, boo hoo, so sad - it's science that put them outside the first 5,000 doses, so now they are making emotional arguments.
Re:Exactly! (Score:5, Informative)
That approach would work if the vaccines were proven to greatly reduce spread. They're not shown to provide sterilizing immunity in airways, so it's possible for a vaccinated person to get an infection and spread it. Their immunization should prevent the virus from spreading to the rest of their body, but the same barriers that keep almost all dust and germs out of your bloodstream also keep almost all of your antibodies and immune cells inside your body.
So until we have enough data to understand the effect of vaccination on the spread of covid, the vaccines should be allocated preferentially to those expected to lose the most if they are not vaccinated.
For reference:
https://www.texastribune.org/2... [texastribune.org]
https://www.nytimes.com/2020/1... [nytimes.com]
https://www.chop.edu/centers-p... [chop.edu]
mod parent up (Score:2)
nailed it
Re: (Score:2)
Exactly.
Young healthy med students are less at risk, lower on the list, and didnt make the cut - tell me more about how hospital workers were ranked, maybe there's an issue - but I bet there isn't.
Re: (Score:2)
You also have to exclude everyone who is not vaccinated from entering said nursing home for any reason, including friends and family.
Also it is far more accurate to say that "the residents can't go anywhere". If you imagine that nursing homes are actual prisons, and that normally people go in and are prevented from leaving for any reason you know nothing about them. People in nursing homes frequently require medical care not provided in the nursing home, if for no other reason.
Re: (Score:2)
Resident means med students in this article, not long-term patients.
Re:Summary contradicts title (Score:4, Insightful)
Most of the public believes the hospital employees on the floor get top priority so I doubt there's be any complaints if they were 1st in line. I know I'd rather there be enough doctors and nurses available if I were to need hospitalization.
Re: (Score:2)
Doctors as a rule deal with patients one on one far less than the nursing staff, but I guess nurses are expendable. The vaccination schedule should always be for the most likely to come into contact with the pathogen. You don't give the
Re: Summary contradicts title (Score:2)
Other than, you know, the possibility that as more experienced but therefore more vulnerable doctors, they could return to work with actual patients.
Re: Summary contradicts title (Score:1)
Re: (Score:1)
That's an easy one, give the vest to the rich guy.
Re: (Score:2)
OTOH, fellows were also left out.
Re: (Score:2)
They have 5k doses and 6.3K workers, 1.3K workers were left off the list, only med students protesting. Sense of entitlement?
Re: (Score:1, Flamebait)
Hey, look! More glib comments from the right!
For the fucking life of me, I just don't understand how the right-wing can't manage to see beyond the illness risk to the individual subject when talking about COVID response and mitigation. A large part of the reason we're vaccinating people is to ensure that they don't spread the virus, and they don't have to quarantine if directly exposed. Yes, they're wearing PPE, but doctors regularly encounter people with severely compromised immune systems (e.g. cancer pat
Re: Summary contradicts title (Score:1)
Re: (Score:2)
... what?
Re: (Score:2)
Hospital with 6,300 workers got 5,000 doses - which older, sicker, staffers are to be put behind young, healthy med school students?
The young, healthy, med student that works in the pediatrics ward and NEVER sees a COVID patient is at profoundly lower risk than an orderly that cleans covid patient rooms or an ICU dictor with comorbidity conditions.
Re: (Score:2)
You don't actually know how any of this works, do you?
Re: Summary contradicts title (Score:3)
Re: (Score:2)
6,300 workers, 5,000 doses - the young, healthy med students, and about a thousand other workers (presumably also young, healthy) didn't make the cut for the first round of vaccines. They feel their education should entitle them to preferential treatment.
Re: (Score:2)
5,000 doses, 6,300 workers - can't vaccinate all workers in first round, simple math.
Young healthy doctors want to bump older, sicker, more at risk non-doctors from the list to ensure they,the young healthy med students are protected.
Wonder how many had COVID in past 90 days and thus have a reduced need for a vaccine?
Re: (Score:2)
"God only knows what facts TFA actually reports."
Indeed.
Since we couldn't possibly RTFA, perhaps a bloody newbie can enlighten us?
Re: (Score:2)
It's a triage situation - not every med student sees covid patients, goes on covid floors, or has comorbidity conditions. They have 5,000 doses and 6,300 workers - which covid floor orderlies get bumped because the med student in the physical therapy room feels he needs covid vaccine first?
Re: (Score:2)
On the other hand, as medical personnel working with COVID patients, they have a very high level of exposure.
Meanwhile, senior faculty working from home (low exposure) were allocated vaccines. While senior, few were likely senior enough to fall into a high risk group (think older, still working not nursing home senior).
Re: (Score:2)
Senior staff - with comorbidity conditions - working from home are at greater risk than spa young healthy med student working in the burn ward that NEVER sees a covid patient or visits the ICU.
Re: (Score:2)
And how is that relevant? residents that were working with COVID patients were skipped over for the vaccine.
Re: (Score:2)
The problem is that they were vaccinating doctors who have been working entirely from home for months, and not vaccinating residents who are directly treating multiple covid patients per day.
Who is more likely to spread or contract Covid, the person working from home, or the person in direct contact with Covid patients?
Re: (Score:2)
Re: (Score:2)
I bet they have exactly SEVEN med students that interact with covid patients, that's why seven med students made the first round of vaccines.
5,000 doses, 6,300 workers, do you assign vaccine on the basis of science (occupation, age, comorbidity) or do you simply go by title (drs, med students, nurses, etc. and bump the orderlies and staff workers)?
Stanford chose science, and it offended the healthy young med students sense of self-importance.
Re: (Score:2)
At this stage most of these frontline workers have probably already been infected and recovered so there is probably a very high level of natural immunity among such people... Unless you take extreme precautions, if you are working with infected patients every day there is a good chance of becoming infected yourself.
More than just Individual Risk (Score:2)
the average, median, and probably 80th percentile resident is under 30 and unlikely to suffer the worst of the covid outcomes.
That may be but how many patients will suffer because they need treatment and there are not enough medical staff to treat them because they are off sick with Covid? How many already sick patients will be exposed to Covid because the doctor treating them had Covid but did not know? (and remember these people are at far more risk since they are already dealing with something else). There is more than the risk to the individual to consider when considering vaccinating medical staff and any decent algorithm sh
Re: (Score:2)
Doctors that attend to covid patients don't also cover the pediatric ward and spread the disease to non-covid patients.
Stanford had 5,000 doses and 6,300 workers, someone was going to miss out on the first round - only med students are protesting.
Re: (Score:2)
Exactly. The clickbait headline reads like only 7 frontline healthcare workers were to be vaccinated, while I'm pretty sure they assigned all 5,000 doses to hospital workers, while only 7 went to young, otherwise healthy RESIDENTS (AKA student doctors) out of an unknown quantity of residents.
From my sofa, I suspect everyone else that was assigned the first round of vaccine was older, had comorbidity conditions, and worked in areas that put them at greater risk.
A young (27 year-old) resident working in pedia
AI Failure (Score:4, Insightful)
Re: (Score:2)
If you don't like this algorithm, I have others.
Management Failure. (Score:2)
They had so little intelligent people in their management, that they resorted to "Artificial intelligence". However, they have none of this also.
Management failure.
Total failure.
Re: (Score:2)
1,300 Stanford workers were not going to be vaccinated in the first round of 5,000 doses - simple math. These med students think they should be in the first round, perhaps they should go thru the list of employees and go tell the 1,300 they feel don't need the vaccine that they, as med students, are more important, more deserving, than the older, sicker, orderly workers on the list.
Failure to Think (Score:2)
Sure, when it fails spectacularly, call it an “algorithm”.
It was not just the algorithm that failed but also the people who chose to use its results without thinking. It's like the idiots who drive cars into lakes, oceans etc. because their GPS told them to. It takes a failure to think as well as a failure of the algorithm.
Re: (Score:2)
Sure, when it fails spectacularly, call it an “algorithm”. When you think it needs lots of VC money, call it “AI"
I don't think the output was all that unreasonable. What's clearly unreasonable is that headline.
reset to healthcare for all mode and rerun (Score:2, Insightful)
reset to healthcare for all mode and rerun
Re: (Score:2, Flamebait)
Re: (Score:1)
so inmates get in front of non inmates and the cost is 0? so usa.
try in Canadian mode
Re: (Score:2)
Issue is constraint of vaccine in first round. How's the national vaccine program working in Canada these days?
Re: (Score:2)
6,300 workers, 5,000 vaccine doses in first round - everyone can't get it, simple math.
We can't vaccinate everyone all at once - we only have tens of millions of doses, not hundreds of millions of doses - patience.
Don't blame the algorithm (Score:5, Insightful)
A committee of humans reviewed the algorithm's results and blessed it. The committee had nobody representing the residents. That's the problem: making decisions without all relevant parties represented in the decision-making.
Re: (Score:2)
A committee of humans reviewed the algorithm's results and blessed it.
Yup. [twitter.com]
Re: (Score:2)
Woah there, calling for representation must be some kind of ism... I mean whenever anyone suggests that representation might be a valid criteria for selecting someone it's considered an affront to the meritocracy and probably sexist against men or something.
Here's hoping that events like this may help a few people understand why it matters.
Re: (Score:2)
I think it's safest to blame everyone. Whoever made the algorithm, whoever used it, whoever blessed it, whoever is proceeding on that basis. ANY of those people could have willfully created this outcome, or it could have been a group effort between all of them.
Re: (Score:2)
The point of the algorithm is to remove 'representation' from influencing the decision.
Ideally, the data provided accurately describes the workforce, and when that data is run through a properly designed - scientifically-based - algorithm, the most at-risk employees will be vaccinated first, if that how the algorithm is structured.
Personally, I'd like to hear more about the other hospital workers that didn't make the cut, there were 1,300 of them, but only the med school residents are protesting.
Let me guess : A.I supported blockhain tech? (Score:2)
They should have taken the gig-economy buzzwords into consideration when making their algorithm
Re:Let me guess : A.I supported blockhain tech? (Score:5, Insightful)
No. The "algorithm" is a scapegoat that makes it sound like people weren't involved at all in setting priorities that put department heads working remotely ahead of residents and fellows actually treating COVID patients.
Re: (Score:2)
Department heads (with comorbidity conditions) are at greater risk than the (young, healthy) med school student working in the burn ward that never goes near a covid patient.
FFS this is round 1, are we pretending the other 1,300 Stanford workers will never be vaccinated? I bet they get sufficient vaccine in a couple weeks.
Re: (Score:2)
Re: (Score:2)
Or, they are younger, healthier, and aside from seven of them don't interact with covid patients at all. There are lots of other areas in the hospital aside from ICU covid cases.
The real questions (Score:1)
Why would someone who claimed doctors are faking covid-19 deaths to make more money be one of the first people to get a vaccination [tumblr.com]?
Why would a healthy 49-year old man, with no known underlying medical conditions, be one of the first people to get a vaccination after spending months claiming the pandemic is a hoax [tumblr.com]?
America doesn't have a healthcare system (Score:1)
Re: (Score:2)
https://64.media.tumblr.com/80... [tumblr.com]
Admittedly off topic, but I wanted to back up the previous posters claim.
Re: (Score:2)
Re: (Score:2)
Thank you Dr. Ezekiel Emanuel.
Frontline? Heroes? (Score:1)
Re: (Score:3, Informative)
They're not working in BSL4 facilities, they're wearing basic masks, etc. if they're lucky. PPE, like birth control, is not perfect. Furthermore, since mask effectiveness is probabilistic, risk increases with duration of exposure; these folks aren't ducking into a restaurant for take-out, they're spending long hours on shift in a hospital, some of them in close contact with infected patients.
"An algorithm" (Score:2)
As far as I am aware that are as yet no sentient artificial intelligences capable of reprogramming themselves beyond the boundaries of their original design. Therefore this "algorithm" didn't just spring out of the grave of Charles Babbage fully formed but was designed by human beings working to a specification which embodied assumptions, norms, and unstated goals as well as written targets.
tl;dr: someone at Stanford wanted the results to come out this way and designed the magic "algorithm" to achieve that
typo in the headline (Score:3)
Stanford Decided To Vaccinate Only Seven of Its Frontline COVID-19 Workers, Out of 5,000 Doses
That should fix it.
Re: (Score:2)
They vaccinated 5,000 out of 6,300 employees and workers IN THE FIRST ROUND of vaccinations. They couldn't get to everyone, and most med students likely don't go near covid patients in ICUs, they likely work in Er, pediatrics, burn wards, physical therapy, cardiovascular, gastrointestinal, etc areas of the hospital and rarely (if ever) actually see covid patients.
Every room in every hospital in every state is not full of covid patients, a fact the press won't likely point out until after Biden is in office.
I wish I had that excuse when I was a kid: (Score:2)
an algorithm ate my homework.
The algorithm seems reasonable (Score:5, Informative)
Re: (Score:2)
Re: (Score:1)
* Posting AC because I applied mod points.
Re: (Score:2)
More details of the algo here: https://www.technologyreview.c... [technologyreview.com]
The main issue seems to have been with the way it calculated risk. Residents were not assigned to any department (they move around) so did not get the department's risk added to their own.
Re: (Score:2)
Maybe, that's ok - since they move around, their risk is reduced compared to an ICU nurse or orderly who is exposed to covid patients every hour of every work day.
Re: The algorithm seems reasonable (Score:2)
That's if your only concern is whether the worker might die. A medical worker that contracts it is out of rotation for 2 weeks, and may end up with long-haul symptoms that generally aren't understood as well as the mortality risks. There's also just the added stress of worrying if you're going to contract it.
It's a poor algorithm that only takes into account mortality here.
Re: (Score:2)
So, do you know the algorithm, or just making up criteria so you can criticize what you imagine the algorithm is?
Re: (Score:2)
5,000 doses, 6,300 workers. 1,300 workers were not included for a variety of reasons IN THE FIRST ROUND of vaccines - there will be a second round.
I'd live to see the med student working in pediatrics go tell a 60 year-old orderly with diabetes and a bad heart condition that the med student will be taking his place on the vaccination list, and the orderly can wait for the next round.
Algorithms are only as good as the designers (Score:1)
Algorithms are only as good as the designers, and often their are implicit presumptions. Or more simply, like Professor Frink in the Simpson's would say "Oops, I forgot the carry." The sad thing is this can seriously impact people's lives and health.
Found the algorithm (Score:2)
Here it is:
getPriority(person) {
if (person.role == 'ADMIN') {
return 1;
} else {
return 120 - person.age;
}
}
It was reviewed by the administration and approved so what's the problem?
there is a bias in the algorithm (Score:2)
...the bias is toward purported equitability.
Decide what you want most.
If it's efficacy, that may not necessarily jibe with equality.
If all your doctors are white, then yes, you may have VERY good reasons for "giving it only to whites" at first.
Science is followed only when convenient (Score:2)
An algorithm was developed, reviewed, and approved.
But people weren't happy with the results because it didn't align with what they wanted to happen. So the algorithm is condemned, set aside, and an alternative will be decided upon ... by people who will find a politically adroit way of placating the loudest complainers.
Think about that. Science continues to be sidelined for political expediency.
Re: (Score:2)
They will put people of color ahead of whites (virus impacts people of color worse than whites, according to countless news stories repeated by countless politicians);
They will put low income workers ahead of high-earners (virus disproportionately impacts low-income workers, according to press and politicians);
They will put first-responders ahead of others, unless they are white or high-income earners (see above);
They will prioritize healthcare workers ahead of others (unless white or high-earners, see abov
Curious (Score:2)
Of the 4,993 other folks, how many are actually seeing patients in person? I've read that many who are getting the vaccine are working from home and not seeing patients. That does seem a little funky even if due to age they are at higher risk. I would generally prioritize anyone seeing patients in person first since if those staff go down, we can't treat people.
First round should have prioritized all in person treatment folks first and then future rounds can roll into the higher risk crowds. That would ensu
Re: (Score:2)
Re: (Score:2)
There's a "handfull" just in my home state. Significantly more suffering "long covid" who may or may not fully recover.
Re: (Score:2)
Unfortunately those people who refuse to wear masks are also clogging up the hospitals when they get sick. So unless we're going to refuse them medical care on the basis of their behavior, we need to vaccinate them too.
Re: (Score:2)
You were so close to having a point, until you added that bit about masks.