How Neglecting Minorities in Medical Research Has Led to Deadly Outcomes
hdphealth.com
hdphealth.com
I did not know that. It's horrifying.
This article gives an overview: https://www.aafp.org/afp/2007/0801/p391.html
Which is a far more pratical approach.
Imagine two worlds:
World #1 (the horrible world): medical research neglects black people. They only bothered to develop an asthma treatment that works on white people.
World #2 (the reasonable world): medical research is careful to make sure they cover all groups. They developed several asthma drugs. None works well across all ethnic groups, but there is a good one available for each ethnic group.
In both of these worlds the most commonly prescribed asthma drug will be one that works on the vast majority of white people and does not work well on black people.
What they need to include to make the point they are trying to make is how the less commonly prescribed asthma drugs do on different ethnic groups.
And that's not the only one.
There's many ways you can rephrase it and still be correct, mind you.
If you're no longer country's minority, you cannot be majority-minority.
You might be hang up on the additional narrative attached to the word “minority”. Consider using “underpriviledged ethnic group", as a. not even today all ethnic groups are equal and b. it is entirely possible some other ethnic group to raise to demographic majority status, and still exhibit the socioeconomic characteristics of a poor population.
>You might be hang up on the additional narrative attached to the word “minority”.
You can't separate words from their "additional narratives" and expect all people to understand what you're attempting to convey. The context and connotation around words is absolutely necessary for communication, and to eschew it leads to poorer communication.
It's important to match the demographics of the population with the demographics of trial users. It's important to prevent as many deaths as we can.
It's important to account for differences between groups when you're developing healthcare.
But please, don't mangle mathematical terms.
You have just demonstrated what is called "attribution of intention" (as in you attributed an intention to the other person), which is a good indicator of a conversation going south.
(More general bias is "attribution bias")
HN guideline:
>Please respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize. Assume good faith.
While we "focus on medical" clinical research isn't just testing new drugs, it's determining if one diet is better than other, it's finding prevention mechanisms and much much more. We work with any center that is running any type of research. And while we focus mostly on cancer treatments (70% of research dollars are spent here) we're building technology and want to expand to cover all types of clinical research as we grow.
Do you have any thoughts on how improvements in these sorts of machine learning algorithms could make clinical trials more fair? I ask because you specifically mention that HDP uses AI in the article.
What we are also working on to make things more fair is helping open critical clinical trials in areas where there is a higher density of minority populations, as well as using demographic data in identifying where minority patients exist that match up to the latest clinical trials to prove to pharma companies that opening them in those areas is not only a moral improvement, but a financial one as well. (Disclaimer we do all of this de-identified).
Would definitely be great to have a larger discussion on this topic!
Would a drug proven effective for whites and ineffective for blacks in the US suddenly become effective for blacks and ineffective for whites if administered in Zimbabwe?
Or do you mean "minorities" in the "there are too many minorities in my waterpark" sense?
(1) Identify various medications that are most likely to be effective between the different races
or
(2) Identify the differences between the races that is causing various medications to not be effective
If I ever figure out which would be better I'll let you know! Or if someone else chimes in.
For research done by private institutions (Pharma, Biotech, etc.) there is no mandate and it's on the researcher to make it a thing.
I'm not in the health field, so I don't know where the regulations stand, but I recall talking to someone who was advocating for this a couple of decades ago to avoid future cases like Thalidomide.
Take the example of hypertension, which is more prevalent among blacks than whites. At the moment, all of the major antihypertensive families are available as generic drugs. The best way to help black people quickly is to do a comparative study of drugs already on the market. No drug company is going to do that because they won't be able to capture the value if the drugs are generic.
Maybe a drug company could try to develop a drug which is particularly good for blacks, but it will take a decade, may not work, won't sell as well (fewer people), and then it will be an expensive new drug and people will complain about the cost.
Further down in the comments somewhere I posted another comment where it's as simple as certain chemo's cause toxicity levels in Asians where they don't in white Americans. So simply lowering the dosage would fix the problem but because that wasn't involved in the original research it wasn't known for quite sometime and Asian populations suffered.
When I say "focus on minorities for research" working any solution is fine, simple or not.
There will be no majority, and whites will (initially, at least) be a plurality.
Essentially, the drug companies can't win. People will find something to complain about. Either minorities are getting neglected, or they are being abused as lab subjects.
Saying "steer clear of minorities" shows complete "superiority" complex and is indicative of why the problem exists.
Interesting. When you say suppose "this" wasn't the case, what do you mean by "this?" Do you mean the neglecting minorities in medical research leading to numerous negative consequences?
From what I understand correctly, you are trying to state the negation of OP's headline. If so, it's unclear how you arrived at that specific headline. Can you walk me through your line of reasoning?
> Essentially, the drug companies can't win. People will find something to complain about.
That's quite a bleak forecast. Can you clarify what it means for drug companies to win?
From my frame of reference, the game is the political arena, and a participant (or political entity) is "winning" if they can ensure they're interests and concerns are considered and upheld. Between 1998 and 2017, Pharmaceutical/Health product industry spent 3.7 billion dollars lobbying on capitol hill [0]. Interestingly, this was the highest amount spent by any industry. Moreover, it was 1.2 billion dollars more than the next highest industry, the insurance industry. If we examine just 2016 (and 2015), we see that Pharmaceutical Research & Manufacturers of America (PRMA) spent 19 million on lobbying [1], up about 1 million from the prior year. PRMA represents companies in the pharmaceutical industry, i.e. Big Pharma. It was surpassed in spending by the Blue Cross, American Hospital Association, Chamber of Commerce, and National Association of Realtors.
Further down that list, we see Pfizer chipped in 9.7 million (up from 7 million in 2015), and the Biotechnology innovation organization chipped 9.2 million (up from 8.3 million in 2015).
At least to me, this doesn't seem like the behavior of someone that is eliminated from a competition. If things weren't going right, I would likely cut losses and try alternative routes, not spend more. So it suggests they are satisfied with the influence they're dollar buys. Presumably, they are able to clearly communicate their interests and concerns to the decision makers on Capitol Hill. Moreover, they are able to communicate these interests and concerns _behind closed doors_ to receptive members of _both parties._
Perhaps my thinking is faulty, but to me, it seems like they're winning, i.e. influencing policy in their favor.
> Either minorities are getting neglected, or they are being abused as lab subjects.
Can you explain why these are the only two options?
[0] https://www.opensecrets.org/lobby/top.php?indexType=i
[1] http://thehill.com/business-a-lobbying/business-a-lobbying/3...
This is a longer article on the subject of poorer health outcomes for minorities, covering the situation for black women and one family's experience, which is very much worth reading, but here are the key points:
Statistics:
* A black woman is 22% more likely to die from heart disease than a white woman
* ...71 percent more likely to perish from cervical cancer
* ...243 percent more likely to die from pregnancy- or childbirth-related causes
Reasons:
* Black women are more likely to be uninsured outside of pregnancy, when Medicaid kicks in, and thus more likely to start prenatal care later and to lose coverage in the postpartum period.
* The hospitals where they give birth are often the products of historical segregation, lower in quality than those where white mothers deliver, with significantly higher rates of life-threatening complications.
* [Black women] are more likely to have chronic conditions such as obesity, diabetes, and hypertension that make having a baby more dangerous.
* Black expectant and new mothers frequently told us that doctors and nurses didn’t take their pain seriously [...] numerous studies that show pain is often undertreated in black patients for conditions from appendicitis to cancer.
* An expanding field of research shows that the stress of being a black woman in American society can take a significant physical toll during pregnancy and childbirth.
* Black women are 49 percent more likely than whites to deliver prematurely (and, closely related, black infants are twice as likely as white babies to die before their first birthday).
As well, many of these outcomes he mentioned are the indirect result of years of research. Think of diabetes, 100 years of research was 99% spent on a white population and that contains prevention, diagnosis, and treatment. Populations that aren't white begin to see detrimental effects that add up over time.
Taking medical checks for foreigners or minorities in japan means the guidance numbers are mostly irrelevant. Doctors might not know very well how to deal with you, a lot of advice for common but non critical illness can be summed up to “you do you”
I'm confused, I thought that over the last ten years a lot of the medical research formerly conducted by U.S. pharma companies has been outsourced to SROs, who in turn test their drugs on poor people in India or whatever. Is that not actually the case?
That's not true at all if you consider PMS (Post Marketing Studies) which are an integral part of clinical development post launch.
It seems a complex question that has many facets. Two tickets that come to mind:
1. When US sociologists and political scientists talk about “systemic and structural racism” this is one of the manifestations.
That in general the well being of Native peoples in the US, of African Americans is devalued. It is witnessed in the exclusion in drug trials, in the diseases that pharma considers worthwhile to address, in the staffing of hospitals, in the access to healthcare,etc.
2. The unique inefficiency of the US healthcare system among those of wealthy countries has been documented in depth so I don’t know if is possible to do an adequate comparison of progressive health systems (e.g. Japan, UK, Finland) against that of the US.
It might be worthwhile to pull in progressive health systems that focus primarily upon Black people —- Botswana comes to mind —- as a point of comparison.
Speaking about medical research specifically:
You seem to say these outcomes are because of white racists "devaluing" black and native peoples' well-being in an evil act of collective racism.
A simpler explanation would be: There are far fewer blacks and natives than whites in America, which means fewer sick people to help and fewer customers, which means their unique illnesses get less research focus.
The same effect happens between common diseases and rare diseases irrespective of race. Common diseases get studied first, because that's where the most good can be done. This isn't because the well-being of people with rare diseases is "devalued".
Frankly I'm concerned you jump so quickly to a mass accusation of collective racial evil (which echoes historic hatred against other high-performing ethnic groups) when a simpler explanation is so obvious.
Do you live in America? Are you familiar with US history? If both of these are true then I dont understand why its so hard for you to grasp the fact that much of this was done out of malice.
https://en.wikipedia.org/wiki/Tuskegee_syphilis_experiment - I guess you think this was a "mistake" too?
In many clinical trials African Americans that contract various conditions at the same rate or higher than White Americans represent only 1% of the clinical trial versus 15% of the population. While 95% of the trial are White Americans but they are only 60% of the population. There is clearly a disparity here.
I also use the word "Neglect" they haven't recently purposefully ignored minorities (although in the past they did), But the people that want to run the trials put the trials in the neighborhoods (read mostly white populations) that they have worked with before and want to cover. Therefore trials aren't being run where Minorities live.
This is where the "systemic racism" comes into play.
Here, you've provided a perfect illustration of why we have the term "structural racism". Structural racism is the emergent discrimination arising from the circumstances that created our status quo. You'd think an audience of computer scientists would have an especially good intuition for emergent systems properties.
Here's a simple explanation for how African Americans can be discriminated against in health care without any of the doctors or nurses involved having overtly racist impulses:
Until the nineteen seventies --- within many of our conscious lifespans! --- African Americans were actively, overtly, deliberately discriminated against in real estate. They were redlined out of white neighborhoods and into low-income neighborhoods. Naturally, once real estate lenders would allow them to buy houses in any neighborhood they wanted, African Americans of means began buying houses anywhere they wanted. Unlike low-income "white" people, low-income "black" people were stuffed into neighborhoods that were first deliberately underfunded, and then further disinvested by the vicious cycle of neighborhood flight ---- like a run on a bank.
The hospitals, doctors offices, pharmacies, and medical service providers available in those neighborhoods are poorer than those in white neighborhoods due to disinvestment.
The unbelievably awful people who designed and executed on redlining are probably long retired by now. Many of them are no doubt deceased. Most of us would recoil from racial barriers in real estate lending. We all believe ourselves to be well-intentioned. Samuel L. Jackson has a retort our best intentions.
>When people try to explain why you're coming to broken conclusions due to broken reasoning, they get attacked as radical leftists for using the straightforward terminology we have for describing the phenomenon we're discussing.
How much of this is caused by people have past experience with selective application of different lines of reasoning.
For example, use the legal's systems racial and sex based discrimination. If we look at racial discrimination, it should be pretty clear that minorities have it much worse than whites. And there is a lot of research on this. If you then look at it based on gender, it appears there is even stronger discrimination based on gender than on race, with males much worse off than females (and a minority male receiving the worst of each). But the treatment of this online seems quite different. While it is a personal anecdote, on multiple occasions I've been told the racial discrimination is caused by structural racism against minorities that treats them worse than whites at every step on the system (from being more likely to be stopped and searched, to being more likely to be convicted given equal evidence, to receiving harsher sentences), and then being told that the gender discrimination is caused by sexism against women, resulting from the legal system treating women as children every step of the way (meaning they are less likely to be stopped and searched, less likely to be convicted, and receive less time). These seem like polar opposite lines of reasoning, yet I've seen both used as the same time.
I think it is at this point you get people who become opposed to the underlying reasoning because it appears that the group using the reasoning is starting with an assumption and then picking the logic that best fits their assumption. And I think many of the people you encounter online who use this reasoning are doing just that. People of every political and other leaning like to manipulate data to fit their world view. Combined with a lack of exposure to the actual scientists who work on this it can paint people's view of the language. To say nothing of scientist being humans and thus there being examples of scientist being very non-scientific about some issue (while I don't know of any examples on this particular issue, I did read through case of correspondences published in a scientific journal dealing with classification of certain behaviors as mental illnesses where some scientist were making some very indefensible arguments concerning evolution of which numerous counter examples were available that basically boiled down to "there is no way trait X evolved because it isn't reproductively advantageous in our environment").
And to be clear on my own stance, I do think that systematic racism exists in our current system, including in sub-systems where there are no racist members. There are agent models that show with even a small in-group bias, completely devoid of any out-group bias, you can have a system where out-group bias is apparent. For example, a system of entities of type A and B where A's has a certain preference for grouping with other A's, but no preference for not grouping with B's, ends up behaving similar to a system where A's have a preference for not grouping with B's.
The problem is you are viewing them as reasoning about the cause from the effect alone rather than reasoning about the cause from a combination of the effect and masses of historical evidence.
Go to your HN profile and turn 'showdead' on. It hasn't been deleted, users flagged it.
Which has always bothered me, since there are other explanations (lifestyle habits, genetics, poverty) that would explain the difference, in whole or in part. It's a politically convenient assumption that goes contrary to Occam's Razor.
"This is not only false, but dangerously false. We are in the process of discovering that certain classes of popularly-prescribed drugs (eg ACE inhibitors for blacks, certain chemotherapy drugs for Asians) are ineffective or even toxic for populations not represented in the relevant drug development research cohorts. It's not identity politics to note that pharmacokinetics can differ between individuals and populations. These differences do not explain all of the population-level morbidity and mortality differences between ethnicities, but they are significant when investigating differences between groups on the same course of treatment."
When you say "your assumption seemed to be" aren't you really talking about your own assumptions about what I'm thinking (but didn't write)?
> since there are other explanations (lifestyle habits, genetics, poverty) that would explain the difference, in whole or in part.
This doesn't contradict the meaning of "structural, systemic racism", but it explains it. When pharmacies are making drugs that are only effective for white people, and not researching effectiveness on black people, that's structural racism almost by definition.
Obviously pharma companies are responding to financial incentives, and if it's not profitable for a company to research treatments specifically helping a minority group then they're probably not going to. Less availability of pharmaceuticals makes treatment harder, causing what is available more expensive or leading to complications that require more further medical treatment (and cost more money); and those who choose not to get treatment will find themselves with further medical conditions later. In the end it would cost the minority more money, which they likely cannot pay for other systemic reasons, so more often they would be denied access to a hospital outright. Everyone involved is responding to natural incentives, but the net result still becomes [minority group] is neglected because of the color of their skin.
> Did you read the whole comment?
When you write like this it feels like you're attacking the character of the person you're talking to, which makes the whole conversation more toxic to follow.
That seems way more likely to me than oft-hinted-at-by-the-"politically-incorrect" "here's a cluster of people who all made the same bad decisions or were victims of the same bad luck in the same way, for no underlying reason other than genetic factors also associated with the color of their skin." That's pretty damn "politically convenient" if you're not in the minority population, too - "hey guys, it's not our fault! They just suck!" Hard to imagine something more politically convenient to the lucky than that.
You dismissed one assumption and named three more. Please explain which applies to Occam's razor.
If you can then explain how your selection isn't related to structural racism and isn't politically convenient for you, I'd appreciate it.
Because there is a direct correlation between observable characteristics like obesity and poverty to health outcomes for people of all races. If a fat, poor white woman in Appalachia has heart problems in her 40s and receives low quality care, is that a result of systemic racism?
And as others have pointed out, we don't know the entire scope of genetic effects, but we know they exist to some degree (which was the whole point of the article).
The centuries-long existence of slavery, segregation (which was brutal oppression, including lynching), and racism isn't an "assumption", but indisputable fact. Occam's Razor is not a real arbiter of truth, but in this case it cuts the other way: Racism is the simpler and blazingly obvious explanation, backed by endless reearch and even the most casual observation. You really have to work to contrive explanations that don't include systemic and structural racism.
> poverty
Another outcome of those centuries.
Things that are wrong can be obvious to individuals and groups of people. It's certainly not obvious to half the country, and that "endless research" is tainted. How long do you get to keep your job in academia if you point out the primary drivers of black misery in the US (out of wedlock births, drugs, and violence) are self inflicted?
To "point out" something (in academia or elsewhere), that something must be a fact.
It's almost as if the sins of the past affects the lives of people in the present somehow.
In any event, if free will isn't a thing, there's no point in trying to make the world a better place, right, so we should just leave things as they are?
You are suggesting that there isn't widespread racism now? What knowledge or basis do you have for all this?
> if free will isn't a thing
So either there is no free will or there are no systemic problems? Are poverty in Somalia and Kirghistan systemic issues, or is it just a failure of the people there that they don't live like people in the Bay Area? In the U.S., is poverty on Native American reservations, and among almost every group that isn't white men, just due to laziness? Society, health care, schools, the economy, racism, etc. - all have no effect?
It's certainly obvious to the most of the U.S., and facts are not subject to a popularity vote regardless.
> "endless research" is tainted
Easily said, but completely unsubstantiated
> How long do you get to keep your job in academia if you point out ...
Can you substantiate that such a thing is true, and then answer your question? One thing that will lose you your job in academia is making intellectually weak, baseless claims. Academia doesn't run an affirmative action program to include all political ideologies; you have to actually have evidence and good arguments.
I seem to recall "scientific consensus" being important on certain things?
https://www.amazon.com/Color-Law-Forgotten-Government-Segreg...
Blaming genetics for such a wide range of negative outcomes is silly if you're looking for a simple explanation. Consider this[1] paper on cervical cancer. There are nine different genes linked to it. Now do that for every relatively worse health outcome. That's the opposite of simple.
Lifestyle is even more nebulous. It encompasses so much that suggesting it's a kind of verbal jujitsu to suggest it's an adequate use of Occam's Razor. There are hundreds of lifestyle factors and you can weight them however you want to get the result you want.
All of those are about as broad and as simple as racism, which you appear to categorically disregard as a plausible explanation.
Yes, it happens sometimes that certains drugs approved in the US fail to make it in Japanese clinical trials. Recent one I can think of is Prozac:
https://clinicaltrials.gov/ct2/show/NCT01808612
It did not show to be better than Placebo in the Phase3 Japan trial. (https://clinicaltrials.gov/ct2/show/results/NCT01808612?sect...)
There could be multiple reasons for that, but in the end the drug is not approved in Japan.
In recent news, Serena Williams had to forcefully talk her nurses and doctors through treating her pulmonary embolism after giving birth to her child:
https://qz.com/1177004/serena-williamss-terrifying-childbirt...
Subheader: "When it comes to maternal health, even the greatest of all time is just another black woman."
Perhaps the author had no evidence of it. Is there any?
If everyone is the same, why bother with diverse studies?
At least within the US, it is currently deemed wrong think for a number of reasons.
First, how are you defining race? The current American definitions of race are imprecise and sociologically constructed. Since the founding of America, the definition of white has expanded to include various groups that were previously excluded (e.g. Polish, Irish, Jewish people, Italians, Greek) [0].
As far as black/African Americans, the current definition (one drop rule) has its limitations. Sub-saharan Africa is not a monolith. It has been found that the genetic diversity between Sub-saharan African ethnicity groups exceeds other regions of the world [1]. Even as far as physical differences, there is a noticeable physical difference between a Tutsi from Rwanda, a Luhya from Kenya, an Amhara from Ethiopia, and a Yoruba from Nigeria. A very noticeable difference. I can't speak for their "genetic difference." However, it's large enough to warrant further investigation on what categories are used. This distinction is important moving forward. Since 1970, voluntary immigration from Africa has surged from the aforementioned regions [2], with many becoming first, and second generation Americans [3], the current racial assumptions and definitions are a tad archaic.
This is just one racial group within the US. I didn't really even mention how ludicrous bucketing 60% of the world's population as "Asian" is.
Second, and most importantly, the US does not have a great track record with biological research targeting black people, Native Americans, and other minority groups. There is a long history of medical abuse where researchers conduct experiments without the knowledge or consent of black patients. A few notable examples in recent history are the Tuskagee Syphilis Experiments between 1932 and 1972 [3] and DoD's non-consensual whole body radiation experiments on black cancer patients between 1960 and 1971 [4][5]. Even today, this sort of non-consensual racial medical experimentation has not stopped. As recently as 2013, Ethiopian Jews (immigrating from Ethiopia) in Israel were coerced into agreeing to injections of long acting birth control drugs by Israeli medical officials [6]. The apprehension is very much justified.
> If everyone is the same, why bother with diverse studies?
You're tone suggests that you don't understand why people are "dancing" around the subject and trying to be politically correct or whatever. I am under the impression that the American medical research community has, through its actions, fostered distrust in "racial" based medical research. Now, it's stuck in a bind, as a direct consequence of its actions.
[0] https://en.wikipedia.org/wiki/Definitions_of_whiteness_in_th...
[1] cshperspectives.cshlp.org/content/6/7/a008524.full
[2] http://www.pewresearch.org/fact-tank/2017/02/14/african-immi...
[3] https://www.bloomberg.com/view/articles/2015-10-13/it-isn-t-...
[4] https://en.wikipedia.org/wiki/Tuskegee_syphilis_experiment
[5] http://www.nytimes.com/2007/01/23/health/23book.html
[7] https://www.forbes.com/sites/eliseknutsen/2013/01/28/israel-...