From a Small Town in North Carolina To Big-City Hospitals, How Software Infuses Racism Into U.S. Health Care (statnews.com) 242
An anonymous reader shares a report: The railroad tracks cut through Weyling White's boyhood backyard like an invisible fence. He would play there on sweltering afternoons, stacking rocks along the rails under the watch of his grandfather, who established a firm rule: Weyling wasn't to cross the right of way into the white part of town. The other side had nicer homes and parks, all the medical offices, and the town's only hospital. As a consequence, White said, his family mostly got by without regular care, relying on home remedies and the healing hands of the Baptist church. "There were no health care resources whatsoever," said White, 34. "You would see tons of worse health outcomes for people on those streets." The hard lines of segregation have faded in Ahoskie, a town of 5,000 people in the northeastern corner of the state. But in health care, a new force is redrawing those barriers: algorithms that blindly soak up and perpetuate historical imbalances in access to medical resources. A STAT investigation found that a common method of using analytics software to target medical services to patients who need them most is infusing racial bias into decision-making about who should receive stepped-up care. While a study published last year documented bias in the use of an algorithm in one health system, STAT found the problems arise from multiple algorithms used in hospitals across the country.
The bias is not intentional, but it reinforces deeply rooted inequities in the American health care system, effectively walling off low-income Black and Hispanic patients from services that less sick white patients routinely receive. These algorithms are running in the background of most Americans' interaction with the health care system. They sift data on patients' medical problems, prior health costs, medication use, lab results, and other information to predict how much their care will cost in the future and inform decisions such as whether they should get extra doctor visits or other support to manage their illnesses at home. The trouble is, these data reflect long-standing racial disparities in access to care, insurance coverage, and use of services, leading the algorithms to systematically overlook the needs of people of color in ways that insurers and providers may fail to recognize.
The bias is not intentional, but it reinforces deeply rooted inequities in the American health care system, effectively walling off low-income Black and Hispanic patients from services that less sick white patients routinely receive. These algorithms are running in the background of most Americans' interaction with the health care system. They sift data on patients' medical problems, prior health costs, medication use, lab results, and other information to predict how much their care will cost in the future and inform decisions such as whether they should get extra doctor visits or other support to manage their illnesses at home. The trouble is, these data reflect long-standing racial disparities in access to care, insurance coverage, and use of services, leading the algorithms to systematically overlook the needs of people of color in ways that insurers and providers may fail to recognize.
I really hate how (Score:5, Insightful)
Words get twisted to shift the Overton window in topics and push agendas. Accidental bias is by definition not racism. Racism requires over malice and is intentional.
Re:I really hate how (Score:4, Insightful)
https://www.merriam-webster.co... [merriam-webster.com] see definition #2
bias, accident or not can be racist in its effects. In this case, algorithmic bias, implemented in software, encodes racist actions/analysis into the practice of medicine.
Re: (Score:3, Insightful)
Re:I really hate how (Score:4, Insightful)
A southern Baptist church decides to help build houses in the poor section of town: obviously these are liberals engaged in virtue signalling!
Maybe what the term "virtue signaling" means is really "I disagree with what they're doing but I can't properly articulate why, so I will resort to name calling." Or it could just be an overly cynical view that no one ever does a good deed without an ulterior motive to try to get more sex or money or power.
Re:I really hate how (Score:5, Insightful)
Doing something to help people isn't virtue signaling. That's just doing something good. However, setting up an ad campaign that cost ten times what was spent on the actual doing of good to promote how awesome and virtuous you are is virtue signaling. It's also virtue signaling to push agendas and legislation that require others to sacrifice to achieve a goal that you are not willing to sacrifice or invest in yourself.
The key aspect of "virtue signaling" is the "signaling" part. It's advertising that you are virtuous rather than just being virtuous.
Re: (Score:3, Insightful)
But the term gets tossed out too often and too lightly, most oftenly used to mean you don't like someone for being a bleeding heart. It's overused and meaningless now, except as a way to mark the speaker's political stance.
Re: (Score:3)
This seems to be true of all words these days. Accurate, nuanced descriptions don't provide enough attention grabbing punch. You get so much more response by using the most extreme language available regardless of whether such language is warranted.
Re: (Score:2)
(Mathew 6:21) Be careful not to practice your righteousness in front of people in order to be noticed by them. If you do, you will have no reward from your Father in heaven.
(Mathew 5:22) But I tell you that anyone who is angry with a brother or sister will be subject to judgment. Again, anyone who says to a brother or sister, 'Raca,' is answerable to the court. And anyone who says, 'You fool!' will be in danger of the fire of hell.
Re: I really hate how (Score:5, Insightful)
And as a general principle, if you're pointing me to a dictionary you yourself have just scrawled on with purple crayon in plain view to back up your point, you're not making an effort at communication; you're just raising a middle finger. Don't be surprised when you get one back.
Comment removed (Score:4, Insightful)
Re: I really hate how (Score:5, Insightful)
Thus, ascribing to wider society the acceptance and common usage of politically charged propaganda invented by a small but vocal minority in the service of an explicitly political agenda to gain power and influence for themselves is definitely not descriptive linguistics.
It may be charitably termed prescriptive linguistics, but I'm not feeling charitable today so I'll call it what it is: sock puppetry, woke proselytizing, and a big fat fuck you of Marxist lies in the service of attaining power and control over us rubes who believe in hard work, personal responsibility, and individual liberty.
Re: I really hate how (Score:2, Insightful)
prejudice
discrimination
bigotry
oppression
and many more, which can each be further refined to indicate degree of intentionality and mental state of the person performing those actions.
Implicit in the distinctions made by the use of those terms is that correlation is not causation and that unequal outcomes are not evidence of oppression: if I say I don't want low income housing to be built in
Re: (Score:3)
I notice you left out racism. Racism is of course tied to the blatant lie of race theory, inherently prejudiced. We can breed to together we are the same race, regional climate differences that alter appearance over time, as in skin colour is not a measure of race, simply of environment.
The story itself is a BLATANT lie, a theory not about race, the lie, but about classism. Where the rich smart class oppress the poor not so smart class.
Did any poor white folk live on the wrong side of the tracks, it was a
Re: I really hate how (Score:5, Insightful)
Racism has always meant one thing: the inherent belief that a race or races are superior to others. If you make the statement "the white race is superior to all other races," you are a racist. If, however, you were to state "I don't like black people" then you are a bigot, not a racist. If you were to state "I don't like BLM protests" then that makes you...someone who doesn't like BLM protests. It neither confirms nor denies any stance on race, being more of a political statement than anything else. Anything above or beyond that is concluding something you lack sufficient information to conclude.
You're confusing/conflating "racism" with what is actually "bigotry" or "prejudice," something distressingly common and all-too-accepted in modern discourse. Words like these have actual meanings and are not fungible. Calling someone racist used to actually mean something, unlike today where it's used as a general-purpose label of anything someone doesn't like.
Re: (Score:3)
I think the problem with an algorithm apparently showing racial bias is that life circumstances are already racially biassed. Put crudely, most brown people are poor, and most rich people are white.
This has apparently had a major influence on the likelihood of infection by coronavirus and of getting worse effects when infected. I think epidemiologists now think that this can be accounted for mostly by economic circumstances, e.g. brown people working in risky jobs like nursing and public transport, living i
Re: I really hate how (Score:2)
As a tautological statement, anything, be it an automobile, an apartment building, a computer program, or a piece of furniture, that is not explicitly designed and implemented to combat racial disparities in X will likely have the effect that racial disparities in X will continue.
Similarly, a machine designed to do anything other than dig an irrigation trench will leave the crops unirr
Re: (Score:2, Interesting)
In what world does skin color denote criminal behavior?
Is shoplifting different because somebody's skin color is different?
Re: (Score:3, Interesting)
It seems you don't understand what I am trying to convey here, so let me clarify.
We are discussing the issue of folks looking at a black person and seeing a criminal, by default. Thanks to our culture and our media portrayals of blacks as predominantly criminal, many people have an inherent bias they don't even know about. So a shopkeeper might follow a black man around the store, even though that black man is, say, a lawyer. The shopkeeper does not see a lawyer, he sees a criminal.
The primary "white privil
Re: (Score:3)
The word we used to use in the past was "prejudiced". It makes more sense as a term as it accurately describes what is happening without implying that there is any malice or ill-intent.
Re: I really hate how (Score:4, Insightful)
So you are saying someone can be prejudiced against black people, but not racist? That simply does not make any sense to me.
Let me just add, people lie. They lie to other people, and they lie to themselves. Especially about things like racism.
People also love to deflect from issues they find uncomfortable, and one of the oldest deflections in the book is to argue semantics.
Re: (Score:3)
You can be prejudiced against people wearing hoodies, keep an extra eye on them in the store to make sure they're not some hooligans trying to get on your lawn. The idea is that you could use the word prejudiced even if it was against a certain ethnic group or against people wearing certain styles of clothing or against a certain religion or assuming that the guy in the lowrider next to you at the stop who is blasting the stereo full blast with gangsta rap is probably not a superior court judge, etc. Racis
Re: (Score:3, Insightful)
Still seems like "prejudiced against black people" is the exact same thing as "racist" even if the word does trigger snowflake racists.
And you might want to take a look at the civil rights movement. Plenty of people who claimed not to have a racist bone in their bodies argued semantics, saying they would support the cause, but those black people were just being too mean about it. They just needed to be more polite, and be patient.
Of course, they were all full of shit. And pretty damn racist, if you ask me.
Re: (Score:3)
That's a disingenuous argument. You understand the difference perfectly well, but set up a straw man to distract from the point and push a conflicting agenda.
You're going well beyond spinning these days and should change your handle to lying propagandist.
Re: (Score:2)
The thing there is that in every example you gave, the characteristic that lead someone to take different action was the person's race. Literally pre-judging someone based on their race. At one time, that would be called racial prejudice but not racism. The term racism was reserved for a conscious belief that one race was actually better than another. But in current fashionable use, racial prejudice is called racism. Either way, it's wrong.
Nobody that I know of is accusing the software of either of those th
Re: (Score:3)
It does. If we insist on calling this racism, it will get a bolted on kludge of a fix that doesn't work very well and leaves a lot of other people still unfairly disadvantaged.
The ACTUAL issue is that the software perpetuates the past (whatever it was). It puts people who were the victims of racism at a disadvantage. It puts people who were victims of poverty at a disadvantage. It puts people who didn't live near healthcare at a disadvantage. Why do you only want to address one of those inequities?
Re: I really hate how (Score:4, Insightful)
While that statement is technically true, it is not a universal truth for all Chinese and all blacks. Ascribing to individuals the characteristics of the extremes of the group is morally wrong. Individuals must be judged on their own words and deeds, not on immutable characteristics of birth they have no control over.
Thus, you would not be morally wrong to run criminal background checks on all prospective tenants (even if most of your prospective tenants or loan applicants are black) , you would be wrong to only run criminal background checks on black applicants. Even if you only run them on guys named "B'lake" or "A-A-ron."
The higher moral truth in our secular religion of democracy is not "be extra nice to folk with more melanin." The higher moral truth is to not hold against people things they have no control over, like their last names or their skin color, and to treat people as they are all children of God, whatever their conception of Him may be.
Re: (Score:2)
Mod parent up. He's the only one here free of racism.
Re:I really hate how (Score:5, Insightful)
Don't even bother using Merriam-Webster as a resource - they've jumped the shark:
https://hotair.com/archives/al... [hotair.com]
"Merriam-Webster Alters Definition Of Sexual “Preference” To Say It’s Offensive After Hirono Attacked Barrett For Using It"
Note also:
https://twitter.com/AGHamilton... [twitter.com]
Re: (Score:2, Offtopic)
How long has that definition been in the dictionaries?
We keep seeing the cycle where people change the definitions of words to win arguments. Racism used to mean the Democrat party founding the KKK, enslaving people simply for being black, and setting up Jim Crow laws.
Now racism is seen in every bit of randomness that ends up being unequal, and the 'fix' always seems to be to add a permanent racial bias to the randomness.
Re:I really hate how (Score:5, Interesting)
Your comment about changing word definitions reminded me of a quote I once read...
"The whole literature of the past will have been destroyed. Chaucer, Shakespeare, Milton, Byron—they'll exist only in Newspeak versions, not merely changed into something different, but actually contradictory of what they used to be."
From some book. An instruction manual on building a better society or something. You can look it up.
Re: (Score:2)
Wow. I've never heard a more relevant reason to keep version control on documents and archive it read-only on a periodic (e.g., yearly) basis.
Re:I really hate how (Score:5, Insightful)
So just because the statistical breakdown of something shows a difference based on race, does not automatically make it racist. In the case of TFA, the bias appears to be due to poverty, not race. It would be a red flag if the U.S. had universal health care (it would mean poorer people are being denied health care they're entitled to). But since the U.S. does not have universal health care, it is not inherently indicative of a bias (poorer people are simply not as able to afford as good health care, so are in poorer health resulting in their care being more expensive for the same condition).
For a charge of racism to stick, you must be able to show intent - that a person was discriminated against specifically because of their race. Intent is incredibly difficult to prove in court. So SJW types have been trying to water down the threshold by advocating the belief that any act which results in a racist outcome is racist. That standard simply doesn't make logical sense, and results in ridiculous cases like a manufacturer being sued for stamping "negro" on a black couch. When the word was put there because it was manufactured in a Spanish-speaking country, and "negro" is the Spanish word for black. There was no racist intent by the manufacturer, it was all imagined by the people seeing the word.
Re:I really hate how (Score:4, Interesting)
>"So SJW types have been trying to water down the threshold by advocating the belief that any act which results in a racist outcome is racist. "
Actually it is worse than that. They are trying to say that any act/system/etc which results in a worse outcome for minorities is racist. Note the difference. If the outcome is better for a minority- not racist. If the outcome is worse for a non-minority- not racist. No other factors matter. Doesn't matter if the system in question was designed by or run by minorities. Doesn't matter if there are actual disparities that exist outside the system. Doesn't matter if the people with worse outcomes have made more bad decisions or life-choices. Doesn't matter what the intent was of the system.
The words racist and racism mean absolutely nothing definable or useful anymore. It is just a label thrown on anything or anyone someone doesn't like, for pretty much any reason. This is the predictable result of identity politics.
At most, the situation in the article is most likely ancillary discrimination by correlation and has absolutely nothing to do with "racism."
Re: (Score:2)
Even so, if an algorithm assesses statistical likelihoods to determine someone's eligibility for a loan or medical care, rather than judging each case solely on its own merits
Re: (Score:2)
As far as I can tell, it discriminates against people who had difficulty accessing healthcare in the past for any reason.
Re:I really hate how (Score:5, Insightful)
bias, accident or not can be racist in its effects. In this case, algorithmic bias, implemented in software, encodes racist actions/analysis into the practice of medicine.
Of course the obvious problem with this is that it is not a racial group being suppressed.
Attempting to alias the group of those who are poor with the group of those who are black is prejudicial bullshit.
https://www.merriam-webster.co... [merriam-webster.com]
These algorithms ultimately are intended to preference people likely to enrich the hospital over those who don't.
You can certainly make the argument this is unfair to poor people which it most certainly is. Yet when you go that extra mile and intentionally alias poor with black a number of problems arise.
1. Poor people who are not black are ignored.
2. Rich people who are black are misclassified.
3. Prejudicial ethnic narratives are perpetuated.
4. The media wins, society loses.
The only reason the media accounts are always couched in racial terms that all but assume black = poor is pure selfish self-enrichment. Nobody cares about stories about the poor treated like crap by huge corporations.... yet if you s/poor/black/ well that gets everyone's attention doesn't it.. Outrage everywhere... read this..clicks and views... ad revenue....$$ $$ $$ $ $ $ $$.
Stat and all of the media doing this should be condemned for promulgating ethnic prejudice simply to make money rather than being rewarded with attention.
Re: (Score:2)
Re: (Score:2)
false dichotomy is built in to the two-party system; the only real choice is which shit sandwich you get to eat
Re: (Score:2)
Thank you for your substantial arguments. You really bring quality content to the debate.
Re: (Score:2, Troll)
We can argue over the meaning of words or as can try to fix this. At least if it's software it is fixable, much easier than trying to counter unconscious bias in thousands of professionals. Question is how much difference will it make with the systemic problems unaddressed.
Re: (Score:2)
We can argue over the meaning of words or as can try to fix this. At least if it's software it is fixable, much easier than trying to counter unconscious bias in thousands of professionals. Question is how much difference will it make with the systemic problems unaddressed.
There are inherent health problems associated with race and ethnicity. Here is a something on how hypertension affects blacks differently. [webmd.com]
And speaking as someone of northern European decent, I was told that my medical issues are common with people of my ethnic background (Master race? Yeah, right!)
Now the scoring system they used may have other factors other than race.
She said she doubts any providers are making referral decisions solely based on cost, but that cost-based risk scores are core features of these products that introduce money — and consequently race — into deliberations where they shouldn’t hold sway.
I wouldn't doubt it. In the USA, medicine is big business. And doctors, hospitals and other providers and suppliers like to get paid and m
Re: (Score:2)
The problem, as others have pointed out, is that hospitals want the patients that pay. Poor people are typically poor so they are unlikely to pay the most for treatment. If you want to try and fix that, go ahead, but I wouldn't bother. The US isn't exactly well known for taking great care of its lower class, and its not exactly fond of ideas that give more power or better living conditions to the same either.
With that said, the algorithm is working as intended.
Re: (Score:2)
It's not really sudden, this meaning has been used for decades.
Re:I really hate how (Score:4, Informative)
"Racism requires over malice and is intentional."
No, that is as fundamentally wrong as claiming that racism can only be perpetrated by a majority against a minority, or that it has to be pervasive to qualify as racism. Racism is a combination of believing in race (which science tells us is not even a thing!) and then going on to ascribe race certain qualities, for example claiming that black people are born lazy or that white people are born with entitlement, because genetics.
Your definition of racism is exactly the same as the one racists use to dismiss claims of racism against them, you might want to watch that. The argument that it's not racism if it's not pervasive is the one used by victims of racism who are also racist in order to claim they aren't also racist.
Re: (Score:2)
I didn't say it wasn't racist if it wasn't pervasive. I said it wasn't racist if it wasn't malicious. Racism is evil. Don't accuse someone of being evil when they make mistakes. It isn't productive and feeds division.
Re:I really hate how (Score:4, Interesting)
No, it's still racist to say certain things even if you're ignorant of why they're racist. If you are genuinely brought up to believe that black people are inferior, it doesn't matter if you say things like, "oh, they can't help being lazy, they're black!" That's a racist comment, even if it's masquerading as something else.
What IS true is that you can say and do racist things and not be a bad person as long as you're willing to try to be better. So if you accidentally say something racist, and are told by someone that it is, accepting it, apologize, move on. The step beyond that is recognizing the times that you were wrong and seeing it in the things other people do, and working against those views being promulgated—being anti-racist.
I'm half-Chinese. I was brought up with a surprising number of racist viewpoints. I've been taught a bunch of racist things about Chinese people. It doesn't make it not racist, but it also doesn't inherently make me a bad person. I'm trying to unlearn a bunch of shit, and I mostly succeed on a day-to-day basis. Really, this is one of those cases where there's value in the effort.
I remember seeing a black man being interviewed on CNN or something, talking about how he was walking past another black guy on the street, and he instinctively tensed up and ran through a bunch of scenarios in his head before he caught himself, and realized that HE had bought into the stereotypes of black men. Society has fed us a bunch of racist lies, and none of us are immune. The problem is that 'racism' and being 'a racist' (that is, someone who actively believes that some races are superior to others) are too easily conflated with merely having racist thoughts. It's understandable that nobody wants to be called racist, and so they fight back against the idea that they have any racist beliefs, and it becomes impossible to change their minds. We need to recognize that there are legitimately good people out there with racist thoughts that need to be called out when they voice them, but they also need to be reminded that it's okay, the racism is in the culture and the institutions, and as long as you're trying to be better, that's what counts.
Re: (Score:3)
And now you are moving the goalposts. Get back to me when you want to have an intellectually honest discussion.
Re: (Score:2)
I think one reason there is so much resistance is that in practice, people are not forgiven even for past racism. Consider how Paula Deen was crucified for admitting that she used the N word as recently as 30 years ago (now 40 years ago). I see no sign there of her efforts or intentions since then counting for much.
Re: (Score:2)
There is no need for malicious intent for something to be racist. The same argument was recently used to defend blackface in The Netherlands. I'm sure you can see the problem there.
As this is slashdot, I refer you to a nice SF story where someone explained how you could commit genocide without anyone ever being really malicious: Falling Free - Miles Vorkosigan Saga.
Re: (Score:2)
Racism is when you make race based assumptions about someone and act upon them. That's it. It doesn't matter what their race is anymore than it does your race. The only thing that matters is if you treated someone of a certain race a certain way based upon said race - that's it.
Re: (Score:2)
Words get twisted to shift the Overton window in topics and push agendas. Accidental bias is by definition not racism. Racism requires over malice and is intentional.
Words keep changing their meaning.
Literally is now just a generic intensifier. It can still be used for its original meaning, but almost no young people do so.
Hack has become so vague that I can't even begin to tell you what it means, but I can tell you that non-IT people use it in bizarre ways that make no sense. Ever put chocolate syrup on ice cream? Now apparently that is a "hack".
Racism has similarly become kind of a synonym for "prejudice" or "bias" and it has lost its original meaning.
Re: (Score:3, Insightful)
It's racist in the sense that the bias isn't accidental - it's a direct result of explicitly racist policy and planning. But that the wording is what you choose to focus on just tells me that you are not vested and/or don't really care about the substance of the consequences.
Re:I really hate how (Score:5, Insightful)
No I care about the consequences and fixing problems. But accusing people of racism when they are at the worst ignorant or negligent is not productive.
Re: (Score:3)
Calling a software algorithm racist isn't accusing anyone. Calling a car red isn't productive either, it's a description of the car and describes the problem if your goal is to not have a red car.
Anyhow, people who take offence to having systems and policies described as racist are not magically going to modify their biases or improve their ignorance regardless of the words used (putting aside direct personal insults.) The meaning whatever words you want to use is always going to end up in the same place -
Re: (Score:2)
The problem here is that they're calling the car red when it's actually blue.
The bias is against anyone who had poor access to healthcare historically.Due to racism among other things, that may disproportionately affect African-Americans, but it's not a race thing. Poor whites would also be affected as would people who faced geographic challenges (if the nearest hospital is hundreds of miles away, you won't go there very often even if it might be a good idea).
Even TFA's use of thetrain tracks was faulty. Th
Re: (Score:2)
The ones creating the algorithm are possible and likely inadvertently perpetuating a situation born out of racism. I hope nobody is accusing them of doing this on purpose, but it's something that data scientists do need to watch out for. As I said in another post, it's a case of overfitting, but quite subtle.
Re:I really hate how (Score:5, Insightful)
If you can't attack the facts, attack the tone.
Sufficiently advanced incompetence is indistinguishable from malice. It doesn't really matter if it's a moustache twirling racist making these decisions or a system built by lazy incompetents which encodes in software racist decisions made by racist people.
Either way the problem is the same, and yet you're trying to drag the conversation away from actual real problems of racial bias into pedantic word games. Why, I wonder are you trying to do that?
Re: (Score:3)
Finding a mistake for which the result is a racial bias and asking for it to be fixed is one thing, and likely to garner support. Calling everyone involved names when there is no evidence of intent is childish and wrong.
Re: (Score:2)
There are all sorts of ways that people have historically tried to launder racism through 'not technically based on race' criteria, as a way of giving themselves plausible deniability. Machine learning has been used in the same way - take your biased history of decisions, train a computer to replicate those decisions, and then claim you're following the computer's advice now, so now your policies are totally independent of race.
Ultimately if your decision criteria treats black people worse, it doesn't matt
Re: (Score:3)
Re: (Score:3)
Words get twisted to shift the Overton window in topics and push agendas. Accidental bias is by definition not racism. Racism requires over malice and is intentional.
Oh please.. Racism is defined by the beholder and has been this way for decades. We've lived in a society where racism has been outlawed and it's been illegal since the 1960 with the passage of the Civil Rights Act. No, that didn't end racists doing bad things, but it made it illegal and empowered the courts to apply appropriate remedies and punish the perpetrators.
What constitutes "Racism" has changed since Dr. King's "I have a Dream" speech from the Lincoln memorial. It used to be active, purposeful, i
Re: (Score:2)
Racist actions have certainly been outlawed as was "separate but equal" Jim Crow laws. In the eyes of the state, there is no distinction and rules are in place making it illegal to discriminate based on a protected class. It's illegal.
Yea, there remain pockets of racist attitudes, you cannot stamp out thoughts, but you can make acting on them illegal. And so we have done.
We have done even more, making the courts available to enforce these civil and criminal rules so if you know of an instance of discrim
Re: (Score:2)
Accidental bias may be accidental, but it's still a bias. It's also bad data science since it's overfitting the data to the current outcomes. By definition overfitting reinforces current outcomes. If they are racist in nature, it reinforces and perpetuates that racism. It becomes a self-fulfilling prophecy.
All of that just means that data scientists have to be very careful in not overfitting their data based on variables that are all strongly correlated with race (in the USA, I may add - elsewhere it may be
Re:I really hate how (Score:5, Informative)
I clicked through until I found one PDF of an actual study: https://www.soa.org/globalasse... [soa.org]
There was not one mention of race, black, or african-american in the whole article. It was misrepresented.
My hot wash analysis is that the OP appears to have submitted a fraud of a story.
Re: (Score:2)
That's one definition. But "racism" is also used as a very common English word to mean "bias based upon race", without requiring malice. That's part of the problem, some people feel absolutely insulted every time the "racist" word is used, they hate being accused of being an evil person. But that is not what the word is doing, it is not assigning blame and it is not accusing anyone of malice.
If there is a problem with algorithms, or a system, or a bureacracy that is leading to unequal outcomes based upon
Re: (Score:2)
if you read the article
Lol. Who does that? This is slashdot, we just start spouting our own opinion here, since we know everything already.
Also if you want to get a +5 mod (Score:2)
You'd think a site full of tech nerds could understand that complex interlocking systems can and would have unpredictable effects. Then again
Re: I really hate how (Score:2)
What services are denied, to similar patients, with similar diagnosis, within the same HC plan?
I want specific examples of, for example, two retired GM workers, getting the same straight-forward hip replacements, their only difference being their race, and the person of color is denied as much PT as the white perso, or the white man being offered painkillers the person of color isn't offered, or their allowed number of days in the hospital recovering are different.
The author is, I believe, relying on hand-w
Not race. Poverty. (Score:5, Insightful)
This seems to be conflating racism with poverty. Yes, minorities are statistically less well off on average than non-minorities, but with the possible exception of skewing caused by a few diseases like sickle cell that are much more common in people of color, those same algorithms should give similarly bad results for poor white people who have received similarly poor medical care in the past because of their poverty.
What this points to is not a need for changing the algorithms, or a problem with the algorithms being biased, but rather a need for changing the fundamental inequality that leads to poorer people being sicker, starting with a guarantee of universal health care. Anything short of that is just putting lipstick on a pig.
Re: (Score:2)
Right, like having the government pay for every thing poor people lack is going to be a net positive on their lives. History shows this idea to be false in the long term. The poor already have Medicaid. It's bare bones, but it's medical care. So we kind of already DO have universal care.
More government isn't the answer here, we already have the necessary programs in place.
Re: (Score:2)
I take it you're not on medicare.
Re: (Score:2)
I take it you're not on medicare.
No, I work for a living still. But I have personally managed someone who was on Medicare and it sucks if you are in any way different. Personally I blame them for the premature death of my mother because they refused to pay for the regular PET scan that would have detected that her cancer was out of remission in time to treat it. Oh no, you've had your lifetime limit of PET scans, you cannot have any more. "One size fits all" doesn't work for everybody and that's what government programs ALWAYS turn into
Re: (Score:2)
Yeah, but redlining and the Massacre in Tulsa and giving lands to white settlers and taking them away from black freed slaves and indigenous people has been the root cause of a tremendous amount of poverty. The GI bill that allowed white vets to take out loans and buy homes was specifically not extended to black people. The statistics for accumulated wealth of black families vs. white families are absurd, and a lot of it comes down to this inability in the past for families to acquire wealth to pass down to
Re: (Score:2)
"Poor kids are just as bright as white kids"
- Biden
RTFA (Score:3)
And go look up the stories about Sammy Davis Jr's accident.
Furthermore it's beyond naïve to suggest that Racism has no effect on poverty [npr.org]
Re: (Score:2)
the article points out that it's not just poverty. The example used is that in one black man's town there was 1 hospital and it was in the White side of the tracks. Black people didn't go there if they wanted to live. You could die from a heart attack or you could die from a lynching.
Even if that were widespread, it can't really explain huge differences in medical care for minorities. 99% of a person's medical care should be at a clinic, not at a hospital. Hospitals are for emergencies, scheduled surgeries, and giving birth. Even if you eliminated the 1% of your care that comes from a hospital entirely from your medical records, it should not make a big dent in the amount of medical data they have on you, which is what this is talking about. Of course, this assumes that people have
I should add (Score:3)
people want to think of problems like racism like they're they Death Star, where you blow up a Bad Thing in one big *boom* and everybody celebrates. Maybe every now and then you've got to blow up another big baddy (Death Star II) but for the most part you're OK.
That's now how the world works. The world is messy and there's a fuck ton of little problems that add up to big problems. And you either tackle them one at a time, making improvements around the edges, or y
Re: (Score:2)
... those same algorithms should give similarly bad results for poor white people ...
I suspect that the algorithms do exactly that, though I have no evidence to prove that. The thing is, nobody is standing up for poor white folks. As far as I am concerned, poverty is poverty, no matter what the colour of your skin.
... a need for changing the fundamental inequality that leads to poorer people being sicker ...
Universal health care in the UK via the NHS has not entirely fixed that, but I am sure it helps a great deal. I think the more fundamental problem is why there are people so poor that they are barely able to live, within a prosperous society where the average person can enjoy wo
we need single payer healthcare for profit needs t (Score:2)
we need single payer healthcare for profit needs to go!
Re: (Score:3, Informative)
I have lived under it and it SUCKS!
Even the Canadians will travel to the USA for healthcare instead of waiting years for the Canadian "single payer healthcare" to treat them. "single payer healthcare" sounds great on paper, with slogans, and sound bites but it definitely provides a lower quality of healthcare.
Don't believe me? Just start asking Senior Citizens about all the baloney that Medicare hoists on them and you'll realize that "single payer
Re: (Score:2)
Mod parent up. Every other developed western country has socialised healthcare. Your insurance system adds a huge burden of inefficiency and waste. I realise those vested interests are paying government to not change, but there also seems to be no real will amongst Americans to change. Why is that?
mod both parents DOWN...
Single payer is not the answer nor is socialized healthcare (beyond what we ALREADY have in Medicaid and Medicare).
One of the things I learned doing IT for doctors (Score:5, Insightful)
This means that more often then not they're just blindly following a guide or in this case an algorithm that tells them how to treat a patient.
Reading the article I can see why that would be a problem here. Black folks use less care overall and therefore cost less. They also have fewer data points to go by because again, less access to care. Both of these things cause the algorithm to swing against them, making them look less risky then they are.
What the software is doing is saying "It'll cost $10,000 to keep Joe Schmoe healthy or $100,000 to keep him out of the hospital, let's spend the money to keep him healthy because it'll save money".
Black folks slip between this crack because they use so much less healthcare they can be written off. What's happening is the algorithm is saying "this guy's too broke to go to the hospital anyway, we can let him die".
Re:One of the things I learned doing IT for doctor (Score:4, Interesting)
is that they're not as smart as everybody thinks they are. They just really, really focused and really, really good at studying.
My wife suffered from dizzy spells and what looked like epileptic fits for a few years. No one knew what was going on, she went to see 4 or 5 different specialists, had MRIs taken as well as multi-day EEGs, went to the epilepsy center, nothing... She asked a few specialists as well as her GP if her medication could be causing this, but that was dismissed out of hand (and what the hell was she doing making suggestions anyway?). In the end we started researching ourselves, and sure enough, it was right there amongst the side effects of her blood pressure medicine. We switched pills and the symptoms disappeared overnight. All these experts just followed a script, and checking her meds apparently wasn't part of that.
Luckily our current GP is more open minded, he doesn't mind doing research or listening to suggestions, and he's a real troubleshooter; a rare quality among doctors apparently. And he pretty much repeated what you just wrote. Like someone here wrote once: "MD stands for Memorized Degree".
Evidence-based medicine doesn't work with no data (Score:2)
The poor tend to cost more because their conditions generally aren't treated in a timely manner. There are a number of probable reasons for this: affordability, fear, apathy. Compliance is a continuous problem as well, even for the wealthy.
The problem becomes that there's no record of any treatment working for that particular ethnicity/group so they're not suggested. It could be that removing ethnicity from the data sets may help, but it's doubtful. Different treatments seem to work on different ethnicities
GIGO you will have with you always (Score:3)
When in Rome (Score:2)
Using a kind of network effect, the dominate culture tends to swamp other cultures out. It's not necessarily intentional, it's just people incrementally shaping gazillion different factors from their own perspective.
No matter who is on the top of the mound, this will happen. We should perhaps look at it as the intersection of human nature with the network effect instead of make it all about blame.
For example, the dominant culture has the luxury of thinking more longer term because they can. Poor people are
Everything's Racist (Score:3, Informative)
Shock horror (Score:5, Interesting)
Insurance companies are biased hugely against the poor.
Noting that the percentage of black Americans who are poor is much higher than the white equivalent.
Solution, address poverty. Do something where those in poverty can get jobs that pay a living wage.
In 2020, everything is racist apparently (Score:3)
Not to point out the obvious but . . .
The entire American Medical Industrial Complex doesn't CARE what color you are.
Neither does Big Pharma.
They only care if you can pay the bill ( cash or insurance ) or not.
If you can, they'll be happy to give you whatever treatments you want.
If you can't, they won't.
It's as simple as that.
If there is any bias here it's towards folks who can afford it vs those who cannot.
( Welcome to one of America's glaring problems. . . . . the income gap. )
Re: (Score:2)
The race question here is why are brown people poor? I will agree that health companies run as businesses do not care what colour you are, as long as you can pay.
I think I have to say here that I am a white male living in the UK, aged sixty something. This is probably relevant to what I feel and say. Apart from anything else, I have never experienced racism directed at me for what I am. There is a risk of being patronising by being offended on behalf of others. However, SJW I am not. I am just trying to get
It isn't just healthcare (Score:2)
WTF are they saying? (Score:2)
The bias is not intentional, but it reinforces deeply rooted inequities in the American health care system, effectively walling off low-income Black and Hispanic patients from services that less sick white patients routinely receive.
OK, so we have 'algorithms' that are denying coverage to certain people. Got, now explain how:
These algorithms are running in the background of most Americans' interaction with the health care system.
What? "Running in the background"? Whatever, continue:
They sift data on patients' medical problems, prior health costs, medication use, lab results, and other information to predict how much their care will cost in the future and inform decisions such as whether they should get extra doctor visits or other support to manage their illnesses at home.
I don't these decision are as personalized as the writer seems to believe. For example, if a woman enters a hospital to have a baby, and there are no complications (to keep the example simple), the covered expenses are standardized - don't pretend insurance companies offer whites 3 days in hospital recovery while women of color get 24 hours, if they have the sam
Re: (Score:2)
Thank the ACA.
Prior to that, our state mandated insurance rates NOT be base upon residence location. Other attributes (like age) applied. But not residence. ACA threw all that out. Shortly after the law passed, my health insurance carrier requested my residence address, "To properly determine your premium." Up to this point, they had only known me by my PO box. So I called the back and asked them for a breakdown of price for each Zip code. "I have numerous residences across the state. I'll just use the one
Re: (Score:2)
Finally some sanity about ACA. I do like that it more or less did something about the absurd pre-existing condition issue -- I kept waiting for simply being born to be a exclusionary pre-existing health care event for all other subsequent ones).
But Jesus H. Christ, it didn't do anything to "reform" health care, it just enshrined most of its worst attributes into law forever.
Re:Oh shut up (Score:4, Interesting)
You guys have made the word "racism" meaningless. This is about poor vs non-poor, not "racism".
Man, you are going to get savaged for that.
Racism has become subjective, it's defined by and need only be seen in the eyes of the aggrieved. There is no objective way to measure it, it just is.
Re:Oh shut up (Score:5, Insightful)
Re: (Score:2)
I see my touch of sarcasm wasn't lost on everybody.
Re: Racism...but not the kind you want it to be... (Score:3, Insightful)
They werent avoiding the white area and the white doctors because they were white. They were avoiding those areas because going into those areas if you were black carried a very real possibility of getting killed. And if that were to happen, and even if the white sheriff decided to arrest the white killer and the white local prosecutor decided to charge the killer, the all white jury would inevitably not convict said killer. More likely, you would just disappear. This shit really happened. My grandfather,
Re: (Score:2)
Congratulations on misreading the article.
It didn't say that they could not go to the white part of town because the doctors were white. It said that they were not allowed to go to the white part of town at any time, for any reason, irrespective of whether they were seeing a doctor or not. Presumably, the white part of town was dangerous for any non-white person.
Now,
Re: (Score:2)
The English people have learned this over the decades with Nationalized Health Care. If you want healthcare, one must take the doctors, etc. that is available.
Actually, my experience in the UK is very variable. I got a useless service out of my GP, to the extent that I changed surgery (marginally better now). I got a thoroughly first class service in hospital being treated for cancer. In the case of my GP, it was not medical expertise that I thought was wrong, but basic admin incompetence. At the hospital, I was treated by some excellent professionals: doctors, nurses and specialists. I forgive Dr Fong for a basic physics mistake, for describing the photon radiot